Chemosensitivity Vol 1

M E T H O D S I N M O L E C U L A R M E D I C I N E TM
Chemosensitivity
Volume 1
In Vitro Assays
Edited by
Rosalyn D. Blumenthal
Chemosensitivity
M E T H O D S I N M O L E C U L A R M E D I C I N E
John M. Walker, SERIES EDITOR

118. Antifungal Agents: Methods and Protocols,
edited by Erika J. Ernst and P. David Rogers,
2005
117. Fibrosis Research: Methods and Protocols,
edited by John Varga, David A. Brenner,
and Sem H. Phan, 2005
116. Inteferon Methods and Protocols, edited by
Daniel J. J. Carr, 2005
115. Lymphoma: Methods and Protocols, edited by
Timothy Illidge and Peter W. M. Johnson,
2005
114. Microarrays in Clinical Diagnostics, edited by
Thomas Joos and Paolo Fortina, 2005
113. Multiple Myeloma: Methods and Protocols,
edited by Ross D. Brown and P. Joy Ho, 2005
112. Molecular Cardiology: Methods and Protocols,
edited by Zhongjie Sun, 2005
111. Chemosensitivity: Volume 2, In Vivo Models, Imaging, and Molecular Regulators, edited
by Rosalyn D. Blumethal, 2005
110. Chemosensitivity: Volume 1, In Vitro Assays,
edited by Rosalyn D. Blumethal, 2005
109. Adoptive Immunotherapy: Methods and
Protocols, edited by Burkhard Ludewig and
Matthias W. Hoffman, 2005
108. Hypertension: Methods and Protocols,
edited by Jrme P. Fennell and Andrew
H. Baker, 2005
107. Human Cell Culture Protocols, Second
Edition, edited by Joanna Picot, 2005
106. Antisense Therapeutics, Second Edition,
edited by M. Ian Phillips, 2005
105. Developmental Hematopoiesis: Methods
and Protocols, edited by Margaret H. Baron,
2005
104. Stroke Genomics: Methods and Reviews, edited
by Simon J. Read and David Virley, 2004
103. Pancreatic Cancer: Methods and Protocols,
edited by Gloria H. Su, 2004
102. Autoimmunity: Methods and Protocols,
edited by Andras Perl, 2004
101. Cartilage and Osteoarthritis: Volume 2,
Structure and In Vivo Analysis, edited by
Frdric De Ceuninck, Massimo Sabatini, and
Philippe Pastoureau, 2004
100. Cartilage and Osteoarthritis: Volume 1,
Cellular and Molecular Tools, edited by
Massimo Sabatini, Philippe Pastoureau, and
Frdric De Ceuninck, 2004
99. Pain Research: Methods and Protocols,

edited by David Z. Luo, 2004
98. Tumor Necrosis Factor: Methods and
Protocols, edited by Angelo Corti and Pietro
Ghezzi, 2004
97. Molecular Diagnosis of Cancer: Methods and
Protocols, Second Edition, edited by Joseph E.
Roulston and John M. S. Bartlett, 2004
96. Hepatitis B and D Protocols: Volume 2,
Immunology, Model Systems, and Clinical
Studies, edited by Robert K. Hamatake and
Johnson Y. N. Lau, 2004
95. Hepatitis B and D Protocols: Volume 1,
Detection, Genotypes, and Characterization,
edited by Robert K. Hamatake and Johnson Y.
N. Lau, 2004
94. Molecular Diagnosis of Infectious Diseases,
Second Edition, edited by Jochen Decker and
Udo Reischl, 2004
93. Anticoagulants, Antiplatelets, and
Thrombolytics, edited by Shaker A. Mousa,
2004
92. Molecular Diagnosis of Genetic Diseases,
Second Edition, edited by Rob Elles and
Roger Mountford, 2004
91. Pediatric Hematology: Methods and
Protocols, edited by Nicholas J. Goulden
and Colin G. Steward, 2003
90. Suicide Gene Therapy: Methods and Reviews,
edited by Caroline J. Springer, 2004
89. The BloodBrain Barrier: Biology and
Research Protocols, edited by Sukriti Nag, 2003
88. Cancer Cell Culture: Methods and Protocols,
edited by Simon P. Langdon, 2003
87. Vaccine Protocols, Second Edition, edited by
Andrew Robinson, Michael J. Hudson, and
Martin P. Cranage, 2003
86. Renal Disease: Techniques and Protocols,
edited by Michael S. Goligorsky, 2003
85. Novel Anticancer Drug Protocols, edited by
John K. Buolamwini
and Alex A. Adjei, 2003
84. Opioid Research: Methods and Protocols,
edited by Zhizhong Z. Pan, 2003
83. Diabetes Mellitus: Methods and Protocols,
edited by Sabire zcan, 2003
82. Hemoglobin Disorders: Molecular Methods
and Protocols, edited by Ronald L. Nagel,
2003
M E T H O D S I N M O L E C U L A R M E D I C I N E
Chemosensitivity
Volume 1
In Vitro Assays
Edited by
Garden State Cancer Center, Belleville, NJ
2005 Humana Press Inc.

999 Riverview Drive, Suite 208
Totowa, New Jersey 07512
www.humanapress.com
All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or
by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise without written permission
from the Publisher. Methods in Molecular Medicine is a trademark of The Humana Press Inc.
The content and opinions expressed in this book are the sole work of the authors and editors, who have
warranted due diligence in the creation and issuance of their work. The publisher, editors, and authors are
not responsible for errors or omissions or for any consequences arising from the information or opinions
presented in this book and make no warranty, express or implied, with respect to its contents.
This publication is printed on acid-free paper.
ANSI Z39.48-1984 (American Standards Institute)
Permanence of Paper for Printed Library Materials.
Cover illustrations: Foreground illustration: Figure 3, from Chapter 10, Chemosensitivity Testing Using
MicroplateAdenosine TriphosphateBased Luminescence Measurements, by Christian M. Kurbacher and
Ian A. Cree. Background illustration: Figure 4, from Chapter 22 (Volume 2), Assessing Growth and
Response to Therapy in Murine Tumor Models, by C. P. Reynolds et al.
Cover design by Patricia F. Cleary.
For additional copies, pricing for bulk purchases, and/or information about other Humana titles, contact
Humana at the above address or at any of the following numbers: Tel.: 973-256-1699; Fax: 973-256-8341;
E-mail: humana@humanapr.com; or visit our Website: www.humanapress.com
Photocopy Authorization Policy:
Authorization to photocopy items for internal or personal use, or the internal or personal use of specific
clients, is granted by Humana Press Inc., provided that the base fee of US $30.00 per copy is paid directly
to the Copyright Clearance Center at 222 Rosewood Drive, Danvers, MA 01923. For those organizations
that have been granted a photocopy license from the CCC, a separate system of payment has been arranged
and is acceptable to Humana Press Inc. The fee code for users of the Transactional Reporting Service is:
[1-58829-345-9/05 $30.00].
Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1
E-ISBN 1-59259-869-2
Library of Congress Cataloging in Publication Data
Chemosensitivity / edited by Rosalyn D. Blumenthal.
v. ; cm. (Methods in molecular medicine ; 110-111)
Includes bibliographical references and index.
Contents: v. 1. In vitro assays v. 2 In vivo models, imaging, and
molecular regulators.
ISBN 1-58829-345-9 (hardcover : alk. paper)
1. CancerChemotherapyLaboratory manuals. 2. Antineoplastic
agentsEffectivenessLaboratory manuals. 3. Cancer cellsLaboratory
manuals. 4. Cancer--Molecular aspectsLaboratory manuals.
[DNLM: 1. Antineoplastic Agentspharmacology. 2. Drug Screening Assays,
Antitumormethods. 3. Drug Resistance, Neoplasm. 4. Models, Animal. 5.
Neoplasmsdrug therapy. QV 269 C5177 2005] I. Blumenthal, Rosalyn D. II.
Series.
RC271.C5C396 2005
616.994061dc22
2004012494
Preface
Chemotherapy is used to treat many types of cancer. A large number of drug
classes are in use, including the vinca alkaloids, taxanes, antibiotics, anthracyclines, DNA alkylators, other DNA damaging agents, hormones, and interferons. More potent analogs of existing drugs and novel agents directed at new
targets are continuously being developed. Over the last few years, agents that
affect COX-2, PPAR, and various signal transduction pathways have received
much attention. To identify which agents are effective for which types of tumors,
it is important to develop accurate in vitro and preclinical in vivo screening systems that can identify the cytotoxic and/or cytostatic potential of an agent on
established tumor cell lines or cells isolated from individual fresh cancer biopsy
specimens removed from cancer patients. Chemosensitivity testing allows the
selection of drugs that appear sensitive in the laboratory, thus offering patients a
better chance of response.
One of the main problems associated with chemotherapy has been that
patient tumors with the same histology do not necessarily respond identically
to the same agent or dose schedule of multiple agents. Identifying the presence
of resistance mechanisms and other determinants for drug sensitivity in order
to classify tumors into response categories has been an ongoing research
effort. Advances in our understanding of the genetic and protein fingerprints of
primary tumors and their metastases has opened a door to the possibility of
customizing therapy to individuals. There is accumulating evidence suggesting that laboratory screening of samples from a patients tumor may help select
the appropriate treatment(s) to administer, thereby avoiding ineffective drugs,
and sparing patients the side effects normally associated with these agents.
The aim of these two volumes on Chemosensitivity of the Methods in Molecular Medicine series, is to comprehensively present protocols that can be used to
(a) assess chemosensitivity in vitro and in vivo, and (b) assess parameters that
modulate chemosensitivity in individual tumors. Volume I presents an overview
in Chapter 1 and then covers In Vitro Measures of Chemosensitivity, includes
clonogenic, colorimetric, fluorometric, and histochemical approaches. Volume
II, Part I, Measurements of DNA Damage, Cell Death, and Regulators of Cytotoxicity, includes methods to detect chromosome loss and breakage, changes in
cell cycle, expression of members of the bcl-2 family of proteins, expression of
caspases and PARP cleavage, metabolic factors influencing sensitivity, measurements of drug retention, expression of drug resistance proteins, and measurements of ceramide and sphingolipids associated with drug sensitivity. Volume
vi
Preface
II, Part II, Genomics, Proteomics, and Chemosensitivity, addresses DNA

microarrays for gene profiling, genetic manipulation to identify genes regulating chemosensitivity, proteomics using 2D-PAGE and mass spectrometry,
and bioinformatics approaches. The last part, In Vivo Animal Modeling of
Chemosensitivity, covers protocols to establish clinically meaningful metastatic
and orthotropic models of solid and liquid tumors, statistical approaches to analyze preclinical data, and animal imaging approaches that can be used to assess
chemosensitivity such as GFP-tagged genes, SPECT using 99mTc-annexin, PET
imaging with 18FDG, and magnetic resonance imaging.
Each chapter is written by someone experienced with the methodology and
contains a detailed introductory section with references of how the technique
has been used in the past, a list of materials and equipment needed to perform
the assay, and a step-by-step set of instructions for each method. At the end of
each chapter a Notes section is included with useful information, helpful
hints, and problems and pitfalls to be aware of, in order to make the assay run
smoothly and allow for easy interpretation of data.
Contents
Preface .............................................................................................................. v
Contributors ..................................................................................................... ix
Contents of Volume 2 ...................................................................................... xi
PART I. OVERVIEW
1 An Overview of Chemosensitivity Testing
Rosalyn D. Blumenthal ......................................................................... 3
PART II. IN VITRO MEASURES
OF
CHEMOSENSITIVITY
2 Clonogenic Cell Survival Assay

Anupama Munshi, Marvette Hobbs, and Raymond E. Meyn .............. 21
3 High-Sensitivity Cytotoxicity Assays for Nonadherent Cells
M. Jules Mattes ................................................................................... 29
4 Sulforhodamine B Assay and Chemosensitivity
Wieland Voight ................................................................................... 39
5 Use of the Differential Staining Cytotoxicity
Assay to Predict Chemosensitivity
Gertjan J. L. Kaspers ........................................................................... 49
6 Collagen Gel Droplet Culture Method
to Examine In Vitro Chemosensitivity
Hisayuki Kobayashi ............................................................................. 59
7 The MTT Assay to Evaluate Chemosensitivity
Jack D. Burton .................................................................................... 69
8 Histoculture Drug Response Assay to Monitor Chemoresponse
Shinji Ohie, Yasuhiro Udagawa, Daisuke Aoki,
and Shiro Nozawa .......................................................................... 79
9 In Vitro Testing of Chemosensitivity in Physiological Hypoxia
Rita Grigoryan, Nino Keshelava, Clarke Anderson,
and C. Patrick Reynolds .................................................................. 87
10 Chemosensitivity Testing Using Microplate Adenosine
TriphosphateBased Luminescence Measurements
Christian M. Kurbacher and Ian A. Cree .......................................... 101
11 High-Throughput Technology:
Green Fluorescent Protein to Monitor Cell Death
Marylne Fortin, Ann-Muriel Steff, and Patrice Hugo ..................... 121
vii
viii
Contents
12 DIMSCAN: A Microcomputer Fluorescence-Based Cytotoxicity

Assay for Preclinical Testing of Combination Chemotherapy
Nino Keshelava, Toms Frgala, Jir Krejsa, Ondrej Kalous,
and C. Patrick Reynolds ................................................................ 139
13 The ChemoFx Assay: An Ex Vivo Cell Culture Assay
for Predicting Anticancer Drug Responses
Robert L. Ochs, Dennis Burholt, and Paul Kornblith ....................... 155
14 Evaluating Response to Antineoplastic Drug
Combinations in Tissue Culture Models
C. Patrick Reynolds and Barry J. Maurer .......................................... 173
15 Image Analysis Using the Fluochromasia Assay
to Quantify Tumor Drug Sensitivity
John F. Gibbs, Youcef M. Rustum, and Harry K. Slocum ................. 185
16 Immunohistochemical Detection of Ornithine Decarboxylase
as a Measure of Chemosensitivity
Uriel Bachrach .................................................................................. 197
17 Immunohistochemistry of p53, Bcl-2 and Ki-67
as Predictors of Chemosensitivity
Mitsuyoshi Itaya, Jiro Yoshimoto, Kuniaki Kojima,
and Seiji Kawasaki ........................................................................ 213
Index ............................................................................................................ 229
Contributors
CLARKE ANDERSON Division Hematology-Oncology, Keck School of Medicine,
University of Southern California, Los Angeles, CA, USA
DAISUKE AOKI Department of Obstetrics-Gynecology, Keio University
School of Medicine, Keio, Japan
URIEL BACHRACH Department of Molecular Biology, Hebrew University
Hadassah Medical School, Jerusalem, Israel
ROSALYN D. BLUMENTHAL Garden State Cancer Center, Belleville, NJ, USA
DENNIS BURHOLT Precision Therapeutics, Pittsburgh, PA, USA
JACK D. BURTON Garden State Cancer Center, Belleville, NJ, USA
IAN A. CREE Department of Histopathology, Queen Alexandria Hospital,
Portsmouth, UK
MARYLENE FORTIN Topigen Pharmaceuticals, Montreal, Quebec, Canada
TOMS FRGALA USC-CHLA Institute for Pediatric Clinical Research,
University of Southern California and Childrens Hospital Los Angeles,
Los Angeles, CA, USA
JOHN F. GIBBS Department of Surgical Oncology, Roswell Park Cancer
Institute, Buffalo, NY, USA
RITA GRIGORYAN Developmental Therapeutics Section, Childrens Hospital
of Los Angeles, Los Angeles, CA, USA
MARVETTE HOBBS Department of Experimental Radiology, University of
Texas M.D. Anderson Cancer Center, Houston, TX, USA
PATRICE HUGO Caprion Pharmaceuticals Inc., Saint Laurent, QC, Canada
MITSUYOSHI ITAYA Department of Hepato-Biliary-Pancreatic Surgery,
Juntendo University, Tokyo, Japan
ONDREJ KALOUS USC-CHLA Institute for Pediatric Clinical Research,
GERTJAN J. L. KASPERS Department of Pediatric Hematology Oncology, VU
University Medical Center, Amsterdam, Netherlands
SEIJI KAWASAKI Department of Hepato-Biliary-Pancreatic Surgery,
NINO KESHELAVA USC-CHLA Institute for Pediatric Clinical Research,
HISAYUKI KOBAYASHI Biochemical Laboratory, Nitta Gelatin Inc.,
Futamata, Yao-City, Japan
ix
Contributors
KUNIAKI KOJIMA Department of Breast and Endocrine Surgery, Juntendo

University, Tokyo, Japan
PAUL KORNBLITH Precision Therapeutics, Pittsburgh, PA, USA
JIR KREJSA USC-CHLA Institute for Pediatric Clinical Research, University of
Southern California and Childrens Hospital Los Angeles, Los Angeles, CA,
USA
CHRISTIAN M. KURBACHER Department of Gynecology & Obstetrics, University
of Cologne, Cologne, Germany
M. JULES MATTES Center for Molecular Medicine and Immunology,
Belleville, NJ, USA
BARRY J. MAURER USC-CHLA Institute for Pediatric Clinical Research,
RAYMOND E. MEYN Department of Experimental Radiology, University of
ANUPAMA MUNSHI Department of Experimental Radiology, University of
SHIRO NOZAWA Department of Obstetrics-Gynecology, Keio University
School of Medicine, Keio, Japan
ROBERT L. OCHS Precision Therapeutics, Pittsburgh, PA, USA
SHINJI OHIE Cancer Research Laboratory, Hanno Research Center, Taiho
Pharmaceutical Co., Saitama, Japan
C. PATRICK REYNOLDS USC-CHLA Institute for Pediatric Clinical Research,
YOUCEF M. RUSTUM Department of Experimental Therapeutics, Roswell
Park Cancer Institute, Buffalo, NY, USA
HARRY K. SLOCUM Department of Experimental Therapeutics, Roswell
Park Cancer Institute, Buffalo, NY, USA
ANN-MURIEL STEFF World Anti-Doping Agency, Montreal, Quebec, Canada
YASUHIRO UDAGAWA Department of Obstetrics-Gynecology, Fujita Health
University School of Medicine, Japan
WIELAND VOIGHT Klinik fr Innere Medizin IVMartin Luther Universitt
Halle, Halle/Saale, Germany
JIRO YOSHIMOTO Department of Hepato-Biliary-Pancreatic Surgery,
Contents of Volume 2
Preface .............................................................................................................. v
Color Plate ....................................................................................................... xi
Contributors .................................................................................................. xiii
Contents of Volume 1 ................................................................................... xvii
PART I. MEASUREMENTS OF DNA DAMAGE, CELL DEATH,

AND REGULATORS OF CYTOTOXICITY
1 In Vitro Micronucleus Technique to Predict Chemosensitivity
Michael Fenech ..................................................................................... 3
2 Cell Cycle and Drug Sensitivity
Aslamuzzaman Kazi and Q. Ping Dou ................................................ 33
3 TUNEL Assay as a Measure of Chemotherapy-Induced Apoptosis
Robert Wieder ..................................................................................... 43
4 Apoptosis Assessment by the DNA Diffusion Assay
Narendra P. Singh ............................................................................... 55
5 PARP Cleavage and Caspase Activity to Assess Chemosensitivity
Alok C. Bharti, Yasunari Takada, and Bharat B. Aggarwal ................ 69
6 Diphenylamine Assay of DNA Fragmentation
for Chemosensitivity Testing
Cicek Gercel-Taylor ............................................................................ 79
7 Immunodetecting Members of the Bcl-2 Family of Proteins
Richard B. Lock and Kathleen M. Murphy.......................................... 83
8 Correlation of Telomerase Activity
and Telomere Length to Chemosensitivity
Yasuhiko Kiyozuka .............................................................................. 97
9 Application of Silicon Sensor Technologies to Tumor Tissue In Vitro:
Detection of Metabolic Correlates of Chemosensitivity
Pedro Mestres-Ventura, Andrea Morguet, Anette Schofer,
Michael Laue, and Werner Schmidt ............................................. 109
10 Overview of Tumor Cell Chemoresistance Mechanisms
Laura Gatti and Franco Zunino ........................................................ 127
11 Flow Cytometric Monitoring of Fluorescent Drug
Retention and Efflux
Awtar Krishan and Ronald M. Hamelik ............................................ 149
xi
xii
Contents of Vol. 2
12 Flow Cytometric Measurement of Functional

and Phenotypic P-Glycoprotein
Monica Pallis and Emma Das-Gupta ................................................ 167
13 Measurement of Ceramide and Sphingolipid Metabolism in Tumors:
Potential Modulation of Chemosensitivity
David E. Modrak ............................................................................... 183
PART II. GENOMICS, PROTEOMICS,
AND
CHEMOSENSITIVITY
14 Gene Expression Profiling to Characterize Anticancer

Drug Sensitivity
James K. Breaux and Gerrit Los ........................................................
15 Identifying Genes Related to Chemosensitivity
Using Support Vector Machine
Lei Bao ..............................................................................................
16 Genetic Manipulation of Yeast to Identify Genes
Involved in Regulation of Chemosensitivity
Giovanni L. Beretta and Paola Perego ..............................................
17 Real-Time RT-PCR (Taqman) of Tumor mRNA
to Predict Sensitivity of Specimens to 5-Fluorouracil
Tetsuro Kubota ..................................................................................
18 Use of Proteomics to Study Chemosensitivity
Julia Poland, Silke Wandschneider, Andrea Urbani,
Sergio Bernardini, Giorgio Federici, and Pranav Sinha ...............
PART III. IN VIVO ANIMAL MODELING
OF
197
233
241
257
267
CHEMOSENSITIVITY
19 Clinically Relevant Metastatic Breast Cancer

Models to Study Chemosensitivity
Lee Su Kim and Janet E. Price ...........................................................
20 Orthotopic Metastatic (MetaMouse) Models
for Discovery and Development of Novel Chemotherapy
Robert M. Hoffman ...........................................................................
21 Preclinical Testing of Antileukemic Drugs
Using an In Vivo Model of Systemic Disease
Richard B. Lock, Natalia L. Liem, and Rachael A. Papa ...................
22 Assessing Growth and Response
to Therapy in Murine Tumor Models
C. Patrick Reynolds, Bee-Chun Sun, Yves A. DeClerck,
and Rex A. Moats ..........................................................................
285
297
323
335
Contents of Vol. 2
xiii
23 Evaluation of Chemosensitivity of Micrometastatses

with Green Fluorescent Protein
Gene-Tagged Tumor Models in Mice
Hayao Nakanishi, Seiji Ito, Yoshinari Mochizuki,
and Masae Tatematsu ................................................................... 351
99m
24
Tc-Annexin A5 Uptake and Imaging to Monitor Chemosensitivity
Tarik Z. Belhocine and Francis G. Blankenberg ............................... 363
25 Magnetic Resonance Imaging
of Tumor Response to Chemotherapy
Richard Mazurchuk and Joseph A. Spernyak ................................... 381
26 Metabolic Monitoring of Chemosensitivity with 18FDG PET
Guy Jerusalem and Tarik Z. Belhocine ............................................. 417
Index ............................................................................................................ 441
I
OVERVIEW
1
An Overview of Chemosensitivity Testing
Summary
This overview chapter presents the importance of chemosensitivity testing for screening new
therapeutic agents, identifying patterns of chemosensitivity for different types of tumors, establishing patterns of cross-resistance and sensitivity in treatment naive and relapsing tumors; identifying genomic and proteomic profiles associated with sensitivity; correlating in vitro response,
preclinical in vivo effect, and clinical outcome associated with a particular therapeutic agent,
and tailoring chemotherapy regimens to individual patients. Various assays are available to
achieve these end points, including several in vitro clonogenic and proliferation assays, cell
metabolic activity assays, molecular assays to monitor expression of markers for responsiveness,
development of drug resistance and induction of apoptosis, in vivo tumor growth and survival
assays in metastatic and orthotopic models, and in vivo imaging assays. The advantages and disadvantages of the specific assays are discussed. A summary of research areas related to chemosensitivity testing is also included.
Key Words
Dose-response curve; IC50 values; imaging; metabolic assays; molecular markers; proliferation
assays.
1. Introduction
Chemosensitivity testing is an ex vivo means of determining the cytotoxic
and/or cytostatic, or apoptosis-inducing effect of anticancer drugs. The emphasis on screening new agents derived from synthetic compound archives, and
from pure natural products and their extracts, for antitumor activity necessitates in vitro evaluation in cell culture and then in vivo in appropriate tumorbearing animal models. If the agent appears effective in this system, then the
drug may be further evaluated in clinical trials. This paradigm seeks to identify
the single best treatment to administer to the average patient with a given form
of cancer through the use of prospective, randomized trials.
From: Methods in Molecular Medicine, vol. 110: Chemosensitivity: Vol. 1: In Vitro Assays
Edited by: R. D. Blumenthal Humana Press Inc., Totowa, NJ
Blumenthal
In addition to the standard, DNA-damaging and metabolic inhibiting agents

in use, many new target classes now exist, such as kinase and nonkinase
enzymes, transcriptional regulators, growth factor receptors, chemokines, angiogenic regulators, and proteinprotein interactions. Identifying which types of
tumors (e.g., breast, lung, pancreas) and which subtypes (e.g., based on morphology, differentiation status, growth rate, drug resistance status, p53 expression) within each tumor category will respond to each new agent is essential.
Experience has shown that individual patients with a similar tumor histology do
not necessarily respond identically to a given agent or set of agents. Defining
the best treatment options, combinations, and schedules using standard agents
alone or combined with new agents for an individual patient is an area of active
investigation that requires both in vitro and in vivo testing. Chemosensitivity
assays can potentially facilitate the individualization of patient treatment plans.
Many retrospective and prospective studies for leukemia and for solid tumors
have shown that drug resistance/sensitivity can be determined accurately by in
vitro drug-response assays. As a consequence, knowing the drug sensitivity of
a given tumor for a particular agent can significantly impact decision making
and treatment planning. By identifying inactive drugs, patients can be spared
standard chemotherapy regimens and their associated toxicities. In cases in
which the patients tumor is unresponsive to chemotherapy, the oncologist may
be able to offer alternative or experimental treatments much sooner, when they
may have a better chance of succeeding. Cell culture drug resistance testing
refers to testing the resistance of a patients own cancer cells in the laboratory
to drugs that may be used to treat the patients cancer (1). Table 1 summarizes
the potential value of predictive chemosensitivity assays.
When addressing the usefulness of an assay method, several questions must
be addressed: (1) Are the drug sensitivity patterns observed in vitro with tumors
of a particular histological type similar to those observed clinically for the same
tumor type? (2) Can chemotherapy, which is selected on the basis of in vitro
studies, improve patient survival? Efforts have been made to correlate in vitro
drug sensitivity at initial diagnosis with 4-yr survival results, or in vitro resistance and early relapse. If this can be achieved, it may allow one to select more
intensive regimens. (3) Do the results obtained in vitro predict those obtained
in vivo? Since the process of tumor resection, transport, and processing for
culture (mechanical or enzymatic disaggregation) introduces tissue stress and
damage, which can perturb cell function and potentiate drug sensitivity, obtaining good correlations can be problematic. Another important consideration is
that tumor cell growth rate in vitro is likely to be much faster than in vivo.
Therefore, the in vitro assay might indicate chemosensitivity in a situation
of rapid cell division, which may be a false positive result, when attempting
to translate results in vivo. Test systems using single-cell suspensions may
Chemosensitivity Testing
Table 1
Summary of Potential Value of Chemosensitivity Assays
Conducting an initial screening of new therapeutic agents
Tailoring chemotherapy regimens to the individual patient: determining tumors that
are likely to respond to a particular agent and eliminating ineffective drugs
Identifying patterns of chemosensitivity for different types and subtypes of tumors
Establishing patterns of cross-resistance and sensitivity in treatment-naive and in
relapsing patients
Identifying genomic and proteomic profiles associated with sensitivity
Correlating in vitro, preclinical in vivo, and clinical response to a therapeutic agent
overestimate sensitivity. Finally, the ability to shrink tumors in vivo with a given
treatment does not necessarily translate into a significant survival benefit.
Prediction of chemosensitivity in the clinic is particularly challenging
because drug responses reflect not only properties intrinsic to the target cell, but
also host metabolic properties. Pharmacokinetic and pharmacodynamic variables that affect drug action in vivo are not considered by in vitro assays.
Because each patient has a unique pharmacogenetic makeup, leading to significant interpatient variations in drug half-life, volume of distribution, types of
metabolites formed, and routes of elimination, correlating in vitro and in vivo
results is often not a straightforward process. Furthermore, because some therapeutic agents (e.g., cyclophosphamide or CPT-11) are prodrugs that require
metabolic activation, the in vitro modeling of in vivo tumor cell drug exposure
becomes even more complex. However, by using in vitro models to address
questions of chemosensitivity, one limits the study to cell-intrinsic properties
found in cultured cells, which simplifies the system and focuses the initial
investigation on tumor cell responsiveness.
Many different methods are available for assessing chemosensitivity (24).
In general, all assays generate dose-response curves where the dose of the drug
is related to the percentage effect, such as tumor cell kill (Fig. 1). Determining
the molar concentration that results in a 50% reduction in cell survival (IC50)
can be used to compare the efficacy of different drugs in one tumor cell system
or the same drug in different cell systems.
The most common in vitro assays can be divided into one of three categories: (1) clonogenic/proliferation assays, (2) assessment of cell metabolic
activity, and (3) measurement of cell membrane integrity. There has been much
debate as to the characteristics of the best chemosensitivity assay. For example, should the assay measure colony formation or tumor cell proliferation?
Should the assay be short term (hours to days) vs long term (days to weeks)?
Blumenthal
Fig. 1. Theoretical dose response curve.
Should the assay measure cytostatic or cytotoxic end points? Should the assay
use a cell suspension or tumor microorgans? Should metabolic or morphological end points be used? Various assays measure different end points of cellular damage. Morphological appearance is generally considered too insensitive.
Measurements of biochemical parameters or reproductive capacity are likely
to be more reliable. Short-term assays may suggest that an agent is cytotoxic,
but cells may recover. Thus, early indicators of drug-induced cell damage may
provide misleading results. For an assay to be useful, the results must correlate
with clinical response and survival; the end point of the assay must detect
effects on cancer cells exclusive of other cellular elements such as fibroblasts,
mesothelial cells, and endothelial cells; the turnaround time must meet clinical
requirements; the test information must be easily interpreted and applied; and
the test must be cost-effective.
2. Clonogenic and Proliferation Assays
The clonogenic method such as the human tumor colony-forming assay (5)
is analogous to antibiotic sensitivity testing in bacteria. Single untreated or
treated tumor cells are grown in Petri dishes in a soft agar system and colonies
are counted after about 2 to 3 wk. Automated image analyzers now make this a
much faster procedure. The use of agar allows most tumor cells to grow but
prevents fibroblast proliferation. A reduction in colony number in treated groups
reflects cytotoxicity of the agent toward the tumor cells. This is the gold standard to which all other predictive assays have been compared for positive predictive reliability (predicting patient response). Controlling the number of
colonies per plate and including only colonies with at least 4050 cells is
important for obtaining accurate information from this assay. However, the
method is complicated by the ability to obtain a single cell tumor suspension,
adequate plating efficiency, proper growth in agar, and sufficient cell numbers
to test multiple concentrations of drug. Variations on this assay include microclonogenic assays on tumor cells grown in a 96-well plate or in suspension
(6). IC50 values are determined based on dose-response curves that fit the data
to a linear quadratic equation.
Several other in vitro short-term growth inhibition assays are in use, which
are based on survival of tumor cell populations that have been in contact with
a chemotherapeutic agent. In these assays, which are experimentally simpler
than clonogenic assays, growth inhibition might not reflect true cell kill and can
result in higher false positive results. However, a measure of chemosensitivity
can be obtained even when plating efficiencies are low. One such assay is the
differential staining cytotoxicity assay (DiSC; [7]), which consists of incubating dissociated cells from biopsy specimens in the presence or absence of a
drug for 46 d, and using a dye such as fastgreen, which permeates only
through dead cells. The ratio of dead cells to total cells is a measure of cell kill.
In general, there has been good qualitative agreement between the DiSC assay
and the clonogenic assay. Duration of the assay is relatively short, and the assay
can be used on the majority of tumor specimens but is labor intensive and subject to individual interpretation. Another method, the Kern assay (8), relies on
uptake of radiolabeled precursor such as 3H-thymidine into the DNA of proliferating cells and is an example of similar assays that measure drug-induced
inhibition of radioactive precursor incorporation into cellular macromolecules (DNA, RNA, or protein) of single-cell suspensions or tumor slices. The
assumption is that thymidine incorporation into DNA reflects cell division.
There is a difference in view as to whether the assay is considered a reliable
means of quantifying drug sensitivity. Some believe that it is not an accurate
measure in biopsies because of contaminating nontumor cells and, in general,
may be problematic because the effect on thymidine incorporation is not the
same as measuring cell kill or growth inhibition (2). Others have reported similar results with the Kern assay to that obtained with clonogenic assays and
significant correlations between clinical responses and depression of DNA synthesis in vitro (9). In general, this assay excels at negative predictive reliability
(predicting drug resistance). A third approach based on proliferation is the collagen gel droplet drug sensitivity test, which is a quick and simple colorimetric quantitative approach using neutral red staining within collagen gel drops
that are imaged on a videomicroscope (10). It affords the advantage of being
able to eliminate the influence of fibroblasts in a mixed tumor/stroma population derived from a biopsy specimen.
Blumenthal
3. Cell Metabolic Activity Assays

Assays that use various surrogate markers of cell number and viability have
been used to measure cell survival in response to a therapeutic. These methods
include bioluminescence of adenosine triphosphate (ATP) (ATP-based tumor
chemosensitivity assay [ATP-TCA]) (11) using the firefly luciferin-luciferase
reagent; fluorometric microculture cytotoxicity assay (FMCA), which measures
fluorescence generated from cellular hydrolysis of nonfluorescent fluorescein
diacetate to fluorescein by viable cells in microtiter plates (12); coloration with
sulforhodamine B (SRB), a general aminoxanthene dye for proteins, which
binds electrostatically to basic amino acids; and coloration with reduction of a
yellow tetrazolium dye such as MTT or MTS (13), substrates for mitochondrial dehydrogenases, resulting in a blue-purple formazan product that reflects
metabolic activity of living cells.
The ATP-TCA is extremely sensitive because ATP levels are linear with the
number of viable cells and correlate with cytotoxicity to a number of antineoplastic agents in vitro. Assay results have also been used to predict clinical
responses (14). This assay is done in a 96-well microplate and can use cell lines,
or specimens from surgical needle biopsies, pleural or ascites fluid, and requires
only 10,00020,000 cells/well (15). The FMCA is also a useful assay in that it
is simple, is rapid, and seems to report clinically relevant cytotoxic drug sensitivity data. An alternative high-throughput fluorescence assay using enhanced
green fluorescent protein (GFP) is available that can be analyzed by flow cytometry or by fluorescence microplate reader (16). It is a sensitive and rapid method
to detect drug-induced cell death that provides results comparable to those
obtained by other traditional apoptosis assays such as annexin-V binding, propidium iodide incorporation, or reactive oxygen species production.
The SRB assay provides a sensitive index of cellular protein content that is
linear over a cell density range of two orders of magnitude, has a stable end
point that does not require immediate range, and compares favorably with the
MTT assay. However, the MTT assay is able to discriminate live cells from
cellular debris, which the SRB assay cannot do. The drawbacks of assays such
as the MTT are the sensitivity to the pH and glucose content of the media;
chemical interferences with dye reduction; the quality of reagents used to solubilize the formazan crystals; the need to read the assay immediately before the
color fades; and the limited dynamic range, with a sensitivity of only 1-log cell
kill (17). In addition, such assays do not differentiate between cytostatic and
cytotoxic effects. In spite of these limitations, the MTT assay has been effective at predicting the sensitivity of patients who have attained remission and of
patient resistance (18). Both SRB and MTT assays provide rapid results and
hold more promise in the screening and evaluation of potential new agents in
established tumor cell lines than in evaluating chemosensitivity of primary

tumor specimens.
Some researchers believe that results derived from single-cell preparations
are not as accurate as those obtained from tumor fragments in which the threedimensional structure of the tumor tissue is maintained (19). Two methods to
assess chemosensitivity using tumor fragments in place of single cells have
been utilized. The first is a variation on the MTT assay, called the histoculture
drug response assay (20). The second approach, the fluorescent cytoprint assay
(FCA), can be directly used on biopsy tissue without cell dissociation (21). In
this assay, the tumor fragments are maintained in a collagen matrix during drug
exposure. They are then treated with a fluorescent dye, which is visualized and
photographed as cytoprints before and after drug treatment. This method has
been shown to have a very high positive and negative predictive accuracy.
Another marketed in vitro assay that has been reported to have a high predictive value is the extreme drug resistance assay (22). Tumor cells are exposed
to suprapharmacological drug concentrations and either precursor incorporation
or colony formation is measured. A drug resistance profile (extreme, intermediate, or low) can be determined based on statistical comparison to a historical
database of tumor specimens tested against the same panel of chemotherapeutic
agents (23).
An alternative approach to measuring cell proliferation or metabolic viability is to measure drug-induced DNA damage at the chromosome level using
morphological criteria. The cytokinesis block micronucleus assay can reliably
measure chromosome loss, chromosome breakage, chromosome rearrangement
(nucleoplasmic bridges), cell division inhibition, necrosis, and apoptosis (24).
All the previously summarized in vitro assays, and which are detailed in the
chapters that follow, must ultimately provide a drug sensitivity index to put the
results into the context of other similar assay results, or calculate a probability
of response. Table 2 provides a summary of correlations of in vitro results with
patient response (25). The most sensitive assays are the DiSc and the FCA,
and the most specific are the clonogenic, MTT, and ATP assays. The ATP assay
also gives the highest positive prediction, and the clonogenic assay, thymidine
incorporation assay, DiSc and FCA are all highly accurate.
4. In Vivo Assays
In vitro assays in general are limited in that they do not take into account insufficient drug absorption, inadequate drug distribution to the tumor owing to poor
vascularization, pharmacological barriers, the need for drug activation, detoxification of drug by general metabolic pathways, differences in tumor growth rate in
vitro and in vivo, or the problem of tumor heterogeneity. Therefore, assessing
10
Blumenthal
Table 2
Correlation of In Vitro Results with Patient Responsea
Type of
assay
Total
(N)
Clonogenic
Thymidine
DiSC
MTT
ATP
FCA
Total
2300
512 1427
494 123 232
510
247
175
326
187
74
129
74
37
333
154
116
4092 1297 2061
TP
TN
FP
FN
226
119
72
37
6
52
512
135
20
16
28
12
11
222
Positive
Accuracy Sensitivity Specificity
prediction (%)
(%)
(%)
(%)
69
51
77
83
93
75
75
91
92
92
73
76
91
86
79
86
94
86
86
93
87
86
66
71
86
86
69
74
a
Summary of clinical correlations from Table 7 in ref. 25. TP = true positivepatients who are
sensitive in vitro and respond to therapy; TN = true negativepatients who are resistant in vitro and
do not respond to therapy; FP = false positivepatients who are sensitive in vitro but resistant clinically; FN = false negativepatients who are resistant in vitro but respond clinically; positive prediction = TP/(TP + FP); accuracy = TN/(TN + FN); sensitivity = tests ability to detect clinically
responsive patients = TP/(TP + FN); specificity = tests ability to detect clinically unresponsive
patients = TN/(TN + FP).
chemosensitivity in preclinical tumor-bearing animal models is essential. Primary tumor models that have been utilized include human tumors grown subcutaneously in athymic nude mice, severe combined immunodeficient mice, or
triple-deficient mice (bg/nu/xid mutations; [26]). Examples of metastasis
models of human tumors include those grown orthotopically (27) or introduced
systemically via iv, ip, intracardiac, intrasplenic, or intrahepatic injection or via
inoculation of a cell suspension or tumor fragment into the mammary pad (28)
or subrenal capsule (29). For primary growth, tumor size is followed by caliper
measurement and tumor growth curves are constructed. Various formulas to
measure tumor volume, area, and diameter have been used (30). Absolute tumor
size, change in tumor size, percentage of growth inhibition, and area under the
tumor growth curve have been reported. Using appropriate statistical methods
to analyze tumor growth experiments that have sufficient power and low type
I error rates is essential when performing in vivo chemosensitivity experiments
(31). For metastasis models, median animal survival time, number and/or size
of metastases after a fixed time, and weight of an organ containing metastases
have been used as measures of response to a therapeutic agent. In addition to
direct measurements of tumor growth, imaging approaches have been developed to assess tumor chemosensitivity (32). These tools include the use of
18
FDG positron emission tomography (33), 99mTc-annexin-V scintigraphy (34),
magnetic resonance imaging (35) and GFP (36).
11
5. Molecular Assays of Chemosensitivity

The range of response to a particular cytotoxic agent can be quite substantial,
and studies to understand the molecular parameters at the gene and at the protein level that regulate chemosensitivity or resistance are an active area of investigation (37). The molecular basis of sensitivity to chemotherapy is a complex
product of cellular and tissue factors (3841). These include expression of the
P-glycoprotein family of membrane transporters (e.g., MDR1, MRP, LRP), which
decrease the intracellular accumulation of drug; changes in cellular proteins
involved in detoxification (e.g., glutathione S-transferase , metallothioneins,
NADP cytochrome P-450 reductase); changes in expression of molecules
involved in DNA repair (e.g., O6-methylguanine DNA methyltransferase, DNA
topoisomerase II, hMLH1, p21WAF1/CIP1); and activation oncogenes such as
Her-2/neu, bcl-2, bcl-XL, c-myc, ras, c-jun, c-fos, MDM2, p210, BCR-abl, or
mutant p53. In addition, expression of growth factors and receptors, proliferation
markers (42), telomerase (43), enzymes that regulate intracellular ceramide
levels (44), positive and negative regulators of apoptosis, and cell-cycle checkpoint controls are all important (45,46). For example, the form of p53 expression
(null, wild type, or mutant and the location of the mutation) is known to affect
chemosensitivity (47). Drug resistance can occur at the onset of the disease or
can be acquired after previous chemotherapy. Methods that identify and detect
expression of drug resistance genes (48) or proteins (49) by flow cytometry
(50), Western blotting, or immunohistochemistry (IHC) (51); or by induction of
apoptosis by the DNA diffusion assay (52), diphenylamine assay to evaluate
DNA fragmentation (53), TUNEL assay (54), flow cytometry with fluorescein
isothiocyanate-annexin-V (34), PARP assay (55), or expression of BCl-2 family
proteins including Bcl-2, Bcl-X(L), Bax, Bad, and Bak (56); or by changes in
cell cycle (57) can be used as indirect measures of chemosensitivity.
6. Genomic and Proteomic Approaches
to Understand Chemosensitivity
Previous efforts to use genetic information to predict drug sensitivity primarily have focused on individual genes that have broad effects, such as the multidrug resistance genes mdr1 and mrp1 (58). There has been an effort to develop
a genomics- and proteomics-based approach for the prediction of drug response
(5961). An example of the clinical application of single-gene analysis as a predictor of chemosensitivity is the evaluation of thymidylate synthase mRNA
expression from tumor core needle biopsies using real-time polymerase chain
reaction analysis. Those patients with a clinical response to 5-fluorouracil (5-FU)
therapy had a significantly lower level of expression of this gene. Expanding
beyond single-gene effects on treatment response, the recent development of
12
Blumenthal
DNA microarrays permits large-scale screening of genes and the simultaneous

measurement of the expression levels of thousands of genes and raises the possibility of an unbiased, genomewide approach to understand the genetic basis
of drug response (62). Algorithms have been developed for the classification
of cell line chemosensitivity based on gene expression profiles. Using oligonucleotide microarrays, the expression levels of genes can be measured in a panel
of cancer cell lines with known chemosensitivity profiles for various chemical
compounds. Genes associated with repair processes, cell-cycle checkpoints,
apoptosis, signal transduction, and metabolism can all be studied. Microarray
analysis of baseline gene expression (63,64) and drug-induced changes in gene
expression (65) have been successfully applied to predicting chemotherapy
response. Further development of this technology will enable those responsible
for treatment planning not only to predict chemosensitivity prior to therapy, but
to answer whether the classifiers are dependent on tumor type or tumor class.
Toward that goal, chemosensitivity prediction studies are being extended beyond
cell line models to include the analysis of primary patient material.
The addition of proteomic studies to genomic studies will further facilitate
the ability to identify a priori sensitive and resistant tumors (66). In the past,
protein analysis of formalin-fixed paraffin-embedded tumor tissue or Western
blotting was used to assess prediction of therapeutic response. Techniques such
as IHC provide semiquantitative information on the level of expression of key
proteins. Similar to mRNA results, thymidylate synthase expression at the protein level has been a consistent predictor of response to 5-FU-based chemotherapy of metastatic colorectal cancer (67). Other protein markers that have been
useful for predicting chemosensitivity include ornithine decarboxylase (68),
HER-2/neu (69), the p27 cell-cycle regulatory protein (70), and Bcl-2 (71).
Other proteins that have been studied but have not been predictive include p53
and cerbB-2 (72). The use of tissue arrays allows IHC to be performed on
0.6-mm core tissue sections from a large panel of tumor samples of varying histotypes (73). The field is now expanding beyond detection and quantification of
single proteins, to include evaluation of a sizable panel of proteins by combining the tools of laser capture detection of tumor cells with silver stained twodimensional gel electrophoresis and generate a tumor phenotype (74). Much
of this effort has focused on the mass spectral identification of the thousands of
proteins that populate complex biosystems. Protein patterns can be analyzed
in hundreds of clinical samples per day utilizing this technology.
7. Future Directions
The availability of predictive in vitro and in vivo assays provides a mean of
addressing many important issues in tumor biology. Several areas of investiga-
13
tion related to chemosensitivity testing are worthy of consideration. First, work

is required to establish gene and protein profiles associated with drug sensitivity (59). Studies are needed to address whether baseline gene/protein expression
or drug-induced changes in expression are most predictive of response to treatment. Second, developing a better understanding of how to use drug combinations and schedules is essential. In vitro (75) and in vivo (76) models have
been developed to assess interactions between multiple therapeutic agents.
Toward that end, studies to determine how the use of one agent influences the
biology of the surviving tumor cells and environmental milieu, and thus affects
sensitivity to the second agent, are needed. Third, how chemosensitivity can
be modulated, such as by using compounds that affect DNA repair (77); pharmacological agents (e.g., small molecule [78]); or genetic methods such as antisense, ribozyme, or RNA interference technology to overcome drug resistance
(79) or antiapoptotic states must be determined. Fourth, the role that circadian
rhythms play in chronotherapy (80) is just beginning to receive attention. In
vivo designs to address how light/dark cycles impact the optimal time of day to
administer a particular drug to achieve maximal efficacy are under investigation. Fifth, tumor heterogeneity within a patient is important. Clinically, this is
seen when a tumor mass in one area responds to chemotherapy, while a separate site may remain stable or progress. Intrapatient response variations have
been evaluated by comparing the in vitro drug response of primary vs metastatic sites from the same patient. Thus, in vitro test results for solid tumors from
one site may not be representative for other sites within the same patient. Additional research to understand the molecular differences in chemosensitivity
between primary and secondary sites is needed. Finally, developing approaches
to individualize drug treatment as a function of tumor cell sensitivity (as has
been done in infectious disease) requires further attention.
References
1. Kaspers, G., Zwaan, C., Pieters, R., and Veerman, A. (1999) Cellular drug resistance in childhood acute myeloid leukemia: a mini-review with emphasis on cell
culture assays. Adv. Exp. Med. Biol. 457, 415421.
2. Robert, J. (1999) Chemosensitivity testing: prediction of response to anticancer
drugs using in vitro assays. Electronic J. Oncol. 2, 198210.
3. Bellamy, W. (1992) Prediction of response to drug therapy of cancer: a review of
in vitro assays. Drugs 44, 690708.
4. Gercel-Taylor, C., Ackermann, M., and Taylor, D. (2001) Evaluation of cell proliferation and cell death based assays in chemosensitivity testing. Anticancer Res.
21, 27612768.
5. Von Hoff, D., Clark, G., Stogdill, B., et al. (1983) Prospective clinical trial of a
human tumor cloning system. Cancer Res. 43, 19261936.
14
Blumenthal
6. Engblom, P., Rantanen, V., Kulmala, J., and Grenman, S. (1996) Paclitaxel and
cisplatin sensitivity of ovarian carcinoma cell lines tested with a 96-well plate
clonogenic assay. Anticancer Res. 16, 17431747.
7. Weisenthal, L., Marsden, J., Dill, P., and Macaluso, C. (1983) A novel dye exclusion method for testing in vitro chemosensitivity of human tumors. Cancer Res. 43,
749757.
8. Kern, D., Drozemuller, C., Kennedy, M., et al. (1985) Development of a miniaturized improved nucleic acid precursor incorporation assay for chemosensitivity
testing of human solid tumors. Cancer Res. 45, 54365441.
9. Sondak, V., Bertelson, C., Tanigawa, N., et al. (1984) Clinical correlations with
chemosensitivities measured in a rapid thymidine incorporation assay. Cancer Res.
46, 17251728.
10. Kobayashi, H., Higashiyami, M., Minamigawa, K., et al. (2001) Examination of in
vitro chemosensitivity test using collagen gel droplet culture method with colorimetric endpoint quantitation. Jpn. J. Cancer Res. 92, 203210.
11. Kangas, L., Gronroos, M., and Nieminen, A. (1984) Bioluminesence of cellular
ATP: a new method for evaluating cytotoxic agents in vitro. Med. Biol. 62,
338343.
12. Csoka, K., Larsson, R., Tholander, B., Gerdin, E., de la Torre, M., and Nygren, P.
(1994) Cytotoxic drug sensitivity testing of tumor cells from patients with ovarian
carcinoma using the fluorometric microculture cytotoxicity assay (FMCA).
Gynecol. Oncol. 54, 163170.
13. Rubinstein, L., Shoemaker, R., Pacell, K., Simon, R., and Tosini, S. (1990) Comparison of in vitro anticancer-drug-screening data generated with a tetrazolium
assay versus a protein assay against a diverse panel of human tumor cells lines.
J. Natl. Cancer Inst. 82, 11131118.
14. Sevin, B., Peng, Z., Perras, J., Panalver, G., and Averette, H. Application of an
ATP bioluminescence assay in human tumor chemosensitivity testing. Gynecol.
Oncol. 31, 191204.
15. Andreotti, P., Cree, I., Kurbacher, C., et al. (1995) Chemosensitivity testing of
human tumors using a microplate adenosine triphosphate luminescence assay: clinical correlation for cisplatin resistance of ovarian carcinoma. Cancer Res., 55,
52765282.
16. Steff, A., Fortin, M., Arguin, C., and Hugo, P. (2001) Detection of a decrease in
green fluorescent protein fluorescence for the monitoring of cell death: an assay
amenable to high-throughput screening technologies. Cytometry 45, 237243.
17. Vistica, D., Skehan, P., Scudiero, D., et al. (1991) Tetrazolium-based assays for
cellular viability: a critical examination of selected parameters affecting formazan
production. Cancer Res. 51, 25152520.
18. Sargent, J. and Taylor, C. Appraisal of the MTT assay as a rapid test of chemosensitivity in acute myeloid leukemia. Br. J. Cancer 60, 206210.
19. Hoffman, R. (1991) Three-dimensional histoculture: origins and applications in
cancer research. Cancer Cells 3, 8692.
15
20. Singh, B., Li, R., Xyu, L., et al. (2002) Prediction of survival in patients with head
and neck cancer using the histoculture drug response assay. Head Neck 24, 437442.
21. Meitner, P. (1991) The fluorescence cytoprint assay: a new approach to in vitro
chemosensitivity testing. Oncology 5, 7581.
22. Kern, D. and Weisenthal, L. (1990) Highly specific prediction of antineoplastic
drug resistance with an in vitro assay using supra-pharmacologic drug exposures.
J. Natl. Cancer Inst. 82, 582588.
23. Haroun, R., Clatterbuck, R., Gibbons, M., et al. (2002) Extreme drug resistance in
primary brain tumors: in vitro analysis of 64 resection specimens. J. Neurooncol.
58, 115123.
24. Fenech, M. (2000) The in vitro micronucleus technique. Mutation Res. 455, 8195.
25. Fruhauf, J. and Bosanquet, A. (1993) In vitro determination of drug response: a
discussion of clinical applications. PPO Updates 7, 121.
26. Clarke, R. (1996) Human breast cancer cell line xenografts as models of breast
cancer: the immunobiologies of recipient mice and the characteristics of several
tumorigenic lines. Breast Cancer Res. Treat. 39, 6986.
27. Hoffman, R. (1999) Orthotopic metastatic mouse models for anticancer drug discovery and evaluation: a bridge to the clinic. Invest. New Drugs 17, 343359.
28. Price, J. (1996) Metastasis from human breast cancer cell lines. Breast Cancer
Res. Treat. 39, 93102.
29. Konovalova, N., Diatchkovskaya, R., Ganieva, L., Volkova, L., Lapshin, I.,
Rudakov, B., Shaposhnikov, Y., and Shapiro, A. (1991) Subrenal capsule assay of
human tumor chemosensitivity. Neoplasma 38, 275284.
30. Tomayko, M. and Reynolds, C. (1989) Determination of subcutaneous tumor size
in athymic (nude) mice. Cancer Chemother. Pharmacol. 24, 148154.
31. Heitjan, D., Manni, A., and Santen, R. (1993) Statistical analysis of in vivo tumor
growth experiments. Cancer Res. 53, 60426050.
32. Gibbs, J., Slocum, H., Cao, S., and Rustum, Y. (1999) Image analysis for quantitation of solid tumor drug sensitivity. Int. J. Surg. Invest. 1, 133138.
33. Kubota, K. (2001) From tumor biology to clinical PET: a review of positron
emission tomography (PET) in oncology. Ann. Nucl. Med. 15, 471486.
34. Belhocine, T., Steinmetz, N., Hustinx, R., et al. (2002) Increased uptake of the
apoptosis imaging agent (99m)Tc recombinant human Annexin V in human tumors
after one course of chemotherapy as a predictor of tumor response and patient
prognosis. Clin. Cancer Res. 8, 27662774.
35. Mazurchuk, R., Glaves, D., and Raghavan, D. (1997) Magnetic resonance imaging
of response to chemotherapy in orthotopic xenografts of human bladder cancer.
Clin. Cancer Res. 3, 16351641.
36. Nakanishi, H., Mochizuki, Y., Kodera, Y., et al. (2003) Chemosensitivity of peritoneal micrometastases as evaluated using green fluorescence protein
(GFP)tagged human gastric cancer cell line. Cancer Sci. 94, 112118.
37. Tewari, K. and Manetta, A. (1999) In vitro chemosensitivity testing and mechanisms of drug resistance. Curr. Oncol. Rep. 1, 7784.
16
Blumenthal
38. el-Deiry, W. S. (1997) Role of oncogenes in resistance and killing by cancer therapeutic agents. Curr. Opin. Oncol. 9, 7987.
39. Fan, S., el-Deiry, W., Bae, I., et al. (1994) p53 gene mutations are associated with
decreased sensitivity of human lymphoma cells to DNA damaging agents. Cancer
Res. 54, 58245830.
40. Bosken, C., Wei, Q., Amos, C., and Spitz, M. (2002) An analysis of DNA repair as
a determinant of survival in patients with non-small-cell lung cancer. J. Natl.
Cancer Inst. 94, 10911099.
41. Zunino, F., Perego, P., Pilotti, S., and Pratesi, G. (1997) Role of apoptotic response
in cellular resistance to cytotoxic agents. Pharmacol. Ther. 76, 177185.
42. Wang, Y., Ashkenazi, Y., and Bachrach, U. In vitro chemosensitivity of hematological cancers: immunohistochemical detection of ornithine decarboxylase. Anticancer Drugs 10, 797805.
43. Park, K., Rha, S., Kim, C., et al. (1998) Telomerase activity and telomere lengths
in various cell lines: changes of telomerase activity can be another method for
chemosensitivity evaluation. Int. J. Oncol. 13, 489495.
44. Modrak, D., Rodriguez, M., Goldenberg, D., Lew, W., and Blumenthal, R. (2002)
Sphingomyelin enhances chemotherapy efficacy and increases apoptosis in human
colonic tumor xenografts. Int. J. Oncol. 20, 379384.
45. Epstein, R. (1990) Drug-induced DNA damage and tumor chemosensitivity.
J. Clin. Oncol. 8, 20622084.
46. Xia, F. and Powell, S. (2002) The molecular basis of radiosensitivity and
chemosensitivity in the treatment of breast cancer. Semin. Radiat. Oncol. 12,
296304.
47. Manahan, K., Taylor, D., and Gercel-Taylor, C. (2001) Clonal heterogeneity of
p53 mutations in ovarian cancer. Int. J. Oncol. 19, 387394.
48. Granjean, F., Bremaud, L., Verdier, M., Robert, J., and Ratinaud, M. (2001)
Sequential gene expression of P-glycoprotein (P-gp), multidrug resistance associated protein (MRP) and lung resistance protein: functional activity of P-gp and
MRP present in the doxorubicin-resistant human K562 cell lines. Anticancer Drugs
12, 247258.
49. Michieli, M., Damiani, D., Ermacora, A., et al. (2000) P-glycoprotein (PGP), lung
resistancerelated protein (LRP) and multidrug resistanceassociated protein
(MRP) expression in acute promyelocytic leukemia. Br. J. Haematol. 108,
703709.
50. Pallis, M., Turzanski, J., Langabeer, S., and Russell, N. (1999) Reproducible flow
cytometric methodology for measuring multidrug resistance in leukaemic blasts.
Adv. Exp. Med. Biol. 457, 7788.
51. Pall, G., Spitaler, M., Hofmann, J., Thaler, J., and Ludescher, C. (1997) Multidrug
resistance in acute leukemia: a comparison of different diagnostic methods.
Leukemia 11, 10671072.
52. Singh, N. (2000) A simple method for accurate estimation of apoptotic cells. Exp.
Cell. Res. 256, 328337.
17
53. Munshi, A., McDonnell, T., and Meyn, R. (2002) Chemotherapeutic agents
enhance TRAIL-induced apoptosis in prostate cancer cells. Cancer Chemother.
Pharmacol. 50, 4652.
54. Maciorowski, Z., Klijanienko, J., Padoy, E., et al. (2001) Comparative image and
flow cytometric TUNEL analysis of fine needle samples of breast carcinoma.
Cytometry 46, 150156.
55. Ogata, S., Okumura, K., and Taguchi, H. (2000) A simple and rapid method for the
detection of poly(ADP-ribose) by flow cytometry. Biosci. Biotechnol. Biochem.
64, 510515.
56. Nita, M., Nagawa, H., Tominago, O., et al. (1998) 5-Fluorouracil induces apoptosis in human colon cancer cell lines with modulation of Bcl-2 family proteins.
Br. J. Cancer 78, 986992.
57. Smith, D., Gao, G., Zhang, X., Wang, G., and Dou, Q. (2000) Regulation of tumor
cell apoptotic sensitivity during the cell cycle [review]. Int. J. Mol. Med. 6, 503507.
58. Sonneveld, P. (2000) Multidrug resistance in hematological malignancies. J. Intern.
Med. 247, 521534.
59. Staunton, J., Slonim, D., Coller, H., et al. (2001) Chemosensitivity prediction by
transcriptional profiling. Proc. Natl. Acad. Sci. USA 98, 10,78710,792.
60. Bao, L. and Sun, Z. (2002) Identifying genes related to drug anticancer mechanisms using support vector machine. FEBS Lett. 521, 109114.
61. McLeod, H. Individualized cancer therapy: molecular approaches to the prediction of tumor response. Expert Rev. Anticancer Ther. 2, 113119.
62. Amundson, S., Myers, T., Scudiero, D., Kitada, S., Reed, J., and Fornace, A.
(2000) An informatics approach identifying markers of chemosensitivity in human
cancer cell lines. Cancer Res. 60, 61016110.
63. Alizadeh, A., Eisen, M., and Dasvis, R. (2000) Distinct types of diffuse large
B-cell lympoma identified by gene expression profiling. Nature 403, 503511.
64. Kihara, C., Tsunoda, T., Tanaka, T., et al. (2001) Prediction of sensitivity of
esophageal tumors to adjuvant chemotherapy by cDNA microarray analysis of
gene-expression profiles. Cancer Res. 61, 64746479.
65. Sotirious, C., Powles, T., Dowsett, M., et al. (2002) Gene expression profiles
derived from fine needle aspiration correlate with response to systemic chemotherapy in breast cancer. Breast Cancer Res. 4, R3.
66. Poland, J., Schadendorf, D., Lage, H., Schnolzer, M., Celis, J., and Siha, P. (2002)
Study of therapy resistance in cancer cells with functional proteome analysis. Clin.
Chem. Lab. Med. 40, 221234.
67. Aschele, C., Debernardis, D., Casazza, S., et al. (1999) Immunohistochemical
quantitation of thymidylate synthase expression in colorectal cancer metastases
predicts for clinical outcome to fluorouracil-based chemotherapy. J. Clin. Oncol.
17, 17601770.
68. Bachrach, U. and Wang, Y. (2003) In vitro chemosensitivity testing of hematological cancer patients: detection of ornithine decarboxylase. Rec. Results Cancer Res.
161, 6270.
18
Blumenthal
69. Konecny, G., Fritz, M., Untch, M., et al. (2001) Her-2/neu overexpression and in
vitro chemosensitivity to CMF and FEC in primary breast cancer. Breast Cancer
Res. Treat. 69, 5363.
70. Yang, Q., Sakurai, T., Yoshimura, G., et al. (2000) Overexpression of p27 protein
in human breast cancer correlates with in vitro resistance to doxorubicin and mitomycin C. Anticancer Res. 20, 43194322.
71. Yang, Q., Sakurai, T., Yoshimura, G., et al. (2000) Expression of Bcl-2 but not
Bax or p53 correlates with in vitro resistance to a series of anticancer drugs in
breast carcinoma. Breast Cancer Res. Treat. 61, 211216.
72. Rozan, S., Vincent-Salomon, A., Zafrani, B., et al. (1998) No significant predictive
value of c-erbB-2 or p53 expression regarding sensitivity to primary chemotherapy
or radiotherapy in breast cancer. Int. J. Cancer 79, 2733.
73. Ginestier, C., Charafe-Jauffret, E., Bertucci, F., et al. (2002) Distinct and complementary information provided by use of tissue and DNA microarrays in the study
of breast tumor markers. Am. J. Pathol. 161, 12231233.
74. Jones, M., Krutzsch, H., Shu, H., et al. (2002) Proteomic analysis and identification of new biomarkers and therapeutic targets for invasive ovarian cancer. Proteomics 2, 7684.
75. Konecny, G., Untch, M., Slamon, D., et al. (2001) Drug interactions and cytotoxic
effects of paclitaxel in combination with carboplatin, epirubicin, gemcitabine, or
vinorelbine in breast cancer cell lines and tumor samples. Breast Cancer Res.
Treat. 67, 223233.
76. Lopez, A., Pegram, M., Slamon, D., and Landaw, E. (1999) A model-based
approach for assessing in vivo combination therapy interactions. Proc. Natl. Acad.
Sci. USA 96, 13,02313,028.
77. Heim, M., Eberhardt, W., Seeber, S., and Muller, M. (2000) Differential modulation of chemosensitivity to alkylating agents and platinum compounds by DNA
repair modulators in human lung cancer cell lines. J. Cancer Res. Clin. Oncol.
126, 198204.
78. Kondratov, R., Komarov, P., Becker, Y., Ewenson, A., and Gudkov, A. (2001)
Small molecules that dramatically alter multidrug resistance phenotype by modulating the substrate specificity of P-glycoprotein. Proc. Natl. Acad. Sci. USA 98,
14,07814,083.
79. Henewisch-Becker, S. (1996) MDR1 reversal: criteria for clinical trials designed to
overcome the multidrug resistance phenotype. Leukemia 10, S32S38.
80. Levi, F., Giacchetti, S., Zidani, R., et al. (2001) Chronotherapy of colorectal cancer
metastases. Hepatogastroenterology 48, 320322.
II
IN VITRO MEASURES
OF
CHEMOSENSITIVITY
2
Clonogenic Cell Survival Assay
Anupama Munshi, Marvette Hobbs, and Raymond E. Meyn
Summary
The clonogenic cell survival assay determines the ability of a cell to proliferate indefinitely,
thereby retaining its reproductive ability to form a large colony or a clone. This cell is then said
to be clonogenic. A cell survival curve is therefore defined as a relationship between the dose
of the agent used to produce an insult and the fraction of cells retaining their ability to reproduce.
Although clonogenic cell survival assays were initially described for studying the effects of radiation on cells and have played an essential role in radiobiology, they are now widely used to
examine the effects of agents with potential applications in the clinic. These include, in addition
to ionizing radiation, chemotherapy agents such as etoposide and cisplatin, antiangiogenic agents
such as endostatin and angiostatin, and cytokines and their receptors, either alone or in combination therapy. Survival curves have been generated for many established cell lines growing in
culture. One can use cell lines from various origins including humans and rodents; these cells can
be neoplastic or normal. Because survival curves have wide application in evaluating the reproductive integrity of different cells, we provide here the steps involved in setting up a typical
experiment using an established cell line in culture.
Key Words
Survival curve; cell survival; plating efficiency; radiation.
1. Introduction
Clonogenic cell survival is a basic tool that was described in the 1950s for
the study of radiation effects. Much of the information that has been generated
on the effect of radiation on mammalian cells has been obtained from clonogenic cell survival assays.
Various mechanisms have been described for cell death; however, loss
of reproductive integrity and the inability to proliferate indefinitely are the
most common features. Therefore, a cell that retains its ability to synthesize proteins and DNA and go through one or two mitoses, but is unable to divide and
21
22
Munshi et al.
produce a large number of progeny is considered dead. This is very commonly

referred to as loss of reproductive integrity or reproductive death and is the
end point measured with cells in culture. On the other hand, a cell that is not
reproductively dead and has retained the capacity to divide and proliferate
indefinitely can produce a large clone or a large colony of cells and is then
referred to as clonogenic. A cell survival curve describes a relationship
between the insult-producing agent and the proportion of cells that survive.
The ability of a single cell to grow into a large colony that can be visualized
with the naked eye is proof that it has retained its capacity to reproduce. The
loss of this ability as a function of dose of radiation or chemotherapy agent is
described by the dose-survival curve. Most laboratories now extensively use
established cell lines for studying the effects of various agents either alone or
in combination. Therefore, the aim of this chapter is to go through the steps
involved in setting up a typical clonogenic cell survival experiment using established cells lines growing as monolayer cultures.
In brief, cells from an actively growing stock culture in monolayer are prepared in a suspension by the use of trypsin, which causes the cells to detach
from the substratum. The number of cells per milliliter in this suspension is
then counted using a hemocytometer or a Coulter counter. From this stock culture, if 50 cells are seeded into a dish, e.g., and the dish is incubated for approx
2 wk, each single cell divides many times and forms a colony that is easily
visible with the naked eye, especially if it is fixed and stained (the steps
involved in this process are briefly outlined in Fig. 1). All the cells that make
up the colony are the progeny of a single cell. For the 50 cells seeded into
the dish, the number of colonies counted may be anywhere from 0 to 50. One
would ideally expect the number to be 50, but that is rarely the case for several
possible reasons, including suboptimal growth medium, errors in counting the
number of cells initially plated, and the loss of cells by trypsinization and general handling. The term plating efficiency (PE) indicates the percentage of cells
seeded into a dish that finally grow to form a colony. Therefore, in the previous example, if there are 25 colonies in the dish, then the PE becomes 50%. If
a parallel dish is seeded with cells exposed to a dose of 6 Gy of gamma rays
and incubated for approx 2 wk before being fixed and stained, then the following may be observed:
1. Some cells may remain single, not divide, and, in some cases, may show evidence of nuclear deterioration as they die by apoptosis. These cells would be
scored as dead.
2. Some cells may go through one or two divisions and form small colonies of
just a few cells. These cells would be scored as dead.
3. Some cells may form large colonies, indicating that the cells have survived the
treatment and have retained the ability to reproduce indefinitely.
23
Fig. 1. Schematic representation of steps involved in setting up a clonogenic cell survival assay.
2. Materials
2.1. Preparation of Cell Lines Prior to Setting Up Clonogenic Assays
1. Cell lines that need to be tested for their ability to form colonies.
2. Complete growth medium as recommended by the manufacturer typically containing 10% fetal bovine serum plus antibiotics (penicillinstreptomycin) and glutamine. For every 500 mL of medium, add 55 mL of serum, 5 mL of 200 mM
L-glutamine, and 5 mL of 10,000 U/mL penicillinstreptomycin solution. Medium
should be stored at 4C but warmed to 37C prior to use.
3. Trypsin-EDTA, to make single-cell suspensions from monolayer cultures. Store
at 4C.
4. Plasticware, for carrying out tissue culture including flasks (T-25 and T-75);
100-mm dishes; and 5-, 10-, and 25-mL pipets.
5. Micropipets and corresponding tips.
6. 70% ethanol, for wiping the surface of the hood as well as the surface of all
medium bottles prior to bringing them into the hood.
7. Cidecon (detergent disinfectant with bactericidal and virucidal properties), to wipe
the surface of the shelf on which the dishes will be incubated.
8. Phosphate-buffered saline (PBS) (calcium magnesium free). Store at 4C.
9. Isoton II (diluent for counting cells using a Coulter counter). Store at room
temperature.
24
Munshi et al.
2.2 Staining of Plates

1. 0.5% Gentian Violet (made up in methanol). Store at room temperature in a dark
bottle. Do not pour it down the sink.
3. Methods
3.1. Cells Growing in Monolayers or Attached Cells
The procedure that we outline in this chapter is a basic protocol for setting
up radiation clonogenic assays. However, this protocol can be modified to test
the effect of different agents on the cell line of interest, either alone or in combination. These could include gene therapy vectors, chemotherapy agents, tyrosine kinase inhibitors, antiangiogenic agentsbasically any agent that has to be
tested for its ability to affect reproductive cell death.
3.2. Preparation of Cell Lines Prior to Setting Up
Clonogenic Cell Survival
1. Label six T-25 flasks in preparation for setting each flask with a known number
of cells as 0, 2, 4, and 6 Gy (depending on the experiment) for the various doses
of radiation to be given. Add 5 mL of growth medium to the flasks and keep them
aside in a hood.
2. Trypsinize the stock flask of cells containing the cells that have to be tested for
their radiosensitivity. Make sure that the cells are in single-cell suspension and
obtain an accurate cell count. We use a Coulter counter to obtain a cell count. If
a Coulter counter is not available, cells can be counted using a hemocytometer.
Using a Pipettman, add 250,000 cells (the cell number can vary depending on the
cell type) to the 5 mL of medium in each T-25 flask. Shake gently to distribute the
cells evenly.
3. Place the flasks in a 37C incubator set at 5% CO2 and be sure to leave the cap
one thread loose so as to allow CO2 exchange (see Note 1).
4. Allow the cells to settle and attach as a monolayer.
3.3. Irradiation of Flasks and Performance of Plating Experiment

for Clonogenic Assay
1. Prepare the hood and clean an incubator shelf. Because these cells are going to be
left untouched in an incubator for up to 2 wk, and the possibility of contamination
is high, clean the shelf thoroughly with Cidecon and 70% ethanol. Keep the
cleaned shelf in the hood. Make sure that more than one bottle of complete
medium is available; this experiment may require >500 mL of medium.
2. Prepare the 100-mm dishes and 15-mL tubes in advance. One will need 100-mm
dishes in triplicate, and two cell numbers will be plated for each dose of radiation.
Because cells will be exposed to six doses of radiation, 36 dishes will be needed.
Label the bottom of each dish as the lid, for the dishes will be loose during staining, and the bottom is where the colonies will form, which is what will actually
3.
4.
5.
6.
7.
8.
9.
10.
25
be counted. Label the first set of triplicate dishes as 0 Gy A and 0 Gy B (A and

B for the two cell numbers to be used). Repeat this for all dose levels: 2, 4, 6, 8,
and 10 Gy. Place 10 mL of complete medium in each dish. Place all dishes,
stacked in threes, on the incubator shelf and put back in the incubator until ready
for plating.
Place 18, 15-mL tubes on a clean rack three deep. Label the first tube 0Gy, 11;
the next one 0 Gy 110; and the next 0 Gy 1100. Repeat this for all dose levels.
Place 4.5 mL of complete medium in the back two tubes but not in the 11 dilution tubes.
Put the flasks on ice. Clean a rectangular tub and fill it halfway with ice. Remove
the flasks from the incubator and close the caps tightly. Place the flasks on ice,
and insert half depth into the ice. Tilt the flasks to the bottom so that the medium
does not rest against the cap. Start a timer for 20 min.
While waiting, prepare the Coulter counting vials. Label each as 0, 2, 4, and 6 Gy.
Place 9.9 mL of Isoton (to be used for counting cells if using a Coulter counter)
into each vial. Place Isoton in a control vial and run through the counter to get a
background measurement; repeat until a satisfactory low background is obtained.
Make sure that the Coulter is set to the appropriate size parameter for the cell line.
When the 20-min time is up, irradiate the flasks using an appropriate irradiator
according to the desired dose.
Return to the hood, keeping the flasks on ice outside the hood. Begin the
trypsinization procedure for each flask. Start with the 0-Gy flask. Aspirate the
medium, rinse the cells gently with PBS, and then trypsinize. Place the harvested
cells in the 15-mL tube labeled 0 Gy, 11. Make sure that you have a good singlecell suspension. From this cell suspension, take 100 L and place in the appropriate counting vial for 0 Gy. Now trypsinize the next flask. While the flask is on
the warming tray, count the previous counting vial and record the counts on a
dilution sheet (see Fig. 2 for a setup of a dilution sheet that we commonly use in
the laboratory). This expedites the experiments and lessens the chance of cell divisions taking place unequally during the time of trypsinization and counting.
Continue this procedure until all the flasks have been trypsinized and counted.
There should now be cells in each of the 11 dilution tubes, with a known number
of cells/milliliter, all documented on the dilution sheet.
Perform serial dilutions for each radiation dose so that the desired number of cells
will be obtained by adding between 100 and 1000 L of volume to the dishes. If
the number of cells needed requires a volume exceeding 1000 L, use a more
concentrated dilution. Plate a number of cells consistent with obtaining a colony
count of 50100. This may require only 100 cells for the control plate whereas at
6 Gy this may require 4000 cells or more. Remember that the larger the insult to
the flask (i.e., increasing radiation dose or increasing drug concentration if using
chemotherapy agents), the lower the plating efficiency and the more cells are
needed to obtain the desired colony count (see Note 2).
It is important to resuspend the cell pellet, which has probably settled to the bottom
of the tube by the time all the flasks have been counted. Place 0.5 mL of the 11
26
Munshi et al.
Fig. 2. (A) Setup of dilution sheet used during clonogenic cell survival assays; (B)
survival curve plotted using hypothetical numbers derived from dilution sheet
11.
12.
13.
14.
15.
dilution into the 110 dilution tube (containing 4.5 mL of medium). Similarly, mix
each dilution well before aliquoting 0.5 mL to the next-higher dilution tube.
When the final dilution tube for each dose has been made, mark the top of each
tube with an X and place the other tubes that are not needed outside the hood. This
will minimize the chance of grabbing the wrong tube at the time of plating.
Place the calculated volume of cell solution slowly into each appropriate 100-mm
dish. Start with the bottom dish and work up. Place the solution on the medium
drop by drop, spreading the drops evenly over the entire surface area to prevent
clumping and overlap of colonies.
After placing the solution in the three 100-mm dishes, rock the plates northsouth
and eastwest to distribute the cells evenly. Avoid swirling the solution, which
allows cells to group at the sides, making counting difficult.
When all of the dishes have been plated, return the shelf to the incubator. Label
the inside door of the incubator with the name of the experiment, initials, date of
the project in and date expected to come out (see Note 3).
As a rule of thumb, incubate for 1012 d for cells with 13 h or less of generation
time, approx 14 d for 14 h or more. We routinely set up clonogenic assays for
these periods of time without any contamination.
27
3.4. Staining of Plates

Staining plates with Gentian Violet is easy, but care must be taken not to
get it on ones clothes, because it is difficult to remove. It is suggested that a
laboratory coat and double glove be used for the staining procedure.
1. Take the shelf from the incubator and place it by the sink. Empty the medium
from six plates into a flask with bleach in the sink. Place 0.5% Gentian Violet
onto each of the six plates. Gentian Violet is diluted from a stock with methanol.
2. Run a gentle stream of lukewarm water into a pan in the sink. Take the stained
plates and transfer the stain from them into an additional six plates. Rinse the
stained plates upside down in the pan to prevent the colonies from loosening and
washing off.
3. Continue steps 1 and 2 until all the plates have been stained. Let the plates air-dry
overnight; they will be ready for counting the next day (see Note 4).
3.5. Counting of Colonies

1. Take the air-dried colonies and count the colonies in each dish to obtain the plating efficiency. We use a dissecting scope to view the colonies under a magnified
field. A cluster of blue-staining cells is considered a colony if it comprises at least
2550 cells. However, it is important to keep the cutoff constant so that there are no
variations introduced between experiments. Count the colonies and note the numbers for both the A and B dilutions in a chart (prepare one similar to the one shown
in Fig. 2).
2. Average the three colony counts for each dilution A and B and divide the mean by
the number of cells plated. This will give the PE:
Number of colonies counted
PE = 100
Number of cells plated
The PE of some cells may be close to 8090%. This is especially true for human
tumor cells of various origins. However, the PE of normal human fibroblasts is
usually very low (ranging from as low as 1 to 1215%).
3. Following determination of PE, calculate the fraction of cells surviving a given
treatment. First, normalize all the plating efficiencies of the treated samples to
that of the control unirradiated plates, considering that to be 100%. The surviving
fraction (SF) is determined by dividing the PE of the treated cells by the PE of the
controls, and then multiplying by 100:
PE of treated sample
SF = 100
PE of control
Plot the data on an Excel spreadsheet with the dose of radiation on the x-axis and
survival on the y-axis.
28
Munshi et al.
4. Notes
1. The CO2 requirement may vary depending on the cell type. If the flasks have lids
with a membrane that allows gas exchange, there is no need to loosen the cap.
2. Note that you can use prior cell survival data, if available, to extrapolate exactly
the number of cells to plate for a desired colony count.
3. If possible, it is a good idea to dedicate an incubator to clonogenic cell survival.
This avoids unnecessary bumping of colonies as people open and close the door
to the incubator. As the colonies grow, bumping the incubator or shelf can cause
the cells to shed and settle as new colonies, thereby leading to an increase in the
colony count and erroneous results.
4. We usually do not pour the used stain down the sink but, instead, collect it in a
bottle dedicated to spent stain.
References
1. Elkind, M. E. and Whitmore, G. F. (1967) In vitro survival curves, in The Radiobiology of Cultured Mammalian Cells, Gordon and Breach, New York, NY,
pp. 53115.
2. Hall, E. J. (2000) Cell survival curves, in Radiobiology for the Radiologist, 5th
ed. (Hall, E. J., ed.), Philadelphia, J.B. Lippincott, pp. 3250.
3. Elkind, M. M and Sutton, H. (1960) Radiation response of mammalian cells grown
in culture. I. Repair of X-ray damage in surviving Chinese hamster cells. Radiat.
Res. 13, 556593.
4. Elkind, M. M., Han, A., and Volz, K. W. (1963) Radiation response of mammalian
cells grown in culture. IV. Dose dependence of division delay and postirradiation
growth of surviving and non surviving Chinese Hamster cells. J. Natl. Cancer.
Inst. 30, 705721.
5. Sinclair, W. K. and Stroud, A. N. (1962) Postirradiation changes in growth, chromosome number and survival properties of cultures of Chinese hamster cells
[abstract]. Radiat. Res. 16, 590.
6. Puck, T. T. and Marcus, P. I. (1956) Action of X-rays on mammalian cells. J. Exp.
Med. 103, 653666.
7. Sinclair, W. K. (1964) X-ray induced heritable damage (small-colony formation) in
cultured mammalian cells. Radiat. Res. 21, 584611.
8. Whitfield, J. F. and Rixon, R. H. (!960) Radiation resistant derivatives of L strain
mouse cells. Exp. Cell Res. 19, 531538.
9. Barendsen, G. W., Beusker, T. L. J., Vergroesen, A. J., and Budke, L. (1960)
Effects of different ionizing radiations on human cells in tissue culture. II. Biological experiments. Radiat. Res. 13, 841849.
10. Alper, T., Gillies, N. E., and Elkind, M. M. (1960) The sigmoid survival curve in
radiobiology. Nature 186, 10621063.
3
High-Sensitivity Cytotoxicity Assays
for Nonadherent Cells
M. Jules Mattes
Summary
High-sensitivity cytotoxicity assays refer to assays that can detect high levels of cell kill, to
many powers of 10, and that can detect, ideally, a single remaining viable cell. Two such assays
are described here, which have been used with Raji B-lymphoma cells, and are applicable to
other nonadherent target cells. The first is a cell-counting assay, performed over a 3-wk period,
which provides a simple, reliable, and sensitive assay of cytotoxicity. By determining the time
required for 16-fold multiplication, the apparent fraction surviving can be calculated. This assay
does not correct for treatment-induced delays in cell division and is dependent on maintaining the
cells in exponential growth. The second assay measures colony-forming units using a limiting
dilution method. Feeder cells are required to obtain a high cloning efficiency. Each dilution is
plated in 48 wells of a 96-well plate, and positive wells are scored rapidly, by eye, after two wk.
Key Words
Cytotoxicity; limiting dilution; B-cell lymphoma.
1. Introduction
A high-sensitivity cytotoxicity assay refers to an assay that can detect high
levels of cell kill, to many powers of 10. Such assays can detect a single remaining viable cell, or at least a small number of viable cells. They are important in
cancer research, because high levels of cell kill are required in order to have a
significant therapeutic effect; a cell kill of 90%, e.g., would be expected to prolong survival for only a few days for rapidly growing tumor cells. In this chapter, I describe two assays that have been used to measure cytotoxicity with
radiolabeled antibodies (Abs). These assays can be used to monitor the cytotoxicity of any type of toxic agent. The first is based on cell counts; because
the counts are continued for 21 d, a single viable cell can be readily detected.
The advantages of this assay are its simplicity and its reliability, but it might be
29
30
Mattes
considered less accurate than a limiting dilution clonogenic assay, which is the
second assay described.
Two factors can be considered to limit the value of the cell-counting assay.
First, the assay does not consider growth delay that may result from sublethal
treatment. For this reason, it may overestimate the fraction of cells that are
killed. However, this does not represent a lack of accuracy, only a problem in
interpretation, and the problem is solved by recognizing that the end point is
not fraction surviving but, rather, apparent fraction surviving. Second, and
more important, is the fact that the calculations assume that the cells are dividing at a constant rate. In fact, doubling times, at least with the cell lines my laboratory has used, are sensitive to small variations in the culture medium, and
only very dilute cells truly grow exponentially. As the cell density increases,
growth rate decreases. Therefore, the calculated time interval required for 16fold multiplication of the cells, which is the basis for calculations, depends, to
some extent, on variables such as the day of the week that the experiment was
started, because for experiments started on Monday, cells counts are obtained
on d 2, 4, and 7, and for experiments started on Tuesday, cell counts are
obtained on d 2 and 6, and for experiments started on Wednesday, cell counts
are obtained on d 2 and 5. Since control cells multiply 16-fold in 4 to 5 d,
depending on the experiment, and are relatively dense at that time, the growth
rate is often significantly lower by d 5. Therefore, this factor will affect the
calculated doubling time, which will, in turn, affect the calculation of the apparent fraction surviving. The calculated doubling time of Raji cells using the
method described herein varied from 20 to 28 h; in other experiments, in which
the cell density was very low, doubling times of 14.4 h have been observed.
Another variable is that killed cells (meaning cells that are unable to divide) are
often still metabolically active and may become markedly enlarged. Such metabolic activity will affect the growth rate of the remaining viable cells. Both of
these problems could be circumvented by maintaining the cells at a lower concentration and by counting them more frequently, but both of these changes
would diminish the convenience of this assay. For these reasons, the limiting
dilution assay was developed. Although the limiting dilution assay provides
greater accuracy, it does require considerably more work than the cell-counting
assay, and it should be emphasized that the cell-counting assay is reliable and
useful for many purposes.
2. Materials
1. Raji human B-lymphoma cell line (American Type Culture Collection, Rockville,
MD).
2. RPMI-1640 medium with 12.5% fetal calf serum, supplemented with penicillin,
streptomycin, glutamine, and pyruvate (Life Technologies, Grand Island, NY).
31
3. Phosphate-buffered saline (PBS) (Dulbeccos PBS without calcium and magnesium, cat. no. D5652; Sigma, St. Louis, MO).
4. Trypan blue (Eastman Kodak, Rochester, NY).
5. Mitomycin C (M4287; Sigma).
6. Hemacytometer, Bright-line (cat. no. 15170-172; VWR, West Chester, PA).
7. Plate-reading mirror (Bel-Art, Pequannock, NJ).
8. Sterile syringe filter, 0.2-m pore size, Corning cat. no. 431219 (purchased from
VWR).
3. Methods
3.1. General Methods
The cells are split every 3 to 4 d, generally 110 or 120, and thus are under
conditions of exponential growth. Cells are tested routinely for mycoplasma
contamination with the Mycotect Assay (Life Technologies). Abs used include
Abs to CD74, CD20, and major histocompatibility complex class II and have
been described previously (1). Radiolabeling with 125I, 131I, 111In, 67Ga, or 90Y
was also described previously (2). Ab solutions are sterilized by passage
through a 0.2-m-pore-size sterile syringe filter.
3.2. Cell Counting Assay (3)
1. Pellet 4.5 106 cells (600g, 5 min) and resuspend in 6 mL of tissue culture
medium.
2. Add 1.1 mL of this suspension to 2.2 mL of radiolabeled Ab at the desired concentration. The Ab concentrations used in our studies were usually based on the
concentration of radioactivity (microcuries per milliliter), but never exceeded
5 g/mL, a near-saturating concentration. Generally, three-fold dilutions of Ab are
tested, and four Ab concentrations are tested in each experiment. Note that the Ab
is diluted 23 due to the addition of the cell suspension. A control tube has no Ab.
3. Plate the cells from each tube in two wells of a 24-well plate, with 1.5 mL/well.
The wells used are dispersed in the plate (leaving empty wells between the wells
that are used), to reduce any chance of crossfire of radiation between the adjacent well; however, no significant crossfire between adjacent wells was detected
in experiments using 131I, intentionally designed to detect it.
4. Count viable cells in the remainder of the control sample after plating; the nominal cell concentration is 2.5 105 cells/mL. Use a twofold dilution (using 50 L
of the cell suspension) with Trypan blue solution (0.1% [w/v] Trypan blue in
PBS), which stains dead cells, for counting in a hemacytometer.
5. Count viable cell counts at d 2. Suspend the cells by repipetting with a Pasteur
pipet, and dilute 30 L with an equal volume of Trypan blue solution.
6. Transfer the entire contents of each well into a T30 flask containing 20 mL of
medium. This transfer is required in order to maintain the cells in exponential
growth. Thus, Ab is kept in the medium for the duration of the experiment, but it
is diluted approx 14-fold on d 2, so most of the uptake is in the first 2 d.
32
Mattes
7. Count viable cells every 25 d thereafter, out to 21 d or until the cells have multiplied 16-fold. At each time point, count one of the two duplicates and alternate
the particular sample counted in subsequent counts, to ensure that the duplicates
behave similarly. This has always been the case, except for some samples in which
near-complete kill was obtained (see example in ref. 3), in which case variation
between the duplicates would be expected to occur sometimes. For the cell count,
count either 100 cells or all nine large squares on the hemacytometer (for cases in
which the cell count is low). In certain cases, in which partial toxicity occurs, the
medium can turn yellowish and cell growth can slow before 16-fold multiplication
is attained. In such cases, dilute an aliquot of cells (usually 4 mL) fivefold in
order to maintain the cells in exponential growth (see Note 1).
8. Calculate the apparent percentage cell kill from the growth curves. This calculation does not take into account any delay in cell division resulting from irradiation. Such delays in division are known to occur in many cases (4,5), so the
calculation may overestimate the percentage kill; therefore, the calculated value is
designated the apparent cell kill. More specifically, the time required for 16-fold
cell multiplication is determined in control wells and in treated wells. This time
interval is calculated by interpolation from the two points on either side of 16-fold
multiplication, using a semilog graph of cell number vs days. The value from
control wells, in each experiment, is used to calculate the doubling time. The time
required for treated cells to multiply 16-fold is expressed in doubling times and is
designated TR (time required). The fraction surviving = 1/2(TR4) (see Note 2).
Representative results are given in Fig. 1, which shows toxicity with 90Y-labeled
Abs (2). Figure 1A shows the growth curves, and Fig. 1B shows the graph of the
calculated fraction surviving as a function of the initial microcuries per milliliter
in the medium.
3.3. Limiting Dilution Clonogenic Assay (2) (see Note 3)

Cells are incubated for 2 d with serial dilutions of radiolabeled Ab, exactly
as decribed above. Typically three, twofold dilutions of each Ab are tested. The
range of concentrations tested is selected to include the concentration at which
the cell kill is approx 9999.9% (based on the cell-counting assay described in
Subheading 3.2.). This level of killing is selected to allow accurate comparison
of potent cytotoxic Abs. Cells from one well of a 24-well plate are used for the
clonogenic assay.
1. Count viable cells in one well of each group, using an aliquot of 30 L.
2. Dilute the remaining cells with 17.3 mL of tisue culture medium.
3. Prepare fourfold serial dilutions by transferring 4.7 mL into 14.1 mL of medium;
make a total of eight dilutions.
4. Add to each dilution 0.9 mL of feeder cells. This was found to markedly
improve the cloning efficiency. The feeder cells are prepared as described in Subheading 3.4. A control tube has only feeder cells; there should be no growth of
only the feeder cells.
33
Fig. 1. (A) Toxicity of 90Y-LL1 (anti-CD74) for Raji cells (solid symbols) in comparison with the nonreactive control Ab MN-14 labeled in the same way (open symbols). Cells were incubated for 2 d with radiolabeled Ab at a starting concentration
of 20 Ci/mL (circles), 10 Ci/mL (squares), 5 Ci/mL (triangles), or 2.5 Ci/mL
(inverted triangles). The growth rate of control, untreated cells is also shown (dotted
line without symbols). The data shown are calculated from cell counts obtained at various times and are representative of two experiments, each done in duplicate. Cells
treated with the highest concentration of LL1 were 100% killed, since no viable cells
were detected after d 6, and growth of a single viable cell would be readily detected in
22 d. (B) The fraction surviving was calculated from the growth curves and plotted vs
the initial microcuries per milliliter for LL1 () or the nonreactive control Ab ().
One hundred percent killing cannot be shown on the exponential y-axis, but the nexthigher concentration of 90Y-LL1, 20 Ci/mL, produced 100% killing. The results shown
are representative of two experiments, each done in duplicate. Other experimental
details were described previously (2). (Reprinted by permission of the Society of
Nuclear Medicine from ref. 2.)
5. Plate each dilution of treated cells plus feeder cells in 48 wells of a 96-well plate,
with 0.2 mL/well. Therefore, approx 104 feeder cells are plated per well (see Note 4).
6. Incubate for 14 d at 37C in a humid incubator with 5% CO2.
7. Using the plate reader, count the number of wells with growing clones. These
wells are easily recognized by their yellow color and by the large, dense colony
34
Mattes
or colonies of cells that are seen, by eye, at the bottom of the well. Control clones
are large and countable after 12 d, but the irradiated cells grow more slowly, so
two additional days are allotted (see Note 5).
8. Calculate the number of colony-forming units per milliliter. This can be done in
various ways. In our published studies to date, we calculated the average number
of cells per well as ln(FN), in which FN = the fraction of wells that are negative
(6,7). Dilutions at which 1090% of the wells were negative were used for calculations. By multiplying by the dilution factor, and by a factor of 1.33 (since
one-fourth of the total cells was used for the serial dilutions), the number of clones
in the original well is calculated. If more than one dilution gives results within the
usable range (1090% of the wells negative), then the mean of the individual
values is used. A better method of calculation is via the QUALITY program that
is on-line at http://ubik.microbiol.washington.edu/computing/ (using the Java
applet). This program gives the concentration of colony-forming units per 0.2 mL
(the volume plated per well) and standard deviation using the 2 method of
Taswell (8). The difference between the two methods of calculation, in our experiments, is minor (see Note 6).
9. Calculate the cloning efficiency of the control, untreated cells from the cell count
at d 2 and the calculated number of colony-forming units. This value ranged from
35 to 94% in our experiments.
3.4. Preparation of Feeder Cells for Limiting Dilution Assay

1. Prepare 2.3 107 Raji cells in 2.3 mL of tisue culture medium (see Note 7).
2. Add 0.23 mL of mitomycin C at 0.5 mg/mL in 10 mM sodium phosphate buffer,
pH 7.6. To prepare the mitomycin C (1.0 mg/vial), dissolve in 2.0 mL of 10 mM
sodium phosphate buffer, pH 7.6, and freeze aliquots of 0.23 mL at 70C. Use
one tube for each experiment.
3. Incubate for 45 min at 37C, with mixing every 15 min.
4. Pellet the cells and wash three times with 10 mL of tissue culture medium. Then
suspend in 26.6 mL.
4. Notes
1. The fact that the cells are examined under a microscope can be considered an
advantage of the cell-counting assay, because additional information is obtained.
Heavily irradiated cells, which subsequently die, become very large. This was
unmistakable at d 5 and was also evident at d 2. Hence, by noting the size of the
cells, the level of toxicity could be reliably estimated at d 5, even though the
assay would not be completed for another 11 d. However, the need for cell counts
might also be considered a disadvantage, because this is time-consuming. As an
alternative, it is possible to use any type of automated cell-counting method. Some
of these methods would also be able to detect the increase in the size of the cells.
However, visual counting is best able to count samples in which considerable cell
2.
3.
4.
5.
35
clustering occurs; although it is simple to count small clusters (of 5 cells) by

eye, none of the automated counting methods can do this. This is a significant
factor, because most of the B-lymphoma cell lines that we have used, e.g., normally form clusters, which may be large. Although such clusters are readily dispersed by repipetting, they also may re-form rapidly, and it is difficult to exclude
the possibility that clusters may be present.
Calculations for the cell-counting assay are correct only if the cells are maintained in exponential growth and if the treatment does not affect the growth rate.
The latter is probably not correct, because it is known that radiation induces
growth delay (4,5). Therefore, the level of toxicity observed is the apparent cell
kill. For this reason, the calculated level of cell kill is sometimes higher than is
possible. For example, a fraction surviving of 108 or 107 has sometimes been calculated in an assay in which only 5 105 cells were used. This can be attributed
to growth delay. The selection of 16-fold multiplication as the end point is arbitrary but was chosen for practical and theoretical reasons. A few cell divisions
might, in some case, be compatible with eventual cell death, but if the cells can
multiply 16-fold (four divisions), then it seems reasonable to conclude that the
cells can multiply indefinitely. To extend the assay to greater than 16-fold multiplication would require more medium and larger volumes, in order to keep the
cells in exponential growth.
The limiting dilution clonogenic assay was developed to provide greater accuracy
in the measure of the fraction surviving. For example, we wanted to determine the
number of radioactive decays required for a particular level of cell kill, and to
compare this value for different radionuclides. Since the actual difference between
some of the radionuclides tested was small, this comparison requires a high level
of accuracy. Representative results are shown in Fig. 2. When the same experiments were performed using the cell-counting assay, which was done previously,
the variation among individual, identical experiments was greater. This difference
appeared to be due to variation in the calculated fraction surviving, rather than to
variation in the decays per cell. These data, then, support our contention that the
limiting dilution assay provides greater accuracy than the cell-counting assay.
The number of dilutions plated can be reduced somewhat, since those dilutions at
which all or none of the wells are positive are not important in the calculations.
Thus, for the control cells, it is necessary to plate only dilutions no. 58. For the
treated cells, it is sufficient to plate seven dilutions, since low levels of cell kill are
of little interest.
Nonadherent cells such as Raji form dense colonies in the wells, which are readily
identified by eye. Moreover, the color of the medium distinguishes wells containing clones, because it changes from pink to yellow. Therefore, wells with growing
clones are readily counted, in a few minutes. This is in contrast to results that would
be obtained with most adherent target cells: these cells form monolayers, which
are not evident by eye, and which often do not turn the medium yellow. Thus, doing
this assay with adherent cells would require either microscopic observation of each
well or some type of staining procedure to make the cells more visible.
36
Mattes
Fig. 2. Relationship between fraction surviving and total disintegrations per cell.
Raji cells were incubated with radiolabeled LL1 (anti-CD74) for 2 d, and the fraction
surviving was determined by a limiting dilution clonogenic assay. Cell-bound counts
per minute were determined at d 1, 2, 3, and 6, or d 1, 2, and 5, and the cumulative disintegrations per initial cell number were calculated. The results are shown for 90Y (),
131
I (), 125I (), 111In (), and 67Ga (). Note that open symbols represent -particle
emitters, and solid symbols represent Auger electron emitters. The results shown are
representative of two experiments performed with each radiolabel. Other experimental
details were described previously (2). (Reprinted by permission of the Society of
Nuclear Medicine from ref. 2.)
6. Statistical analysis of limiting dilution assays is relatively complex. In our published data, we calculated the concentration of colony-forming units per milliliter
using the formula given under Subheading 3.3., step 8. The means and standard
deviations of the fraction surviving were calculated from replicate experiments,
and statistical comparisons between two treatment groups could then utilize students t-test. However, there are more sophisticated methods of calculation and
error analysis (611). Although we are not aware of commercial software that
performs these calculations, a program on the Internet is available, as described
under Subheading 3.3., step 8. This program uses the 2 method of Taswell (8).
Although it has been argued that the maximum likelihood method is optimal (11),
the difference between these two methods is small (11), and we are not aware of
readily available software that utilizes the maximum likelihood method. A calculation method that can more readily be done manually, because it does not require
iterative calculations, is the weighted mean method of Taswell (8).
7. The feeder cells that we have used are the same cells used as targets but treated
with mitomycin. Cloning efficiency was approx 10-fold lower in the absence of
feeder cells. Since our cloning efficiency with Raji cells was usually >50%, we
37
can conclude that the feeder cells used were adequate. However, different target
cells may require different types (or different amounts) of feeder cells. Although
it is often convenient to use the same cell line as both target and feeder, this may
not always be optimal. Some target cells may not require feeder cells at all. Such
determinations can only be made by trial and error.
Acknowledgments
We are grateful to Rosana B. Michel for technical support and to Dr. Andre
Rogatko for assistance with the statistics. This work was supported in part by
USPHS grant CA87059 and USDOE grant ER63191.
References
1. Ong, G. L., Elsamra, S. E., Goldenberg, D. M., and Mattes, M. J. (2001) Singlecell cytotoxicity with radiolabeled antibodies. Clin. Cancer Res. 7, 192201.
2. Govindan, S. V., Goldenberg, D. M., Elsamra, S. E., Griffiths, G. L., Ong, G. L.,
Brechbiel, M. W., Burton, J., Sgouros, G., and Mattes, M. J. (2000) Radionuclides linked to a CD74 antibody as therapeutic agents for B-cell lymphoma: comparison of Auger electron emitters with beta-particle emitters. J. Nucl. Med. 41,
20892097.
3. Griffiths, G. L., Govindan, S. V., Sgouros, G., Ong, G. L., Goldenberg, D. M., and
Mattes, M. J. (1999) Cytotoxicity with Auger electron-emitting radionuclides delivered by antibodies. Int. J. Cancer 81, 985992.
4. Little, J. B. and Williams, J. R. (1977) in Handbook of Physiology, Sect. 9 (Lee,
D. H. K., Falk, H. L., Murphy, S. D., and Geiger, S. R., eds.), American Physiological Society, Bethesda, MD, pp. 127155.
5. Warters, R. L. and Hofer, K. G. (1977) Radionuclide toxicity in cultured mammalian cells. Elucidation of the primary site for radiation-induced division delay.
Radiat. Res. 69, 348358.
6. Finney, D. (1978) Statistical Methods in Biological Assay, 3rd ed., Charles Griffin
& Co., London.
7. Fazekas de St. Groth, S. (1982) The evaluation of limiting dilution assays.
J. Immunol. Methods 49, R11R23.
8. Taswell, C. (1981) Limiting dilution assays for the determination of immunocompetent cell frequencies. I. Data analysis. J. Immunol. 126, 16141619.
9. Johnson, E. A. and Brown, B. W. (1961) Biometrics 27, 7988.
10. Lefkovits, I. and Waldmann, H. (1979) Limiting Dilution Analysis of Cells in the
Immune System, Cambridge University Press, Cambridge.
11. Strijbosch, L. W. G., Buurman, W. A., Does, R. J. M. M., Zinken, P. H., and Groenewegen, G. (1987) Limiting dilution assays. Experimental design and statistical
analysis. J. Immunol. Methods 97, 133140.
4
Sulforhodamine B Assay and Chemosensitivity
Wieland Voigt
Summary
The sulforhodamine B (SRB) assay was developed by Skehan and colleagues to measure
drug-induced cytotoxicity and cell proliferation for large-scale drug-screening applications. Its
principle is based on the ability of the protein dye sulforhodamine B to bind electrostatically and
pH dependent on protein basic amino acid residues of trichloroacetic acidfixed cells. Under
mild acidic conditions it binds to and under mild basic conditions it can be extracted from cells
and solubilized for measurement. Results of the SRB assay were linear with cell number and cellular protein measured at cellular densities ranging from 1 to 200% of confluence. Its sensitivity is comparable with that of several fluorescence assays and superior to that of Lowry or
Bradford. The signal-to-noise ratio is favorable and the resolution is 10002000 cells/well. It
performed similarly compared to other cytotoxicity assays such as MTT or clonogenic assay.
The SRB assay possesses a colorimetric end point and is nondestructive and indefinitely stable.
These practical advances make the SRB assay an appropriate and sensitive assay to measure
drug-induced cytotoxicity even at large-scale application.
Key Words
Sulforhodamine B; trichloroacetic acid; optimal cell number; cytotoxicity.
1. Introduction
The sulforhodamine B (SRB) assay as first described by Skehan and colleagues was developed for use in the disease-orientated, large-scale anticancer
drug discovery program of the National Cancer Institute (NCI) that was
launched in 1985 (1). The SRB assay is based on the ability of the SRB dye to
bind electrostatically and pH dependent on protein basic amino acid residues.
Under mild acidic conditions, SRB binds to protein basic amino acid residues
of trichloroacetic acid (TCA)fixed cells. It can be quantitatively extracted from
cells and solubilized for optical densitiy (OD) measurement by weak bases
such as Tris base. Results of the SRB assay were linear with the number of
39
40
Voigt
cells and the cellular protein measured at cellular densities in 96-well microtiter
plates ranging from 1 to 200% of confluence (1). The sensitivity of the SRB
assay is comparable with that of several fluorescence assays and is superior to
that of Lowry or Bradford. It has a signal-to-noise ratio at 5000 cells/well of
4.83 and a resolution of 10002000 cells/well (1). The end point of the SRB
assay is colorimetric, nondestructive, and indefinitely stable. The SRB assay
represents an appropriate and sensitive assay to measure drug-induced cytotoxicity and is useful to quantify clonogenicity. In addition, the SRB method
is rapid and inexpensive. In comparison with tetrazolium assays [2,3-Bis
(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide (XTT) or
3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) assay] or
clonogenic assay, the SRB assay performed similarly when data were limited to
the inhibitory 50% concentration (IC50) (25). More important, since the endpoint measurement is not time critical, the SRB assay possesses practical advantages over the tetrazolium assays. Once automated using a microplate washer
and microplate reader, it is suitable for high-throughput screening approaches.
SRB is an anionic bright pink aminoxanthene protein dye with two sulfonic
groups. Its molecular formula and molecular weight are C27H30N2O7S2 and
558.66, respectively. The optimal wavelength for measurement of the OD of SRB
is 564 nm. Curves of OD vs dye concentration were linear up to 1.52 OD units.
When linearity range is exceeded, it is necessary to dilute an aliquot or to use
a suboptimal wavelength (490550 nm). SRB fluoresced with laser excitation
at 488 nm, and measuring SRB fluorometrically increases sensitivity about
threefold (1).
When establishing the SRB assay, Skehan et al. (1) determined the SRB
binding as a function of time, dye concentration, number of destaining washes,
and dye volume per unit area of cell culture. Based on their results, it is recommended that a 96-well culture plate be stained with 100 L of 0.4% SRB
solution per well for 30 min. The number of acetic acid washes to remove
unbound dye should be at least four (1).
Most researchers use the IC50 as a measure of in vitro cytotoxicity. This is
defined as the concentration that yields 50% less cells than the drug-free control. At the NCI, the GI50 (concentration that causes 50% growth inhibition)
was defined. The GI50 is corrected for the cell count at time zero (start of drug
exposure) and follows the following equation:
100 (T T0) / (C T0)
in which T is the OD after exposure to a certain concentration of a drug, T0 is

the OD at the start of drug exposure, and C is the OD of the control. To present
response patterns of a drug in a cell line panel such as the NCI 60 panel, a
mean graph could be constructed (6).
41
The clinical relevance of in vitro cytotoxicity data is a critical issue. To estimate potential clinical activity of a drug based on in vitro data, we and others
defined the relative antitumor activity (RAA) as peak plasma concentration of
a drug/IC50 (7).
Overall, it is assumed that in vitro drug screening was approx 60% reliable
for predicting in vivo sensitivity and 90% reliable for in vivo resistance.
2. Materials
2.1. Seeding of Microtiter Plates and Drug Exposure
1. 96-Well culture microtiter plates.
2. Conventional cell culture equipment and reagents such as cell culture microscope
and cell incubator, sterile tubes (different sizes), growth medium and fetal calf
serum (according to the requirements of cultured cell lines), trypsin.
3. Cell count chamber or other cell count device such as a Coulter counter.
4. Eight-channel multipipet.
5. Sterile gauzes.
6. Sterile plastic troughs.
2.2. SRB Procedure

1.
2.
3.
4.
5.
6.
7.
8.
10% (w/v) TCA.

1% Acetic acid.
0.4% (w/v) SRB dissolved in 1% acetic acid (store at 4C protected from light).
Eight-channel multipipet.
10 mM Unbuffered Tris base (pH 10.0).
Deionized water.
Automated microplate washer.
Automated microplate reader (enzyme-linked immunosorbent assay [ELISA]
reader).
3. Methods
All cell culture work must be performed under sterile working conditions
and under standard cell culture conditions (humidified air, 5% CO2, and 37C)
or adapted to requirements of the specific cell lines, if necessary.
3.1. Seeding of Microtiter Plates for Growth Kinetics
For cytotoxicity experiments, it is critical to ensure exponential cell growth
for the entire duration of the assay. In particular, this is dependent on the
number of cells seeded per well in the microtiter plate at the initiation of the
experiment. Therefore, prior to cytotoxicity experiments, it is important to
determine the optimal cell count ensuring exponential cell growth for the entire
period of the assay and an OD at the end of the experiment in a range of
1.52.0.
42
Voigt
3.1.1. Day 1
1. Harvest cells by trypsinization and determine the cell count according to standard cell culture procedures (see Note 1).
2. Seed serial dilutions of cells (should cover a broad range, e.g., 10050,000 cells/
well) in microtiter plates in 100 L of growth medium per well at d 1 (keep one
eight-well row as a blank containing only growth medium, usually row 1). The
number of culture plates seeded depends on the desired length of the cytotoxicity
experiment (usually 57 d).
3.1.2. Day 2
1. Gently remove the growth medium of one culture plate (generally, we flip the
culture plate and soak up residual medium with a sterile piece of gauze while
keeping the culture plate upside down).
2. Fix cellular protein by the addition of 100 L of 10% TCA (4C) per well.
3. Add 100 L/well of growth medium to the remaining plates only at this time
point.
4. Stop another culture plate by fixating cellular protein as described in steps 1 and
2 every 24 h until all the plates are fixed. Fixed culture plates must be kept at 4C
for at least 1 h but can be stored for up to several days prior to analysis by the
SRB assay (keep in mind that storage for several days may lead to a slight
increase in background levels).
At this time point, perform the SRB assay as described next. Analyze OD
values graphically as illustrated in Fig. 1. From these growth kinetics plots,
the optimal number of cells per well can be depicted for use in further cytotoxicity experiments.
3.2. Choice of Drug Dose Range and Seeding of Culture Plates
for Cytotoxicity Assays
The design of the culture plates depends on the intended experiment. For
example, for a regular dose-response curve of a single drug, rows 1 and 12 are
kept as blank rows containing only growth medium (to determine the nonspecific background). The choice of doses to be tested and the exposure time to the
drug are the most critical points. It is conventionally agreed that a drug, when
singly administered, must be used at concentrations around the clinically
achievable peak plasma concentration in cancer patients. Dose- and timeresponse curves must also be defined for each drug in order to identify the concentration and the exposure time capable of inhibiting cell growth by 50%
(IC50). We usually do half log step dilutions, ensuring a homogeneous distribution of data points on the growth curve and coverage of a broad activity range
(see Fig. 2). This further allows extrapolation of data concerning drug dose at
high levels of activity, e.g., 90% of growth inhibition or higher.
43
Fig. 1. Determination of optimal cell number per well for cytotoxicity experiments
by cell growth kinetics. A typical cell growth kinetics is illustrated. Various numbers of
cells/well were seeded in 96-well culture plates and fixed at the indicated time points.
OD was measured at 570 nm subsequent to performing the SRB assay. OD values were
evaluated graphically, and the optimal cell number for further growth experiments could
be determined from the plots. According to the two criteria, exponential cell growth for
the entire assay period and OD 1.52 at the end of the 120-h assay time, the optimal
cell number for further cytotoxicity experiments is 1600 cells/well.
3.2.1. Day 1
1. Seed in 100 L of growth medium a cell linespecific number of cells (see Subheading 3.1.) per well in 96-well culture plates. Include blanks depending on the
experimental design, and allow 24 h of cell growth.
3.2.2. Day 2
Exponentially growing cells are exposed to serial dilutions of the drug for
the desired times (we usually choose a 2- or 96-h drug exposure time).
1. Control cellular growth, distribution of cells in the well, and cellular density in
different wells using a cell culture microscope (see Note 2).
2. Add 100 L/well of drug-free growth medium to the blank and control rows (usually rows 1, 12, and 2 in our laboratory, respectively).
3. Add 100 L/well of growth medium containing different drug concentrations to
rows 311 (starting with the lowest concentration and finishing with the highest).
44
Voigt
4. Stop one culture plate per cell line to estimate the cell count/OD at the time of
adding the drug (zero h plate).
5. Allow the cells to grow depending on the desired drug exposure time (e.g., 96 h).
6. Remove gently the growth medium by flipping the culture plate (be cautious;
medium contains potentially toxic drugs).
7. Remove residual medium by utilizing a sterile piece of gauze while keeping the
culture plate upside down.
8. Fix the cellular protein using 100 L/well of 10% TCA (4C), and store the plates
at 4C for at least 1 h prior to analysis by the SRB assay.
In contrast to the original method as described by Skehan and colleagues, we

remove growth medium before the TCA fixation step because this significantly
reduces the signal-to-noise ratio. Cell loss is negligible and assay results are
unaffected (8).
In the case of short-term drug exposure (e.g., 2 h), follow steps 6 and 7.
Then include a washing step by adding 200 L/well of either sterile growth
medium or phosphate-buffered saline. Again perform steps 6 and 7. Subsequently, add 200 L/well of complete growth medium and allow further cell
growth for, e.g., 94 h to reach a total assay time of 96 h (see Note 3).
Depending on experience and cell line characteristics, one may risk a high
and variable cell loss during the washing procedure leading to high variations
in the experimental data. In our hands, as determined by wash kinetics, cell
loss approximates cell line dependence (1020%). However, some cell lines
might have an only weak attachment to the plastic bottom of the culture well
and, thus, might be unsuitable for this kind of experiment.
Using the washing procedure as described, one can expect a drug dilution of
about 1400 to 11000, depending on the drain of residual medium or washing
solution. Keep this in mind when using highly active drugs or cell cycle
specific drugs; the remaining substance could flaw the results. If this is a critical point in the experiment, consider including an additional washing step.
However, this may increase cell loss significantly.
3.3. SRB Procedure
The SRB assay outlined next represents the method described by Skehan
and colleagues with minor modifications. Prior to beginning the SRB assay,
cellular protein must be fixed with 10% TCA for at least 1 h at 4C.
1. Rinse TCA using an automated 96-well plate washer and add five washing cycles
using deionized water (200 L/well).
2. Sharply flick the plates over a sink and remove roughly the residual wash solution
with a piece of gauze.
3. Pipet 100 L of SRB solution into each well of the culture plate using a multichannel pipet. Allow a 30-min staining period.
45
4. Remove the SRB from the culture plates with an automated plate washer by five
washing cycles using 1% acetic acid (remember to prefill the automated plate
washer with the appropriate wash solution) (see Note 4).
5. Flick the culture plates over a sink and remove the residual wash solution with a
piece of gauze.
6. Air-dry the culture plates until no further moisture is visible. (When air-dried,
both TCA-fixed and SRB-stained culture plates can be stored indefinitely. Trissolubilized SRB is also stable for extended periods of time, provided that evaporation does not occur.)
7. Solubilize the protein-bound dye with 100 L of 10 mM Tris base per well. Shake
the plates for at least 10 min on a gyratory shaker to homogenize the dye solution.
8. Measure the OD by using an automated 96-well plate reader (ELISA reader) at a
wavelength of 564 nm (see Note 5).
3.4. Data Processing and Interpretation of Results

Generally, the reproducibility of the SRB assay is high and the standard
deviation (SD) relatively low. Processing of the raw data can easily be performed by programs such as Excel or similar. It should include quality control
of the data (SD? OD in correct range? [see Note 6]), subtraction of background
staining, and subtraction of the OD at the time of adding the drug (zero h
plate). The data of each eight-well row should be averaged and the SD calculated. Finally, the percentage of growth inhibition should be calculated as the
percentage of the untreated control. We developed an Excel spreadsheet that
processes the data as outlined in one step. Subsequently, plots of the processed
data can be generated by graphic programs such as Sigma plot or similar. From
these plots, one can graphically determine growth inhibitory concentrations
such as IC50 (Fig. 2). Alternatively, mathematical curve fitting can be performed, which allows a more precise estimation of the IC50. Models used are
often based on the Hill equation. The topic of curve fitting and estimation of
drug interaction is not part of this protocols description. For extended reviews,
see refs. 9 and 10.
For clinical correlation of in vitro drug sensitivity data, we use the model of
RAA (7). Potential clinical activity can be assumed if RAA exceeds 1.
4. Notes
1. It is important to obtain and maintain a homogeneous cell suspension. Cellular
clumps/aggregates should be avoided. To disaggregate cells, it is often sufficient
to pipet up and down the cell suspension with a regular cell culture pipet. Otherwise, we sometimes use a syringe and a 26-gauge needle. If this is still insufficient, changing the trypsinization procedure might help. Depending on the cell
line used, one may have a varying fraction of dead cells. To estimate this fraction,
we usually add Trypan blue when counting the cells in a cell chamber. It is advisable to correct the cell count according to the fraction of dead cells.
46
Voigt
Fig. 2. Graphic determination of IC50 from dose-response plots. A typical doseresponse plot is illustrated. For graphic determination of the IC50 (drug concentration
that yields 50% less cells than the drug-free control), a bar (P) parallel to the x-axis and
intersecting the point 50% on the y-axis was constructed. In the next step, a bar was
plotted parallel to the y-axis that starts from the point of intersection of P with the
dose-response plot. The IC50 could then be directly determined at the point of intersection with the x-axis.
2. To ensure optimal assay results and a low SD, it is important to have homogeneously distributed cells in a well and an approximately similar amount of cells in
each well of each culture plate of the same cell line. Inhomogeneous cell distribution might be owing to the method of pipeting or unsuitable culture plates (the
plastic of the culture plates is a critical issue in our opinion and should be kept
constant for the entire experimental series). Sometimes it is useful to gently shake
the plates on the shelf of the incubator after seeding. If the numbers of cells vary
obviously between different wells, one should consider not continuing the experiment because of the high deviation that is to be expected. Reasons might be
an inhomogeneous cell suspension while pipeting, a malfunction of the eightchannel multipipet, a bad batch of culture plates, or an early sign of microbiological contamination.
3. Try to pipet as gently as possiblethis helps to reduce cell loss. Do not process
too many plates per time, particularly while learning the assay; one may easily run
out of time and not be able to hold the time points. Additionally, culture plates
cool down while being stored in the hood and pH of the growth medium changes
47
dramatically. Sometimes, it might be beneficial to store plates on a warming

bench.
4. Carefully watch the cell monolayer at the bottom of the 96-well culture plates. If
the needles of the 96-well plate washer are poorly adjusted, they may scratch the
monolayer. This could result in partial or even complete loss of TCA-fixed cells.
Sometimes, some needles of the plate washer become clotted with debris and so
on. This may lead to inhomogeneous washing results.
5. If one does not have the appropriate filter for the microplate reader, measurements can be made at 570 nm or at a suboptimal wavelength ranging from 490 to
550 nm.
6. Usually, results of the SRB assay are highly reproducible. SD between independent experiments should not exceed 2030%. SD of an eight-well row should not
exceed 20% and is usually the highest in the range of the IC50. If the SD greatly
exceeds these values, this may be caused by additional washing steps or problems outlined in Notes 14. The OD at the end of the experiment is a critical
issue. First, measurement is not linear if OD values exceed 2. Second, if culture
wells were overgrown, this may lead to exhaustion of growth medium, partial cell
arrest, and cell death, particularly in control wells and wells with lower drug concentration. Taken together, this may lead to a significant change in IC50 (mostly an
underestimation).
References
1. Skehan, P., Storeng, R., Scudiero, D., Monks, A., McMahon, J., Vistica, D., Warren,
J. T., Bokesch, H., Kenney, S., and Boyd, M. R. (1990) New colorimetric cytotoxicity assay for anticancer-drug screening. J. Natl. Cancer Inst. 82, 11071112.
2. Rubinstein, L. V., Shoemaker, R. H., Paull, K. D., Simon, R. M., Tosini, S.,
Skehan, P., Scudiero, D. A., Monks, A., and Boyd, M. R. (1990) Comparison of in
vitro anticancer-drug-screening data generated with a tetrazolium assay versus a
protein assay against a diverse panel of human tumor cell lines. J. Natl. Cancer
Inst. 82, 11131118.
3. Perez, R. P., Godwin, A. K., Handel, L. M., and Hamilton, T. C. (1993) A comparison of clonogenic, microtetrazolium and sulforhodamine B assays for determination of cisplatin cytotoxicity in human ovarian carcinoma cell lines. Eur. J. Cancer
29A, 395399.
4. Griffon, G., Merlin, J. L., and Marchal, C. (1995) Comparison of sulforhodamine B,
tetrazolium and clonogenic assays for in vitro radiosensitivity testing in human ovarian cell lines. Anticancer Drugs 6, 115123.
5. Keepers, Y. P., Pizao, P. E., Peters, G. J., van Ark-Otte, J., Winograd, B., and
Pinedo, H. M. (1991) Comparison of the sulforhodamine B protein and tetrazolium (MTT) assays for in vitro chemosensitivity testing. Eur. J. Cancer 27,
897900.
6. Hodes, L., Paull, K., Koutsoukos, A., and Rubinstein, L. (1992) Exploratory data
analytic techniques to evaluate anticancer agents screened in a cell culture panel.
J. Biopharm. Stat. 2, 3148.
48
Voigt
7. Voigt, W., Bulankin, A., Muller, T., Schoeber, C., Grothey, A., Hoang-Vu, C., and
Schmoll, H. J. (2000) Schedule-dependent antagonism of gemcitabine and cisplatin
in human anaplastic thyroid cancer cell lines. Clin. Cancer Res. 6, 20872093.
8. Papazisis, K. T., Geromichalos, G. D., Dimitriadis, K. A., and Kortsaris, A. H.
(1997) Optimization of the sulforhodamine B colorimetric assay. J. Immunol.
Methods 208, 151158.
9. Berenbaum, M. C. (1989) What is synergy? Pharmacol. Rev. 41, 93141.
10. Greco, W. R., Bravo, G., and Parsons, J. C. (1995) The search for synergy: a critical review from a response surface perspective. Pharmacol. Rev. 47, 331385.
5
Use of the Differential Staining Cytotoxicity Assay
to Predict Chemosensitivity
Gertjan J. L. Kaspers
Summary
The differential staining cytotoxicity (DiSC) assay is one of the total cell-kill assays that can
be used for drug resistance testing. Numerous publications have demonstrated the clinical value
of chemosensitivity testing with that assay (and similar ones). The DiSC assay is successful in
the majority of malignant samples of which the cells can be brought in suspension (not necessarily as single-cell suspension). Although the assay is laborious and requires skilled technicians, it requires few cells, can be used for proliferating and nonproliferating cell populations,
and can discriminate between malignant and contaminating nonmalignant cells. The latter is a
major advantage of the DiSC assay. This chapter describes the practical aspects of this assay, several topics that need to be taken into account, and potential pitfalls. As such, it is not an extensive
review of studies in which the DiSC assay was used.
Key Words
Differential staining cytotoxicity assay; dye exclusion assay; drug resistance; chemosensitivity testing; leukemia; solid tumors; childhood; prognosis; clonogenic assay.
1. Introduction
The success of chemotherapy in cancer may be limited by at least three factors (1). First, unfavorable pharmakinetics may result in inappropriately low
drug exposure of the malignant cells. Second, increased cellular drug resistance may explain the lack of total cell kill despite appropriate drug exposure.
Third, increased regrowth potential of remaining minimal residual disease may
result in a relapse. The extent of cellular drug sensitivity or resistance of malignant cells can be measured in vitro by different methods, and this provides
clinically relevant information in numerous malignancies, including childhood
leukemia (16). These different methods include clonogenic and nonclonogenic
49
50
Kaspers
assays, as reviewed by several researchers (1,3,68). Clonogenic assays have

the theoretical advantage that clonogenic cells are being studied, which might
be the cells responsible for a possible relapse (a more frequent problem in,
e.g., leukemia, in which an apparent complete remission normally is obtained,
than in the case of more drug-resistant malignancies). However, they have disadvantages as well, such as having a (very) low plating efficiency and being
very laborious and not highly reproducible. Moreover, several investigators
have reported similar results using clonogenic and nonclonogenic assays
(912). Therefore, nonclonogenic assays such as the differential staining cytotoxicity (DiSC) assay, tetrazolium-based colorimetric assays, and flow cytometry assays have recently been used more often. Several investigators reported
similar results with different nonclonogenic assays under specific conditions,
especially a high purity of the samples (1316). Each assay has its own advantages and disadvantages, and pitfalls.
This chapter focuses on the DiSC assay, based on dye exclusion assays,
which was probably most well studied by Weisenthal et al. (e.g., see ref. 12),
Bird et al. (17), and Bosanquet et al. (e.g., see ref. 2). The main advantages of
the DiSC assay are that relatively few cells are needed (as compared with, e.g.,
colorimetric assays) and especially that the results can be determined regarding
only the malignant cells. The latter is especially important in the case of a significant amount of contaminating nonmalignant cells, which cannot be discriminated with the often used colorimetric assays. Kaspers et al. (18) reported
that in childhood acute lymphoblastic leukemia (ALL) samples, this became a
significant problem in the case of more than 30% of contaminating nonmalignant cells in the control experiments at the time of the actual measurements,
because contaminating nonmalignant cells are more drug resistant (19). The
DiSC assay can overcome this problem. However, the DiSC assay is laborious
and has several practical disadvantages and pitfalls, as described in Table 1.
Nevertheless, the DiSC assay has provided clinically relevant information. Its
results are therefore being used for risk-group stratified treatment of pediatric
ALL cases (not eligible for the 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide [MTT] assay) in ongoing studies of the German COALL Study
Group (20), and for DiSC assayguided therapy in chronic lymphatic leukemia
according to an MRC trial (21). These types of studies may demonstrate the
clinical benefit of cellular drug resistance assays.
2. Materials
1. Lymphoprep (density: 1.077 g/mL; Axis-Shield, Norway), to isolate mononuclear
or malignant cells by density gradient centrifugation.
2. Falcon polystyrene tubes, 14-mL (Becton Dickinson, Belgium) and 50-mL
polypropylene tubes (Sarstedt, Germany).
DiSC Assay for Chemosensitivity Testing
51
Table 1
Advantages and Disadvantages of DiSC Assaya
Advantages
Needs few cells, results within 47 d
Discrimination between malignant
and contaminating nonmalignant
cells possible
Clinically relevant correlations reported
Not restricted to single-cell suspensions
Suitable for different tumor types,
and for both proliferating and
nonproliferating cells
Disadvantages
Laborious
Discrimination sometimes difficult
No proven benefit on patient survival
No dose-dependent cytotoxicity for some
drugs (e.g., methotrexate) (as with
similar assays)
Note that other cellular drug resistance assays have partly similar disadvantages as well.
3. Immunomagnetic Dynabeads M-450 (Dynal, Norway) (18) to remove contaminating lymphocytes.

4. RPMI culture medium plus fetal calf serum (FCS) (Integro, The Netherlands) and
appropriate supplements (e.g., insulin, glutamine, and transferrin) (see Note 1).
5. Anticancer agents (see Note 2).
6. 96-Well round-bottomed microculture plates (Greiner, The Netherlands).
7. Cytospin columns (Shandon, UK).
8. Duck red blood cells (DRBCs) (Sigma, The Netherlands).
9. May-Grunwald-Giemsa (MGG) (Merck, Germany).
10. Cytospin slides (Menzo & Glzer, Germany).
11. Gauze with 70-m pores (Becton-Dickinson).
3. Methods
3.1. Culture of Leukemic Samples
1. For isolation of leukemic cells, dilute bone marrow or peripheral blood 11 with
RPMI-1640, or more in the case of >30 106 cells/mL.
2. Carefully layer up to 8 mL of the diluted sample onto 4 mL of lymphoprep at
room temperature in a 14-mL Falcon tube.
3. Carry out density gradient centrifugation for 20 min (room temperature, at 540g)
when cells are handled the same day as the sampling, or for 15 min (room temperature, at 380g) when cells are handled at least 1 d or later after sampling. The
acceleration time is 90 s, the break at 3.
4. Collect the interphase cells in a 50-mL tube, and add wash medium (RPMI-1640
plus 1% heat-inactivated FCS [heat inactivation: incubation of FCS for 20 min at
56C]) up to a maximum of 35 mL.
5. Take a sample of 100 L for cell counting using 2% trypan blue (in phosphatebuffered saline [PBS]), to determine the percentage and total number of viable
52
6.
7.
8.
9.
10.
11.
Kaspers
cells, and to check the percentage of malignant cells after making a cytospin and
MGG staining (see below).
Meanwhile, centrifuge the tube at 540g for 10 min at room temperature, break 7
(mainly to remove contaminating platelets); remove the supernatant, and resuspend the pellet in wash medium (1020 mL, depending on the amount of cells).
Centrifuge the tube at 300g for 10 min at room temperature, break 5.
Remove the supernatant and resuspend the pellet in culture medium (see step 9 and
also Note 1). If one wishes to remove contaminating nonmalignant lymphocytes,
then cells should be resuspended in culture medium: in the case of 20 106 cells,
add 0.5 mL of culture medium; in the case of >20 106 cells, add 1 mL of culture
medium. This procedure is a truly separate subject, and the reader is referred
to ref. 18.
Suspend leukemic cells in specific culture medium as described below, in the
case of ALL at 2 106 cells/mL, and in the case of acute myelogenous leukemia
at 0.81.4 106 cells/mL (1.4 in the case of FAB type M0, M1, and M2). This
can also be done with chronic myeloid and lymphocytic leukemia cells, at 1 and
2 106 cells/mL, respectively. It can be done with solid tumor cells as well, but
we do not have enough experience with these specimens in the DiSC assay ourselves to advise on the cell concentrations needed. However, in the MTT assay, we
use 0.5 106 cells/mL. If one aims at using the DiSC assay anyhow, the percentage of malignant cells does not matter at this point. Moreover, one may decide to
use fewer cells, because the DiSC assay allows the detection of fewer cells. However, if one prefers the use of, e.g., the MTT assay if possible, removal of contaminating nonmalignant cells may be indicated. This is a separate procedure, and
the reader is referred to ref. 18.
Dispense 80 L of the leukemic cell suspension into each well of the 96-well
plates, except for the outer wells (evaporation, not reliable). Fill these outer wells
with 100 L of RPMI-1640. These wells already contain 20 L of drug dilution
(see Notes 3 and 4) (or RPMI-1640 in the case of the blanks), with the plates
being stored at 90C until use. Next the culture and drug exposure period follows (in our laboratory always 4 d, although this is not necessarily the gold
standard).
Determine the number of contaminating nonleukemic cells at the end of 4 d of
culture after making four cytospins from this suspension. For this purpose, remove
the cell suspension from the two lower control wells (i.e., the wells located at the
greatest distance from the wells with the highest drug concentrations) of a plate.
We always use the same plate (what we call plate I) for this purpose. Similarly,
make a cytospin before culture to determine the percentage of malignant cells.
For this purpose, make a cell suspension of 0.5 106/mL in PBS plus 0.1%
bovine serum albumin (BSA) plus 5% human serum albumin (HSA). Prewet the
cytospin slides in the cytospin centrifuge by adding one drop of PBS plus 0.1%
BSA to each cytospin cup with a Pasteur pipet and centrifuge at 60g for a few seconds. Then add 50 L of cell suspension to each cup, followed by one drop of
PBS plus 0.1% BSA. Centrifuge at 60g for 7 min (low acceleration, no break).
53
Air-dry the slides for several minutes before staining. For any immunocytochemical staining, the slides are first silicagel dried for at least 24 h. However, this
discussion concerns determining the percentage of malignant cells. For that purpose, fix cytospin slides in 100% methanol for 3 min, subsequently stain in standard MayGrnwald solution for 3 min, rinse in tap water, stain for 1015 min in
Giemsa (time depending on the freshness of the solution; Giemsa is made fresh
every week, diluted 120 in tap water), and rinse again with tap water. Dry the
slides before microscopic evaluation. Then determine the percentage of malignant cells by microscopy.
3.2. DiSC Assay (see Notes 5 and 6)

There are several important points to consider when using the DiSC assay,
some of which are more in general relevant to all cell culture and total cell-kill
assays, whereas others are more specific for the DiSC assay. Although the
DiSC assay is labor-intensive and subjective, a main advantage is that malignant and contaminating nonmalignant cells can be discriminated. The results
have been reported in numerous publications and appear to have clinical relevance in several aspects (2,3,22). A team of dedicated and experienced technicians and investigators is required for meaningful experiments with the DiSC
assay.
1. After the 4 d of culture and drug exposure, remove the cell suspension from the
two lower control wells (i.e., the wells at the greatest distance from the wells with
the highest drug concentrations) of a plate. We always use the same plate (what
we call plate I) for this purpose. If one wishes to perform the DiSC assay anyhow,
independent from the percentage of malignant cells, this and step 2 are not necessary.
2. Use 50 L of cell suspension for cytospin centrifugation and conventional MGG
staining. (See Subheading 3.1., step 11 for an explanation of how to perform
this procedure.)
3. If the sample shows 70% or less malignant cells, add 20 L of DRBC suspension
(see Note 6) to the remaining 80-L cell suspension, and make another cytospin
slide and stain. The DRBC suspension is made by adding 0.5 mL of HSA, 0.5 mL
of PBS, and 0.1 mL of DRBCs (10 106/mL) together and is used only after filtration through gauze with pores of 70-m. This cytospin ideally shows about
2030 leukemic cells/100 DRBCs, and the final amount of DRBC suspension to
be added to all (except blank) wells is chosen such that this is achieved. This
should be done as quickly as possible and with continuous mixing of the DRBC
suspension. Aliquots from the wells are mixed and then cytocentrifuged at 690g
for 7 min to deliver the cells onto cytospin slides that have been cleaned thoroughly with alcohol. After centrifugation, clean the equipment.
4. Air-dry the resultant cytospin slides, fix with methanol (100%), and counterstain
with MGG (see Subheading 3.1., step 5).
54
Kaspers
5. Have an experienced technician determine the ratio of living (MGG-stained)

leukemic cells over DRBCs by using a light microscope and counting at least 300
(normally 500) DRBCs per slide (see Notes 7 and 9). The viable leukemic
cells/DRBC ratio of a treated well is again expressed as the percentage of the
untreated control wells and then called leukemic cell survival.
4. Notes
1. The culture medium being used depends on the type of cells being studied. In
general, cell lines can be cultured in basic culture medium such as RPMI-1640
supplemented with 10% FCS (end concentration in the culture of 8%). Partly
depending on how cells grow, other supplements can be added. Other researchers
use serum-free culture medium, also for patient samples, partly depending on the
type of study questions (23). In the case of patient cells, we normally use RPMI1640 (Gibco, Dutch modification, without L-glutamine) with 20% FCS in the cell
suspension (end concentration in the culture of 16%), supplemented with 2 mM
glutamine (Flow, The Netherlands), 5 g/mL of insulin, 5 g/mL of transferrin,
and 5 ng/mL of sodium selenite (all from Sigma). To prevent culture infections,
we also add 100 IU/mL of penicillin, 100 g/mL of streptomycin, 0.125 g/mL
of fungizone, and 200 g/mL of gentamycin (end concentrations in the culture
medium; from Flow and ICN, The Netherlands). It may be easy to make 10-mL
aliquots of penicillin, streptomycin, and fungizone and store these at 20C. Similarly, one may decide to make 0.6- or 1.2-mL aliquots of insulin, transferrin, and
sodium selenite.
2. Handling of these drugs is extremely important for reliable test results. Stability
of the drug may depend on light exposure, temperature, solvents used, storage
conditions of the drug and its stock solutions.
3. Some drugs may not show dose-dependent cytotoxicity in (mainly) nonproliferating cell cultures, especially methotrexate. This theoretically is also a problem for
other drugs, of which the cytotoxicity might be circumvented by the uptake of
specific counteracting products through one or more salvage pathways. However,
this has not been convincingly demonstrated for drugs other than methotrexate
and is not a problem in proliferating models such as cell lines (24). This problem
is not unique for the DiSC assay but is relevant for all nonclonogenic total cell-kill
assays. Sensitivity or resistance to methotrexate may be measured indirectly by the
thymidylate synthase inhibition assay (24).
4. Drug instability is a concern for all in vitro drug resistance assays and should be
taken into account. Bosanquet and Bell (25) have written an extensive review on
this. Another pitfall is the use of nonactive (pro-) drugs such as prednisone instead
of prednisolone, ifosfamide instead of 4-hydroperoxy-ifosfamide, and cyclophosphamide instead of mafosfamide.
5. The DiSC assay is laborious. This fact in itself makes it more appealing to use
other assays, such as colorimetric or flow cytometric ones, provided that the
malignant samples are pure enough (i.e., that the percentage of malignant cells is
high enough). The procedure may improve by making multispotted cytospins, and
55
by reducing the number of analyses, such as by calculating the number of living

cells at only one drug concentration. Alternatively, one may quickly screen for
the two drug concentrations at which the cell survival is just below and above
50%, respectively, to calculate an LC50 value without the need to actually determine cell survival at all drug concentrations. Under these circumstances, triple
instead of duplicate experiments should be considered. The conditions when the
DiSC assay would be recommended or preferred are highlighted in Table 1.
6. An internal standard is needed. Since dying and dead cells may disappear completely but unpredictably, it is important to include an internal standard in all
wells, to avoid an underestimation of the cytotoxicity of a given drug. DRBCs
are suitable.
7. Recognition of malignant cells can be difficult. Therefore, one needs dedicated
and especially skilled and experienced technicians for the actual microscopic
examinations, sometimes helped by the use of additional immunological stainings. This subjectivity also makes the DiSC assay less reproducible, especially if
several technicians are involved in the analyses.
8. Infection is a problem not unique to the DiSC assay but may in fact occur in all
cell culture assays. The routine prophylactic use of antibiotics and an antifungal
drug in the culture medium is indicated.
Acknowledgments
I thank all the research technicians who performed DiSC assays in the
research laboratory of pediatric oncology at the VU university medical center
(Amsterdam, The Netherlands), and A. H. Loonen for critical comments on the
manuscript.
References
1. Kaspers, G. J. L. and Veerman, A. J. P. (2003) Clinical significance of cellular
drug resistance in childhood leukemia. Recent Results Cancer Res. 161, 196220.
2. Bosanquet, A. G. (1991) Correlations between therapeutic response of leukaemias
and in-vitro drug-sensitivity assay. Lancet 337, 711714.
3. Freuhauf, J. P. and Bosanquet, A. G. (1993) In vitro determination of drug
response: a discussion of clinical applications. Principles Pract. Oncol. 7(12),
116.
4. Kaspers, G. J. L., Veerman, A. J. P., Pieters, R., et al. (1997) In vitro cellular drug
resistance and prognosis in childhood acute lymphoblastic leukemia. Blood 90,
27232729.
5. Kaspers, G. J. L., Zwaan, C. M., Veerman, A. J. P., et al. (1999) Cellular drug
resistance in acute myeloid leukemia: literature review and preliminary analysis of
an ongoing collaborative study. Klin. Pdiatr. 211, 239244.
6. Veerman, A. J. P. and Pieters, R. (1990) Drug sensitivity assays in leukaemia and
lymphoma. Br. J. Haematol. 74, 381384.
56
Kaspers
7. Pieters, R., Kaspers, G. J. L., and Veerman, A. J. P. (1997) Drug resistance culture
assays in childhood leukemia: a review of the results and applications. Int. J. Pediatr. Hematol. Oncol. 4, 531541.
8. Weisenthal, L. M. and Lippman, M. E. (1985) Clonogenic and non-clonogenic in
vitro chemosensitivity assays. Cancer Treat. Rep. 69, 615632.
9. Bird, M. C., Godwin, V. A. J., Antrobus, J. H., and Bosanquet, A. G. (1987) Comparison of in vitro drug sensitivity by the differential staining cytotoxicity (DiSC)
and colony-forming assays. Br. J. Cancer 55, 429431.
10. Carmichael, J., deGraff, W. G., Gazdar, A. F., Minna, J. D., and Mitchell, J. B.
(1987) Evaluation of a tetrazolium-based semi-automatic colorimetric assay:
assessment of chemosensitivity testing. Cancer Res. 47, 936942.
11. Laurent, G., Kuhlein, E., Casellas, P., et al. (1986) Determination of sensitivity of
fresh leukemia cells to immunotoxins. Cancer Res. 46, 22892294.
12. Weisenthal, L. M., Marsden, J. A., Dill, P. L., and Macaluso, C. K. (1983) A novel
dye exclusion method for testing in vitro chemosensitivity of human tumors.
Cancer Res. 43, 749757.
13. Alley, M. C., Scudiero, D. A., Monks, A., et al. (1988) Feasibility of drug screening with panels of human tumor cell lines using a microculture tetrazolium assay.
Cancer Res. 48, 589596.
14. Finlay, G. J., Wilson, W. R., and Baguley, B. C. (1986) Comparison of in vitro
activity of cytotoxic drugs towards human carcinoma and leukaemia cell lines.
Eur. J. Cancer Clin. Oncol. 22, 655662.
15. Pieters, R., Huismans, D. R., Leyva, A., and Veerman, A. J. P. (1989) Comparison
of the rapid automated MTT-assay with a dye exclusion assay for chemosensitivity testing in childhood leukemia. Br. J. Cancer 59, 217220.
16. Twentyman, P. R., Fox, N. E., and Rees, J. K. H. (1989) Chemosensitivity testing of
fresh leukaemia cells using the MTT colorimetric assay. Br. J. Haematol. 71, 1924.
17. Bird, M. C., Bosanquet, A. G., Forskitt, S., and Gilby, E. D. (1986) A novel dye
exclusion method for testing in vitro chemosensitivity of haematological malignancies. Leuk. Res. 10, 445449.
18. Kaspers, G. J. L., Veerman, A. J. P., Pieters, R., et al. (1994) Mononuclear cells
contaminating leukemic samples tested for cellular drug resistance using the
methyl-thiazol-tetrazolium assay. Br. J. Cancer 70, 10471052.
19. Kaspers, G. J. L., Pieters, R., Van Zantwijk, C. H., et al. (1991) In vitro drug sensitivity of normal peripheral blood lymphocytes and childhood leukaemic cells
from bone marrow and peripheral blood. Br. J. Cancer 64, 469474.
20. Janka-Schaub, G. E., Harms, D. O., den Boer, M. L., Veerman, A. J., and Pieters,
R. (1999) In vitro drug resistance as independent prognostic factor in the study
COALL-O5-92 treatment of childhood acute lymphoblastic leukemia; two-tiered
classification of treatments based on accepted risk criteria and drug sensitivity profiles in study COALL-06-97. Klin. Pdiatr. 211, 233238.
21. Bosanquet, A. G., Johnson, S. A., and Richards, S. M. (1999) Prognosis for fludarabine therapy of chronic lymphocytic leukaemia based on ex vivo drug response
by DiSC assay. Br. J. Haematol. 106, 7177.
57
22. Staib, P., Lathan, B., Schinkothe, T., et al. (1999) Prognosis in adult AML is precisely predicted by the DISC-assay using the chemosensitivity-index Ci. Adv. Exp.
Med. Biol. 457, 437444.
23. Duyn, A. E. J., Kaspers, G. J. L., Pieters, R., et al. (1999) Effects of interleukin 3,
interleukin 7 and B-cell growth factor on proliferation and drug resistance in vitro
in childhood acute lymphoblastic leukemia. Ann. Hematol. 78, 163171.
24. Rots, M. G., Pieters, R., Kaspers, G. J. L., et al. (1999) Differential methotrexate
resistance in childhood T- versus common/pre-B acute lymphoblastic leukemia can
be measured by an in situ thymidylate synthase inhibition assay, but not by the
MTT assay. Blood 93, 10671074.
25. Bosanquet, A. G. and Bell, P. B. (1993) Handling requirements to achieve active
drugs in in vitro drug sensitivity and resistance assays, in Drug Resistance in
Leukemia and LymphomaThe Clinical Value of Laboratory Studies (Kaspers,
G. J. L., ed.), Harwood Academic, Chur, Switzerland, pp. 227255.
6
Collagen Gel Droplet Culture Method to Examine
In Vitro Chemosensitivity
Hisayuki Kobayashi
Summary
For effective cancer chemotherapy, chemosensitivity testing of anticancer drugs should be
performed using fresh surgical specimens obtained from the cancer. We have developed a new in
vitro chemosensitivity test named the collagen gel droplet embedded culture drug sensitivity test
(CD-DST). The CD-DST method consists of a collagen gel droplet embedded culture step, exposure and washout of anticancer drug, a serum-free culture step, and evaluation of anticancer
effect by image analysis. This method has many advantages including a high success rate for primary culture, the need for only a small number of cells for the test, easy quantification of the
anticancer effects without contamination with fibroblasts by using an image analysis system, a
good correlation between in vitro and in vivo results, and simplicity and speed. The CD-DST
method can be performed in the laboratory using a system kit Primaster.
Key Words
Collagen gel droplet embedded culture drug sensitivity test; anticancer agents; primary cancer
cell; image analysis system.
1. Introduction
For effective cancer chemotherapy, chemosensitivity testing of anticancer
drugs should be performed using fresh surgical specimens obtained from the
cancer. Various chemosensitivity tests have been developed and studied (116).
However, none of these methods have been adopted clinically for various reasons. We have, therefore, developed an in vitro chemosensitivity test that satisfies the following requirements: a high success rate for primary culture, the
need for a small number of cells for the test, easy quantification of the antitumor effects without contamination with fibroblasts, and a good correlation
between the results of in vitro and clinical response. We have named this
59
60
Kobayashi
Fig. 1. Process of CD-DST method.
method the collagen gel droplet embedded culture drug sensitivity test (CDDST) (1720) and reported its clinical usefulness (1826). In this chapter, we
report the protocol and major procedure (Fig. 1) of the CD-DST using fresh
surgical cancer specimens.
2. Materials
1. Primaster (Nitta Gelatin, Japan); this is composed of the following reagents:
a. Collagen Drop Culture Kit (Cellmatrix Type CD, 10X F12 medium, Reconstitution Buffer).
b. Cell Dissociation Enzyme EZ (Nitta Gelatin).
c. CG (Collagen Gel-coated) Flask (Nitta Gelatin).
d. Pre-Culture Medium PCM-1 (Nitta Gelatin).
e. Serum-Free Medium PCM-2 (Nitta Gelatin).
f. Neutral Red Solution (Nitta Gelatin).
2. DF(10) medium: DF medium with 10% fetal bovine serum, 100 g/mL of penicillin, 100 g/mL of kanamycin, and 10 mM HEPES buffer.
3. 1 mM EGTA (cat. no. 152-14; Nakarai Tesque, Japan). 0.03% trypsin (T-8003;
Sigma-Aldrich Japan) solution.
CD-DST Method
4.
5.
6.
7.
8.
9.
10.
11.
12.
61
10% Formalin neutral buffer solution (cat. no. 062-01661; Wako, Japan).
Conical beaker covered with nylon mesh.
Mild mixer (PR-12; Taitec Co., Japan).
Apple PowerMacintosh G3 or G4.
NIH-Image (Ver. 1.56) Macro (software).
Gray-scale image digitizer (LG-3/PCI; Scion).
Videomicroscope (VH-5910 + [VH-W50]; Keyence, Japan).
Device for lighting (MegaLight100 + FGF6F1000-30 30, Hoya-Schott, Japan).
Macro program Primage (for NIH-IMage) (Nitta Gelatin).
3. Methods
3.1. Collection of Primary Cancer Cells
Using Cell Dissociation Enzyme EZ, it becomes possible to digest stroma
from the tumor and dissociate tumor cells without any reduction in the viability of cancer cells in human tumor tissues. This enzyme maintains the viable
cells better than when collagenase treatment is used to disrupt tumor tissues.
This step is followed by growth of tumor cells on collagen gelcoated flasks
(CG flasks) in PCM-1 medium, allowing living cells to adhere and selectively
removing dead cells, blood, and noncell elements.
3.1.1. Pretreatment of Tumor Tissue
1. Transfer a tumor tissue specimen into a 10-cm plastic dish and add 10 mL of
Hanks solution.
2. Cut the tumor tissue into small pieces, approx 5 5 mm, using forceps and scissors.
3. Transfer the pieces of tissue into a new 10-cm dish and mince the pieces into
paste form with a razor blade (see Note 1).
4. Add 10 mL of Hanks solution to the paste form tumor, and disperse the tumor by
pipetting.
3.1.2. Dissociation of Tumor Cells

1. Transfer all the tissue dispersion into a 50-mL centrifuge tube.
2. Add more Hanks solution to a total volume of 3045 mL. Spin the tube at 250g
for 3 min.
3. Remove the supernatant, and add 5 mL of Dulbeccos modified Ealges medium
(DMEM) solution and 0.5 mL of Cell Dissociation Enzyme EZ to the pellet (see
Note 2). Incubate for 2 h at 37C with rotation (using a mild mixer) (see Note 3).
4. Remove the supernatant, and treat the pellet with EGTA-trypsin.
5. Add 30 mL of Hanks solution. After pipetting gently, filter the cells through
300-m nylon mesh overlaid on a conical beaker.
3.1.3. Preculture Using CG Flask

1. After centrifuging, remove the supernatant and add 5 mL of PCM-1 medium to
the cell pellet and pipet gently to generate cell suspension.
62
Kobayashi
2. Inoculate 5 mL of the suspension into a 25-cm2 CG flask, and preculture for

1224 h at 37C in CO2 incubator.
3.1.4. Collection of Viable Cells

1. After culturing, remove the preculture medium in the CG flask.
2. Add 2 mL of DMEM solution and 100 L of Cell Dissociation Enzyme EZ solution to the CG flask (the final concentration of the EZ solution is 0.05%), and
shake the flask at 37C.
3. When the gel in the CG flask has been completely dissolved, transfer the cell
suspension in the flask into a 50-mL centrifuge tube and perform centrifugation.
4. Remove the supernatant, add 1 mL of EGTA-trypsin solution, and then filter the cell
through using a conical beaker provided with 125-m nylon mesh (see Note 4).
5. Repeat steps 3 and 4 twice, and then add 100500 L of DF(10) medium to the
pellet.
3.2. Collagen Gel Droplet Embedded Culture Drug Sensitivity Test

The CD-DST method is a new in vitro chemosensitivity test that combines
the collagen gel droplet culture method and image analysis system. Cancer
cells and fibroblast cells respectively show specific growth morphologies in the
collagen gel droplet. The CD-DST method quantifies only cancer cells in the
image by utilizing the differences of morphology and stains in the collagen gel
droplet.
3.2.1. Collagen Gel Droplet Culture
1. Transfer 4 mL of Cellmatrix-type CD into an ice-cold 50-mL centrifuge tube.
Add 0.5 mL of 10X F-12 medium to the solution and mix well. Add 0.5 mL of
reconstitution buffer to prepare a collagen mixture (see Note 5).
2. Add the cell suspension to the collagen mixture, and mix well (see Notes 6
and 7).
3. Place three droplets (30 L/droplet) of the collagen mixture containing cells into
each well of a six-well plate, using a micropipettor.
4. Incubate the plate for 1 h at 37C to gelate the collagen gel (Fig. 2).
5. After gelation, overlay each well with 3 mL of DF(10) medium and incubate
overnight.
3.2.2. Exposure to Anticancer Agents
The exposure conditions of anticancer agents should be designed to reproduce the clinical area under the curve. The concentration should be similar to
the physiological concentration (27).
1. Add 30 L/well (1/100 vol of the medium in each well) of an anticancer agent
solution, using a micropipettor.
2. Incubate the plate for 24 h in an incubator.
CD-DST Method
63
Fig. 2. Three collagen gel droplets placed on a multiwell.
3. Remove the medium from each well in the plate and overlay with 4 mL of DME
medium for each well.
4. Incubate the plate for 10 min with agitation at 37C in a CO2 incubator. Repeat
this step twice to remove the anticancer agent.
3.2.3. Serum-Free Culture
After the anticancer agents are removed, apply 4 mL of serum-free medium

PCM-2 to each well and incubate the plate for 7 d at 37C in a CO2 incubator.
3.2.4. Staining and Fixation
1. After cultivation for 7 d, add 40 L (1/100 vol of the medium) of a neutral red
solution to each well.
2. Stain the cell for 2 h while gently shaking the plate in a CO2 incubator at 37C
(see Note 8).
3. Two hours after the staining, remove the medium containing neutral red.
4. Dispense 4 mL of phosphate-buffered saline (PBS) in each well, shake the plate
at room temperature, and then remove the PBS.
5. Apply 4 mL of 10% neutral buffered formalin to each well. Allow the plate to
stand for 45 min at room temperature.
64
Kobayashi
Fig. 3. Procedure for quantification by an image analysis system.

6. After removing the formalin, soak the plate in a tray filled with water and wait
10 min (see Note 9).
7. Drain off the water and air-dry the plate.
3.2.5. Quantification by Image Analysis
Using a measuring apparatus consisting of a personal computer, an image

digitizer, and a videomicroscope, density on images is measured according to
the procedure in Fig. 3. Cumulative value of density on the binary image is
calculated, and image-optical density (A) is obtained according to the following
formula:
A = log10 [(WC BC) / (TC BC)]
in which WC is the cumulative density value of the blank image, BC is the

cumulative density value in the covering light, and TC is the cumulative density value of the test sample image.
3.2.6. In Vitro Chemosensitivity
The in vitro sensitivity is expressed as the percentage T/C ratio, in which T

is the imageoptical density of the treated group and C is the imageoptical
CD-DST Method
65
Fig. 4. Collagen gel droplet stained in neutral red (after being air-dried).
density of the control group; a T/C ratio of 50% or less is regarded as being
sensitive in vitro (Fig. 4).
4. Notes
1. Mince the tissue pieces quickly before they become dried.
2. The volume of the solution can be increased up to 10 mL of D-MEM + 1 mL of
Cell Dissociation Enzyme EZ, if the tumor volume is large.
3. When any large pieces of tissue are found at this step, collect those pieces and
digest them again with Cell Dissociation Enzyme EZ.
4. Prolonged treatment with EGTA-trypsin leads to overreduction in size of the cell
clusters and lowers the cellular activity.
5. Cool the tube sufficiently to avoid rapid gelation.
6. Be sure to add the cell suspension in a 1/10 or less volume of the collagen
mixture.
7. Adjust the cell density to 2 105 to 5 105 cells/mL.
8. Inadequate shaking will lead to uneven staining.
9. The gel should be handled gently; it will detach from the plate if the plate is handled roughly.
Acknowledgments
This method was developed through the collaboration of Dr. M. Koezuka, K.
Tanisaka, and many staff of the biochemical laboratory of Nitta Gelatin Inc. I
appreciate their advice on this work and cooperation. I also thank Mr. K.
Minamigawa, T. Takano, and Ms. N. Yoneda for excellent technical supports.
References
1. Osieka, R., Houchens, D. P., Goldin, A., and Johnson, R. K. (1977) Chemotherapy
of human colon cancer xenografts in athymic nude mice. Cancer 40, 26402650.
66
Kobayashi
2. Countenay, V. D. and Mills, J. (1978) An in vitro colony assay for human tumours
grown in immune-suppressed mice and treated in vivo with cytotoxic agents. Br. J.
Cancer 37, 261268.
3. Fujita, M., Fujita, F., and Taguchi, T. (1984) Chemosensitivity of human gastrointestinal and breast cancer xenografts in nude mice and predictability to clinical
response of anti-cancer agents: immune-deficient animals, in 4th International
Workshop on Immune-Deficient Animals in Experimental Research, Karger, Basel,
Germany, pp. 311315.
4. Rygaard, J. and Povlsen, C. O. (1969) Heterotransplantation of a human malignant
tumour to nude mice. Acta Pathol. Microbiol. Scand. 77, 758760.
5. Bogden, A. E., Kelton, D. E., Cobb, W. R., and Esber, H. J. (1978) A rapid screening method for testing chemotherapeutic agents against human tumor xenografts,
in Symposium on the Use of Athymic (Nude) Mice Cancer Research. Gustav New
York, New York, pp. 231250.
6. Bogden. A. E., Griffin, W., Reich, S. D., Costanza, M. E., and Cobb, W. R. (1984)
Predictive testing with the subrenal capsule assay. Cancer Treat. Rev. 11, 113124.
7. Salmon, S. E., Hamburger, A. W., Soehnlen, B., et al. (1978) Quantitation of
differential sensitivity of human-tumor stem cells to anticancer drugs. N. Engl.
J. Med. 298, 13211327.
8. Von Hoff, D. D., Cowan, J., Harris, G., and Reisdorf, G. (1981) Human tumor
cloning: feasibility and clinical correlations. Cancer Chemother. Pharmacol. 6,
265271.
9. Rozencweig, M., Hofmann, V., Sanders, C., Rombaut, W., Fruh, L. J., and Martz,
G. (1984) In vitro growth of human malignancies in a cloning assay. Recent
Results Cancer Res. 94, 17.
10. Tanigawa, N., Kern, D. H., Hikasa, Y., and Morton, D. L. (1982) Rapid assay for
evaluating the chemosensitivity of human tumors in soft agar culture. Cancer Res.
42, 21592164.
11. Kern, D. H., Drogemuller, C. R., Kennedy, M. C., Hildebrand-Zanki, S. L., Tanigawa, N., and Sondak, V. K. (1985) Development of miniaturized, improved
nucleic acid precursor incorporation assay for chemosensitivity testing of human
solid tumors. Cancer Res. 45, 54365441.
12. Tanigawa, N., Morimoto, H., Dohmae, N., Shimomatsuya, T., Takahashi, K., and
Muraoka, R. (1992) In vitro growth ability and chemosensitivity of gastric and
colorectal cancer cells assessed with the human tumor clonogenic assay and thymidine incorporation assay. Eur. J. Cancer 28, 3134.
13. Kondo, T., Imamura, T., and Ichihashi, H. (1966) In vitro test for sensitivity of
tumor to carcinostatic agents. Jpn. J. Cancer Res. 57, 113121.
14. Mosmann, T. (1983) Rapid colorimetric assay for cellular growth and survival:
application to proliferation and cytotoxicity assays. J. Immunol. Methods 65,
5563.
15. Carmichael, J., DeGraff, W. G.. Gazdar, A. F.. Minna, J. D., and Mitchell, J. B.
(1987) Evaluation of a tetrazolium-based semiautomated colorimetric assay:
CD-DST Method
67
16. Shoemaker, R. H., Wolpert-DeFilippes, M. K., Kern. D. H., et al. (1985) Application of human tumor colony-forming assay to new drug screening. Cancer Res. 45,
21452153.
17. Koezuka, M., Kondo, N., Kobayashi, H., et al. (1993) Drug sensitivity test for primary culture of human cancer cells using collagen gel embedded culture and
image analysis. Int. J. Oncol. 2, 953959.
18. Kobayashi, H., Tanisaka, K., Kondo, N., et al. (1995) Development of new in vitro
chemosensitivity test using collagen gel droplet embedded culture and its clinical
usefulness. Jpn. J. Cancer Chemother. 22(13), 19331939 (in Japanese).
19. Kobayashi, H., Tanisaka, K., Doi, O., et al. (1997) An in vitro chemosensitivity test
for solid tumors using collagen gel droplet embedded cultures. Int. J. Oncol. 11,
449455.
20. Kobayashi, H., Higashiyama, M., Minamigawa, K., et al. (2001) Examination of in
vitro chemosensitivity test using collagen gel droplet culture method employing
colorimetric endpoint quantification. Jpn. J. Cancer Res. 92, 203210.
21. Yasuda, H., Takada, T., Wada, K., et al. (1998) A new in-vitro drug sensitivity test
(collagen-gel droplet embedded-culture drug sensitivity test) in carcinomas of pancreas and biliary tract: possible clinical utility. J. Hepatobiliary Pancreat. Surg. 5,
261268.
22. Araki, Y., Isomoto, H., Matsumoto, A., et al. (1999) An in vitro chemosensitivity
test for colorectal cancer using collagen-gel droplet embedded cultures. Kurume
Med. J. 46, 163166.
23. Hanatani, Y., Kobayashi, H., Kodaira, S., et al. (2000) An in vitro chemosensitivity test for gastric cancer using collagen gel droplet embedded culture. Oncol. Rep.
7, 10271033.
24. Higashiyama, M., Kodama, K., Tokouchi, H., et al. (2001) Cisplatin-based
chemotherapy for postoperative recurrence in non-small cell lung cancer patients:
relation of the in vitro chemosensitive test to clinical response. Oncol. Rep. 8,
279283.
25. Kawamura, M., Inoue, Y., Oyama, T., and Kobayashi, K. (2002) Chemosensitivity
test for unresectable non-small cell lung cancer. Nihon geka-gakkai zassi. 103,
229232 (in Japanese).
26. Takamura, Y., Kobayashi, H., Taguchi, T., Motomura, K., Inaji, H., and Noguchi,
S. (2002) Prediction of chemotherapeutic response by collagen gel droplet embedded culture-drug sensitivity test. Int. J. Cancer 98, 450455.
27. Kobayashi, H. (2003) Development of a new in vitro chemosensitivity test using
collagen gel droplet embedded culture and image analysis for clinical usefulness.
Recent Results Cancer Res. 161, 4861.
7
The MTT Assay to Evaluate Chemosensitivity
Jack D. Burton
Summary
The assessment of the degree or rate of cellular proliferation and cell viability is critical to the
assessment of the effects of drugs, antibodies, or cytokines on both normal and malignant cell
populations. This can be accomplished by either direct or indirect counting methods. Direct
counting by manual or automated methods, using a hemacytometer or particle counter, respectively, allows for serial cell counting at multiple time points, but these are low-throughput
approaches. High-throughput and robust alternatives to direct counting utilize either radiotracers
(e.g., 3H-thymidine) or dye compounds, which can be adapted to multiwell culture plate formats. This chapter focuses on the use of tetrazolium-type indicator dyes, of which the compound
3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) has been the most widely
utilized. Newer tetrazolium dyes that yield water-soluble products and offer added flexibility,
increases in sensitivity, and fewer steps, which are offset by increased costs, are also covered.
Key Words
Tetrazolium dye; MTT, proliferation; optical density; 50% inhibitory concentration; test
agent/drug.
1. Introduction
Assessment of the effects of drugs, antibodies, and cytokines on the proliferation and viability of specific cell types grown in culture is a critical initial
step toward understanding and quantifying the effects of such agents. The ability to evaluate the effects of these test agents in vitro allows for the screening
of a larger number of agents to identify those with the desired activity. In the
case of anticancer agents such as chemotherapy or other small-molecule drugs,
those with the greatest antiproliferative activity are typically selected for in
vivo testing in appropriate animal models of human cancer. This is also true
for antitumor antibodies that are being evaluated as anticancer agents. For
immunological studies involving either cytokines or antibodies, agonist or
69
70
Burton
growth-promoting effects are usually of interest. In this case, normal immune

cell populations or factor-dependent cells lines are used to assess such effects.
In all of these cases, the effect of these agents on total, viable cell numbers at
the end of the assay period is one of the key endpoints. Tetrazolium dyes have
been shown to be sensitive, accurate, and efficient in the in vitro evaluation of
anticancer or immunological agents prior to preclinical and, ultimately, clinical testing.
3-(4,5-Dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) is a
yellow, water-soluble tetrazolium dye, which crosses both plasma and mitochondrial membranes. In the latter organelle of viable and glycolytically active
cells, MTT is reduced by the action of NADH- or NADPH-dependent dehydrogenases. Cellular reduction of MTT produces an insoluble, purple formazan.
The amount of the formazan product that results from this reaction is dependent
on the number of cells and their viability. As this reaction goes to completion,
residual viable cells die, the formazan product can be solubilized, and the optical density (OD) can be measured, which reflects the number of cells present
at the end of the assay period. This approach to identifying viable cells as well
as estimating cellular metabolic activity was first described using another tetrazolium dye, 2,3,4-triphenyl tetrazolium chloride (TTC) (13). It was used to aid
microscopic visualization of viable cells grown in soft agar in experiments
designed to assess in vitro growth conditions and cellular responses to drug
treatment (2,3). A quantitative approach to determining the metabolic activity of
tracheal explants was developed by adding TTC to these cultures, dissolving the
resulting formazan precipitates in an organic solvent, followed by determining
OD of the solubilized formazan product (1).
Mosmann (4) first described the design and setup of an assay using the MTT
dye, in which cells are added to replicate wells of 96-well plates. The plates are
placed in an incubator for a sufficient time interval (usually 37 d) to allow
cells to undergo several cell division cycles. At the end of the incubation period,
MTT is added to each well followed by an additional incubation period of at
least 4 h to allow for complete bioreduction of MTT to the final formazan
product. The OD of each well is determined in a plate reader, allowing a highthroughput, accurate, and sensitive method to estimate cell numbers. The original work with this dye was with growth factordependent cell lines, but it was
later shown to be useful for assessing growth characteristics and responses to
chemotherapy drugs of cancer cell lines (57). Two groups of researchers
addressed the factors that affect the performance of MTT-based assays (6,7). As
described in more detail below, the limitation of MTT and other earlier tetrazolium dyes is the water insolubility of the formazan product. This led to the
development and evaluation of other tetrazolium dyes whose formazan biore-
MTT Assay to Evaluate Chemosensitivity
71
duction products are water soluble. The first of this class of tetrazolium dyes to
be described was 2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium5-carboxanilide (XTT) (8,9). It was noted that with XTT (and later with other
members of this class) it was necessary to add an electron-coupling reagent
(typically phenazine methosulfate [PMS]) to achieve optimal absorbance values.
When XTT/PMS was compared with MTT, similar absorbance values were
obtained with both methods, but the former method did not require either the
centrifugation or solubilization steps of the MTT method (8). Although XTT has
this advantage, it has limited solubility and requires prewarming of the medium
to at least 37C; it is also necessary to prepare XTT/PMS fresh just prior to
adding it to the culture plate. Thus, other dyes in this class were synthesized and
evaluated. The dye 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)2-(4-sulfophenyl)-2H-tetrazolium (MTS) was the next one to be described
(10,11). As with XTT, PMS needs to be added to the MTS solution prior to its
addition to cells. Unlike XTT, MTS solutions are stable, as are PMS solutions;
thus, individual stock solutions of each can be stored and mixed just prior to
adding to cells. MTS is available in a kit format containing stock solutions
of both MTS and PMS from Promega. (Madison, WI). Another dye with
improved characteristics over XTT has also been described, (4-iodophenyl)3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium (WST-1). It has been
shown to perform at least as well as MTT and XTT (1215). Like MTS,
WST-1 is stable in solution. It also requires the use of PMS. This dye is available in kit format from Roche Diagnostics, GmbH (Penzberg, Germany). XTT,
MTS, and WST-1 are all considerably more expensive than MTT, but the
time that is saved, as well as the increased sensitivity and utility of these
dyes with a wider range of cell lines, often makes them a fairly cost-effective
alternative.
2. Materials
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Sterile laminar flow hood or biosafety cabinet.

Incubator (usually kept at 37C with an internal atmosphere of 95% air/5% CO2).
Cell culture medium and required additives (see Note 1).
Phosphate-buffered saline (PBS) (see Note 2).
Trypsin/EDTA solution.
Sterile disposables including pipets, pipet tips, and tissue culture flasks.
Repeating Eppendorf-type pipettor and compatible, sterile Combi-Tips (also helpful is an electronic pipettor with repetitive dispensing capability).
96-Well plate reader with adjustable, visible wavelength settings.
Source of cells (usually continuously growing cell lines) (see Note 3).
Chemotherapeutic drugs (see Note 4).
MTT.
72
Burton
3. Methods
3.1. Design of MTT Assays
The MTT assay is useful for measuring the effect of a wide range of compounds on the in vitro growth of either normal or cancer cell lines. The assay
is set up in a 96-well, flat-bottomed polystyrene microtiter plate.
Cells are suspended in appropriate growth medium, and the cells are added
to replicate wells (triplicates are usually preferred) (see Note 5). It is preferable
to add the cells to the required number of wells in the plate prior to adding the
drugs or agents to be tested. After the cells are added to the plate, it can be
placed on the incubator to allow cells to settle and attach (in the case of adherent cells) while the agents to be tested are being prepared.
Drugs or other compounds are added at defined concentrations to each set of
replicate wells (see Notes 6 and 7). Agents to be tested must be properly solubilized. Aqueous solubility does not pose a problem for protein agents such as
antibodies and cytokines. Many small-molecule compounds, however, have limited water solubility. Most of these compounds can be dissolved in dimethyl
sulfoxide (DMSO) or dimethyl formamide (DMF)based solvents. DMSO is
preferable over DMF, because it tends to have slightly less toxicity for cells
in culture. Since most compounds of interest demonstrate antiproliferative
activity at or below the micromolar level, they should be dissolved at concentrations of 10100 mM. This should result in a stock solution that has a concentration of 10005000 times the highest concentration that will be tested in
the assay. This results in final concentrations of DMSO or DMF that usually
exhibit minimal cellular toxicity. It is critical, however, to prepare control solutions of the identical dilutions of the solvent used for the stock solution for
each cell line to rule out or control for any solvent effects.
The choice of the concentration range to be tested depends on what is known
about the agent in question and the cell line(s) to be tested. If little is known
about the agent to be tested, a high initial concentration should be selected followed by approximately five serial dilutions to cover a range of at least 100fold. To accomplish this, serial dilutions of threefold or higher are needed. The
procedure for the preparation of serial dilutions is shown in Fig. 1. Once initial
experiments define the boundaries of the dynamic range of antiproliferative
activity for a compound, a narrower concentration range should be evaluated.
This can be accomplished by using approximately twofold dilutions. For a
series of twofold dilutions, the highest concentration is selected and prepared in
a volume of at least 2 mL of the appropriate tissue culture medium (higher
volumes are needed if a large number of plates are being set up). If 2 mL is
selected for the final volume of the first tube, then 1 mL of medium is added
to each of the subsequent tubes in the series of dilutions. After the first tube is
73
Fig. 1. Procedure for preparation of serial dilutions.
prepared by adding the correct volume of test agent from its stock solution followed by thorough mixing, 1 mL from the first tube is added to the second
tube. Tube 2 is then mixed, and 1 mL from this tube is added to tube 3 and so
on to reach the required number of tubes (dilutions).
Cells are allowed to grow under incubator conditions (usually 37C with
supplemental CO2). The planned duration of the assay determines the appropriate number of cells to add to each well. For the assessment of the antiproliferative effects of a wide range of compounds, an incubation period of 57 d
is reasonable (see Note 8).
To maximize the dynamic range of this assay, cells that are untreated, which
serve as controls, should be allowed to proliferate until the level of cellular
confluence (the estimated percentage of the total area of the well that is covered
by cells) is 7090%. This parameter should be monitored by inspecting plates
daily using an inverted microscope (see Note 9).
3.2. Processing of Plates for Adherent Cells
Once cells in the untreated wells have reached a confluency of 7090%,
5 mg/mL MTT is added to the plate (see Note 10). The approach to adding
MTT and harvesting plates depends upon whether the cell lines or populations
are adherent or nonadherent in their growth patterns. While adherence to the
plastic surface of microplate wells varies among cell lines, the line of demarcation with respect to these assays is whether trypsin treatment is needed to
release cells from the surface of culture flasks. For such adherent cell types
(usually epithelial and fibroblast-like cells), the processing procedure may be
carried out as outlined next.
1. Prepare a stock solution of MTT in PBS. This final part of the assay does not
require sterile conditions.
2. Remove the microplate from the incubator, and invert it over a container with
mild shaking to remove most of the culture medium. While the plate is still
inverted, remove the remaining medium by blotting it on a small stack of paper
towels.
74
Burton
3. Prepare a working solution of MTT by diluting the stock solution 110 (final
concentration: 0.5 mg/mL) using tissue culture medium (RPMI-1640-based
medium should be used). Using a repeating pipettor, add 100 L of this working
solution to each well.
4. Cover the plate and place back in the incubator for at least 4 h to allow full conversion of the MTT.
5. After incubation, centrifuge the plate to pellet the precipitated formazan dye. To
accomplish this, a standard tabletop centrifuge with an appropriate microplate
carrier is needed. Centrifuge the plates at 1000g for 10 min at ambient temperature, and then invert the plate and blot onto paper towels to remove the bulk of the
liquid.
6. Solubilize the MTT product (formazan) by adding 100 L of DMSO to each well.
To speed the rate of solubilization of the dye, the plate can be returned to the 37C
incubator, with gentle tapping of the plate every 510 min. Solubilization is usually
complete within 30 minutes. An alternative approach to solubilize the MTT product in plates containing adherent cell types is to use an acid-isopropanol solvent as
follows. At the end of the initial culture period, the plate is inverted and blotted to
remove the culture medium. In this case, the MTT working solution should be prepared using RPMI-1640 medium without added fetal bovine serum (FBS) at a
110 dilution as described in step 3 and added to the plate at 100 L/well. After
the requisite 4-h (or more) incubation period, the MTT product is solubilized by the
direct addition of an equal volume (in this case, 100 L) of 0.04 N HCl in isopropanol to each well. Since residual protein may precipitate if the resulting solution is allowed to stand more than 30 min, plates should be read 515 min after
addition of the acid-isopropanol.
3.3. Processing of Plates for Nonadherent Cells

Nonadherent cells cover the gamut of both normal and malignant cell lines
of hematopoietic origin. While these cells settle and distribute fairly evenly
over the surface of the plates, changing medium in plates containing nonadherent cells requires either removing the medium with a multichannel pipettor
or centrifuging the plates followed by inverting and blotting them to remove the
medium. The latter approach is much quicker. To facilitate the process, a modification of the MTT method described in Subheading 3.2. is useful. (This
method may also be used with adherent cells.)
1. Dilute the same MTT stock solution (5 mg/mL) with an equal volume of tissue
culture medium, and add 25 L of this solution directly to each well with a
repeating pipettor.
2. As with the adherent-cell method, return the plates to the incubator for a period of
at least 4 h.
3. Centrifuge the plates at 1000g for 10 min at ambient temperature, followed by
inversion of the plates and blotting of excess medium.
75
4. Use DMSO to solubilize the MTT formazan product as described in Subheading

3.2., step 6.
3.4. Obtaining of OD Values

A standard microplate reader with adjustable wavelengths across the visible
spectrum is required. The detection wavelength for MTT should be between
560 and 570 nm (see Note 11).
The OD values obtained for each set of triplicates corresponding to a specific
concentration of a test agent can then be transferred into a spreadsheet program such as Excel as described below.
3.5. Quantification and Calculation of Inhibitory Activity
The antiproliferative activity of agents is expressed as the 50% inhibitory
concentration (IC50). This may be expressed as a molar concentration or on a
mass-per-volume basis such as micrograms per milliliter or milligrams per liter.
The former mode of expression is preferable; it can be readily calculated using
the molecular mass of the agent in question. Thus, the IC50 is defined as the
concentration of the test agent that results in a 50% decrease in the baseline
or control level of proliferation. Since the readout of the MTT assay is OD, the
baseline for calculation of IC50 values is the OD value of the corresponding
control wells in the MTT assay plate. To perform this analysis, the appropriate
control OD value is essential. This control should be exactly matched for the
dilution of the solvent that was used to prepare the stock solution of the test
agent.
Using the control OD values, the percent inhibition at each concentration of
the test agent can be calculated by dividing the observed OD value by the corresponding control and multiplying by 100 as shown in the following equation:
1 (ODobserved / ODcontrol) 100 = % inhibition
Once the percent inhibition values corresponding to each concentration of the

test agent are calculated, these values can be plotted on a graph to allow for
the interpolation of the IC50 value. To obtain a linear or nearly linear plot, the
percent inhibition values should be log-transformed to obtain a linear equation,
y = ax. The r 2 value for the plot will indicate the goodness of fit for this equation.
4. Notes
1. Cell culture medium is usually supplied according to specific, published formulations, such as RPMI-1640 or Dulbeccos modified Eagles medium. Such widely
used and published media formulations can usually be obtained from any established vendor.
76
Burton
2. Vendors also supply standard stock solutions in sterile cell culturetested formulations such as PBS (note that calcium/magnesium-free PBS is usually preferred)
and trypsin/EDTA (needed to detach adherent cells). The latter two solutions as
well as other additives are supplied in concentrated stock solutions that either are
diluted in sterile deionized water (in the case of PBS concentrates) or sterile PBS
(in the case of trypsin/EDTA concentrates), or are added directly to media to
achieve the desired final concentrations (in the case of antibiotics and L-glutamine).
For any solutions that require dilution in deionized water, a source of highly pure
water from which contaminating ions and volatile organics (activated charcoal
adsorbed and exhibiting 18 m of resistance) have been sufficiently removed is
critical.
3. Cell lines can be either obtained from individual investigators or purchased from
repositories such as the American Type Culture Collection (ATCC; Manassas,
VA). The published requirements of the cell lines (provided by the investigator
who established the cell line or by the repository [e.g, ATCC]) will determine the
additives that are needed for the culture medium. Typically, basic cell culture
medium needs to be supplemented with FBS; antibiotics; L-glutamine; and, potentially, other additives. The specific preferred vendor from which to purchase these
components is specified by the investigator or the repository such as ATCC.
4. For drugs that are not water soluble, organic solvents are required to prepare stock
solutions. DMSO is the most useful and least toxic of these solvents. Other alternatives are DMF and ethanol. These should be obtained either in tissue culturetested formulations or in their most pure forms from any established supplier
of fine chemicals. Although microbial contamination is unusual in organic solvents, it is still important that they be sterile filtered. To achieve an acceptable
level of sterility, solutions must be filtered through filters with a cutoff of 0.45
or less. For these organic solvents, a solvent-resistant syringe-tip filter is preferred. For filtration of other aqueous solutions, syringe-tip or vacuum-assisted
filters may be used, depending on the volume of the solution that needs sterile filtration. Obtain information about the stability of the test agent or drug. If there is
a question, prepare a new stock solution of it. Be sure to control as much as possible for the effects of the organic solvents used to prepare drug stock solutions on
the growth of the cells by setting up control wells containing equal dilutions of
solvent without the test agent or drug.
5. Cell lines should be maintained in culture medium to which they are adapted,
and that supports consistent cellular proliferation. They should have maximum
viability (>90% at the start of the assay) and be in the logarithmic phase of growth
at the time of harvest. For most cell lines, this is either a basal or an enriched
medium (e.g., RPMI-1640, Dulbeccos modified Eagles medium, or Iscoves
modified Eagles medium) that is supplemented with bovine or other serum or
purified serum components as well as antibiotics and L-glutamine. The culture
medium in which cell line(s) are routinely carried should be used for setting up
the MTT assay.
77
6. The final volume of medium in each well should be 200 L. It is common to

assess binary combinations of agents to examine the interaction of the two agents,
which may be additive, synergistic, or antagonistic. In most instances, no more
than two agents need to be added to each well. Thus, it is convenient to first plate
the cells in a volume of 100 L of culture medium. This allows for a single agent
to be added to each of the wells in a volume of 100 L or, for two agents, in a
volume of 50 L for each agent.
7. To increase precision and reduce the setup time for assays, it is very helpful to utilize repeating pipettors. Commonly used repeating pipettors are those that are
similar in design to one of the original instruments of this type, the Eppendorf
repeater. This pipettor requires specialized tips, which are often referred to as
Combi-Tips. These must be sterile and are usually available in individually packaged and sterile form. These tips may also be used to add the agents to be tested
to the individual wells, but it is more economical over time to use an electronic
pipettor with a repeat pipetting mode. A 1000-L size of such a pipettor will
allow for the precise delivery of up to 20 separate 50-L vol with a single standard pipet tip. A standard manual pipettor may also be used.
8. For plates being incubated for 5 d or longer, an incubator chamber that is humidified should be used to minimize evaporative effects. To reduce the tendency
toward evaporation in the wells on the perimeter of the plate, the outer columns
and rows should not be used. To further reduce evaporation of the inner wells,
200250 L/well of sterile water or PBS should be added to the outer perimeter
wells.
9. For most cell lines, a total of 39 104 cells/well will result in the target level of
confluence of 7090%. To reach the target number of cells within the planned
time frame (57 d), the doubling time of the cell line(s) needs to be estimated
from their current growth and maintenance characteristics. Typical doubling times
for most mammalian cell lines range from 20 to 60 h. Thus, the number of cells
that typically needs to be added to the wells of the plate at the start of the assay
ranges from 1.5 to 7.5 103/well. Estimating an appropriate cell number for initial plating often requires repeat experiments to calibrate this parameter.
10. Observe the plate under an inverted microscope on a daily basis to assess
the level of confluence and the color of the medium. If the medium has taken on
a yellowish color, it is depleted, and the MTT (or other dye) should be added
soon.
11. If it is not possible to achieve maximum control OD values of >0.9 with MTT, try
substituting one of the dyes that yields water-soluble end product (formazan),
such as MTS or WST-1.
References
1. Gabridge, M. G. and Polisky, R. B. (1976) Quantitative reduction of 2,3,4-triphenyl
tetrazolium chloride by hamster trachea organ cultures: effects of Mycoplasma
pneumoniae cells and membranes. Infect. Immun. 13, 8491.
78
Burton
2. Douglas, W. H., McAteer, J. A., Dellorco, R. T., and Phelps, D. (1980) Visualization of cellular aggregates cultured on a three dimensional collagen sponge
matrix. In Vitro 16, 306312.
3. Alley, M. C., Uhl, C. B., and Lieber, M. M. (1982) Improved detection of drug
cytotoxicity in the soft agar colony formation assay through use of a metabolizable
tetrazolium salt. Life Sci. 31, 30713078.
5563.
5. Carmichael, J., DeGraff, W. G., Gazdar, A. F., Minna, J. D., and Mitchell, J. B.
(1987) Evaluation of a tetrazolium-based semiautomated colorimetric assay:
6. Twentyman, P. R. and Luscombe, M. (1987) A study of some variables in a tetrazolium dye (MTT) based assay for cell growth and chemosensitivity. Br. J. Cancer
56, 279285.
7. Vistica, D. T., Skehan, P., Scudiero, D., Monks, A., Pittman, A., and Boyd, M. R.
(1991) Tetrazolium-based assays for cellular viability: a critical examination of
selected parameters affecting formazan production. Cancer Res. 51, 25152520.
8. Scudiero, D. A., Shoemaker, R. H., Paull, K. D., et al. (1988) Evaluation of a soluble tetrazolium/formazan assay for cell growth and drug sensitivity in culture
using human and other tumor cell lines. Cancer Res. 48, 48274833.
9. Roehm, N. W., Rodgers, G. H., Hatfield, S. M., and Glasebrook, A. L. (1991) An
improved colorimetric assay for cell proliferation and viability utilizing the tetrazolium salt XTT. J. Immunol. Methods 142, 257265.
10. Cory, A. H., Owen, T. C., Barltrop, J. A., and Cory, J. G. (1991) Use of an aqueous soluble tetrazolium/formazan assay for cell growth assays in culture. Cancer
Commun. 3, 207212.
11. Buttke, T. M., McCubrey, J. A., and Owen, T. C. (1993) Use of an aqueous soluble tetrazolium/formazan assay to measure viability and proliferation of lymphokine-dependent cell lines. J. Immunol. Methods 157, 233240.
12. Ishiyama, M., Shiga, M., Sasamoto, K., Mizoguchi, M., and He, P. (1993) A new
sulfonated tetrazolium salt that produces a highly water-soluble formazan dye.
Chem. Pharm. Bull. 41, 11181122.
13. Takenouchi, T. and Munekata, E. (1995) Trophic effects of substance P and betaamyloid peptide on dibutyryl cyclic AMPdifferentiated human leukemic (HL-60)
cells. Life Sci. 56, PL479PL484.
14. Teruya, K., Yano, T., Shirahata, S., et al. (1995) Ras amplification in BHK-21 cells
produces a host cell line for further rapid establishment of recombinant protein
hyper-producing cell lines. Biosci. Biotechnol. Biochem. 59, 341344.
15. Ishiyama, M., Tominaga, H., Shiga, M., Sasamoto, K., Ohkura, Y., and Ueno, K.
(1996) A combined assay of cell viability and in vitro cytotoxicity with a highly
water-soluble tetrazolium salt, neutral red and crystal violet. Biol. Pharm. Bull.
19, 15181520.
8
Histoculture Drug Response Assay
to Monitor Chemoresponse
Shinji Ohie, Yasuhiro Udagawa, Daisuke Aoki, and Shiro Nozawa
Summary
We provide a detailed explanation of the procedure of the histoculture drug response assay
(HDRA) with 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) end point
among several modified HDRA procedures. Fresh surgical specimens are cut into approx 1- to
2-mm3 pieces and put on a gelatin sponge infiltrated with culture medium containing a test drug.
After incubation for 7 d, cell viability is assessed by the MTT assay. HDRA uses cancer tissue
fragments with cells growing in three dimensions, with maintenance of intercellular contact and
interactions with stromal cells. Therefore, it seems that HDRA can assess the sensitivity of tumor
cells to anticancer drugs in conditions similar to those in vivo and, consequently, shows high prediction rate.
Key Words
Histoculture drug response assay; chemoresponse; chemosensitivity; surgical specimen;
gelatin sponge; collagen matrix.
1. Introduction
Hoffman and colleagues developed the histoculture drug response assay
(HDRA) and applied it to three-dimensional culture of tissue fragments of
tumor using a collagen gel matrix and a [3H]thymidine uptake end point (15).
Many conventional drug sensitivity tests reported use isolated tumor cells
obtained after enzymatic digestion. By contrast, the HDRA technique uses
cancer tissue fragments with cells maintained in their native tissue architecture, which can grow in three dimensions, with maintenance of intercellular
contact and interactions with stromal cells. Therefore, HDRA can assess the
sensitivity of tumor cells to anticancer drugs in conditions similar to those present in vivo. Kubota et al. (6) demonstrated that in vitro response in HDRA
79
80
Ohie et al.
with the [3H]thymidine end point is correlated with survival of patients with
gastric cancer. Furukawa et al. (79) demonstrated that the response in HDRA
with 3-(4,5-dimethyl-2-thiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT)
reduction as an endpoint is correlated with patient drug response and survival.
Ohie et al. (10) modified the HDRA procedure to be able to perform it more
easily without measuring the weights of tumor fragments aseptically. Eightyfive (97%) cases were evaluable, and a high prediction rate (87%) was found
when HDRA results for CDDP were compared to clinical response to combination chemotherapy containing CDDP in patients with ovarian cancer (10).
Of several modified HDRA procedures, we provide a detailed explanation of
the HDRA procedure with MTT as the end point that we have used for assessing the drug sensitivities of ovarian cancer specimens.
2. Materials
1. Gelatin sponge (Gelfoam Sterile Sponge no. 100, Pharmacia; or Spongostan
70 50 10 mm Standard, Health Design, Rochester, NY) (see Note 1).
2. Hams F-12 medium (Gibco-BRL, Gaithersburg, MD) (see Note 2).
3. Heat-immobilized fetal bovine serum (FBS).
4. Hanks solution.
5. Phosphate-buffered saline (PBS).
6. MTT.
7. Sodium succinate.
8. Dimethyl sulfoxide (DMSO).
9. Membrane filter (pore size: 0.2 m diameter).
10. 24-Well culture plates.
11. 96-Well culture plates.
12. Antibiotics (kanamycin, penicillin, gentamycin, and so on) (see Note 3).
13. Surgical knife (Disposable Scalpel no. 11; Feather Safety Razor, Medical
Division).
14. Surgical scissors.
15. Microplate reader.
16. Clean bench (laminar flow chamber).
17. CO2 incubator (5% CO2, 37C).
18. Gelatin sponge (Gelfoam; Pharmacia).
3. Methods
The HDRA procedure is shown in Fig. 1. From excision of specimens to
generation of MTT-formazan in cells, specimens must be kept wet and sterile.
3.1. Preparation of Tissue Specimens
1. Obtain fresh surgical specimens in the operating room as aseptically as possible.
2. Wash the surgical specimens with Hanks solution containing antibiotic (100 U/mL
penicillin, 100 g/mL gentamycin, and the like), if necessary (see Note 3).
HDRA to Monitor Chemoresponse
81
Fig. 1. HDRA procedure.
3. Cut the tissue specimens with sterile surgical scissors into approx 1- to 2-mm3
pieces in Hanks solution on a clean bench (see Note 4).
3.2. Preparation of Assay Plate and Culture

1. Cut the gelatin sponge into approx 1-cm3 pieces with a sharp surgical knife under
sterile conditions on a clean bench.
82
Ohie et al.
Fig. 2. Illustration of HDRA. The side and top views of 1 well of a 24-well plate are
shown.
2. Place one piece of cut gelatin sponge in each well of a 24-well culture plate.
3. Drugs are dissolved at varying concentrations in Hams F-12 medium containing
20% heat-immobilized fetal bovine serum and kanamycin (80 g/mL). Then sterilize this solution by filtration with a 0.2-m-pore membrane filter.
4. Pour medium containing the test drug into each well of the plate (1 mL/well).
Use four wells at each concentration of test drug (n = 4) (see Note 5).
5. Allow the culture medium to infiltrate sufficiently (see Note 6).
6. Place the pieces of cancer tissue on the gelatin foam (see Note 7).
7. Incubate the plate for 7 d at 37C under 5% CO2 (see Fig. 2).
3.3. Estimation of Number of Live Cells

Cell viability is assessed by MTT assay at the end of incubation.
1. MTT is dissolved in PBS (5 mg/mL) containing 100 mM sodium succinate.
2. Aseptically add MTT solution to each well of the 24-well plate (100 L/well).
Sink the fragment in the medium completely (see Fig. 1).
3. Incubate the plate for about 4 h at 37C under 5% CO2.
4. Stain the pieces of tumor tissue violet with generation of MTT-formazan in proportion to the activities of succinyldehydrogenase in cancer cells.
5. Transfer the stained tumor fragments to another new 24-well plate containing
DMSO at a rate of 1 mL/well to extract the MTT-formazan (see Fig. 3).
6. Place the extract in a 96-well microplate (100 L/well).
7. Measure the absorbance of each well at 540 nm using a microplate reader.
8. Measure the weights of the residual tumors after extraction of MTT-formazan.
9. Calculate the absorbance per gram of residual tumor (see Note 8).
3.4. Data Processing and Analysis

1. Calculate the tumor inhibition rate (%), relative to the untreated control group,
using the following equations:
83
Fig. 3. Twenty-four-well plate after extraction of MTT-formazan with DMSO from

ovarian cancer fragments treated with CDDP.
T/C =
Absorbance/g tumor in treated group
Mean of absorbance/g tumor in untreated group
Tumor inhibition rate (%) = (1 T/C) 100

2. At each concentration, average the inhibition rates for the four wells and draw
the dose-response curve (see Note 5).
3. Calculate the concentration yielding 50% inhibition of tumor growth (IC50).
3.5. Estimation of Optimal Cutoff IC50 Value

1. Draw the cumulative efficacy rate curve for the IC50 values as determined by
HDRA. The vertical axis represents the cumulative efficacy rate, and the horizontal axis is the IC50 value.
2. Obtain the equation representing an approximate curve by curve fitting using computer processing.
3. Apply the clinical response rate of the tested drug to this equation to calculate the
optimal cutoff IC50 value (see Note 9).
3.6. Evaluation of Results

When the IC50 value estimated by HDRA is lower than the cutoff IC50 value,
the case is determined to be sensitive to test drug and expected to exhibit complete or partial response clinically, whereas cases not sensitive to test drug with
84
Ohie et al.
IC50 values higher than the cutoff IC50 value are not expected to exhibit clinical response.
If one cannot obtain a specimen of sufficient volume and are obliged to use
a simplified HDRA procedure, one must set the concentration of test drug at
the cutoff IC50 value. When the absorbance per gram of tumor in the drugtreated group is lower than 50% of absorbance per gram of tumor in the control group, the case is determined to be sensitive to test drug.
4. Notes
1. The gelatin sponges introduced here are sterile and dry. If one uses pure collagen
gel containing water prepared in the laboratory, adjustment of the final concentration of drug is extremely difficult and too expensive.
2. Dr. Hoffman has used Eagles minimal essential medium and Dr. Kubota and
many researchers have used RPMI-1640 medium for gastric cancer, colon cancer,
liver cancer, and so on. Before starting HDRA on a large scale, one must preliminarily determine a medium suitable for examination of the cancer concerned.
3. We have used F-12 medium containing kanamycin (80 g/mL) for culture and
transport from the operating room to the laboratory. During HDRA for ovarian
cancer and uterine cancer, contamination with bacteria has only rarely been
observed. However, for specimens at greater risk of contamination with bacteria,
such as colon cancer specimens, surgical specimens should be washed with
Hanks solution containing penicillin (100 U/mL) and gentamycin (100 g/mL)
and the culture medium containing these two antibiotics should be used at the
same concentrations. However, the effects of such antibiotics on cell growth and
interaction among the antibiotics and test drug must be considered.
4. When the tissues are cut to fragments of the same volume as precisely as possible, good results are obtained although absorbance is normalized to fragment
weight. This appears to be due to diffusion of drug in the tissue fragment.
5. To maintain a high level of confidence in the results obtained, we recommend
using four wells at each concentration of test drug (n = 4). Values for out of range
can then be omitted. For the same reason, one must draw the dose-response curve
using inhibition rates for four or more drug concentrations and should confirm that
inhibition of growth of cells is caused by the test drug.
6. In many cases, infiltration of the culture medium into the gelatin sponge requires
a long period of time. In this case, time can be saved by putting the plate in a
37C CO2 incubator and turning the gelatin sponge upside down in the same well
aseptically more than 10 min after the medium has been poured.
7. If one has a clean room pressurized with air filtered through high-efficiency particle filters and can weigh the fragments precisely and aseptically, several cut
fragments of tissue can be placed on a gelatin sponge up to about 10 mg/well of
24-well plates at the start of the HDRA. If this procedure is chosen, an MTT
solution with collagenase type I (Sigma, St. Louis, MO) is applied and incubated
at 37C under 5% CO2 for 8 h. Then the gelatin sponge is digested and MTT-
85
formazan is generated. The medium is carefully removed with a micropipet, and

1.0 mL of DMSO is poured to extract the MTT-formazan.
8. We have confirmed a good correlation between fresh tissue fragments and the
residual tumors after extraction of MTT-formazan for results obtained for ovarian
and uterine cancers, and we recommend preliminarily confirming the correlation
for results between fresh tissue fragments and residual tumor.
9. Furukawa et al. (11) reported that 2 g/mL for MMC, 300 g/mL for
5-fluorouracil, 15 g/mL for ADM, and 20 g/mL for CDDP are the optimal
cutoff IC50 values for gastric and colon cancers. Ohie et al. (10) reported that
25 g/mL for CDDP was the optimal cutoff IC50 value for ovarian and uterine cancers. These IC50 values are very high compared to the peak blood levels of drugs.
Maehara et al. (12) reported that the reason for this is that the activity of succinate
dehydrogenase measured by the MTT procedure remained even after cells died.
References
1. Freeman, A. E. and Hoffman, R. M. (1986) In vivo-like growth of human tumors
in vitro. Proc. Natl. Acad. Sci. USA 83, 26942698.
2. Vescio, R. A., Redfern, C. H., Nelson, T. J., Ugoretz, S., Stern, P. H., and Hoffman, R. M. (1987) In vivo-like drug responses of human tumors growing in
three-dimensional gel-supported primary culture. Proc. Natl. Acad. Sci. USA 84,
50295033.
3. Hoffman, R. M., Connors, K. M., Meerson-Monosov, A. Z., Herrera, H., and Price,
J. H. (1989) A general native-state method for determination of proliferative capacity of human normal and tumor tissues in vitro. Proc. Natl. Acad. Sci. USA 86,
20132017.
4. Vescio, R. A., Connors, K. M., Kubota, T., and Hoffman, R. M. (1991) Correlation
of histology and drug response of human tumors grown in native-state threedimensional histoculture and in nude mice. Proc. Natl. Acad. Sci. USA 88,
51635166.
5. Hoffman, R. M. (1991) Three-dimensional histoculture: origins and application in
cancer research. Cancer Cells 3, 8692.
6. Kubota, T., Sasano, N., Abe, O., Nakao, I., Kawamura, E., Saito, T., Endo, M.,
Kimura, K., Demura, H., and Sasano, H. (1995) Potential of the histoculture drug
response assay to contribute to cancer patient survival. Clin. Cancer Res. 1,
15371543.
7. Furukawa, T., Kubota, T., Watanabe, M., et al. (1992) High in vitroin vivo correlation of drug response using spongegel-supported three-dimensional histoculture
and the MTT end-point. Int. J. Cancer 51, 489498.
8. Furukawa, T., Kubota, T., Watanabe, M., et al. (1992) Chemosensitivity testing of
clinical gastrointestinal cancers using histoculture and the MTT end-point. Anticancer Res. 12, 13771382.
9. Furukawa, T., Kubota, T., and Hoffman, R. M. (1995) Clinical applications of the
histoculture drug response assay. Clin. Cancer Res. 1, 305311.
86
Ohie et al.
10. Ohie, S., Udagawa, Y., Kozu, A., Komuro, Y., Aoki, D., Nozawa, S., Moossa, A. R.,
and Hoffman, R. M. (2000) Cisplatin sensitivity of ovarian cancer in the histoculture drug response assay correlates to clinical response to combination
chemotherapy with cisplatin, doxorubicin and cyclophosphamide. Anticancer Res.
20, 20492054.
11. Furukawa, T., Suzuki, K., Yuasa, S., Izumo, M., Kozakai, K., Yano, T., Harada,
N., and Kubota, T. (1996) Clinical application of histoculture drug response assay
(HDRA) for gastrointestinal cancers with reference to cumulative efficacy rate
curve. J. Jpn. Soc. Cancer Ther. 31(2), 116126.
12. Maehara, Y., Anai, H., Tamada, R., and Sugimachi, K. (1987) The ATP assay is
more sensitive than the succinate dehydrogenase inhibition test for predicting cell
viability. Eur. J. Cancer Clin. Oncol. 23, 273276.
9
In Vitro Testing of Chemosensitivity
in Physiological Hypoxia
Rita Grigoryan, Nino Keshelava, Clarke Anderson,
and C. Patrick Reynolds
Summary
Highly aggressive, rapidly growing tumors are often hypoxic, owing to an inadequate supply
relative to consumption of oxygen (O2) in the expanding tumor mass, or growth in tissues with
physiologically low O2 concentrations (such as bone marrow). Selection of tumor cells that can
grow or survive under hypoxic conditions appears from both experimental and clinical studies to
impact tumor progression, response to therapy, and to increase resistance to radiation and to certain cytotoxic drugs. Therefore, the predictive value of preclinical testing of anticancer agents in
cell culture might be improved by conducting testing in conditions of physiological hypoxia.
We review the impact of hypoxia on anticancer drug cytotoxicity and the methods used in our
laboratory to asses the cytotoxic activity of single antineoplastic drugs under conditions of physiological hypoxia.
Key Words
Tumor hypoxia; drug resistance; hypoxia-targeted therapy; cytotoxicity assay; digital image
microscopy.
1. Introduction
Solid tumors often have areas in which circulation is compromised because
of structurally disorganized blood vessels and tumor cells that grow faster than
the developing tumor capillary network (1,2). The fraction of solid tumors that
are hypoxic can vary from 0.2 to >50% (3,4). The degree of hypoxia in tumors
is highly variable, with the PO2 generally <1030 mmHg (13% O2), in contrast
to a PO2 of 5080 mmHg in most normal tissues (4) (Table 1). The microenvironment found in hypoxic tumors leads to low extracellular pH from lactic acid,
low glucose, genomic instability, and selective pressure to adapt and survive
(5,6). In addition, hypoxia is associated with local increases in tumor interstitial
87
88
Grigoryan et al.
Table 1
O2 Tensions in Various Tissues, Tumors, and Cell Cultures
PO2 (mmHg)a
Normal tissues
Murine brain
Murine muscle
Bone marrow (children)
Bone marrow (adult)
Normal liver
Normal breast tissue
Subcutaneous tissue
Normal cervix
Normal head and neck tissue
Tumors
Breast carcinoma
Solid tumors
Murine Fsall fibrosarcoma
Cancer of cervix
Head and neck cancers
Soft-tissue sarcomas
Neuroblastoma xenograft
in athymic (nu/nu) mouse
Tissue Culture
RPMI-1640 medium with
10% FBS incubated at 2% O2
RPMI-1640 medium with 10%
FBS incubated at 5% O2
RPMI-1640 medium with
10% FBS incubated at 20% O2
~60
~42
~44
~4050
~55
~65
~38
~48
~43
~28
~2.5
~5
12
10
10 and 22
References
Unpublished data
Unpublished data
28
21
29
30,31
20,32,33
33
33
30
11,20,34
33
26,33
20,32
1,20
~~7
Unpublished data
~12.6
Unpublished data
~47
Unpublished data
149
Unpublished data
a
PO2 values vary from study to study based on the methods used for PO2 level measurements.
However, these relatively constant values are from studies done using similar methods to measure PO2 levels in normal and tumor tissues (computerized PO2 and polaragraphic electrodes). The
% O2 was calculated based on correlations between mm Hg and % O2, done by Stone (31). FBS,
fetal bovine serum.
fluid leading to microthrombi and increased blood viscosity (6). This microenvironment has prognostic implications, because cells in hypoxic areas are less
vulnerable to ionizing radiation and cytotoxic drugs, and tumors with substantial hypoxia metastasize more efficiently (1,2). Hypoxia is a prognostic variable
for unfavorable outcome, because it provides a mechanism by which tumors
can selectively promote a more aggressive phenotype, recruit a nutrient supply,
Chemosensitivity Testing in Hypoxia
89
Fig. 1. Hypoxia (2% O2) significantly antagonizes L-PAM cytotoxicity. The effect is
most pronounced in the p53-functional SMS-SAN cell line (established at diagnosis)
but is also seen in the TP53-mutated multidrug resistant cell line, SK-N-BE(2), estabalished at relapse after chemotherapy (p < 0.01). Similar results were obtained with
other neuroblastoma cell lines.
and may promote essential metabolic adaptations that improve tumor survival
(1,2,7). Hypoxia is also considered to protect cells in solid tumors from
chemotherapeutic agents (3,4,79), although there are only a limited number of
direct studies supporting this concept, mostly with doxorubicin (9,10) and
methotrexate (10). This induced resistance can be explained by decreased drug
concentrations, because of limited drug penetration into tumor masses;
decreased drug activity in areas where tumor cells are slowly growing or nonproliferating due to hypoxia and the associated alterations in nutrient supply
and utilization; and direct antagonism of drug action by hypoxia and the associated acidosis (11).
Many chemotherapeutic agents are dependent on cellular oxygenation for
maximal efficacy. Cytotoxic alkylating agents, such as the nitrogen mustard
alkylating agent melphalan (L-PAM), are a class of chemotherapeutic drugs
that act by transferring alkyl groups to DNA during cell division. Following
this, the DNA strand breaks or crosslinking of the two strands occurs, preventing subsequent DNA synthesis (12). Teicher et al. (13) showed that tumor cells
grown in normoxic conditions were more sensitive to L-PAM than in hypoxic
conditions, and we have observed similar results (Fig. 1). Under hypoxic
90
Grigoryan et al.
conditions, alkylating agents may have less activity for a variety of reasons,
including an increase in intracellular glutathione, which may compete with the
target DNA for alkylation (11,14).
Hypoxia also causes cells to slowly cycle and induces pre-DNA-synthetic
(pre-S-phase) arrest in cells (1,7). Therefore alkylating agents have reduced
efficacy against slow cycling tumor cells under hypoxic conditions, because
they have increased activity in apoptosis induction during highly active cell
proliferation (14). Other drugs directly affected by hypoxia include the
podophyllotoxin derivative etoposide (12,15), presumably due to free-radical
scavengers, dehydrogenase inhibitors, and dehydrogenase substrates, which prevent the formation of single-strand breaks and decrease the cytotoxic effect of
etoposide (12). DNA-damaging chemotherapeutic agents may also have compromised function due to increased activity of DNA repair enzymes under
hypoxic conditions (16).
An example of a drug with significant antagonism by hypoxia is the glutathione (GSH) synthesis inhibitor buthionine sulfoximine (BSO), which as a
single agent was highly cytotoxic against a wide range of neuroblastoma cell
lines under standard cell culture conditions (20% O2) (17). Physiological
hypoxia as found in solid tumors (2% O2) abrogates the cytotoxic effect of
BSO for neuroblastoma by diminishing the increase in reactive oxygen species
(ROS) caused by BSO (18), and that antagonism appeared to be more pronounced in cell lines lacking functional p53 (18).
The difference between hypoxic cancer cells and normal cells could give
researchers a basis on which to design drugs or drug combinations. Cytotoxic
drugs active in hypoxia are currently under investigation (19), including the
bioreductive agent tirapazamine (TPZ) (2). In the presence of oxygen, TPZ is
a stable, nontoxic parent molecule. However, TPZ is transformed by intracellular reductases into a highly reactive and toxic radical in the presence of
hypoxia (2). Because TPZ is relatively ineffective at killing aerobic cells, this
agent alone is not well studied (2). We have shown that TPZ synergistically
reversed inhibition of BSO-mediated cytotoxicity in hypoxia for human neuroblastoma cell lines, by increasing the formation of ROS, decreasing mitochondrial membrane potential (m), depleting GSH, and increasing apoptosis
(18). These data suggest that combining BSO with TPZ could have clinical
activity against neuroblastoma in hypoxic sites (18).
Oxygen levels are typically quite heterogeneous, both among patients and
within individual tumors. The oxygenation status of tissues has primarily been
measured using either polarographic oxygen electrodes or biochemical techniques that rely on antibody detection of nitroimidazole-based adducts in
hypoxic tissue (20). A comparison of tissue oxygenation levels to clinical tumor
behavior suggests an advantage for tumors in hypoxia. At diagnosis, neuro-
91
blastoma most commonly develops in relative hypoxic sites (e.g., the PO2 in
bone marrow is 4050 mmHg) (21) (Table 1) such as the retroperitoneum,
bone, and bone marrow. By contrast, recurrent neuroblastomas more commonly
metastasizes to high-oxygen-tension sites such as brain and lungs. Therefore,
low oxygen tension could provide a sanctuary site for tumors leading to selection for oxygen tolerance and drug resistance (22).
Multiple lines of evidence point toward profound differences in the behavior
of tumor cells under conditions of high levels of oxygen relative to conditions
of physiological hypoxia. This suggests that determination of anticancer drug
activity in standard culture conditions (i.e., 20% O2, which is 10-fold greater
than the O2 levels of many tumors, and 4- to 5-fold greater than bone marrow)
may cause an artifactually high estimation of drug activity. For that reason, it
is important to validate drug activity in hypoxic conditions, and such an
approach may increase the predictive value of cell culture drug testing. Here,
we describe methods used to evaluate cytotoxic activity of various drugs and
drug combinations in hypoxic conditions.
2. Materials
1. Fluorescein diacetate (FDA) (Sigma, St. Louis, MO): 1 mg/mL solution prepared
in dimethyl sulfoxide, aliquoted into 1.5-mL Eppendorf tubes, and kept at 20C
in the dark. Avoid repeated thawing and refreezing.
2. Eosin Y: 1% stock solution (w/v) prepared in 0.9% NaCl and kept in an amber
bottle at room temperature.
3. DIMSCAN system (see Chapter 12). This system consists of an inverted microscope, a stepper motor scanning stage, a stage controller, a charge-coupled device
(CCD) camera, and a Pentium 4 microcomputer (Microsoft Windows 2000)
running the main application software (DIMSCAN 3.0, developed at Childrens
Hospital Los Angeles, CA), which controls stage movement and processes CCD
camera images (23,24). An inverted microscope, Olympus IX50, is equipped with
a 103-W mercury vapor lamp (HBO-103 W/2), optical filters (Omega Optical
[Woburn, MA] XF22 filter set for FDA or BCECF [excitation: 490 nm; emission:
525 nm] and Omega Optical XF05 filter set for Hoechst 33342 [excitation:
345 nm; emission: 475 nm]) and 4 high N.A. objective lens. It is also equipped
with a motorized Prior Pro Scan stage with two stepper motors for stage movements in the horizontal plane (x and y) and one stepper motor for focusing
(z-axis). The stage controller communicates with the computer through a serial
port. A Qimaging Microimager II CCD camera is attached to the standard trinocular head with an 80/20 beam splitter. Maximum resolution of the camera CCD
chip is 1024/768 pixels, and the camera has internal cooling, enabling long-term
use without degrading image quality. The camera is connected to the PC through
IEEE1394 FireWire interface, which enables a high-rate data transfer to the PC.
4. Data analyzer software. Following incubation with drugs and scanning of microwell
plates by DIMSCAN, the Data Analyzer software (developed at Childrens Hospital
92
Grigoryan et al.
Los Angeles, CA) calculates fractions of affected cells (Fa) (Fa = 1 RFdrug /
RF control), survival fractions (calculated as RFdrug / RF control), standard deviations (SDs), confidence levels, and standard errors using the relative fluorescence (RF) values obtained during the DIMSCAN. SigmaPlot (Jandell, San
Rafael, CA) can be used to create dose-response graphs, and Excel (Microsoft
Office) to determine fractions of affected cells, survival fractions, and SDs.
5. Nitrogen cylinders (Praxair, Danbury, CT): Refrigerated liquid nitrogen cylinders
containing 160 L (3690 ft3) liquid nitrogen at intracylinder pressure of 230 psi
deliver nitrogen gas to the incubators at 1015 psi. Custom gas mixtures with
0.6%5% oxygen, 5% CO2, and the balance N2 are provided in cylinders with 25 ft3
under 2000 psi. Nitrogen cylinders must be equipped with proper regulators in
order to deliver nitrogen or a mixture of gases at a desirable pounds per square
inch (psi).
6. Tissue culture incubators for O2 level control: There are three major approaches to
maintaining a designated hypoxic environment inside incubators: using incubators
with an integrated O2 control system (Fig. 2), placing a sealed modular incubation
chamber with a separate CO2 and O2 control system inside a standard host incubator (Fig. 3), and placing inside the host incubator a hermetically sealed chamber that has been flushed with the appropriate O2 concentration (Fig. 4). We
present here examples of these three different approaches.
a. CO2- and O2-Controlled water-jacketed incubator. The following applies to the
Model 3110 Incubator from ThermoForma (Marietta, OH). Other vendors, such
as Sanyo (Itasca, IL), also make O2-regulated incubators. The Forma 3110
water-jacketed incubator has separate regulators for O2 and for CO2, allowing
the user to define the exact atmospheric conditions desired. The O2 setpoint
range is 121%. The O2 control sensor is located on the blower scroll plate in
the chamber unit (see Fig. 2F). This sensor is a fuel cell that puts out a linear
millivolt signal based on O2 content of the chamber. The O2 sensor fuel cell
depletes over time, depending on required O2 levels; therefore, the system
should be calibrated at least every 6 mo. Two methods are available to calibrate
the O2 system: the preferred method is calibration of the system to the known
ambient O2 value of 20.7%, which checks the life of the sensor; the second
method allows the system to be calibrated to an independent reference instrument by entering an offset.
The Forma 3110 incubator also has a CO2 thermal conductivity (T/C) sensor.
All T/C CO2 cells are precalibrated at the factory at 37C, high humidity, and
10% CO2. Therefore, if a temperature set point of 37C is entered, the humidity
pan filled, and the CO2 control is to run between 0 and 10% with a T/C CO2
sensor, the CO2 set point may be entered immediately. This sensor is not only
affected by the quantity of CO2 present, but also by the air temperature and the
water vapor present in the incubator atmosphere. In monitoring the effects of
CO2, air temperature and absolute humidity must be held constant so that any
change in thermal conductivity is caused only by a change in CO2 concentration.
93
Fig. 2. CO2 water-jacketed oxygen and eight CO2-regulated incubator (Model 3110
from ThermoForma, Marietta, OH), which has eight smaller inner doors separating the
incubator into eight chambers. (A) Temperature display; (B) % O2 display; (C) humidity pan; (D) water jacket fill port; (E) HEPA filter; (F) O2 sensor.
The Forma 3110 incubator has a HEPA filter (see Fig. 2E) to minimize
accumulation of microbial contaminants inside the incubator. During automatic
calibration, the CO2 display is blanked out and HEPA-filtered room air is
pumped through the CO2 sensor. Replacement of the HEPA filter can be set for
a specific amount of time, from 1 to 12 mo of actual unit running time. When
the allotted time has run out, REPLACE HEPA appears in the display and
the visual alarm flashes.
For best operation of the incubator, sterilized distilled, demineralized, or
deionized water should be used in the humidity pan (see Fig. 2C). Water
94
Grigoryan et al.
Fig. 3. PROOX model 110 (Reming, Redfield, NY). (A) CO2 controller; (B) O2
controller; (C) hypoxia chamber inside a host incubator; (D) outer door; (E) inner door.
should always be sterilized or treated with a decontaminant that is safe for use
with stainless steel and nonvolatile so as not to cause toxicity to the cells. The
humidity pan should be filled to within 0.5 in. of the top with sterile, distilled
water and placed directly on the incubator floor (see Fig. 2C) to ensure optimum humidity and temperature response. In addition, the water jacket should be
filled with 11.7 gal (42.5 L) of distilled water having a resistance range of 50 K
to 1 m/cm using silicone tubing connected directly to the fill port (Fig. 2D).
In addition to the standard outer and inner glass door, the 3110 incubator
can be equipped with eight smaller inner doors, providing separate inner doors
for each shelf of the incubator (Fig. 2), separating the incubator into eight
chambers. The advantage of using such an inner door kit is to decrease the
gas-mixture disturbance, otherwise inevitable with repeated opening and closing of the incubator door. Incubators should be cleaned and sterilized every
2 to 3 mo.
b. O2-controlled insert chambers for incubators. The PROOX model 110 from
Reming (Redfield, NY) (Fig. 3) is a versatile and compact gas oxygen controller for oxygen-sensitive work. The PROOX chamber can be placed inside
95
Fig. 4. O2-controlled chamber (Billups-Rothenberg, Del Mar, CA). The sealed,

humidified, modular incubation chamber is flushed with a gas mixture and sealed to
obtain hypoxic conditions. (A) Inlet and outlet tubes; (B) sterile, distilled water.
a standard host incubator, and the O2 and CO2 controller panels located outside
the host incubator. The host incubator provides the necessary 37C temperature. An O2 sensor measures oxygen tension inside the chamber, and if needed
infuses nitrogen to displace air and thus lower the oxygen level. Oxygen or
oxygen-enriched gas is used to raise oxygen concentrations if hyperoxic conditions are desired. The working range of the incubator is 199% O2. The
oxygen controller is connected to the chamber via infusion tubing, and CO2
levels are maintained by the PROOX chamber via a CO2 controller connected
to a separate CO2 tank. Control gas must be supplied through a 1/8-in. id hose
to the back panel of the PROOX and delivery pressure set at 525 psi. The
PROOX chamber must be humidified autonomously by placing a stainless steel
pan on the chamber floor and filling it periodically with sterile distilled water.
The chamber must not be totally sealed, because it can develop a positive pressure, which can damage the sensors. It is important that only limited numbers
of culture containers be placed in the chamber to allow for gas circulation
needed to establish the hypoxic environment. Repeated access of the chamber
can introduce sufficient O2 to disturb critical experiments.
c. Static (noncontrolled) incubator chambers. Incubator bubbles that can be flushed
with a gas mixture and sealed, such as those made by Billups-Rothenberg (Del
Mar, CA), provide an inexpensive way to maintain a hypoxic cell culture environment (Fig. 4) (25). The chamber has inlet and outlet tubes, allowing con-
96
Grigoryan et al.
nection to a gas cylinder containing the desired oxygen level (0.65% O2),
5% CO2, and balance nitrogen (9599.4% N2). The modular incubation chamber is flushed for 90 s at 10 psi with the gas mixture to displace atmospheric
air from the chamber, and then both inlet and outlet tubes are closed simultaneously to avoid underpressure accumulation inside the chamber. The pressure
should be approx 5 psi over atmosphere to avoid leaking of O2 into the chamber. Disadvantages of these chambers are the time required to get in and out of
the chamber; the space that they consume, because they are round; and the
need for custom gas mixtures. The chamber can be placed in any 37C incubator. Sterile distilled water should be added in the floor of the chamber (under
the plate grating) in order to keep it humidified. It is important that only limited numbers of culture containers be placed in the chamber to allow for gas
circulation needed to establish the hypoxic environment and to minimize
trapped room air from disturbing the desired O2 levels in the chamber.
7. Eppendorf PO2 histograph. We used the Eppendorf Histograph (Model KIMOC6650) to measure in vitro and in vivo oxygen tension. The method (well described
in the literature) (26) utilizes a sterile polarographic electrode consisting of a gold
wire contained within a 0.3-mm steel casing. Prior to use, the probe was calibrated by pure, sterile nitrogen bubbled through sterile normal saline. When
tumors, murine tissue, or pediatric bone marrow were examined, a maximal track
length of 0.5 cm was used.
8. Falcon 96-well microtiter plates (Becton Dickinson, Lincoln Park, NJ).
3. Methods
3.1. Drug Cytotoxicity Assays
1. Harvest and resuspend cells in their appropriate medium, and plate (total volume
of 100 L) in Falcon 96-well microtiter plates at 500030,000 viable cells/well
(determined by trypan blue dye exclusion), depending on growth characteristics
and tumor type; for example, solid tumor cell lines should be seeded at lower cell
concentrations (100015,000) than leukemia (up to 50,000) due to the appreciable difference in cell size, or the doubling time of a given cell line (slowergrowing cells are plated at higher numbers than fast-growing lines).
2. Allow cells to recover overnight. Then add drugs to final concentrations in
50100 L of medium, 812 wells per drug concentration, with appropriate drug
vector added to control wells, and incubate the plates at 37C for 47 d, depending on growth properties (27). At least three drug concentrations must be tested to
be able to calculate lethal drug concentrations. For evaluation of the cytostatic
effect of a drug, plates should be incubated for longer periods (1421 d) and are
seeded with fewer cells in order to avoid overgrowth of cells in control wells (22).
However, using the highest number of cells that will not cause overgrowth of controls is advisable to provide the largest dynamic range.
3. To minimize the potential for cytotoxicity, add FDA to the 96-well plates with a
final concentration of 10 g/mL, and incubate for 20 min. One cannot stain more
than three to four plates at once with FDA and eosin Y.
97
4. Add 50 L of eosin Y (0.5% in normal saline) to quench background fluorescence of the FDA in the medium and in nonviable cells.
5. Measure relative fluorescence by DIMSCAN as described in Subheading 2., item 4
and express the results as the fractional survival cells compared with control cells.
3.2. Drug Testing Under Reduced Oxygen Conditions

1. For assays in hypoxia, seed cells into plates or flasks and place into one of the three
oxygen-controlled systems discussed in Subheading 2., item 6 incubated at 37C.
Under these conditions using 2% O2, medium in plates and flasks attains a PO2 of
approx 15 mmHg (25). This level of oxygen is below the degree of hypoxia found
in bone marrow (21) and in the range of hypoxia found in tumor tissue (Table 1).
3. Dilute drugs from concentrated stock or make freshly in medium preincubated
overnight in reduced-oxygen chambers or incubator and in tubes preflushed
and capped with reduced-oxygen gas.
4. Rapidly add the drugs to plates to the final concentrations, and quickly place the
plates back inside insert hypoxia chambers (Fig. 3), a host incubator, a modular
incubator chamber flushed and sealed with the appropriate gas mixture (Fig. 4),
or a hypoxia incubator (Fig. 2).
5. If using a modular sealed chamber, after drug addition, reflush with an appropriate O2 mixture for 90 s, and seal the chamber at approx 5 psi over pressure to
reduce atmospheric leaks, incubate at 37C, and reflush with the appropriate O2
mixture every other day until assayed.
6. Use the same drug mix tubes to make plates to be studied also in normoxic
atmospheres.
4. Notes
1. Cells should be incubated at least overnight in a hypoxia chamber or a hypoxia
incubator before any experiment is performed, in order to make cells hypoxic.
2. To minimize the exposure of cells to atmospheric O2 during drug addition, one
should carefully plan experiments and conduct them rapidly.
3. One should be careful not to add more pressure in a hypoxia chamber than is
required.
4. All studies determining the effect of hypoxia should be done simultaneously with
the experiments performed in normal atmospheric conditions as a control.
5. Incubator chambers must not be fully sealed, because the sensors cannot tolerate
positive pressure.
6. Chlorinated tap water should not be used for humidifying an incubator or chambers, because chlorine can deteriorate the stainless steel and may react with drugs.
Tap water may also have a high mineral content, which would produce a buildup
of scale in the reservoir.
References
1. Hockel, M. and Vaupel, P. (2001) Tumor hypoxia: definitions and current clinical,
biologic, and molecular aspects. J. Natl. Cancer Inst. 93, 266276.
98
Grigoryan et al.
2. Brown, J. M. (2002) Tumor microenvironment and the response to anticancer therapy. Cancer Biol. Ther. 1, 453458.
3. Teicher, B. A. (1995) Physiologic mechanisms of therapeutic resistance: blood
flow and hypoxia. Hematol. Oncol. Clin. North Am. 9, 475506.
4. Moulder, J. E. and Rockwell, S. (1987) Tumor hypoxia: its impact on cancer therapy. Cancer Metastasis Rev. 5, 313341.
5. Hockel, M. and Vaupel, P. (2001) Biological consequences of tumor hypoxia.
Semin. Oncol. 28, 3641.
6. Baronzio, G., Freitas, I., and Kwaan, H. C. (2003) Tumor microenvironment and
hemorheological abnormalities. Semin. Thromb. Hemost. 29, 489497.
7. Brown, J. M. and Koong, A. (1991) Therapeutic advantage of hypoxic cells in
tumors: a theoretical study [see comments]. J. Natl. Cancer Inst. 83, 178185.
8. Kennedy, K. A. (1987) Hypoxic cells as specific drug targets for chemotherapy.
Anticancer Drug Des. 2, 181194.
9. Kalra, R., Jones, A. M., Kirk, J., Adams, G. E., and Stratford, I. J. (1993) The
effect of hypoxia on acquired drug resistance and response to epidermal growth
factor in Chinese hamster lung fibroblasts and human breast-cancer cells in vitro.
Int. J. Cancer 54, 650655.
10. Sanna, K. and Rofstad, E. K. (1994) Hypoxia-induced resistance to doxorubicin
and methotrexate in human melanoma cell lines in vitro. Int. J. Cancer 58,
258262.
11. Vaupel, P., Thews, O., and Hoeckel, M. (2001) Treatment resistance of solid
tumors: role of hypoxia and anemia. Med. Oncol. 18, 243259.
12. Shannon, A. M., Bouchier-Hayes, D. J., Condron, C. M., and Toomey, D. (2003)
Tumour hypoxia, chemotherapeutic resistance and hypoxia-related therapies.
Cancer Treat. Rev. 29, 297307.
13. Teicher, B. A., Holden, S. A., and Jacobs, J. L. (1987) Approaches to defining the
mechanism of enhancement by Fluosol-DA 20% with carbogen of melphalan antitumor activity. Cancer Res. 47, 513518.
14. Colvin, O. M. (1994) Mechanisms of resistance to alkylating agents. Cancer Treat.
Res. 73, 249262.
15. Teicher, B. A. (1994) Hypoxia and drug resistance. Cancer Metastasis Rev. 13,
139168.
16. Harris, A. L. (2002) Hypoxiaa key regulatory factor in tumour growth. Nat. Rev.
Cancer 2, 3847.
17. Anderson, C. P., Tsai, J. M., Meek, W. E., Liu, R. M., Tang, Y., Forman, H. J., and
Reynolds, C. P. (1999) Depletion of glutathione by buthionine sulfoxine is cytotoxic for human neuroblastoma cell lines via apoptosis. Exp. Cell Res. 246,
183192.
18. Yang, B., Keshelava, N., Anderson, C. P., and Reynolds, C. P. (2003) Antagonism
of buthionine sulfoximine cytotoxicity for human neuroblastoma cell lines by
hypoxia is reversed by the bioreductive agent tirapazamine. Cancer Res. 63,
15201526.
99
19. Bremner, J. C., Stratford, I. J., Bowler, J., and Adams, G. E. (1990) Bioreductive drugs and the selective induction of tumour hypoxia. Br. J. Cancer 61,
717721.
20. Kizaka-Kondoh, S., Inoue, M., Harada, H., and Hiraoka, M. (2003) Tumor
hypoxia: a target for selective cancer therapy. Cancer Sci. 94, 10211028.
21. Grant, J. L. and Smith, B. (1963) Bone marrow gas tensions, bone marrow blood
flow, and erythropoiesis in man. Ann. Intern. Med. 58, 801809.
22. Anderson, C. P. and Reynolds, C. P. (2002) Synergistic cytotoxicity of buthionine
sulfoximine (BSO) and intensive melphalan (L-PAM) for neuroblastoma cell lines
established at relapse after myeloablative therapy. Bone Marrow Transplant. 30,
135140.
23. Proffitt, R. T., Tran, J. V., and Reynolds, C. P. (1996) A fluorescence digital image
microscopy system for quantifying relative cell numbers in tissue culture plates.
24. Krejsa, J., Frgala, T., Alfaro, P., and Reynolds, C. P. (2002) DIMSCAN 3.0, a new
generation of flourescence digital image microscopy system for measuring cytotoxicity in microplates. Proc. Am. Assoc. Cancer Res. 43.
25. Maurer, B. J., Metelitsa, L. S., Seeger, R. C., Cabot, M. C., and Reynolds,
C. P. (1999) Increase of ceramide and induction of mixed apoptosis/necrosis by
N-(4-hydroxyphenyl)-retinamide in neuroblastoma cell lines. J. Natl. Cancer Inst.
91, 11381146.
26. Wong, R. K., Fyles, A., Milosevic, M., Pintilie, M., and Hill, R. P. (1997) Heterogeneity of polarographic oxygen tension measurements in cervix cancer: an evaluation of within and between tumor variability, probe position, and track depth. Int.
J. Radiat. Oncol. Biol. Phys. 39, 405412.
27. Keshelava, N., Zuo, J. J., Chen, P., Waidyaratne, N. S., Luna, M. C., Gomer, C. J.,
Triche, T. J., and Reynolds, C. P. (2001) Loss of p53 function confers high-level
multi-drug resistance in neuroblastoma cell lines. Cancer Res. 61, 51035105.
28. Anderson, C. P., Castro, C. S., Meeks, W. M., et al. (1999) Physiologic bone
marrow oxygen tension (O2T) contributes to buthionine sulfoximine (BSO) resistance in neuroblastoma. Med. Pediatr. Oncol. 33(3), 266.
29. Kessler, M., Lang, H., Sinagowitz, E., Rink, R., and Hoper, J. (1973) Homeostasis of oxygen supply in liver and kidney. Adv. Exp. Med. Biol. 37A, 351360.
30. Greco, O., Marples, B., Joiner, M. C., and Scott, S. D. (2003) How to overcome (and exploit) tumor hypoxia for targeted gene therapy. J. Cell. Physiol.
197, 312325.
31. Stone, H. B., Brown, J. M., Phillips, T. L., and Sutherland, R. M. (1993) Oxygen
in human tumors: correlations between methods of measurement and response to
therapy: summary of a workshop held November 1920, 1992, at the National
Cancer Institute, Bethesda, Maryland. Radiat. Res. 136, 422434.
32. Nordsmark, M., Bentzen, S. M., and Overgaard, J. (1994) Measurement of human
tumour oxygenation status by a polarographic needle electrode: an analysis of
inter- and intratumour heterogeneity. Acta Oncol. 33, 383389.
100
Grigoryan et al.
33. Teicher, B. A. (1995) Physiologic mechanisms of therapeutic resistance: blood

flow and hypoxia. Hematol. Oncol. Clin. North Am. 9, 475506.
34. Weiss, H. R. and Winbury, M. M. (1972) Intracoronary nitroglycerin, pentaerythritol trinitrate and dipyridamole on intramyocardial oxygen tension. Microvasc.
Res. 4, 273284.
10
Chemosensitivity Testing Using Microplate Adenosine
TriphosphateBased Luminescence Measurements
Christian M. Kurbacher and Ian A. Cree
Summary
During the last two decades, novel nonclonogenic methods for pretherapeutic chemosensitivity testing have been developed that are likely to overcome major technical limitations of older
assays such as low evaluability rates, low degree of standardization and reproducibility, lack of
technical robustness, and poor methodological efficacy. Among these, the microplate adenosine
triphosphate (ATP)based tumor chemosensitivity assay (ATP-TCA) has gained particular merits
for ex vivo chemosensitivity testing of native nonhematological tumors including cancers of the
breast, ovary, gastrointestinal tract, cervix and corpus uteri, and lung; malignant melanomas;
gliomas; sarcomas; and mesotheliomas. For this indication, the ATP-TCA can now be considered
the best documented and validated technology. This assay, which is now commercially available, provides a highly reproducible, easy-to-handle kit technique; low technical failure rates; and
a high methodological efficacy requiring only 1 106 tumor cells to test four to six different
drugs or combinations. In ovarian and breast carcinomas, the predictive accuracy is >90%, with
a positive predictive value of 8590% and a negative predictive value near 100%, respectively. In
primary ovarian cancers, the ATP-TCA has been found to accurately predict both clinical
response and survival. In two prospective clinical trials in patients with heavily pretreated ovarian cancer, chemotherapy individually selected by the ATP-TCA has been found to triple the
response rates and nearly double the survival compared to empirically chosen regimens. Consequently, this assay, which is now under phase III evaluation, has successfully been used in new
agent development to screen for novel chemotherapy regimens for the treatment of patients with
breast and ovarian carcinoma and melanoma, respectively. This chapter highlights the recent preclinical and clinical experience with this promising technology and gives a detailed description
of all the technical aspects of the ATP-TCA.
Key Words
Adenosine triphosphatebased tumor chemosensitivity assay; breast cancer; chemoresistance;
chemotherapy; colorectal cancer; luminescence; malignant melanoma; nonclonogenic assay;
ovarian carcinoma.
101
102
Kurbacher and Cree
1. Introduction
Because of the considerable toxicity of antineoplastic agents in conjunction
with their often uncertain clinical efficacy, the concept of individualized
chemotherapy based on pretherapeutic chemosensitivity testing has attracted
attention for almost five decades (13). In particular, the 1980s saw an intensive
preclinical and clinical research in this field mainly focused on clonogenic
assays stimulated by the work of Hamburger and Salmon and Courtenay in the
late 1970s (see Chapter 2). A number of clinical trials using these technologies
in various tumor entities have been conducted, sometimes providing promising
results, but none were able to provide indisputable evidence of the superiority
of assay-directed chemotherapy over empirical treatments (46). As a consequence, the validity of the whole concept of individualized chemotherapy was
principially questioned by many oncologists. However, many of these studies
were unable to unequivocally demonstrate a clinical benefit owing to severe
technical limitations of the utilized assays such as low evaluability rates, long
incubation periods, the requirement of large numbers of tumor cells, and a low
degree of reproducibility (3,5,6). Nevertheless, it is quite clear from these early
studies that individualized chemotherapy is extremely unlikely to worsen a
patients clinical outcome (6). During the last two decades, a number of newer
nonclonogenic assays have been developed that seem to overcome technical
problems associated with older technologies. Among these, the adenosine
triphosphate (ATP)based tumor chemosensitivity assay (ATP-TCA) may be
regarded as having the best track record for testing native tumor cells derived
from nonhematological human malignancies (3,7).
ATP is the ubiquitary end point of all energy-gaining processes in eukaryote
cells. Moreover, after lethal cell demage, ATP levels drop to zero within a few
milliseconds owing to the hydrolytic activity of intracellular adenosine triphosphatases (8). Thus, ATP may be regarded as the most sensitive marker for cell
viability that can easily be measured by the luciferinluciferase (Lu-Lu) bioluminescence reaction using the luciferase of various firefly species. Because of
its extreme sensitivity, the lu-lu bioluminescence reaction is able to determine
exactly the cellular ATP content of no more than 10 viable cells (911) Moreover, the relationship between the intracellular ATP and the luminescense
responsei.e., the emission of visible lightremains linear over a broad concentration range (7,11).
In the early 1980s, several groups combined bioluminescence measures that
were initially used to test microbes with common cell culture methods in order
to assess the viabilty of eukaryote cells, namely permanently growing tumor
cell lines (810,12). Subsequently, these techniques have been adopted to
Microplate ATP-TCA
103
develop novel nonclonogenic assays for pretherapeutic chemosensitivity testing

of native tumor cells. Currently, three major types of ATP-based chemosensitivity assays exist, all incorporating the principles of nonadherence and/or
selective culture conditions to guarantee the exclusive growth of neoplastic
cells: the ATP cell viability assay (ATP-CVA) and its microplate modification,
the serum-free ATP assay (SF-ATPA), and the ATP-TCA (7,1315). All three
types use techniques of both single cell and tissue culture simultaneously. Two
major components guarantee the exclusive growth of neoplastic cells: the use of
nonadherent culture plates in the ATP-CVAs and selective cell culture media in
the SF-ATPA. The ATP-TCA, which was originally designed with respect to the
requirement of a routine laboratory, combines both principles (3,7). To provide
a higher degree of standardization, agar-coated culture vessels, which are used
in the ATP-CVA, have been substituted by nonadherent polypropylene
microplates. Another important advantage of the ATP-TCA is its suitability for
both solid tumor samples and malignant effusions (3,7).
A number of different tumor entities have since been investigated in depth
using the ATP-TCA such as breast, ovarian, renal cell, pancreatic, colorectal,
and lung cancer; malignant melanoma (cutaneous and choroidal type); retinoblastoma; and mesothelioma (3,7,1625). The ATP-TCA provides evaluability
rates of about 90% and a highly standardized methodology that makes this
technique particularly suitable for multicenter investigations (3,7,17,18,20,22,
23). Comparison of the assay results with the clinical outcome of individual
patients demonstrates a predictive accuracy of >90% for the majority of tumors
tested, with a negative predictive value approaching 100% and a positive predictive value between 85 and 95%, respectively (17,26,27). Moreover, the ATPTCA has been found to accurately predict the survival of patients with primary
International Federation of Gynecology and Obstetrics stage III ovarian carcinomas undergoing postoperative platinum-based chemotherapy (28). Because
of these technical advantages, the assay has already been used successfully to
screen for novel chemotherapy regimens to be subsequently used in patients
with advanced or recurrent ovarian, breast, or colorectal cancer and malignant
melanoma (20,27,29,30). Additionally, the ATP-TCA appears to be a valuable
tool in the preclinical development of investigational cytostatics using both
native tumor cells and permanently growing cell lines that can be tested with a
slightly modified methodology (19,3133).
During the last decade, we conducted a number of prospective clinical trials
of ATP-TCA-directed chemotherapy in patients with pretreated solid neoplasms. In a first case-control intervention study in patients with recurrent ovarian cancer, individual chemotherapy produced both a threefold increase in
clinical responses and a doubled progression-free and overall survival (Fig. 1)
104
Kurbacher and Cree
Fig. 1. Survival curves of ATP-TCA-directed chemotherapy in patients with recurrent ovarian cancer vs empirically chosen chemotherapy: (A) overall survival; (B) progression-free survival. (Reprinted from ref. 34 with permission of the publisher.)
(34). Two subsequent phase II trials in heavily pretreated patients with recurrent
ovarian cancer that were independently run in both the United Kingdom and
Germany showed response rates of 59 and 68% and a median overall survival
of 15 and 21 mo, respectively (26,35,36). The most exciting finding of both
these trials was the fact that patients with platinum-refractory disease did as
well as those with potentially platinum-sensitive tumors, a result that is clearly
unique in comparison to empirical chemotherapy for recurrent ovarian carcinoma. As a consequence, we set up a phase III trial of ATP-TCA-directed vs
empirically chosen chemotherapy in patients with platinum-refractory relapsed
ovarian carcinoma; recruitment was stopped in February 2003. The first follow-
Microplate ATP-TCA
105
up data, which will be available at the end of 2004, are eagerly awaited.
Recently, additional clinical trials of ATP-TCA-directed chemotherapy in breast
and ovarian cancer and malignant melanoma are actively recruiting patients in
different European countries and will help to determine further the role of individually tailored chemotherapy as a routine procedure in clinical oncology.
2. Materials
2.1. ATP-TCA Kits
For the ATP-TCA, commercially available test kits (TCA 100, Sartoritest,
DCS Innovative Diagnostik Systeme, Hamburg, Germany) providing the following materials, reagents, and medium are used:
1. Two sterile disposable scalpels (no. 11).
2. Disposable syringe (10 mL) and 0.22-m cellulose acetate filter.
3. Sterile serum- and glutathione (GSH)-free complete assay medium (CAM)
(250 mL) at 28C.
4. Sterile round-bottomed 96-well polypropylene microplates (Costar no. 4690).
5. Lyophilized tumor dissociation enzymes (TDE); these must be reconstituted with
10 mL of CAM. TDE solution can be stored briefly at 28C. Reconstituted TDE
can be stored longer at 20C. Reagent that has been frozen and thawed should
not be frozen again.
6. Sterile maximum ATP inhibitor reagent (MI) (4 mL); store at 28C.
7. Sterile buffered tumor cell extraction reagent (TCER) (20 mL); store at 28C.
8. Sterile reconstitution buffer, pH 7.8 (dilution buffer) (20 mL).
9. Luciferinluciferase light reagent (Lu-Lu): a lyophilized mixture of D-luciferin
and luciferase (EC 1.13.12.7) must be reconstituted with 15 mL of dilution buffer
and is ready to use after subsequent incubation for 30 min at room temperature.
Reconstituted Lu-Lu can be stored in the dark at 28C for a maximum of 14 d.
Longer storage is possible at 20C. Lu-Lu that has been frozen and thawed must
not be frozen again.
10. ATP lyophilizate (500 ng) to be dissolved in 2 mL of dilution buffer. Reconstituted ATP standard can be stored briefly at 28C or for a longer period at 20C.
2.2. Additional Materials

The following reagents, medium, and materials not included in the ATPTCA kit are also required:
1.
2.
3.
4.
Sterile conical centrifuge tubes (15 and 50 mL).

Sterile disposable pipets (1, 5, and 10 mL).
Adjustable automated pipets (0200 L, 2001000 L) with sterile tips.
Adjustable automated multichannel pipets (6-, 8-, or 12-fold) at 0200 L with
sterile tips.
5. Sterile disposable Petri dishes (100 15 mm).
106
Kurbacher and Cree
6. Round-bottomed polysterol tubes (12 75 mm) (no. 55476, Sarstedt, Nmbrecht,

Germany) to be used in chain luminometers or 96-well white microplates (Costar
no. 3912/13) to be used in microplate luminometers.
7. Neubauer hemocytometer (0.1 mm).
8. Sterile destilled water.
9. Sterile preservative-free sodium-heparin (Vetren, Promonta, Hamburg, Germany).
10. Sterile Ficoll-Hypaque (1077 g/mL).
11. Trypan blue solution (0.14%).
2.3. Laboratory Equipment

The ATP-TCA can be performed in a routine cell culture laboratory providing the following equipment:
1.
2.
3.
4.
5.
6.
7.
Centrifuge (minimum: 400g).

Inverted and phase-contrast microscopes.
Class II laminar airflow cabinet.
Humidified incubator (37C, 5% CO2).
Chain or plate luminometer.
Shaker water bath, 37C (facultative).
Personal computer.
2.4. Drugs
The ATP-TCA uses commercial formulations of most common antineoplastic agents except cyclophosphamide (CPA) and ifosfamide (IFO) and hormonal
agents. CPA and IFO are inactive in vitro and must be substituted either
by their activated metabolites 4-hydroperoxy-cyclophosphamide (4-HC) and
4-hydroperoxy-IFO, respectively, or by mafosfamide cyclohexylamine, which
is able to generate 4-HC after dissociation in watery solution. All three active
derivatives are provided by ASTA Medica (Frankfurt, Germany). Hormonal
agents should be used as free substances and must be dissolved in absolute
ethanol prior to use. Stock solutions of drugs should be prepared and stored
according to the manufacturers instructions and to previous publications (7,37).
If information about the adequate solvent is not provided by the manufacturer,
stock solutions should be prepared with CAM except melphalan (L-PAM;
methanol 100%) and nitrosoureas (100% ethanol). Detailed information is summarized in Table 1.
3. Methods
3.1. Sampling of Tumor Specimens
The methodology of the ATP-TCA is illustrated in Fig. 2. Native tumor cells
derived from both solid tumors and effusions can be tested with the ATP-TCA
(see Notes 1 and 2). Strict asepsis is mandatory during all steps of cell collec-
Microplate ATP-TCA
107
Table 1
Preclinical Characteristics of Single Agents Tested in ATP-TCA
Drug
Actinomycin D
Bleomycin
Carboplatin
Carmustine
Cisplatin
4-OOH-cyclophosphamide
Cytosine arabinoside
Dacarbacin
Daunorubicin
Docetaxel
Doxorubicin
Epirubicin
Etoposide
Ca-folinate
5-Fluorouracil
Gemcitabine
Irinotecan
Melphalan
Methotrexate
Mitomycin C
Mitoxantrone
Oxaliplatin
Paclitaxel
Tamoxifen
Thiotepa
Topotecan
Treosulfan
Vinblastine
Vincristine
Vindesine
Vinorelbine
a
Clinical dosage Stock solution

(mg/m2)a
(mg/mL)b
100% TDC
(g/mL)
Storage of
stock solutionc
1110.51
30 mg abs.
14001.1
5 100
11001.1
10001.1
0.1
3.0
15.8
4.0
3.8
3.0
1.0
3.0
10.0
3.3
2.0
5.0
70C/6 mo
28C/28 d
28C/5 d
28C/2 d
28C/2 d
20C/3 mo
11001.1
5 250
11601.1
11001.1
11751.1
11901.1
3 100
14501,1
10001.1
10001,1
13501,1
4 10
11401.1
11121.1
11121.1
11301.1
11751.1
20 mg daily
11601.1
1111.25
50001..
11141..
111111.
111411.
113011.
2.4
20.0
0.4
11.3
0.5
0.5
48.0
1.2
22.5
25.0
100.0
1.8
2.8
0.23
0.6
5.0
13.6
20.0
2.0
0.75
20.0
0.5
0.5
0.4
1.94
20.0
10.0
1.0
5.0
2.0
2.0
20.0
10.0
50.0
40.0
20.0
0.5
2.5
1.0
2.0
5.0
6.0
50.0
5.0
1.0
50.0
1.0
1.0
1.0
10.0
28C/8 d
28C/4 d
28C/28 d
70C/3 mo
28C/28 d
28C/2 d
25C/28 d
28C/28 d
25C/28 d
70C/6 mo
20C/6 mo
28C/28 d
28C/14 d
25C/28 d
20C/6 mo
28C/28 d
70C/6 mo
28C/5 d
70C/3 mo
70C/3 mo
28C/30 d
28C/14 d
28C/28 d
28C/28 d
Clinical standard dosage

Stocks of drugs should be prepared in sterile water or saline according to the manufacturers
instructions.
c
See ref. 37 or refer to the manufacturers instructions.
b
108
Kurbacher and Cree
Fig. 2. Methodology of ATP-based tumor sensitivity assay.
tion and processing. If possible, solid material for chemosensitivity testing

should be separated intraoperatively provided that histopathological diagnosis
and staging are not compromised. If the surgeon cannot decide which part of
the tumor should be reserved for ATP-TCA, it is also appropriate to transfer the
whole tissue to the pathologist prior to separation. One should be aware, however, that this may increase the risk of contamination.
1. After sampling, place solid specimens into a sterile tightly closed transport vessel
containing Hanks balanced salt solution or serum-free Dulbeccos modified
Eagles medium (DMEM) with antibiotics (200 IU/mL of penicillin, 100 g/mL
of streptomycin). In potentially infected probes (i.e., tumors derived from the skin,
upper aerodigestive tract, bronchi, large bowel, or lower female genital tract),
transport medium should additionally contain 2.5 g/mL of amphotericin B and
1 g/mL of metronidazole (20) (see Note 3).
2. Collect malignant effusions by paracenthesis.
3. Coagulate all exudates with 10 IU/mL of sodium heparin.
4. Do not fix or freeze samples. Perfrom transport and storage at room temperature;
gentle cooling may be appropriate at temperatures exceeding 25C. Process all
specimens as soon as possible, preferably within 24 h after sampling.
3.2. Preparation of Cell Suspensions

The ATP-TCA utilizes cell suspensions containing both single cells and
small organoid aggregates composed of tumor cells and stroma. All preparation
steps are carried out under a laminar airflow hood.
Microplate ATP-TCA
109
1. Place solid specimens in a 10-cm Petri dish to remove necrotic parts, fat, and
connective tissue using sterile scissors and scalpels.
2. Add a few milliliters of serum-free DMEM, and then cut the tumor tissue into particles of approx 2 mm in diameter.
3. Pour the tumor fragments into a 15-mL test tube and mix with 10 mL of TDE.
4. Depending on the consistency of the individual material, the following enzymatic
digestion can be performed either in a 37C shaker bath for 24 h or overnight at
37C in a humidified 5% CO2 atmosphere.
5. Pass the tumor lysate through an 80-m mesh filter gauze in order to remove the
undissociated particles.
6. Wash the enzymes by two centrifugation steps at 200g for 10 min after adding
10 mL of DMEM.
7. Fill 50-mL test tubes with aliquots of liquid samples and then centrifuge for
10 min at 200g in order to remove all the serum components.
8. Resuspend the pellets with 10 mL of DMEM.
9. Purify all tumor preparations by Ficoll-Hypaque density gradient centrifugation
and subsequently resuspend in 15 mL of CAM.
10. Determine the quality and viabilty of the resulting tumor suspensions by cytological examination using Papanicolau or Wright-Giemsa staining and Trypan
blue dye exclusion.
11. Add another amount of CAM, and then adjust the suspensions to a final concentration of 1 to 2 105 viable tumor cells/mL.
3.3. Final Drug Concentrations (see Note 4)

All drugs and drug combinations are tested at six test drug concentrations
(TDCs) ranging from 6.25 to 200%. Both therapeutic and supratherapeutic concentrations are included, with 100% TDC referring to either the plasma peak or
a clinically relevant equivalent of the area under the plasma elimination curve.
For each assay, 800% TDCs are freshly prepared in 10 mL of CAM from the
aforementioned stock solutions (see Table 1) with drug combinations being made
up by simply mixing an adequate amount of individual single agents. Final TDCs
of both single agents and combinations are prepared directly on the culture plates
by serial 11 dilutions. Each cytostatic regimen is tested in triplicates.
3.4. Preparation of Culture Plates
All of the following preparation steps prior to incubation are performed
directly on culture plates. Each plate comprises 72 test drug wells with each
cytostatic regimen set up in triplicate and two internal controls in 12 wells of
the remainder, a no-inhibition control (M0) with blank CAM, and a maximum
inhibition control with MI instead of medium or drugs. The design of a culture
plate is shown in Fig. 3. Preparation of culture plates should be performed as
follows:
110
Kurbacher and Cree
Fig. 3. Design of a 96-well microplate to test four drugs or drug combinations in

triplicates. MI, maximum inhibition control; M0, no inhibition control.
1. Add 100 L of CAM to 84 wells of each microplate using a multichannel pipet

and change the pipet tips.
2. Add 100 L of MI to the remaining 12 wells and change the tips.
3. Add 100 L of each drug or combination 800% TDC to either of three wells,
and mix gently by repeated aspiration. Four regimens can be set up on every plate;
starting wells should be positioned in the same row as shown in Fig. 3.
4. Transfer 100 L of mixture to the subsequent three wells and mix gently.
5. Repeat step 4 until the last row is reached and mix gently. Aspirate 100 L of
the resultant dilution and discard it together with the pipet tips. Each well of the
culture plate now contains 100 L of blank CAM, MI, or drugs at serial 11
dilutions.
6. Add 100 L of cell suspension to each well and mix gently.
3.5. Incubation
Prepared microplates are covered and then transferred to a humidified 5%
CO2 incubator with subsequent routine incubation for 6 (range: 57) d. Both to
guarantee complete disappearance of stromal components and to allow the
tumor cells time for sufficient DNA repair, the incubation period should not
be shorter than 4 d. On the other hand, nutritive exhaustion may falsify the test
results at incubation periods exceeding 8 d. To avoid drying artifacts, plates
may be placed in a sterile wet chamber.
3.6. ATP Extraction
At the end of incubation, another cytological examination should be performed using cells from three to six M0 wells in order to confirm tumor cell
Microplate ATP-TCA
111
enrichment. From the remaining wells, cellular ATP is extracted and stabilized
by the following procedure:
1. Add 50 L of TCER to each MI and M0 control well using a multichannel pipet,
mix gently by repeated aspiration without producing bubbles or foam, and change
the pipet tips.
2. Add another 50 L of TCER to each test well and mix gently.
3. Incubate the uncovered plates for 2030 min at room temperature.
4. Immediately after ATP extraction and incubation, perform measurement as
described in Subheading 3.7. Alternatively, extracted microplates can be frozen
and stored at 20C. Plates that have been frozen and thawed must be immediately processed for ATP measurement and should not be frozen again.
3.7. ATP Measurement by Firefly Luminescence Reaction

After lysing the cells by adding TCER, ATP measurement should be performed within 1 h. Alternatively, lysed plates can be frozen and stored at 20C
without addition of cryoprotectants. Frozen plates should be thawed completely
by appropriate incubation at 37C. Prior to ATP determination, Lu-Lu must be
dissolved in 15 L of dilution buffer. After gently mixing, the reagent should
be incubated at room temperature for 30 min and covered with aluminum foil
in order to avoid loss of activity. For ATP determination, both chain and plate
luminometers can be used. When using a chain luminometer, the procedure is
as follows:
1. Aspirate 50 L of each cell lysate separately and place it in a 12 75 mm counting tube.
2. Connect the bottle with Lu-Lu with the injector system of the luminometer.
3. Put five empty tubes (washing tubes) in the luminometer to guarantee complete
preload of the injection system.
4. Leave one space blank and add another empty tube.
5. Put all the prepared tubes into the luminometer in the correct sequence of different regimens investigated.
6. Close the luminometer tightly and start the measurement; the counting time
should be 1015 s with a 4-s delay.
7. After ATP determination, wash the injection system five times with both 70%
ethanol and distilled water in order to avoid microbial contamination.
Generally, the use of a plate luminometer is less time-consuming. Both devices

with and without injectors are suitable. All lysates must first be transferred to a
96-well microplate. When using a luminometer with an injector, 50 L of Lu-Lu
will be automatically added to each well. The injector must be washed with
Lu-Lu prior to ATP measurement according to the manufacturers instructions
and afterward cleaned with both 70% ethanol and distilled water. When using
devices without an injector, as we do in our laboratories, 50 L of Lu-Lu must
112
Kurbacher and Cree
be manually pipetted to each well of the microplate. The plate must then be
immediately placed into the counting chamber. For all types of plate luminometers, the counting time is 1 s. Luminescence response (i.e., the amount of
visible light) is expressed as relative light units (RLU = photons/10).
3.8. Evaluation and Interpretation of Assay (see Note 5)
For each tumor and drug, respectively, relative tumor growth inhibition
(TGI) is determined by the following formula:
TGI (%) = [1 (RLUTest RLUMI) / (RLUM0 RLUMI)] 100
(1)
The six TGI values of each regimen are then transferred into individual inhibition curves from which the following parameters are drawn by using DCS evaluation software based on Microsoft Excel:
AUIC = area under the inhibition curve, calculated by a trapezoidal rule
IC50, IC90 = relative 50 and 90% inhibitory concentration (in % TDC)
IndexSUM = natural logarithmic sum index calculated as 600 (TGI6.25 + + TGI200)
When regarded separately, both the AUIC and IndexSUM can easily distinguish between principally resistant (AUIC < 12,500 and/or IndexSUM 300)
and potentially sensitive tumors (AUIC 12,500; IndexSUM < 300). However, a
more detailed differentiation is possible by using a semiquantitative score that
integrates the other aforementioned parameters (29,34). A typical example is
shown in Fig. 4:
High sensitivity (S): AUIC 12,500/IndexSUM < 300; IC90 90% TDC;
IC50 25% TDC
Partial sensitivity (P): AUIC 12,500/IndexSUM < 300; IC90 > 90% TDC;
IC50 25% TDC
Weak sensitivity (W): any other with AUIC 12,500/IndexSUM < 300
Resistance (R ): AUIC < 12,500/IndexSUM 300
When testing drug combinations, it may be of interest to obtain information

about possible drug interactions, which requires separate testing of all single
agents included. In the ATP-TCA, all components are tested in natural logarithmic dilutions. Therefore, the sigmoid action model described by Pch et al.
(38), which compares the actuarial dose-response curves with a theoretical one
assuming that the components act independently (or additively), is considered
the most reliable procedure to analyze drug interactions observed in this assay.
This analysis may be performed by using the appropriate spreadsheet included
in the DCS evaluation software. Automated interaction analysis can be per-
Microplate ATP-TCA
113
Fig. 4. Typical dose-response curves achieved with different drugs tested in the ATPTCA. The different types of ex vivo chemosensitivity are adapted from ref. 29.
formed only if both the combination and the different single agents involved are
set up on the same microplate. Figure 5 demonstrates different types of drug
interactions ranging from antagonism to potentiation (or synergism).
3.9. ATP Standard Curve
An ATP standard measurement should be performed as follows every day
prior to the first assay evaluation in order to guarantee the linearity of all subsequent ATP determinations:
1. Mix the ATP lyophilisate with 2 mL of reconstitution buffer to prepare the ATP
stock. Shake gently and wait a few minutes. The resultant stock ATP concentration is 250 ng/mL.
2. Add 300 L of reconstitution buffer to each of nine 12 75 mm test tubes.
3. Add 150 L of the ATP stock to the first tube, mix carefully, and transfer 150 L
to the second tube. Continue the procedure until the ninth tube. The final ATP
concentrations of the resultant 13 dilution series are 83.33, 27.78, 9.26, 3.09,
1.03, 0.34, 0.11, 0.04, and 0.01 ng/mL.
4. Transfer a 50-L aliquot of each dilution to either of three wells of a 96-well
microplate or to three round-bottomed luminometer tubes (if a chain luminometer is used). Add both Reconstitution Buffer without ATP and the ATP stock solution to another three wells or tubes. Thus, the luminescence vs ATP concentration
curve will range between 0 and 250 g/mL.
5. Measure the ATP following the instructions in Subheading 3.7. When using a
chain luminometer, it is preferred that the ATP measurement be performed using
decreasing concentrations.
114
Kurbacher and Cree
Fig. 5. Drug interactions in two-drug combination analyzed with sigmoid model of

Pch (38). Combination C, potentiation (synergism); combination D, independent action
(additivity); combination E, antagonism. The dashed line represents the theoretical
dose-response curve for the combination assuming that both drugs act independently.
6. Graph the mean of each triplicate against the ATP concentrations. The Pearsons
correlation coefficient of the resulting dose-response plot should be at a minimum of 0.975.
3.10. Quality Control (see Notes 6 and 7)

Particularly when using the ATP-TCA as a guide to select chemotherapy
regimens for individual patients pretherapeutically, each assay must meet the
following criteria to be considered as providing valid results (3,7,34):
1. Prior to being tested with the ATP-TCA, the malignant nature of the sample must
be uneqivocally confirmed by morphological (histopathological or cytological)
analysis. Additional immunohistochemistry or immunocytochemistry may prove
useful in selected cases.
2. Tumor cell enrichment during the 5- to 7-d incubation period should be confirmed
by pre- and postassay cytology. Preassay examination can be performed using the
original cell suspension prior to seeding. Postassay evaluation must use cells
derived from three M0 wells before TCER is added to the microplate. Cytological confirmation of tumor cell enrichment is particularly important in cases with
a low initial tumor cell content and/or a low viabilty in order to demonstate the
proliferative capacity of the specimen under culture conditions.
Microplate ATP-TCA
115
3. Viable tumor cells must be able to proliferate in the ATP-TCA system. Low M0
values after 57 d of culturing may indicate a reduced proliferative activity, which
often causes artificially high chemosensitivity ex vivo. Therefore, assays with a
mean M0 <50,000 RLU (measured in an LB953 chain luminometer) should be
considered nonevaluable. When using a microplate luminometer, M0 values below
this threshold may be appropriate provided that the assay meets the other quality
criteria mentioned in this section.
4. Clear sigmoidal dose-response curves must be detectable for at least one single
agent or drug combination per plate. Exclusive appearance of so-called flat curves,
which demonstrate a high TGI at low concentrations but never reach total cell
kill even at maximum doses (which thus are to be considered as partial sensitivity), may indicate the presence of cells that have been sublethally damaged prior
to being assayed (eventually owing to preceding chemo- or radiotherapy). These
dying cells exhibit a higher chemosensitivity compared with those retaining their
whole viability and may thus cause artificial hypersensitivity. Assays exclusively
showing flat curves should therefore be regarded as invalid.
5. Since all pro- and eukaryote cells contain ATP, which is the functional end point
of the ATP-TCA, assays evidencing gross microbiological contamination at the
end of the incubation period are extremely unlikely to provide valid results and
must be discarded.
6. Another good indicator of contamination is high MI values even in cases without
macroscopic signs of infection. As a consequence, all assays with an MI/M0
>0.01 should be regarded as being potentially infected and thus nonevaluable.
7. When performing ATP-TCA-based multicenter trials in which multiple laboratories are involved, quality control evaluations preferentially using easy-to-grow cell
lines and simultaneous testing of a standard panel of drugs are strongly recommended.
4. Notes
1. Recently, the ATP-TCA has been regarded as the most advanced and best documented biological system for pretherapeutic chemosensitivity testing of native
organ tumors. As mentioned in Subheading 1., clinical trials using the ATP-TCA
to individually select chemotherapy all gave promising results. Numerous different tumor types including carcinomas of the breast, lung, stomach, pancreas, gallbladder, liver, colon, ovary, endometrium, uterine cervix, vulva, pharynx, larynx,
and kidney have now been successfully tested with the ATP-TCA, as have various
sarcomas, malignant brain tumors, mesotheliomas, and malignant melanomas of
both the skin and the uvea. The assay may also be suitable to test hematological
malignancies. However, growth characteristics and culture requirements of
leukemias and lymphomas are only poorly elucidated so far (39).
2. Blood samples should be processed according to the recommendations for malignant effusions. Repeated density-gradient centrifugation is strongly recommended
in these cases. In blood samples, the proportion of blasts should exceed 70%.
Assuming the high intrinsic chemosensitivity of hematological malignancies, it
116
3.
4.
5.
6.
Kurbacher and Cree

may be appropriate to use 10-fold lower TDCs compared with the concentrations
used for solid tumors to obtain reasonable test results.
Another important problem is the prevention of microbial infection of the culture
plates. Strictly aseptic conditions are therefore a must for all methodological steps
of the ATP-TCA. Potentially contaminated samples should normally not be further
processed. However, a recent publication has shown that infection of colon cancer
specimens can effectively be prevented by adding both amphotericin B and
metronidazole to the transport and culture media without deteriorating the tumor
cell viability and biology (20). It may thus be important to routinely add these
drugs to the media if the material is potentially contaminated and performing the
ATP-TCA is mandatory in the individual case.
Liposomal formulations of anthracyclines and other cytostatics are attracting
increasing interest in modern oncology. Unfortunately, it is uncertain whether they
can be tested directly in the ATP-TCA system. It is well known, however, that
liposomal anthracyclines achieve a three- to fivefold higher intracellular accumulation compared with the parent compounds. Therefore, testing a threefold higher
concentration of free anthracyclines may be regarded as a good substitute for the
corresponding liposomal preparations (40,41).
Strict quality criteria, as mentioned in Subheading 3.10., are mandatory to obtain
valid test results with the ATP-TCA. Among these, demonstration of doseresponse relationships is particularly important. Dose-response curves should be
sigmoid shaped, especially for alkylators, platinum analogs, and intercalating
agents. Flat curves may indicate the absence of cells that have retained their complete proliferative capacity. However, some drugs are known to frequently produce
such atypical dose-response curves. Antimetabolites such as methotrexate and
vinca alkaloids may demonstrate flat curves as well. Moreover, cytidine analogs
such as cytosine arabinoside and gemcitabine are internalized via a transmembranal energy-dependent transporter system that is refractory beyond a certain
drug concentration (generally 50% TDC). Dose-response curves for these drugs
therefore show remarkable plateau effects above this threshold concentration (30).
To exclude artifacts, it is therefore recommended that both drugs with presumably
typical curves and those often producing atypical curves be simultaneously incubated on the same culture plate.
Among different other in vitro or ex vivo predictive assays, the ATP-TCA can be
regarded as a true chemosensitivity assay owing to its exceptionally high positive predictive accuracy. This should be taken into account when deciding what
chemotherapy regimens should be tested in an individual patient. It is of particular importance that drug combinations be assayed when the patient is planned to
be treated with combination chemotherapy, because single agent activity ex vivo
is unlikely to accurately predict its clinical efficacy when used as part of a combination regimen. Drug interactions may be more important for the clinical utility
of a drug combination than the isolated activity of its components. Good examples
are combinations of gemcitabine or cytosine arabinoside and platinum, or treo-
Microplate ATP-TCA
117
sulfan or mitoxantrone and paclitaxel. These combinations display features of true

clinical synergism, which may be related to resistance-modulating effects. As a
consequence, these regimens have been shown to produce major preclinical and
clinical activity even when both of the single agents included were completely
inactive (26,29,30). Exclusive testing of single agents, which has been preferred
in the past, is therefore appropriate only in cases in which combination chemotherapy is abandoned.
7. With slight modifications, the ATP-TCA can be used to test cell lines that may
sometimes be preferred. However, note that cell lines including their resistant
variants display a remarkably higher chemosensitivity compared to native tumors.
This may be due to the absence of tumor stroma in permanent growing cell lines
that are known to exhibit important chemomodulating effects. Moreover, the heterogeneity in cell lines is generally inferior to that in clinical tumors. By using
modern nonclonogenic techniques such as the ATP-TCA, testing a large number
of native samples per entity in a considerably short time may provide a more reliable estimate of the activity of the drug under investigation than evaluating a panel
of different tumor cell lines (42). These methods are thus of particular interest in
future preclinical development of novel agents and combinations (41,43).
References
1. Bellamy, W. T. (1992) Prediction of response to drug therapy of cancer: a review
of in vitro assays. Drugs 44, 690708.
2. De Vita, V. T. (1997) Principles of cancer management: chemotherapy, in Cancer:
Principles and Practice of Oncology (DeVita, V. T., Hellman, S., and Rosenberg,
S. A., eds.), Lippincott-Raven, Philadelphia, pp. 333347.
3. Cree, I. A. and Kurbacher C. M. (1997) Individualizing chemotherapy for solid
tumorsis there any alternative? Anti-Cancer Drugs 8, 541548.
4. Von Hoff, D. D. (1990) Hes not going to talk about in vitro predictive assays
again, is he? J. Natl. Cancer Inst. 82, 97101.
5. Brown, E. and Markman, M. (1996) Tumor chemosensitivity and chemoresistance
assays. Cancer 77, 10201025.
6. Cortazar, R. and Johnson, B. E. (1999) Review of the efficacy of individualized
chemotherapy selected by in vitro drug sensitivity testing for patients with cancer.
J. Clin. Oncol. 17, 16251631.
7. Andreotti, P. E., Cree, I. A., Kurbacher, C. M., et al. (1995) Chemosensitivity testing of human tumors using a microplate adenosine triphosphate luminescence
assay: clinical correlation for cisplatin resistance of ovarian carcinoma. Cancer
Res. 55, 52765282.
8. Kangas, L., Grnroos, M., and Nieminen, A. (1984) Bioluminescence of cellular
ATP: a new method for evaluating cytotoxic agents in vitro. Med. Biol. 62, 338343.
9. Kuzmits, R., Rumpold, H., Muller, M., and Schopf, G. (1986) The use of bioluminescence to evaluate the influence of chemotherapeutic drugs on ATP levels of
malignant cell lines. J. Clin. Chem. Clin. Biochem. 24, 293298.
118
Kurbacher and Cree
10. Maehara, Y., Anai, H., Tamada, R., and Sugimachi, K. (1987) The ATP assay is
more sensitive than the succinate dehydrogenase inhibition test for predicting cell
viability. Eur. J. Cancer Clin. Oncol. 23, 273276.
11. Petty, R. D., Sutherland, L. A., Hunter, E. M., and Cree, I. A. (1995) Comparison
of MTT and ATP-based assays for the measurement of viable cell number.
J. Biolum. Chemilum. 10, 2934.
12. Ahmann, F., Garewal, H., Schifman, R., Celniker, A., and Rodney, S (1987) Intracellular adenosine triphosphate as a measure of human tumor cell viability and
drug modulated growth. In Vitro Cell. Dev. Biol. 23, 474480.
13. Sevin, B.-U., Peng, Z., Perras, J., Penalver, G., and Averette, H. (1988) Application
of an ATP bioluminescence assay in human tumor chemosensitivity testing.
14. Koechli, O. R., Avner, B. P., Sevin, B.-U., et al. (1994) Application of the ATP-Cell
Viability Assay in human breast cancer chemosensitivity testing: a report on the
first results. J. Surg. Oncol. 54, 119125.
15. Tereshima, M., Ikeda, K., Kawamura, H., et al. (1994) A comparative study on
the ATP assay in serum-free media and the adhesive tumor cell culture system as
a drug sensitivity test for human esophageal cancer. Cancer Res. Ther. Control 3,
297304.
16. Hunter, E. M., Sutherland, L. A., Cree, I. A., Dewar, J. A., Preece, P. E., and
Andreotti, P. E. (1993) Heterogeneity of chemosensitivity in human breast carcinoma: use of an adenosine triphosphate (ATP) chemiluminescence assay. Eur.
J. Surg. Oncol. 19, 242249.
17. Cree, I. A., Kurbacher, C. M., Untch, M., et al. (1996) Correlation of the clinical
response to chemotherapy in breast cancer with ex vivo chemosensitivity. AntiCancer Drugs 7, 630635.
18. Kurbacher, C. M., Cree, I. A., Brenne, U., et al. (1996) Heterogeneity of in vitro
chemosensitivity in perioperative breast cancer cells to mitoxantrone versus doxorubicin evaluated by a microplate ATP bioluminescence assay. Breast Cancer
Res. Treat. 41, 161170.
19. Kurbacher, C. M., Nagel, W., Mallmann, P., et al. (1994) In vitro activity of
titanocenedichloride in human renal cell carcinoma compared to conventional antineoplastic agents. Anticancer Res. 14, 15291534.
20. Whitehouse, P. A., Knight, L. A., Di Nicolantonio, F., Mercer, S. J., Sharma, S.,
and Cree, I. A. (2003) Heterogeneity of chemosensitivity of colorectal adenocarcinoma determined by a modified ex vivo ATP-tumor chemosensitivity assay. AntiCancer Drugs 14, 369375.
21. Lee, J.-M., Bruckner, H. W., Szrajer, L., and Brenne, U. (1995) Taxol inhibits
growth of mesothelioma xenografts. Anticancer Res. 15, 693696.
22. Myatt, N., Cree, I. A., Kurbacher, C. M., Foss, A. J. E., Hungerford, J. L., and
Plowman, P. N. (1997) The ex vivo chemosensitivity profile of choroidal
melanoma. Anti-Cancer Drugs 8, 756762.
23. Cree, I. A., Neale, M. H., Myatt, N. E., et al. (1999) Heterogeneity of chemosensitivity of metastatic cutaneous melanoma. Anti-Cancer Drugs 10, 437444.
Microplate ATP-TCA
119
24. Ugurel, S., Tilgen, W., and Reinhold, U. (2003) Chemosensitivity testing in malignant melanoma. Recent Results Cancer Res. 161, 8192.
25. Di Nicolantonio, F., Neale M., Onadim, Z., Hungerford, J. L., Kingston, J., and
Cree, I. A. (2003) The chemosensitivity profile of retinoblastoma. Recent Results
Cancer Res. 161, 7380.
26. Kurbacher, C. M., Grecu, O. M., Stier, U., et al. (2003) ATP chemosensitivity testing in ovarian and breast cancer: early clinical trials. Recent Results Cancer Res.
161, 221230.
27. Untch, M., Ditsch, N., Langer, E., et al. (2003) Chemosensitivity testing in gynecologic oncologydream or reality? Recent Results Cancer Res. 161, 146158.
28. Konecny, G., Crohns, C., Pegram, M., et al. (2000) Correlation of drug response
with the ATP tumor chemosensitivity assay in primary FIGO stage III ovarian
cancer. Gynecol. Oncol. 77, 258263.
29. Kurbacher, C. M., Bruckner, H. W., Cree, I. A., et al. (1997) Mitoxantrone combined with paclitaxel as salvage therapy for platinum-refractory ovarian cancer:
laboratory study and clinical pilot trial. Clin. Cancer Res. 3, 15271533.
30. Neale, M. H., Myatt, N., Cree, I. A., et al. (1999) Combination chemotherapy for
choroidal melanoma: ex vivo sensitivity to treaosulfan with gemcitabine or cytosine arabinoside. Br. J. Cancer 79, 14871493.
31. Kurbacher, C. M., Mallmann, P., Kurbacher, J. A., et al. (1994) In vitro activity of
titanocenedichloride versus cisplatin and doxorubicin in primary and recurrent
epithelial ovarian cancer. Anticancer Res. 14, 19611966.
32. Neale, M. H., Charlton, P. A., and Cree, I. A. (2000) Ex vivo activity of XR5000
against solid tumors. Anti-Cancer Drugs 11, 471478.
33. Kurbacher, C. M., Bruckner, H. W., Andreotti, P. E., Kurbacher, J. A., Sa, G.,
and Krebs, D. (1995) In vitro activity of titanocenedichloride in four ovarian carcinoma cell lines evaluated by a microtiter plate ATP bioluminescence assay. AntiCancer Drugs 6, 697704.
34. Kurbacher, C. M., Cree, I. A., Bruckner, H. W., et al. (1998) Use of an ex vivo ATP
luminescence assay to direct chemotherapy for recurrent ovarian cancer. AntiCancer Drugs 9, 5157.
35. Kurbacher, C. M., Stier, U., Jant, M.-M., Cree, I. A., and Bruckner, H. W. (2001)
ATP-assay directed chemotherapy for recurrent ovarian cancer: mature results of an
ISCO clinical study group trial. Proc. Am. Soc. Clin. Oncol. 20, 185b.
36. Sharma, S., Neale, M. H., Di Nicolantonio, F., et al. (2003) Outcome of ATPbased tumor chemosensitvity assay directed chemotherapy in heavily pretreated
ovarian carcinoma. BMC Cancer 3, 19.
37. Hunter, E. M., Sutherland, L. A., Cree, I. A., Subedi, A. M. C., Hartmann, D.,
Linder, D., and Andreotti, P. E. (1994) The influence of storage on cytotoxic drug
activity in an ATP-based chemosensitivity assay. Anti-Cancer Drugs 5, 171176.
38. Pch, G. (1993) Combined Effects of Drugs and Toxic Agents: Modern Evaluation
in Theory and Practice, Springer, Vienna, Austria.
39. Andreotti, P. E., Linder, D., Hartmann, D. M., et al. (1993) ATP tumor chemosensitivity assay: application for solid tumor and leukemia, in Bioluminescence and
120
40.
41.
42.
43.
Kurbacher and Cree
Chemiluminescence: Current Status (Szalay, A. A., Stanley, P. E., and Kricka, L. J.,
eds.), John Wiley & Sons, Chichester, UK, pp. 271275.
Neale, M. H., Lamont, A., Hindley, A., Kurbacher, C. M., and Cree, I. A. (2000)
The ex vivo effect of high concentrations of doxorubicin on recurrent ovarian
carcinoma. Anti-Cancer Drugs 11, 865871.
Cree, I. A. and Kurbacher, C. M. (1999) ATP based tumour chemosensitivity
testing: assisting new agent development. Anti-Cancer Drugs 10, 431435.
Andreotti, P. E., Linder, D., Hartmann, D. M., Cree, I. A., Pazzagli, M., and Bruckner, H. W. (1994) TCA-100 Tumor Chemosensitivity Assay: differences in sensitivity between cultured tumor cell lines and clinical studies. J. Biolum. Chemilum.
9, 373378.
Cree, I. A. (2003) Chemosensitivity testing as an aid to anti-cancer drug and regimen development. Recent Results Cancer Res. 161, 119125.
11
High-Throughput Technology
Green Fluorescent Protein to Monitor Cell Death
Marylne Fortin, Ann-Muriel Steff, and Patrice Hugo
Summary
Reliable assessment of cell death is now pivotal to many research programs aiming at generating new antitumor compounds or at screening cDNA libraries to identify genes with pro- or
antiapoptotic functions. Such approaches need to rely on reproducible, easy handling, and rapid
microplate-based cytotoxicity assays that are amenable to high-throughput screening technologies. We describe here a method for the direct measurement of cell death, based on the detection
of a decrease in fluorescence observed following death induction in cells stably expressing
enhanced green fluorescent protein (EGFP). Our data clearly show that such a decrease in EGFP
fluorescence after cell death induction happens in various cell types, including those routinely
used in anticancer drug screening (i.e., murine and human, lymphoid, fibroblastic, or epithelial
cell lines). Moreover, the decrease in EGFP fluorescence is observed in cells induced to die
by a variety of apoptosis-inducing agents, such as glucocorticoids (dexamethasone), DNAdamaging agents (etoposide, cisplatin), microtubule disorganizers (paclitaxel), protein kinase C
inhibitors (staurosporine), or a caspase-independent apoptotic stimulus (CD45 crosslinking). A
decrease in fluorescence can be assessed either by flow cytometry or with a fluorescence
microplate reader. The kinetics and specificity of this EGFP-based assay were comparable
with those of other conventional techniques used to detect cell death. This novel EGFP-based
microplate assay combines sensitivity and rapidity and is amenable to high-throughput setups,
making it an assay of choice for evaluation of cell cytotoxicity.
Key Words
Green fluorescent protein; apoptosis; cell death; high-throughput screening; flow cytometry;
fluorescence plate reader; cytotoxicity.
1. Introduction
Several approaches have been developed to assess cellular viability based
on the targeting of various cellular components. For instance, measurement of
121
122
Fortin et al.
the ability of a cell to uptake or retain vital probes (e.g., propidium iodide)
(13), measurement of the cellular metabolic activity by mitochondrial activity
(e.g., 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide assay, rhodamine 123 staining) or labeled nucleotide (e.g., radioactive thymidine) incorporation into the DNA (48), the release into the culture medium of cellular
contents (e.g., 51Cr-release assay, intracytoplasmic enzymerelease assay) (9),
the remodeling of chromatin by labeling of DNA with dyes (e.g., DAPI,
Hoechst) (1012), or the specific labeling of nicked DNA (i.e., TUNEL assay)
(13) have conventionally been used to assess cell viability. One drawback
shared by these assays is that they all require the addition of staining reagents
and washing of the cells. Moreover, after these assays are carried out, cells
cannot be reused, thus preventing dynamic studies. A high number of samples
is therefore necessary to cover a range of time points for each drug to be tested.
A simple alternative to these assays would be to use fluorescent probes directly
synthesized within the cells, which could be used to monitor viability of the
cells without affecting their potential to grow further in culture. We provide
here detailed information on a method for the direct evaluation of cell viability,
based on the fact that cells transfected with the enhanced green fluorescent
protein (EGFP) gene display a loss in EGFP-mediated fluorescence after the
induction of cell death.
The GFP from Aequoria victoria has been shown to be an ideal reporter
gene for in vitro (14) and even in vivo assays (15). This protein exhibits a spontaneous fluorescence in diverse cell types that does not require the presence of
cofactors or substrates. Furthermore, GFP fluorescence can be easily monitored and quantified by a conventional flow cytometry unit or microplate-based
fluorometric readers (16). The EGFP-based assay described here for the monitoring of cell death cumulates several advantages:
1. It only requires to derive a cell line of interest stably expressing high levels of
EGFP.
2. It does not necessitate the addition of reagents, substrates, or buffers, thus minimizing time, expenses, and quality control steps.
3. It is as sensitive as other methods for measuring cytotoxicity.
4. It can be used at the cellular or population levels.
5. It is easily amenable to automation in high-throughput screening (HTS) setups.
6. It allows real-time and kinetic measurements of cell death, because the detection
of the loss of EGFP fluorescence does not damage or alter the cells.
Accordingly, applications for this GFP-based cytotoxicity assay are diverse,

ranging from anticancer drug screening to the screening of cDNA libraries for
the identification of new pro- or antiapoptotic genes. Although the technique
described in this chapter is applicable to diverse cell types, including adherent
123
and nonadherent cells, a murine T-cell hybridoma cell line stably expressing
EGFP is used to illustrate this method for detection of apoptosis.
2. Materials
2.1. Generation of Cell Line Stably Expressing EGFP
1. Retroviral vector encoding EGFP (LZRS-pBMN-IRES-EGFP; a gift from Dr. H.
Spits) (17). EGFP is a Ser65 Thr, Phe64 Leu GFP mutant (BD BioSciences,
Clontech, Palo Alto, CA; [18]).
2. Ecotropic BOSC 23 packaging cell line (American Type Culture Collection no.
CRL-11270).
3. Iscoves modified Eagles medium (IMDM) cell culture medium supplemented
with 10% fetal calf serum (FCS) (Wisent, St-Bruno, Quebec, Canada).
4. Freezing medium: 90% (v/v) FCS and 10% (v/v) dimethyl sulfoxide.
5. OPTI medium (Invitrogen, Burlington, ON, Canada).
6. Lipofectamine and Plus reagent (Invitrogen).
7. Phosphate-buffered saline (PBS).
8. Trypsin-EDTA solution: 0.25% trypsin, 0.03% EDTA (Wisent).
9. Puromycin (Sigma Aldrich, Oakville, Ontario, Canada).
10. DO11.10 cells (murine T-cell hybridoma) (19).
11. RPMI-1640 supplemented with 20% FCS (Wisent).
12. Coulter XL flow cytometer (Beckman Coulter, Ville St-Laurent, Quebec, Canada).
13. FACStar cell sorter (Becton Dickinson, San Jose, CA).
2.2. Flow Cytometric Detection of Apoptosis Through Monitoring

of the Decrease in EGFP Fluorescence in Cells Induced
to Undergo Apoptosis
1.
2.
3.
4.
5.
RPMI-1640 cell culture medium supplemented with 5% FCS (Wisent).

Dexamethasone (Sigma Aldrich).
Propidium iodide (PI) (Sigma Aldrich).
Dihydroethidium (Molecular Probes, Eugene, OR).
AnnexinV (BioDesign, Kennebunk, ME).
2.3. Adaptation of GFP-Based Assay to High-Throughput Settings

1. Black/clear-bottomed 96-well plate (Falcon Microtest, Optilux 1; Fisher, Ontario,
Canada).
2. Analyst HTS assay detection system (LJL BioSystem, Sunnyvale, CA) equipped
with 485- and 530-nm excitation and emission filters, respectively.
3. Methods
3.1. Generation of Cell Line Stably Expressing EGFP
The procedure used to generate DO11.10 cells stably expressing EGFP
relies on retroviral infection performed with the LZRS-pBMN-IRES-EGFP
124
Fortin et al.
bicistronic vector (kindly provided by G. Nolan). This vector is a murine

Moloney leukemia virusbased retroviral vector encoding the EGFP located
downstream of an internal ribosome entry site and under the control of the viral
long terminal repeat promoter. The EGFP protein is a Ser65 Thr, Phe64
Leu GFP mutant displaying high fluorescence intensity. The vector also contains ampicillin and puromycin resistance genes for selection into bacteria and
mammalian cells, respectively. Generation of the EGFP-expressing cell lines
involves introduction of the retroviral vector into a packaging cell line (BOSC
23) by standard lipofectamine transfection procedures, and the subsequent
infection of a cell line with the supernatant of the retrovirus-producing packaging cell line, as described in more detail in Subheadings 3.1.1. and 3.2.2.,
respectively. Transduction of the murine T-cell hybridoma DO11.10 serves as
an example to illustrate the technique. Stable EGFP expression can also be
achieved using other standard transduction methods (see Notes 1 and 2).
3.1.1. Transfection of BOSC 23 Packaging Cell Line
Thaw fresh BOSC 23 cells and maintain them in IMDM cell culture medium
supplemented with 10% FCS. Because the viral production capabilities of these
cells diminish after a few weeks in culture, plan to perform transfections as
soon as the cells have recovered (i.e., 4 or 5 d after thawing). We recommend
not to split the cells at a dilution higher than 15 in order to maintain homogeneity and avoid the generation of clonal variants. Do not let the cells become
overconfluent, for cell clump formation might result in reduced transfection
efficiency. It is important to freeze a large number of backup vials (50100) of
BOSC 23 cells at early passages; this will allow uniform and efficient virus
production over several years. To freeze the cells, spin 5 106 cells at 4C,
remove the supernatant, and add 1 mL of cold freezing medium. Allow the
cells to freeze slowly by placing the vials in a Styrofoam holder in a 80C
freezer for 48 h; then place the vials in a liquid nitrogen tank.
3.1.1.1. DAY 0
1. Split the BOSC 23 cells into wells of a six-well plate.
2. Prepare three wells with different cell numbers (e.g., 3 105, 6 105, and 1 106
cells per well), in order to be able to choose the best cell confluence for transfection the following day.
3. Let the cells adhere overnight. The highest transfection efficiencies are obtained
with BOSC 23 cells that are 7080% confluent at the time of transfection.
3.1.1.2. DAY 1
1. Select the well where confluence has reached 7080%. Carefully rinse the cells to
be transfected with OPTI medium or with other serum-free medium (make sure
that the cells do not detach by gently pipetting the medium on the side of the well).
125
2. Prepare the two following solutions:

a. Solution A: Bring 4 g of DNA (at 1 g/L) up to 80 L with OPTI medium, and
then add 20 L of lipofectamine Plus reagent (do not add the Plus reagent directly
to the DNA; it may precipitate out). Let sit at room temperature for 15 min.
b. Solution B: Combine 8 L of lipofectamine and 92 L of OPTI, and let sit at
room temperature for 15 min.
3. Combine solutions A and B and incubate for 1530 min at room temperature
(solution C).
4. Add 800 L of OPTI medium to solution C and dispense into wells containing
cells to be transfected. Keep one well of cells without DNA as a negative control
(solution C without DNA).
5. Incubate for 3 h at 37C under gentle rocking.
6. Add 4 mL of IMDM medium supplemented with FCS to reach a 20% final concentration of serum and culture overnight.
3.1.1.3. DAY 2
1. Remove the medium; rinse with 4 mL of PBS; and add 1 mL of 0.25% trypsin,
0.03% EDTA solution. Incubate the plate at 37C until the cells detach. Resuspend the cells in fresh IMDM 10% FCS medium, and transfer into a 10-cm Petri
dish. Keep an aliquot of cells for determination of the percentage of EGFP+ cells
by flow cytometry analysis. Prepare an aliquot of nontransfected cells as a negative control. Transfection efficiencies usually range between 5 and 15%. Retrovirus-producing cells must be handled in Level 2 facilities (for more details on
safety issues related to the use of retroviruses, see Biosafety in Microbiological
and Biomedical Laboratories, http://www.cdc.gov/od/ohs/pdffiles/4th%20BMBL.
pdf).
2. Begin selection by culturing the cells in IMDM 10% FCS medium supplemented
with 1 g/mL of puromycin. Selection of puromycin-resistant cells takes from 4 to
6 d. During the selection process, it is possible that dead cells will accumulate and
that the culture medium will become yellow. In this case, simply aspirate the medium
to remove the nonadherent dead cells and add fresh medium containing puromycin.
However, do not trypsinize the cells until the selection process is complete.
3.1.1.4. DAYS 68
1. When the cells are confluent, trypsinize and transfer the cells into a 15-cm Petri
dish with 10 mL of IMDM 10% FCS medium supplemented with 1 g/mL of
puromycin. Keep an aliquot of the cells for determination of the percentage of
EGFP+ cells by flow cytometry analysis (use nontransfected cells as a negative
control). A high percentage of EGFP+ cells (80% or more) is required to ensure
maximal viral production. If the percentage of EGFP+ cells is higher than 80%,
plan to collect the viral supernatant the next day or 2 d later, i.e., when the cells
reach 75% confluence in the 15-cm Petri dish. If the percentage of EGFP+ cells
has not reached 80%, maintain the selection with 1 g/mL of puromycin for an
additional 35 d and split the cells when necessary. If this does not result in an
126
Fortin et al.
increase in EGFP+ cells, the cells should be discarded and a new transfection with
freshly thawed cells should be performed.
3.1.2. Infection of DO11.10 Murine T-Cell Hybridoma

with Retroviral Supernatant
3.1.2.1. DAY 1
1. Rinse the BOSC 23 producer cells twice with 10 mL of OPTI medium.
2. For viral production, add 10 mL of OPTI medium and incubate for 6 h at 37C
with gentle rocking.
3. Harvest the viral supernatant, centrifuge for 10 min at 200g, and filtrate the supernatant on an acetate filter (0.45 m) to remove any cell debris. Prepare 2-mL
aliquots and freeze at 80C or use immediately. Frozen supernatant can be kept
for up to 3 mo; however, fresh viral supernatant gives better transduction efficiencies.
4. Keep the producer cells with IMDM 10% FCS supplemented with puromycin for
a second viral production the following day; then discard the producer cells.
5. Take a 2-mL aliquot of viral supernatant, add 2 L of lipofectamine, and incubate
at room temperature for 15 min (solution D).
6. Wash the DO11.10 cells (target cells) with PBS, and resuspend 2 105 cells in
2 mL of solution D (viral supernatant plus lipofectamine).
7. Transfer the cells to a well in a six-well plate and culture overnight at 37C under
gentle rocking. Plate the cells without viral supernatant as a negative control.
3.1.2.2. DAY 2
1. Add 2 mL of RPMI-1640 supplemented with 20% FCS to each well. Culture
overnight at 37C.
3.1.2.3. DAY 3
1. Transfer the cells of a well into a T75 flask. Keep an aliquot of cells for the determination of the percentage of EGFP+ cells by flow cytometry, using noninfected
cells as a negative control. Transduction efficiencies usually range between 2 and
15% (for this specific cell type). At this stage, puromycin selection can be used in
order to increase the proportion of EGFP+ cells before cell sorting (the optimal
puromycin concentration must be determined for each specific cell type). Plan to
accumulate large quantities of retrovirally transduced cells for cell sorting by fluorescence-activated cell sorting (FACS).
3.1.2.4. DAYS 47
1. Sort the EGFP+ cells by FACS. Many successive rounds of cell sorting might be
required in order to obtain cell populations in which nearly 100% of the cells are
EGFP+ (see Note 3). It is also important to obtain cells expressing high levels of
EGFP, because the higher the intensity of fluorescence, the easier will be the discrimination of dead cells in future assays. It is recommended that EGFP-expressing and EGFP cells display at least one log difference in fluorescence intensity.
127
2. When a cell culture in which nearly 100% of the cells express high levels of EGFP
is obtained (DO11.10-EGFP), prepare a large number of vials (2550) of frozen
cells for subsequent uses (see Subheading 3.1.1. for the freezing procedure).
3.2. Flow Cytometric Detection of Apoptosis Through Monitoring

of Decrease in EGFP Fluorescence
in Cells Induced to Undergo Apoptosis
This section describes the EGFP-based assay for the monitoring of cell death
by flow cytometry and provides validation of the technique by comparison with
other commonly used markers for detection of apoptosis. The induction of
apoptosis by dexamethasone treatment in DO11.10-EGFP cells is used as an
example to illustrate the method. Other cell types and other apoptosis-inducing
agents can be used as well (see Note 4).
3.2.1. Induction of Apoptosis
1. Thaw fresh DO11.10-EGFP cells and maintain them in RPMI-1640 cell culture
medium supplemented with 5% FCS. Because the susceptibility to apoptosis induction increases with the time in culture, use only freshly thawed cells (less than
10 passages). In addition, we recommend using exponentially growing cells (less
than 1 106 cells/mL). Make sure that the percentage of EGFP+ cells is near 100%
(among viable cells) by flow cytometry analysis before starting an experiment.
2. Plate DO11.10-EGFP cells at a concentration of 4 105 cells/mL in the wells of
a flat-bottomed 96-well plate (8 104 cells/well in 200 L of RPMI-1640 5%
FCS). Prepare triplicate wells for untreated and for treated cells.
3. Using a stock solution of dexamethasone at 103 M (diluted in ethanol), prepare
intermediate dilutions and add dexamethasone to the wells to reach a final concentration of 107 M. The final concentration of ethanol in the culture medium
should not exceed 1% (v/v). Add an equivalent concentration of ethanol to
untreated control cells.
4. Incubate at 37C, 5% CO2 for 16 h.
5. Harvest and transfer the cells into a conical 96-well plate, centrifuge at 85g for
3 min, and remove the supernatant.
6. Wash the cells twice with 200 L of PBS.
7. Resuspend the cells in 400 L of PBS and transfer into microtiter tubes for direct
flow cytometry analysis. No additional procedure is required for EGFP detection
by flow cytometry. The cells can be kept on ice for up to 2 h before analysis.
3.2.2. Measurement of Apoptosis Induction Through Monitoring

of Decrease in EGFP Fluorescence by Flow Cytometry
1. Using noninfected, parental DO11.10 cells, set the parameters of the flow cytometer. First, adjust the forward side (FS) and side scatter (SS) parameters in order to
be able to see both viable (high FS, low SS) and dead (low FS, high SS) cells on
an FS/SS dot plot (see Fig. 1B, top panel). Second, set the FL-1 voltage such that
EGFP (noninfected cells) appear in the first log decade on an FL-1 histogram.
128
Fortin et al.
Fig. 1. Dead cells exhibit decreased EGFP fluorescence. (A) DO11.10-EGFP cells
were treated with dexamethasone (107 M for 16 h). EGFP fluorescence was detected by
flow cytometry in treated (dashed line) or untreated (solid line) cells. (B) DO11.10-EGFP
cells were treated as in (A). Electronic gating was performed to evaluate EGFP fluorescence in either viable (R1 region) or dead (R2 region) cells, as determined by a reduction
in forward scatter. Data are representative of at least five independent experiments.
129
Using untreated DO11.10-EGFP cells, make sure that at least one log difference
in fluorescence intensity is detected between EGFP-expressing and EGFP cells.
2. For determination of the percentage of apoptotic cells in dexamethasone-treated
DO11.10-EGFP cells, draw a region (R1) on the FS/SS dot plot that includes both
viable (high FS, low SS) and dead (low FS, high SS) cells, excluding cell debris
that may appear on the lower left corner of the FS/SS dot plot. Two well-discriminated peaks should be visible on the FL-1 histogram that is gated on R1;
the FL-1-negative peak (EGFP-low) represents dead cells and the FL-1-positive
peak (EGFP-high) represents viable cells. The percentage of apoptotic cells can
then be obtained by drawing a region for the negative peak.
An example of apoptosis induction in DO11.10-EGFP cells by dexamethasone is presented below. In DO11.10-EGFP cells, induction of apoptosis with
dexamethasone causes a dramatic decrease in EGFP fluorescence readily
detectable by flow cytometry (Fig. 1A, dashed line). DO11.10-EGFP cells displayed the same sensitivity to apoptosis induction as parental DO11.10 cells
(data not shown). The small proportion of EGFP cells that can be seen in nontreated cells (Fig. 1A, solid line) is owing to background cell death in the culture (<10%). To verify whether EGFP cells represent dead cells, a gate was
drawn around cells displaying reduced forward scatter (FS), a characteristic of
dead cells (Fig. 1B, R2 region in top panel). DO11.10-EGFP cells included in
the R2 gate were no longer fluorescent (Fig. 1B, bottom panel). By contrast, in
the remaining viable DO11.10-EGFP cells, as determined by unchanged FS
(Fig. 1B, R1 region in top panel), high levels of EGFP fluorescence were
detected (Fig. 1B, middle panel). This shows that in a mixture of live and dead
cells, only the latter display reduced EGFP fluorescence. EGFP fluorescence in
dead cells returned to background levels and was indistinguishable from autofluorescence of noninfected, EGFP parental cells (Fig. 1B, bottom panel and
data not shown). Thus, the loss of EGFP fluorescence parallels the loss of cell
viability as determined by cell shrinkage in this lymphocytic cell line.
3.2.3. Validation of EGFP-Based Assay by Correlation with Other Markers
Used for Detection of Apoptosis by Flow Cytometry
We examined in more details the relationship between the loss in EGFP fluorescence and cell death using other conventional apoptotic markers. For this
purpose, DO11.10-EGFP cells were treated with 107 M dexamethasone for
20 h and stained with either (1) PI, a vital dye that is retained in dead cells; (2)
annexin V, a protein that binds to membrane phosphatidylserine (PS) residues,
which are externalized during the early phases of apoptosis (20); or (3) dihydroethidium, which is oxidized to red fluorescent ethidium (21) when reactive
130
Fortin et al.
Fig. 2. Loss of EGFP fluorescence correlates well with acquisition of conventional

apoptosis markers after cell death induction. DO11.10-EGFP cells (8 104) were
treated with 107 M dexamethasone for 20 h; harvested; and stained with either PI,
annexin V, or dihydroethidium (for ROS detection). EGFP fluorescence and each of
these apoptotic markers were examined simultaneously by flow cytometry. Note that all
the dying or dead cells, according to conventional apoptotic markers, are also EGFP
(i.e., located in the upper-left quadrant, with very few cells appearing in the upperright quadrant). Percentages of dead cells (upper-left quadrant) and viable cells (lowerright quadrant) are indicated for untreated and dexamethasone-treated cells.
oxygen species (ROS) are produced in apoptotic cells. All these alternative procedures for the detection of cell death were conducted as described in the manufacturers guidelines. Figure 2 shows that according to these apoptotic
markers, all dead cells in the dexamethasone-treated samples, i.e., PI+, annexin
V+, and ROS+ cells, also exhibited a dramatic loss in EGFP fluorescence (see
upper-left quadrants in Fig. 2). Moreover, <5% of the cells stained by one of
131
these conventional methods remained positive for EGFP fluorescence (see

upper-right quadrants in Fig. 2 for dexamethasone-treated cells). Similarly, <8%
of the cells with low EGFP fluorescence remained negative for apoptosis markers (see lower-left quadrants in Fig. 2 for dexamethasone-treated cells).
Together, these results indicate that, under these experimental conditions, loss
in EGFP fluorescence displays a very strong correlation with traditional apoptotic markers for the detection of dead cells.
The exposure of PS residues at the outer layer of the cell membrane is recognized to be a rapid event following apoptosis induction, which precedes other
apoptotic features such as cell shrinkage or retention of vital dyes (20). We
wished to verify whether loss in EGFP fluorescence was also an early event
in the apoptotic process. To this end, we performed a kinetic study (depicted in
Fig. 3A), in which DO11.10-EGFP cells were treated with 107 M dexamethasone for various durations, and then stained with annexin V to detect PS. The
percentage of annexin V+ cells is well correlated with the percentage of EGFPlow cells for each time point, indicating that the decrease in EGFP fluorescence in DO11.10-EGFP cells is as rapid as PS exposure after treatment with
dexamethasone. Furthermore, the EGFP-based assay is as sensitive as annexin V
staining to detect small proportions of apoptotic cells generated at low doses of
dexamethasone treatment (Fig. 3B) (see Notes 5 and 6).
3.3. Adaptation of EGFP-Based Assay to High-Throughput Settings
Easy handling of samples combined with sensitivity makes the EGFP
biosensor a readout of choice for microplate-based cytotoxicity assays that
could be used in high-throughput settings for various applications. As a proof
of concept for this approach, we tested whether apoptosis could be detected in
EGFP-expressing cells using a microplate fluorescence reader. A dose-response
experiment in which the DO11.10-EGFP cells are induced to die by dexamethasone treatment is used to illustrate the method.
3.3.1. Induction of Apoptosis
1. Plate DO11.10-EGFP cells at a concentration of 4 105 cells/mL in a flat-bottomed 96-well plate (8 104 cells/well in 200 L of RPMI-1640 10% FCS). Prepare triplicate wells for each dose of apoptosis-inducing agent.
2. Add dexamethasone to the wells to reach final concentrations ranging from 109
M to 106 M. Prepare cells without dexamethasone as a negative control.
3. Incubate at 37C, 5% CO2 for 16 h.
4. Harvest and transfer the cells into a conical 96-well plate, centrifuge, and remove
supernatant.
5. Wash the cells twice with 200 L of PBS (see Note 7).
132
Fortin et al.
Fig. 3. Loss of EGFP fluorescence is as rapid and sensitive as annexin V staining for
the detection of cell death. DO11.10-EGFP cells were incubated with 107 M dexamethasone for various times (A) or with different doses of dexamethasone for 20 h (B).
Cells were harvested and stained with annexin V. Mean percentages (SD) of cells displaying low EGFP fluorescence () or positive annexin V staining (), both determined by flow cytometry, are plotted.
3.3.2. Fluorescence Plate Reader Analysis of Apoptosis Induction

1. Resuspend the cells in 200 L of PBS and transfer into a black/clear-bottomed 96well plate.
2. Measure the fluorescence emitted by the total cell population with an automated
fluorescence microplate reader, using excitation and emission filters of 485 and
530 nm, respectively. Figure 4A shows arbitrary fluorescence units plotted against
the doses of dexamethasone used. The total fluorescence emitted by the cell population decreases with increasing doses of dexamethasone, which is indicative of
cell death induction.
133
3.3.3. Validation of Fluorescence Plate Reader Data by Correlation

with Flow Cytometric Data
To provide validation for the EGFP-based assay for apoptosis detection by

automated microplate fluorescence reading, we have compared the data
obtained by the plate reader with flow cytometric data obtained by annexin V
staining of apoptotic cells. After treatment of DO11.10-EGFP cells with
increasing doses of dexamethasone (see Subheading 3.3.1.), half of the cells
were stained with annexin V and analyzed by flow cytometry, and the other half
was transferred to a black/clear-bottomed plate and analyzed by a fluorescence
microplate reader. As shown in Fig. 4B, a very high correlation (R2 = 0.98)
was found between these two different methods of detecting cell death, thus
indicating that the EGFP-based method is amenable to the detection and quantification of cell death in a high-throughput screening setup (see Note 8).
4. Notes
1. Stable expression of EGFP can be achieved using different vectors or other transduction methods, such as conventional transfection with expression vectors encoding for EGFP (pEGFP-N1; Clontech). However, it should be kept in mind that a
near 100% EGFP+ cell population and high EGFP expression levels are required
for optimal application of the apoptosis detection assay. In our hands, retroviral
infection was more efficient than standard plasmid transfection methods (for the
specific cell types used here) and, therefore, required a reduced number of successive rounds of cell sorting. For the same reason, transient transduction methods
(e.g., with adenoviral vectors) are not recommended.
2. We have only carried out this EGFP-based apoptosis assay with the EGFP (Clontech, BD BioSciences) variant of GFP. Other variants of GFP exist (e.g., the
SuperGlo Green Fluorescent Protein variant from QBiogene) that could possibly
be used for the detection of apoptosis; however, this remains to be verified. Alternatively, other fluorescent proteins (such as yellow fluorescent protein or red fluorescent protein) could represent interesting alternatives and might allow
combinations with other fluorochromes.
3. One problem that we have encountered with this technique is that for some cell
types high intensity of EGFP expression is difficult to obtain. This results in poor
discrimination between EGFP+ and EGFP cell populations by flow cytometry.
Precise determination of the percentage of dead cells is therefore hampered. This
difficulty could be overcome by using brighter GFP variants for transduction or by
sorting the cells in order to keep only the brightest ones.
4. The EGFP-based assay described here should work with a variety of different cell
types. We have used it with NIH 3T3 (murine fibroblasts) and human cancer cell
lines such as HEC-1A (endometrial), MCF-7 (breast), and PC-3 (prostate) (22).
Moreover, we detected the loss in EGFP fluorescence following apoptosis induction with a variety of drugs, such as paclitaxel, etoposide, cisplatin, and stau-
134
Fortin et al.
Fig. 4. Fluorescence plate reader analysis of apoptosis induction. (A) After treatment with various doses of dexamethasone as in Fig. 3, the cells were transferred to a
black/clear-bottomed microplate and fluorescence was assessed on a fluorescence
microplate reader. Arbitrary fluorescent units are plotted for each dose of dexamethasone used (mean of triplicates). (B) After treatment as in (A), half of the cells were
stained with annexin V and analyzed by flow cytometry; the other half was transferred
to a black/clear-bottomed microplate and fluorescence was assessed on a fluorescence
microplate reader. Percentages of annexin V+ cells and arbitrary fluorescent units are
plotted for each dose of dexamethasone used (mean of triplicates). Linear regression
and correlation coefficient are indicated.
rosporine (22). We have also found that loss of EGFP fluorescence can be used to
detect apoptosis induced by a caspase-independent cell death stimulus (CD45
crosslinking on murine lymphocytes) (22). The selection of the appropriate cell
line depends on the objectives set forth in the investigation.
5. An observation made by Strebel et al. (23) suggested that apoptotic and necrotic
cells could be distinguishable, based on the fact that apoptosis led to a partial
135
decrease in EGFP fluorescence, whereas necrosis led to a complete loss in fluorescence. However, such discrimination between apoptotic and necrotic cells was
not observed in our system, because cells that were proapoptotic (annexin V+ and
PI) had already completely lost EGFP fluorescence.
6. The mechanism by which EGFP fluorescence decreases under these conditions is
not yet elucidated and is beyond the scope of this chapter. However, we ruled out
the possibility that EGFP could leak out of the cells, because Western blot analysis indicated that the amount of EGFP was similar in living or dying cells.
Another possibility was that EGFP could be degraded by proteases such as caspases. However, the decrease in EGFP fluorescence was observed even when cell
death was induced in the absence of caspase activation (i.e., in the presence of
caspase inhibitors). This does not rule out that other proteases such as calpains
could participate in EGFP degradation, but as detected by Western blot, the EGFP
levels did not seem to be reduced and EGFP did not appear to be cleaved in cells
undergoing cell death. The fact that EGFP fluorescence is quenched by modifications in the intracellular milieu is another possibility. For instance, it has been
shown that fluorescence of EGFP and other GFP mutants is reduced with cytosolic acidification (24). Such a decrease in pH (about 0.4 units) is observed during
the apoptotic process (25,26). However, for the dramatic loss of EGFP fluorescence observed here to occur, changes in pH would have to be far below physiological levels (pH <6.0). Yet another possibility relies on the fact that chromophore
formation is posttranslationally regulated and depends on the oxidation of Tyr66,
which requires molecular oxygen (14). It is therefore conceivable that the redox
changes occurring during the apoptotic process (27) are responsible for the loss of
EGFP fluorescence.
7. Induction of cell death can also be performed directly in sterile black/clearbottomed 96-well plates to eliminate the need for harvesting and washing the
cells. The only requirement to perform apoptosis induction and fluorescence reading in the same microplate is that the apoptosis-inducing agent must not interfere
with detection of EGFP fluorescence. After measurement of fluorescence, cells
can be readily returned to culture, because they have not been manipulated.
8. The EGFP biosensor assay described here would have to be improved for HTS
setups, which measure the global fluorescence in microplate wells, in order to
clearly make the distinction between reduced fluorescence owing to either cytotoxic (i.e., death) or cytostatic (i.e., lack of growth) effects. Hence, kinetic measurements of the decrease in EGFP fluorescence would allow distinction between
induction of cell death and cell growth arrest. Given that detection of the loss in
EGFP fluorescence does not alter the cells, such kinetic studies would be easily
feasible.
Acknowledgment
Portions of this chapter appeared in Cytometry (22). Used by permission of
Wiley-Liss Inc., a subsidiary of John Wiley & Sons, Inc.
136
Fortin et al.
References
1. Nieminen, A. L., Gores, G. J., Bond, J. M., Imberti, R., Herman, B., and Lemasters,
J. J. (1992) A novel cytotoxicity screening assay using a multiwell fluorescence
scanner. Toxicol. Appl. Pharmacol. 115, 147155.
2. Rat, P., Korwin-Zmijowska, C., Warnet, J. M., and Adolphe, M. (1994) New in
vitro fluorimetric microtitration assays for toxicological screening of drugs. Cell.
Biol. Toxicol. 10, 329337.
Cancer Res. 43, 749757.
4. Larsson, R., Nygren, P., Ekberg, M., and Slater, L. (1990) Chemotherapeutic drug
sensitivity testing of human leukemia cells in vitro using a semiautomated fluorometric assay. Leukemia 4, 567571.
5563.
6. Pavlik, E. J., Flanigan, R. C., van Nagell, J. J., et al. (1985) Esterase activity, exclusion of propidium iodide, and proliferation in tumor cells exposed to anticancer
agents: phenomena relevant to chemosensitivity determinations. Cancer Invest. 3,
413426.
7. Scudiero, D. A., Shoemaker, R. H., Paull, K. D., Monks, A., Tierney, S., Nofziger,
T. H., Currens, M. J., Seniff, D., and Boyd, M. R. (1988) Evaluation of a soluble
tetrazolium/formazan assay for cell growth and drug sensitivity in culture using
human and other tumor cell lines. Cancer Res. 48, 48274833.
8. Skehan, P., Storeng, R., Scudiero, D., Monks, A., McMahon, J., Vistica, D.,
Warren, J. T., Bokesch, H., Kenney, S., and Boyd, M. R. (1990) New colorimetric cytotoxicity assay for anticancer-drug screening. J. Natl. Cancer Inst. 82,
11071112.
9. Korzeniewski, C. and Callewaert, D. M. (1983) An enzyme-release assay for natural cytotoxicity. J. Immunol. Methods 64, 313320.
10. Douglas, R. S., Tarshis, A. D., Pletcher, C. H. Jr., Nowell, P. C., and Moore, J. S.
(1995) A simplified method for the coordinate examination of apoptosis and surface phenotype of murine lymphocytes. J. Immunol. Methods 188, 219228.
11. Darzynkiewicz, Z., Bruno, S., Del Bino, G., et al. (1992) Features of apoptotic
cells measured by flow cytometry. Cytometry 13, 795808.
12. Dive, C., Gregory, C. D., Phipps, D. J., Evans, D. L., Milner, A. E., and Wyllie, A.
H. (1992) Analysis and discrimination of necrosis and apoptosis (programmed cell
death) by multiparameter flow cytometry. Biochim. Biophys. Acta 1133, 275285.
13. Gorczyca, W., Gong, J., and Darzynkiewicz, Z. (1993) Detection of DNA strand
breaks in individual apoptotic cells by the in situ terminal deoxynucleotidyl transferase and nick translation assays. Cancer Res. 53, 19451951.
14. Kain, S. R. (1999) Green fluorescent protein (GFP): applications in cell-based
assays for drug discovery. Drug Discov. Today 4, 304312.
137
15. Hoffman, R. M. (1999) Orthotopic metastatic mouse models for anticancer drug
discovery and evaluation: a bridge to the clinic. Invest. New Drugs 17, 343359.
16. Tsien, R. Y. (1998) The green fluorescent protein. Annu. Rev. Biochem. 67,
509544.
17. Blom, B., Heemskerk, M. H., Verschuren, M. C., et al. (1999) Disruption of alpha
beta but not of gamma delta T cell development by overexpression of the helixloop-helix protein Id3 in committed T cell progenitors. EMBO J. 18, 27932802.
18. Yang, T. T., Cheng, L., and Kain, S. R. (1996) Optimized codon usage and chromophore mutations provide enhanced sensitivity with the green fluorescent protein.
Nucleic Acids Res. 24, 4592, 4593.
19. Haskins, K., Kubo, R., White, J., Pigeon, M., Kappler, J., and Marrack, P. (1983)
The major histocompatibility complexrestricted antigen receptor on T cells.
I. Isolation with a monoclonal antibody. J. Exp. Med. 157, 11491169.
20. Martin, S. J., Reutelingsperger, C. P., McGahon, A. J., et al. (1995) Early redistribution of plasma membrane phosphatidylserine is a general feature of apoptosis
regardless of the initiating stimulus: inhibition by overexpression of Bcl-2 and Abl.
J. Exp. Med. 182, 15451556.
21. Carter, W. O., Narayanan, P. K., and Robinson, J. P. (1994) Intracellular hydrogen
peroxide and superoxide anion detection in endothelial cells. J. Leukoc. Biol. 55,
253258.
22. Steff, A. M., Fortin, M., Arguin, C., and Hug, P. (2001) Detection of a decrease in
green fluorescent protein fluorescence for the monitoring of cell death: an assay
amenable to high-throughput screening technologies. Cytometry 45, 237243.
23. Strebel, A., Harr, T., Bachmann, F., Wernli, M., and Erb, P. (2001) Green fluorescent protein as a novel tool to measure apoptosis and necrosis. Cytometry 43,
126133.
24. Matsuyama, S., Llopis, J., Deveraux, Q. L., Tsien, R. Y., and Reed, J. C. (2000)
Changes in intramitochondrial and cytosolic pH: early events that modulate caspase activation during apoptosis. Nat. Cell. Biol. 2, 318325.
25. Barry, M. A. and Eastman, A. (1992) Endonuclease activation during apoptosis: the
role of cytosolic Ca2+ and pH. Biochem. Biophys. Res. Commun. 186, 782789.
26. Gottlieb, R. A., Nordberg, J., Skowronski, E., and Babior, B. M. (1996) Apoptosis
induced in Jurkat cells by several agents is preceded by intracellular acidification.
Proc. Natl. Acad. Sci. USA 93, 654658.
27. Buttke, T. M. and Sandstrom, P. A. (1995) Redox regulation of programmed cell
death in lymphocytes. Free Radic. Res. 22, 389397.
12
DIMSCAN
A Microcomputer Fluorescence-Based Cytotoxicity Assay
for Preclinical Testing of Combination Chemotherapy
Nino Keshelava, Toms Frgala, Jir Krejsa, Ondrej Kalous,
and C. Patrick Reynolds
Summary
DIMSCAN is a semiautomatic fluorescence-based digital image microscopy system that
quantifies relative total (using a DNA stain) or viable (using fluorescein diacetate [FDA]) cell
numbers in tissue culture multiwell plates ranging from 6 to 384 wells per plate. DIMSCAN is
a rapid and efficient tool for conducting in vitro cytotoxicity assays across a 4 log dynamic
range. The specificity of detecting viable cells with FDA is achieved by using digital image processing and chemical quenching of fluorescence in nonviable cells with eosin Y. Average scan
time for the most commonly used format, a 96-well plate, is 6 min. Cytotoxicity for neuroblastoma cell lines measured by DIMSCAN was found to be comparable to manual Trypan blue
dye exclusion counts or colony formation in soft agar, but with a significantly wider dynamic
range, which enables drug combination studies used to detect synergistic or antagonistic interactions. The linearity of DIMSCAN was validated (r2 = 0.99967 0.0003) for cells stained with
FDA deposited using a fluorescence-activated cell sorter, documenting a dynamic range >4 logs,
and the ability to detect a single viable cell in a well 93% of the time. DIMSCAN has been
used to demonstrate preclinical activity of cytotostatic and cytotoxic drugs and drug combinations
that have subsequently shown activity in clinical trials.
Key Words
DIMSCAN; cytotoxicity assay; cytotostatic assay; digital image microscopy; fluorescence;
fluorescein diacetate; 2456-tetrabromofluorescein; Hoechst 33342; eosin Y.
1. Introduction
A number of different assays have been developed to measure in vitro cytotoxicity using tumor cell lines to assess the activity of antineoplastic drugs.
139
140
Keshelava et al.
Table 1
Dynamic Range of Cytotoxicity Assays
Assay
Dynamic
range
References
Clonogenic assay in semisolid medium

Dye exclusion assay (Trypan blue)
MTT colorimetric assay
Sulforhodamine-B colorimetric assay
WST-1 + Neutral Red + Crystal Violet assay
Adenosine triphosphate cell viability assay
ChemoSelect Test (cytosensor microphysiometer)
Differentiated staining cytotoxicity assay
Collagen gel embedded culture and image analysis
[3H]-Uridine assay
Fluorometric microculture cytotoxicity assay (FMCA)
Flow cytometric quantification of FDA/PI viable cell number
DIMSCAN assay
12 logsa
2 logsa
2 logsa
2 logsa
2 logsa
2 logsa
2 logsa
2 logsa
2 logsa
23 logsa
23 logsa
4 logsa
4 logsa
2,30
31
24
9
32
33
34
35
36
8
37
11
15,18
a
When linear scale was used for data presentation rather than a common log scale, the dynamic
range was assessed as no greater than 2 logs.
Colony-forming assays in semisolid medium and dye exclusion assays are

labor-intensive, are subject to observer error, and have a limited dynamic range
(can detect only 1 to 2 logs of cell kill). Assays employing 96-well microplates
allow automation and the use of standard tissue culture conditions to improve
clonogenicity of tumor cell lines, and such assays enable the exchange of
medium for drug washout experiments or multiweek assays. Cell numbers in
96-well plates can be quantified using radioactive isotopes (51Cr release, or
3
H-uridine incorporation), colorimetric measurement of substrates altered by
viable cells such as 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) (17), or protein-binding dyes such as sulforhodamine (24,8,9).
However, many of these assays have a limited (usually not greater than 2 logs)
dynamic range (Table 1).
Alternatively, cytotoxicity can be measured with fluorochromes, such as fluorescein diacetate (FDA) or 2,7-bis(carboxymethyl)-5,6-carboxyfluorescein
(BCECF), which accumulate selectively in viable cells. Such fluorescent dyes
are used either alone or in combination with DNA fluorescent dyes that do not
penetrate viable cells (1014). Fluorescence-based systems offer a number of
advantages, such as a lack of radioactive waste, short incubation times, the
DIMSCAN Cytotoxicity Assay
141
ability to brightly stain individual cells, and the ability to rapidly measure cell
numbers directly in microplates using a 96-well fluorescence reader (1517).
However, commercially available 96-well fluorescence readers provide a limited dynamic range, because they cannot discriminate between the fluorescence
from viable cells and background fluorescence, a problem only partially overcome by the use of expensive custom filtration plates (12).
DIMSCAN is a new method developed in our laboratory at Childrens Hospital Los Angeles and is superior to other in vitro microwell cytotoxicity assays
owing to its sensitivity and linearity over a range >4 logs. This rapid, semiautomated 96-well cytotoxicity assay employs the fluorescent dye FDA, which is
converted into fluorescein by intracellular esterases in viable cells and becomes
trapped within cells with intact membrane integrity. The wide dynamic range is
achieved because DIMSCAN decreases background fluorescence using both
digital image processing (15) and quenching of fluorescence in medium and
dead cells with 2456-tetrabromofluorescein (eosin Y) (18). This system has
been proven to be useful for both attached and suspension cell cultures of a
variety of cell types, including normal lymphocytes and cell lines from
leukemias and a wide variety of solid tumors (1921).
The sensitivity and linearity of the system were validated with various numbers of FDA-stained CEM leukemia cells deposited per well using a fluorescence-activated cell sorter, eight replicates per plate. A linear relation (linear
regression of the first order) between the number of viable cells deposited and
DIMSCAN quantitation was observed (r 2 = 0.99967 0.0003), documenting a
dynamic range >4 logs. In similar experiments using neuroblastoma cells,
slightly less linearity was observed owing to cell clumping, but the dynamic
range remained >4 logs. The linearity and dynamic range were preserved
during repeated scanning up to five times with a correlation coefficient of
r 2 >0.999. Detection of single viable cells by DIMSCAN was achieved in 93%
of wells seeded without any false-positive wells (22).
The wide linear range offered by DIMSCAN allows the user to more accurately measure concentrations of drugs lethal for 90 or 99% of treated cells
(i.e., LC90 or LC99 values, also expressed as 1 log or 2 logs of cell kill), as
opposed to more commonly used LC50 values. Obtaining at least a 2 log cytotoxicity is accepted as the amount of cell kill necessary to achieve a partial
response (23). This points toward the importance of using LC99 values at a clinically achievable level as criteria for defining a potentially active agent based on
in vitro data. Equally important, the 4 log dynamic range renders DIMSCAN
an invaluable tool for determining synergistic, additive, and antagonistic effects
for tested drug combinations. Such assessments are limited in systems with a
dynamic range of 1 to 2 logs.
142
Keshelava et al.
DIMSCAN can be used as a cytostatic assay if lower numbers of cells are

seeded and the growth inhibitory effect of a tested agent is measured 14 21 d
after initiating the experiment (vs 37 d as a cytotoxicity assay).
As a cytostatic assay, DIMSCAN was used to develop 13-cis-retinoic acid
for neuroblastoma, which led to successful phase I and phase III clinical trials of
that drug (24). Results from DIMSCAN cytotoxicity testing have been used to
identify promising new agents for neuroblastoma that have shown responses in
clinical trialsbuthionine sulfoximine (BSO) + melphalan (25,26) and low-dose
(oral) fenretinide ([27]; unpublished data)and also to develop additional agents
currently in or about to enter phase I trials [prolonged infusion of pyrazoloacridine (28), or high-dose (iv) fenretinide (27), and fenretinide + safingol (21)].
2. Materials
1. FDA (Sigma, St. Louis, MO). Prepare a 1 mg/mL solution in dimethyl sulfoxide,
aliquot into 1.5-mL Eppendorf tubes, and store at 20C in the dark. Avoid
repeated thawing and refreezing.
2. Hoechst 33342 (Calbiochem, San Diego, CA). Dissolve in double-distilled water
to make a 1 mg/mL stock solution, and store in 1.5-mL Eppendorf tubes in the
dark at 4C.
3. Eosin Y. Prepare a 1% stock solution (w/v) in 0.9% NaCl and store in an amber
bottle at room temperature.
4. Precision 2000 robot pipettor (Bio-Tek).
5. DIMSCAN system (Fig. 1; www.DIMSCAN.com). This consists of an inverted
microscope; a stepper motor scanning stage; a stage controller; a charge-coupled
device [CCD] camera; and a Pentium 4 microcomputer (Microsoft Windows
2000) running the main application software (DIMSCAN 3.0, developed at Childrens Hospital Los Angeles), which controls stage movement and processes CCD
camera images ([15,22]; unpublished). An inverted microscope, Olympus IX50, is
equipped with a 103-W mercury vapor lamp (HBO-103 W/2), optical filters
(Omega Optical [Woburn, MA] XF22 filter set for FDA or BCECF [excitation:
490 nm; emission: 525 nm] and an Omega Optical XF05 filter set for Hoechst
33342 [excitation: 345 nm; emission: 475 nm]), and a 4, N.A. = 0.16 objective lens. It is also equipped with a motorized Prior Pro Scan stage with two stepper motors for stage movements in the horizontal plane (x and y) and one stepper
motor for focusing (z-axis). The stage controller communicates with the computer
through a serial port. A Qimaging Microimager II CCD camera is attached to the
standard trinocular head with an 80/20 beam splitter. Maximum resolution of the
camera CCD chip is 1024/768 pixels, and the camera has internal cooling,
enabling long-term use without degrading image quality. The camera is connected
to the PC through IEEE1394 FireWire interface, which enables a high rate of data
transfer from the camera to the computer memory.
6. DIMSCAN 3.0 software (Fig. 2). This is a user-friendly interface for all the tasks:
plate definition, scan control, camera settings, autofocus, well-image reconstruc-
143
Fig. 1. DIMSCAN system: (A) A computer equipped with IEEE1394 FireWire PCI
card running DIMSCAN application; (B) Olympus IX50, inverted microscope,
equipped with mercury lamp, optical filters, and high N.A. 4 objective lens; (C)
Qimaging Microimager II CCD camera; (D) motorized stage Prior Pro Scan; (E) stage
controller, which communicates with computer through serial port; (F) focus drive.
tion. It automatically controls stage motion, quantifies fluorescence for individual
wells, performs image processing to reduce background fluorescence (digital
thresholding) (15), and stores images of scanned wells for repeated scan or elimination from final analysis of selected wells ([15,22]; unpublished).
7. Data analysis software. Following incubation with drugs and scanning of microwell
plates by DIMSCAN, the DataAnalyzer software (developed at Childrens Hospital
Los Angeles) calculates fractions of affected cells (Fa) (Fa = 1 RFdrug / RF control),
survival fractions (calculated as RFdrug / RF control), standard deviations (SDs), confidence levels, and standard errors (SEs) using the relative fluorescence (RF) values
obtained during the DIMSCAN. Fractions of affected cells, survival fractions, and
SDs can be calculated using Microsoft Excel. This program can also be used for
creating dose-response graphs. Analyzed data can be copied to SigmaPlot (Jandell,
San Rafael, CA). to create publication-quality dose-response graphs.
8. Dose-effect analysis with microcomputers software. This program utilizes doseeffect data to compute the median-effect dose (i.e., the lethality [LC50], effect
[ED50], inhibition [IC50]), whether the dose-effect relationships conform to the
mass-action principle (i.e., r value); the dose that is required to produce a given
effect (lethal drug concentration for 50% [LC50], 90% [LC90], or 99% [LC99], and
so on of treated cells), and the effect that can be produced by a given dose. The
144
Keshelava et al.
Fig. 2. Layout of DIMSCAN software window at the end of scan. The main window
displays the menu bar and three panels: plate control panel (shown on the left side of the
window), live camera panel (not shown), and thumbnail panel (shown on the right of
the window), displaying an image of a representative well of a 96-well microplate with
viable cells. Cells are stained with FDA, and background fluorescence is quenched
with digital thresholding and eosin Y. Dead cells are not detectable. The plate control
panel displays a picture of the whole plate, miniatures of images of already scanned
wells. It enables the user to select the wells for scan, and it also contains a virtual joystick, which allows the user to control the stage movement, a necessary practice prior to
scanning a plate to check uniformity of thresholding throughout the well. Also shown are
output datathe sum of pixel intensities per well. The live camera panel allows the user
to control the camera output (gain, exposure time, threshold). These parameters must be
set prior to the scan and should not be changed during the scan. The thumbnail panel displays reconstructed images of individual wells for the users revision at the end of the
scan. Unsatisfactory images can be rescanned or deleted from further analysis.
program also computes a combination index (CI), which determines synergism,
summation, or antagonism at different dose levels (29).
3. Method
1. Harvest cells that are 7580% confluent and seed the cells in complete medium
into 96-well plates. Cell number is determined by tumor type; for example, solid
2.
3.
4.
5.
145
tumor cell lines should be seeded at lower cell concentrations (100015,000) than
leukemia (up to 50,000) owing to the appreciable difference in cell size, or the
doubling time of a given cell line (slower-growing cells are plated at higher numbers than fast-growing lines), and the length of an experiment (experimental
design for a cytotoxicity assay may vary from 3 to 7 d). If one desires to assess
the cytostatic effect of a drug, then longer-term (1421 d) assays should be
employed and fewer cells should be seeded. If inappropriately high cell concentrations are used, the control wells will overgrow and cause erratic results. However, using the highest number of cells that will not cause overgrowth of controls
is advisable to provide the largest dynamic range.
To calculate the 96-well plate format for the single-drug cytotoxicity assay
(see Note 1), first determine the total amount of cells per plate (e.g., n cells/
well 80 wells/plate). Then determine the amount of complete medium
with which to seed the cells (e.g., 150 L/well 80 wells/plate = 12 mL/
plate). Often, only 60 wells are utilized, and the wells on the edges are filled with
water because evaporation of medium can result in a change in drug concentrations in the outer wells. To avoid a shortage of cell/medium or drug/medium suspensions, make calculations for 10% more wells than the expected use. Use a
multichannel pipettor or Precision 2000 robot pipettor for seeding cells or drugging the plates.
After overnight incubation, for a single-drug experiment, add four different drug
concentrations in complete medium to each well (see Note 2). At least three drug
concentrations must be tested to be able to calculate lethal drug concentrations.
Each drug concentration is added to two consecutive columns of wells comprising 12 replicates. Drug concentrations are chosen to cover the established clinically achievable levels. This range includes the lowest drug levels up to a
clinically unachievable high drug concentration to demonstrate possible drug
activity. The maximum volume used is 250 L/well (150 L/well was deposited
when seeding the cells; an additional 100 L/well is used for delivering drug[s]),
or 4 mL/condition (250 L/well 16 wells/plate). Therefore, dissolve drug in
100 L/well 16 wells/plate = 1.6 mL/condition/plate. Use vehicle solution in
controls. Pipet drugs into the columns in increasing drug concentrations. Discard
leftover drug/medium suspension from the reservoir in between drug conditions to
avoid dilution of desired drug concentrations.
For drug combination assays, set up separate plates for Drug A, Drug B, and Drug
A + Drug B (see Note 3). In the latter case, the first two columns will carry cells
treated with vehicle solution; the next two columns will carry Drug Acondition1 +
Drug Bcondition1, then Drug Acondition2 + Drug Bcondition2, Drug Acondition4 + Drug
Bcondition4. It is preferable to set Drug A and Drug B in such a fashion that the ratio
between them is maintained throughout all the concentrations (e.g. drug concentrations of Drug A = 0, 3, 6, 9, and 12 M, and drug concentrations of Drug B =
0, 1, 2, 3, and 4 M gives a ratio of 31).
For drug resistance modulation experiments, determine a dose-response curve for
one drug against one fixed dose of a modulator that can be obtained clinically
146
6.
7.
8.
9.
10.
11.
12.
13.
Keshelava et al.
(e.g., 500 M BSO + increasing melphalan concentrations, or 2 g/mL of the
cyclosporin analog SDZ PSC 833 + increasing etoposide concentrations.).
Incubate plates for 37 d, depending on the type of cells and the total time of
drug exposures desired. For drugs with a very short t1/2, in vitro incubation times
days longer than the t1/2, may be needed to allow for maximal cell death to occur.
Prior to DIMSCAN analysis (2030 min), turn on the mercury bulb of the DIMSCAN system and allow it to warm up and stabilize, turn on the stage controller
with CCD camera, and then run the DIMSCAN application software.
After incubation of cells with tested drugs, add to each well 0.5% eosin Y and
FDA (the final FDA concentration should be 10 g/mL) using a multichannel
pipettor (see Notes 4 and 5). Incubate the plate for an additional 20 min (see
Note 6) in the dark at room temperature and analyze total fluorescence measured
by the DIMSCAN assay. One can stain three to four plates at once with FDA and
eosin Y. Eosin Y considerably reduces background fluorescence (especially in
dead cells). However, fluorescence still remains in the medium and is easily controlled by digital thresholding. For cytostatic experiments one may find it preferable to use Hoechst 33342, a supravital DNA stain, instead of FDA (see Note 7).
Before scanning or reading the plate, define the type of plate (e.g., 96-well,
384-well) from the menu bar by choosing File, Basic default, New plate, and
then 96-well plate. Scanning options are Overscan (covering 100% of well and
scanning nine frames), Optiscan I (covering 89% of well and scanning four
frames), Optiscan II (covering 98% of well and scanning six frames), and
Underscan (covering 42% of well and scanning two frames). The Autofocus feature is implemented for automation of the reading (see Note 8). Choose a well by
double-clicking on the plate map from the Plate control panel for thresholding.
Digital thresholding is provided on the Live camera panel. Using the arrows
from the Stage movement (virtual joystick), make sure that background fluorescence is eliminated. Check other representative wells for background fluorescence.
Highlight the wells to be scanned. Choose Scan from the Plate control panel. In
addition, this panel provides options such as Abort, Pause/Continue (to pause,
terminate, or continue the scan), Bad well, Rescan, and Go to well. At the end of
the scan, go to the Thumbnail panel, and by navigating through the Plate Control panel, inspect well by well the reconstructed thumbnail images. Wells with
unsatisfactory images can be rescanned, or labeled as bad well. Bad wells
are not used in further data processing.
Store fluorescence values of a cytotoxicity/cytostatic assay as a .txt file using the
DIMSCAN software.
If multiple plates are scanned, select Another plate from the File. Making this
selection will avoid the plate definition process.
At the end of the DIMSCAN session, exit the DIMSCAN software and then turn
off the mercury bulb, the stage controller, and the CCD camera.
Once a week conduct stage alignment of the DIMSCAN as follows: From the
menu bar select File, Basic default, New plate, and then 96-well plate. Doubleclick on A1 well. Place on the stage a 96-well plate modified for stage alignment
147
(it has a very fine cross in the center of the A1 well). Turn on an incandescent
light. Choose Start from the Plate control panel. One should see a fine cross in
the middle of Live camera panel. From Tools choose Reset stage center wizard.
The following message will appear:
This is stage center reset wizard.

If stage lost its position (wells are not centerd) it is
necessary to reset stage center. To do it use following steps:
1. Turn on the microscope incandescent light
2. Put the calibration plate onto the scope
3. Adjust light to get reasonable image of the plate
4. Press Next.
Cancel
Next >>
Press Next. Then follow the instructions of the next message:

1. Move the stage using joystick (or virtual joystick) to the
left upper well of the calibration plate.
2. Match the cross-hair of the camera image with the
cross-hair image on the well.
3. Press Finish to reset the stage center.
Cancel
Finish
Then the last message will appear:
Stage center resetCommand successful

OK
148
Keshelava et al.
14. Use the DataAnalyzer software to analyze fluorescence values obtained by the
DIMSCAN system. The software opens DIMSCAN files containing RF values
and the user maps of the treatment groups. The software will first calculate average fluorescence values per condition and apply these to compute affected fractions, surviving fractions, SDs, SEs, and 95% confidence intervals. Alternatively,
Microsoft Excel can be used to make these calculations.
15. Copy the surviving fractions SD to SigmaPlot 2000 or Microsoft Excel. Doseresponse curves are then created. The drug cytotoxic/cytostatic effect is plotted on
a log scale (y-axis) against increasing drug concentrations on a linear scale (x-axis).
16. Determine the LC50, LC90, or LC99 values (drug concentration lethal for 50, 90,
and 99% of treated cells, respectively) using the software Dose-Effect Analysis
with Microcomputers. This is achieved by entering the drug dose and a corresponding Fa into the software. The program computes these lethal drug concentrations along with r value (mass-action relationship). The r value should be close
to 1, and if low values are obtained, then one should consider repeating an experiment. Calculate CI values to establish synergistic, additive, or antagonistic interaction between drugs for studies of drug combinations. The CI values can be
calculated by choosing the Multiple drug-effect analysis. Enter the drug dose
and corresponding Fa value for both of the drugs separately and then in combination. The software will compute various lethal drug concentrations for both
single agents and drug combinations, and CI values. When CI = 1, summation
effect is assumed; when CI < 1, synergism is assumed; when CI > 1, antagonism
is assumed. The CI value at any effect level can be calculated (see Fig. 3 and
Table 2).
4. Notes
1. Similar calculations are used when setting up multiweek cytotostatic experiments.
However, once a week washout and exchange of medium is necessary to avoid
medium exhaust effect.
2. It is preferable to seed cells uniformly and then drug the plates, rather than prepare cell suspensions with individual drug concentrations and then plate, because
it will introduce less plating error, and it allows cells to adapt (and attach for
some cell lines) prior to drug exposure.
3. If Drug B is separately added from Drug A, then Drug A should be added with
Drug B to maintain the desired drug concentration.
4. For a 96-well microplate format, if each well contains 250 L of medium, add
50 L/well of 0.5% eosin Y (4 mL/plate) and a final FDA concentration of
10 g/mL (or 240 L of stock solution/plate).
5. One can add 3050 L of 0.5% eosin Y in between the wells to eliminate occasional artificial fluorescence that occurs from the edge of the wells.
6. Do not incubate cells longer than 3040 min with FDA to diminish toxicity from
the dye and artificially high background owing to hydrolysis of FDA to fluorescein in the medium by esterases.
149
Fig. 3. Dose-response curves of 4-hydroperoxycyclophosphamide (4-HC) () and

etoposide ( ), and combination of 4-HC + etoposide () tested by DIMSCAN. A representative synergistic drug interaction is shown using the multidrug-resistant human
neuroblastoma cell line CHLA-171 (28). The cytotoxic effect of 4-HC, an active in
vitro metabolite of the alkylating agent cyclophosphamide, in combination with etoposide, a topoisomerase inhibitor, was greater than the summation effect of these drugs
alone at all concentrations tested; CI values were <1.
Table 2
Dose-Response of 4-HC, Etoposide, and Their Combination
in the CHLA-171 Neuroblastoma Cell Line Tested by DIMSCAN
Dose
4-HC
(g/mL)
(r = 0.98)
Etoposide (g/mL)
(r = 0.99)
4-HC + etoposide
(r = 0.99)
0.25
1.25
2.25
4.25
8.25
0.25
1.25
2.56
5.25
10.25
0 + 0.25
1 + 1.25
2 + 2.55
4 + 5.25
8 + 10.2
Fraction
affected
Surviving
fraction
SD
0.0000
0.2453
0.7017
0.9176
0.9549
0.0000
0.8052
0.9627
0.9882
0.9982
0.0000
0.9616
0.9975
0.9995
0.9999
1.0000
0.7547
0.2983
0.0825
0.0451
1.0000
0.1949
0.0373
0.0118
0.0018
1.0000
0.0384
0.0025
0.0005
0.0001
0.2402
0.2733
0.1847
0.1101
0.0680
0.2902
0.2465
0.0276
0.0081
0.0032
0.3396
0.0724
0.0027
0.0009
0.00002
LC and
CI values
LC50 = 1.5 g/mL

LC90 = 4.4 g/mL
LC99 = 14.3 g/mL
LC50 = 0.7 g/mL
LC90 = 1.7 g/mL
LC99 = 4.9 g/mL
CI = 0.5
CI = 0.3
CI = 0.3
CI = 0.3
150
Keshelava et al.
7. Hoechst 33342 is a membrane-permeant minor groove DNA-binding fluorescent

dye. Light emitted by this dye measures total fluorescence of both viable and
nonviable cells (15). Quantification of fluorescence from Hoechst 33342 provides
an accurate measure of total cell number and is useful for assaying the cytostatic
effect of a tested drug.
8. At times it is necessary to adjust the focus manually. In such instances, one must
turn off the Autofocus function to avoid fracture of a stepper motor, and then turn
it on again.
Acknowledgments
We thank Eddie Gowharrizi for meticulous help in preparing the illustrations. This work was supported in part by the Neil Bogart Memorial Laboratories of the T.J. Martell Foundation for Leukemia, Cancer, and AIDS
Research, and by National Cancer Institute grants CA82830, CA81403, and
CA102990.
References
1. Pieters, R., Huismans, D. R., Leyva, A., and Veerman, A. J. (1989) Comparison of
the rapid automated MTT-assay with a dye exclusion assay for chemosensitivity
testing in childhood leukaemia. Br. J. Cancer 59, 217220.
2. Rubinstein, L. V., Shoemaker, R. H., Paull, K. D., et al. (1990) Comparison of in
vitro anticancer-drug-screening data generated with a tetrazolium assay versus a
protein assay against a diverse panel of human tumor cell lines. J. Natl. Cancer
Inst. 82, 11131118.
3. Sobottka, S. B. and Berger, M. R. (1992) Assessment of antineoplastic agents by
MTT assay: partial underestimation of antiproliferative properties. Cancer
Chemother. Pharmacol. 30, 385393.
4. Vistica, D. T., Skehan, P., Scudiero, D., Monks, A., Pittman, A., and Boyd, M. R.
(1991) Tetrazolium-based assays for cellular viability: a critical examination of selected parameters affecting formazan production. Cancer Res. 51,
25152520.
5. Gerson, S. L., Phillips, W., Kastan, M., Dumenco, L. L., and Donovan, C. (1996)
Human CD34+ hematopoietic progenitors have low, cytokine-unresponsive O6alkylguanine-DNA alkyltransferase and are sensitive to O6-benzylguanine plus
BCNU. Blood 88, 16491655.
6. Lee, C. K., Harman, G. S., Hohl, R. J., and Gingrich, R. D. (1996) Fatal
cyclophosphamide cardiomyopathy: its clinical course and treatment. Bone
Marrow Transplant. 18, 573577.
7. Dhar, S., Nygren, P., Csoka, K., Botling, J., Nilsson, K., and Larsson, R. (1996)
Anti-cancer drug characterisation using a human cell line panel representing
defined types of drug resistance. Br. J. Cancer 74, 888896.
151
8. Ford, C. H., Richardson, V. J., and Tsaltas, G. (1989) Comparison of tetrazolium

colorimetric and [3H]-uridine assays for in vitro chemosensitivity testing. Cancer
Chemother. Pharmacol. 24, 295301.
9. Skehan, P., Storeng, R., Scudiero, D., et al. (1990) New colorimetric cytotoxicity
assay for anticancer-drug screening. J. Natl. Cancer Inst. 82, 11071112.
10. Leeder, J. S., Dosch, H. M., Harper, P. A., Lam, P., and Spielberg, S. P. (1989)
Fluorescence-based viability assay for studies of reactive drug intermediates. Anal.
Biochem. 177, 364372.
11. Ross, D. D., Joneckis, C. C., Ordonez, J. V., et al. (1989) Estimation of cell survival by flow cytometric quantification of fluorescein diacetate/propidium iodide
viable cell number. Cancer Res. 49, 37763782.
12. Wierda, W. G., Mehr, D. S., and Kim, Y. B. (1989) Comparison of fluorochromelabeled and 51Cr-labeled targets for natural killer cytotoxicity assay. J. Immunol.
Methods 122, 1524.
13. Nishio, K., Ishida, T., Arioka, H., et al. (1996) Antitumor effects of butyrolactone
I, a selective cdc2 kinase inhibitor, on human lung cancer cell lines. Anticancer
Res. 16, 33873395.
14. Bramson, J., OConnor, T., and Panasci, L. (1995) Effect of alkyl-N-purine DNA
glycosylase overexpression on cellular resistance to bifunctional alkylating agents.
Biochem. Pharmacol. 50, 3944.
15. Proffitt, R. T., Tran, J. V., and Reynolds, C. P. (1996) A fluorescence digital image
microscopy system for quantifying relative cell numbers in tissue culture plates.
16. Wang, X. M., Terasaki, P. I., Rankin, G. W. Jr., Chia, D., Zhong, H. P., and Hardy,
S. (1993) A new microcellular cytotoxicity test based on calcein AM release. Hum.
Immunol. 37, 264270.
17. Jensen, P. B., Holm, B., Sorensen, M., Christensen, I. J., and Sehested, M. (1997)
In vitro cross-resistance and collateral sensitivity in seven resistant small-cell lung
cancer cell lines: preclinical identification of suitable drug partners to taxotere,
taxol, topotecan and gemcitabin. Br. J. Cancer 75, 869877.
18. Reynolds, C. P. and Frgala, T. (2002) Fluorescence Digital Imaging Microscopy
System (Chemical quenching of non-cellular fluorescence from fluorescein diacetate in cytotoxicity assays using eosin y and other similar vital dyes). USPTO
patent no. 6,459,805.
19. Keshelava, N., Seeger, R. C., Groshen, S., and Reynolds, C. P. (1998) Drug resistance patterns of human neuroblastoma cell lines derived from patients at different
phases of therapy. Cancer Res. 58, 53965405.
20. ODonnell, P. H., Guo, W. X., Reynolds, C. P., and Maurer, B. J. (2002) N-(4hydroxyphenyl)retinamide increases ceramide and is cytotoxic to acute lymphoblastic leukemia cell lines, but not to non-malignant lymphocytes. Leukemia 16,
902910.
21. Maurer, B. J., Melton, L., Billups, C., Cabot, M. C., and Reynolds, C. P. (2000) Synergistic cytotoxicity in solid tumor cell lines between N-(4-hydroxyphenyl)retinamide
and modulators of ceramide metabolism. J. Natl. Cancer Inst. 92, 18971909.
152
Keshelava et al.
22. Krejsa, J., Frgala, T., Alfaro, P., and Reynolds, C. P. (2002) DIMSCAN 3.0, a new
generation of flourescence digital image microscopy system for measuring cytotoxicity in microplates. Proc. Am. Assoc. Cancer Res. 43, 586.
23. Harrison, S. (2002) Perspective on the history of tumor models, in Tumor Models
in Cancer Research (Teicher, B. A., eds.), Humana, Totowa, pp. 319.
24. Reynolds, C. P. (2000) Differentiating agents in pediatric malignancies: retinoids
in neuroblastoma. Curr. Oncol. Rep. 2, 511518.
25. Anderson, C. P., Seeger, R. C., Bailey, H. H., and Reynolds, C. P. (2002) Pilot
of buthionine sulfoximine (BSO) combined with non-myeloablative melphalan
(L-PAM) against refractory neuroblastoma (NB). Proc. Am. Soc. Clin. Oncol. 21,
298a.
135140.
27. Reynolds, C. P., Wang, Y., Melton, L. J., Einhorn, P. A., Slamon, D. J., and
Maurer, B. J. (2000) Retinoic-acid-resistant neuroblastoma cell lines show altered
MYC regulation and high sensitivity to fenretinide. Med. Pediatr. Oncol. 35,
597602.
28. Keshelava, N., Tsao-Wei, D., and Reynolds, C. P. (2003) Pyrazoloacridine is active
in multidrug-resistant neuroblastoma cell lines with nonfunctional p53. Clin.
Cancer Res. 9, 34923502.
29. Chou, J. and Chou, T. C. (1987) Multiple drug-effect analysis (program B), in
Dose-Effect Analysis with Microcomputers (Chou, J. and Chou, T. C., eds.),
Biosoft, Cambridge, UK, pp. 1929.
30. Ellwart, J. W., Kremer, J. P., and Dormer, P. (1988) Drug testing in established
cell lines by flow cytometric vitality measurements versus clonogenic assay.
Cancer Res. 48, 57225725.
Cancer Res. 43, 749757.
32. Ishiyama, M., Tominaga, H., Shiga, M., Sasamoto, K., Ohkura, Y., and Ueno, K.
(1996) A combined assay of cell viability and in vitro cytotoxicity with a highly
water-soluble tetrazolium salt, neutral red and crystal violet. Biol. Pharm. Bull.
19, 15181520.
33. Sevin, B. U., Peng, Z. L., Perras, J. P., Ganjei, P., Penalver, M., and Averette, H. E.
(1988) Application of an ATP-bioluminescence assay in human tumor chemosensitivity testing. Gynecol. Oncol. 31, 191204.
34. Metzger, R., Deglmann, C. J., Hoerrlein, S., Zapf, S., and Hilfrich, J. (2001)
Towards in-vitro prediction of an in-vivo cytostatic response of human tumor cells
with a fast chemosensitivity assay. Toxicology 166, 97108.
35. Bosanquet, A. G. and Bell, P. B. (1996) Enhanced ex vivo drug sensitivity testing
of chronic lymphocytic leukaemia using refined DiSC assay methodology. Leuk.
Res. 20, 143153.
153
36. Tanigawa, N., Kitaoka, A., Yamakawa, M., Tanisaka, K., and Kobayashi, H. (1996)
In vitro chemosensitivity testing of human tumours by collagen gel droplet culture
and image analysis. Anticancer Res. 16, 19251930.
37. Csoka, K., Tholander, B., Gerdin, E., de la Torre, M., Larsson, R., and Nygren, P.
(1997) In vitro determination of cytotoxic drug response in ovarian carcinoma
using the fluorometric microculture cytotoxicity assay (FMCA). Int. J. Cancer 72,
10081012.
13
The ChemoFx Assay
An Ex Vivo Cell Culture Assay for Predicting
Anticancer Drug Responses
Robert L. Ochs, Dennis Burholt, and Paul Kornblith
Summary
We provide a detailed description of the ChemoFx Assay, a phenotype-based cell culture
assay for predicting anticancer drug responses in individual cancer patients. The ChemoFx Assay
is based on the outgrowth and short-term primary culture of epithelial cells derived from pieces
of solid tumor adenocarcinomas that are obtained at the time of tumor resection. Malignant
epithelial cells are grown attached in wells of microtiter plates and treated with six escalating
doses of chemotherapeutic drug. Using an operator-controlled automated image analysis system,
cell kill is measured microscopically by counting the number of live cells remaining after dead
cells have detached and are subsequently rinsed away. A dose-response graph is automatically
generated by comparing the number of cells in drug-treated wells with those in control wells.
Key Words
Cancer; primary cell culture; chemoresponse testing; phenotypic ex vivo assay.
1. Introduction
The ChemoFx Assay is a tissue culturebased assay that measures the individual response of a patients tumor-derived cells to anticancer chemotherapeutic
agents in vitro (14). The assay can be utilized to test the effects of multiple anticancer agents against cells that are grown ex vivo from an individual patients
tumor that is obtained at the time of surgical resection, thus predicting a patients
response to subsequent chemotherapy. The rationale of the ChemoFx Assay is
based on the adherent properties of malignant epithelial cells derived from carcinomas. In culture, these epithelial cells attach and spread to form cell monolayers. When these same epithelial cells are killed in response to chemotherapeutic
155
156
Ochs et al.
drugs in vitro, they undergo cell death by apoptosis and subsequently detach from
the substratum (58). Therefore, counting of live cells remaining attached after
drug treatment in vitro is a quantitative measure of cell kill.
In the ChemoFx Assay, adherent cells are cultured from ex vivo solid tumor
tissue explants or from cells recovered from fluid specimens such as ascites or
pleural fluids. Once a sufficient number of cells are grown in culture, the cells
are trypsinized and counted, so that a specific number of cells are delivered to
each well of a microtiter plate. After overnight incubation to allow adherence
to the bottom of the wells, cells are then treated with multiple concentrations of
a particular anticancer agent, using exposure times and concentrations that simulate in vivo drug treatment. The drugs are then removed, and after a 72-h
incubation period, the plates are rinsed, fixed, and stained, thereby removing
the dead cells and staining the remaining live cells with a fluorescent nuclear
stain. The number of surviving cells is then counted using an automated microscopic image analysis system, and a cytotoxic index (CI) (equivalent to percentage of cell kill) is calculated for each drug concentration tested. From these
data, a dose-response graph is derived that provides a physician with information on the patients response to a given anticancer agent.
Chemoresponse assays vary not only in the kinds of information provided, but
also in the way in which the effects of drugs are measured (reviewed in refs.
911). Different assays have evolved based on measures of cell proliferation
(11), cell metabolism (1214), cell viability (15), and cell death (16). Compared
with these other ex vivo chemoresponse assays, the ChemoFx Assay is unique
in that tumor cells are isolated and maintained in short-term culture before drug
testing and their epithelial identity is verified by immunohistochemical staining
for the presence of cytokeratin (4,17). Another unique aspect is that both drug
sensitivity and resistance are measured by using a range of six drug doses, two
that are well below the literature values for peak plasma concentration, two that
are within the peak plasma concentration interval, and two that are well above
the peak plasma concentration values. Furthermore, by measuring only the
number of live cells remaining after drug treatment, the ChemoFx Assay measures both drug-induced cell killing and inhibition of proliferation.
2. Materials
2.1. Cell Culture Media and Supplies
1.
2.
3.
4.
5.
Media: RPMI-1640, Hams F-10, McCoys 5A, MEGM (Clonetics).

Fetal bovine serum (FBS).
Hanks balanced salt solution (HBSS), with and without calcium and magnesium.
Trypsin and trypsin neutralizer.
Antibiotic/antimycotic solution: penicillin, streptomycin, ciprofloxicin, gentamycin,
amphotericin B, nystatin.
The ChemoFx Assay

6.
7.
8.
9.
157
Type I collagenase (Worthington).

Vitrogen 100 (Cohesion Technologies).
Cell culture flasks and 60-well microtiter plates (Greiner).
Sterile scalpel and scissors.
2.2. Chemicals and Reagents

1.
2.
3.
4.
5.
Tris-buffered saline (TBS).

Methanol, ethanol.
Anticytokeratin (AE1/AE3, CAM5.2, CK7) mouse monoclonal antibodies (Dako).
Secondary antimouse antibody conjugated to Alexa 488 (Molecular Probes).
DAPI dilactate (Molecular Probes).
2.3. Safety Supplies

1.
2.
3.
4.
Biohazard and chemotherapy waste containers.

Sharps waste container.
Chemotherapy gowns and gloves.
Absorbent pads.
2.4. Equipment
1.
2.
3.
4.
5.
6.
Class IIB laminar flow hoods.

CO2 incubators.
Vacuum pump.
Multichannel pipettor.
Fluorescence microscope.
Image analysis microscope and software (Zeiss).
3. Methods
The major steps of the ChemoFx Assay are illustrated in the flow diagram in
Fig. 1.
3.1. Acquisition and Transport of Specimen
This section provides an overview of instructions for the proper processing
of freshly excised tumor tissue and fluid specimens (see Notes 13).
Solid tumor specimens submitted for testing should be representative of
tissue removed by the physician for therapeutic or diagnostic purposes. Ideally, they should be adjacent to specimens submitted for histology. A portion of
the resected tumor should be aseptically transferred to a specimen transport
bottle. This tumor specimen should be free of all adherent normal tissue and of
any necrotic or fibrotic regions. Transfer to medium should be done directly in
the operating suite within 25 min of tumor removal. If this is impractical
because of logistics or local procedures, the tumor specimen should be placed
in a sterile Petri dish or sterile vial with 510 mL of sterile 0.9% normal saline.
158
Ochs et al.
Fig. 1. Flow diagram of individual steps comprising ChemoFx Assay.
The specimen can then be removed from the operating room and subsequently
transferred to the specimen bottle using aseptic procedures in a biomaterials
processing hood. For larger specimens, it is often appropriate to mince the
tumor with a sterile pair of scissors before transferring it to the specimen bottle.
This will minimize hypoxia and autolytic processes that will normally appear
in the center of larger specimens. Mincing need not be excessive (2- to 3-mm
pieces are adequate). Target mass for a submitted solid tumor specimen is from
100 mg to 1 g (a pencil eraser size to 5 mm3). For fluid specimens, 100500 cc
of ascites or pleural fluid should be collected, if available. The fluid should be
transferred promptly into an evacuated container under sterile conditions. Specimens should be shipped promptly by overnight carrier.
3.2. Explanting of Tissue and Initiation of Cultures
3.2.1. Processing of Solid Tumor Specimens
1. Equally divide the shipping medium from the tumor specimen bottle into two
50-mL sterile conical tubes. Pellet the shipping medium by centrifugating at
500800g for 5 min.
2. Pour or use a pipet to discard the shipping medium into a waste container, being
careful not to disturb the cell pellet.
3. Using a fresh pipet, gently resuspend the cell pellet in the appropriate volume
of growth medium for the size of tissue culture flask to be used. Pipet the cell
suspension into the culture flask.
4. Remove the tumor specimen from the shipping container and place it in a sterile
tissue culture dish (see Notes 4 and 5).
5. Cut the tumor sample into the smallest fragments possible using sterile scalpels or
scissors.
The ChemoFx Assay
159
6. Based on the size and type of the tumor specimen (see Table 1), determine the
type of media, vitrogen 100 requirement, number, and size of tissue culture flasks
for the explantation procedure.
7. Using a sterile pipet containing the desired amount of growth media, draw up the
cell suspension and tissue fragments and place in the appropriate-size flask(s).
8. Separate the fragments over the bottom surface of the flask by gently swirling
the suspension, or use a sterile pipet to disperse the fragments along the growth
surface of the flask.
9. Tilt the flask at a slight angle, allowing the media to pool at the bottom edge of
the flask, leaving tissue explants attached to the growth surface of the flask. After
1520 min, carefully return the flask to a flat position without dislodging the
attached tissue explants. Tumor explantation is carried out in a minimal volume of
growth media to enhance tissue attachment and tumor cell outgrowth.
3.2.2. Processing of Fluid Tumor Specimens

1. Equally divide the fluid specimen into two 200-mL sterile conical centrifuge
tubes. Pellet the cells by centrifugating at 500800g for 10 min.
2. Pour or use a pipet to discard the supernatant into the waste container, being careful not to disturb the cell pellet.
3. Using a fresh pipet, gently resuspend the cell pellet with the appropriate volume
of growth media and transfer the cell suspension to the corresponding flask.
3.3. Monitoring of Cultures: Preassay Growth Phase

Successfully explanted primary tumor cell cultures will grow as a monolayer attached to the bottom of the culture flask (see Fig. 2). Cultures must be
monitored microscopically on a regular basis, noting cell growth (confluency),
cell type, cell number, and possible microbial contamination.
3.3.1. Macroscopic Examination
1. Examine the color of the medium. The medium should be neutral pH (reddish
orange). Yellow medium indicates acidic conditions (low pH), and pink medium
indicates alkaline conditions (high pH). The medium may need to be changed in
either of these conditions.
2. Examine the turbidity of the medium. The medium should be clear. Cloudy
medium could be an indication of microbial contamination.
3. Examine the size and density of the explants.
3.3.2. Microscopic Examination

1. Bacterial contamination: If the medium is slightly pink and/or turbid, check for
bacteria or yeast under 20 (or higher) magnification.
2. Cellular debris: Examine the bottom of the flask under 10 magnification. Cellular debris appears dark and granular under phase contrast and is usually attached
to the bottom of the flask. It is generally larger than bacteria and is not motile.
160
Ochs et al.
Table 1
Growth Conditions According to Tumor Type
Tumor type
Appendix
Astrocytoma
Bladder
Bone marrow
Breast
Cervical
Colon
Craniopharyngioma
Endometrial
Ependyoma
Esophageal
Extraovarian mllerian
Gallbladder
Histiocytoma
Glioblastoma
Glioma
Granulosa cell
Hepatoma/liver
Laryngeal
Leiomyosarcoma
Lung/bronchial
Melanoma
Meningioma
Nasopharyngeal
Oligodendroglioma
Ovarian
Pancreatic
Pituitary
PNET
Prostate
Pseudomyxoma
Renal
Sarcoma
Sertoli-Leydig
Stomach
Culture
medium
RPMI-1640
Hams F-10
RPMI-1640
RPMI-1640
MEGM
RPMI-1640
RPMI-1640
Hams F-10
RPMI-1640
Hams F-10
McCoys 5A
McCoys 5A
RPMI-1640
RPMI-1640
Hams F-10
Hams F-10
McCoys 5A
RPMI-1640
McCoys 5A
RPMI-1640
RPMI-1640
RPMI-1640
Hams F-10
McCoys
Hams F-10
McCoys 5A
RPMI-1640
Hams F-10
Hams F-10
PREGM
RPMI-1640
RPMI-1640
RPMI-1640
McCoys 5A
RPMI-1640
FBS
Vitrogen 100 Antibiotic/antimycotic
concentration (%)a requirement
treatmentb
10
20
10
10
0
2
2
20
10
20
10
10
10
10
20
20
10
10
10
10
2
10
20
10
20
10
10
20
20
0
10
10
10
10
10
Yes
No
Yes
No
Yes
Yes
Yes
No
No
No
Yes
No
Yes
No
No
No
No
Yes
Yes
No
Yes
No
No
Yes
No
No
No
No
No
Yes
No
Yes
No
No
Yes
No
No
No
No
No
Yes
Yes
No
No
No
Yes
No
No
No
No
No
No
No
Yes
No
Yes
No
No
Yes
No
No
No
No
No
No
No
No
No
No
Yes
a
The concentration of FBS used for the establishment of primary cultures is determined by the
specimen site (potential for fibroblast contamination) and not the tumor type.
b
The need for antibiotic/antimycotic treatment is dependent on the specimen site (potential for
contamination) and not the tumor type.
The ChemoFx Assay
161
Fig. 2. Explantation of tumor tissue: (A) small explant (Ex) piece with tumor cells
that have migrated out to form a monolayer on bottom of culture flask (B). Magnifiation: (A) 200; (B) 1000.
3. Crowding: Examine the flask under 10 magnification. Look for excess explant
tissue, including cell clumps in suspension.
4. Confluency: Under 10 magnification, determine the percentage of confluency
of the flask by scanning the whole flask and estimating cell growth.
5. Cell type: Under 10 magnification, observe the morphology of the cells. Make a
note if the cells appear fibroblastic or if the culture is a mixture of epithelial and
fibroblast cells. Fibroblast cells appear elongated, spindle-like and tend to form
swirls when densely packed.
3.3.3. Flask Actions
Once all of the flasks of an individual tumor culture have been examined, a
decision must be made as to what action each flask requires:
1. Medium should be changed at least once a week but can be done more frequently
if necessary.
2. Excess tumor tissue may be transfered to a new culture flask.
3. Regarding collagenase, if there is no adherent cell growth after 5 to 6 d, specimens may be enzymatically dissociated (see Note 6).
4. Differential trypsinization may be attempted if there is a mixed population of
epithelial and fibroblastic cells (see Subheading 3.5.2.).
5. Unattached cells can be transferred to Vitrogen 100-coated flasks (see Note 7).
6. If enough cells of the right type are present, plating into microtiter plates is
initiated.
3.4. Immunohistochemistry
The malignant cell type expressed in solid tumor adenocarcinomas is the
epithelial cell. The other major cell type present in solid tumors is the nonmalignant stromal fibroblast. Therefore, when cells are cultured from solid tumors,
162
Ochs et al.
Fig. 3. IHC for detection of cytokeratin. (A) Phase and (B) indirect immunofluorescence for the detection of epithelial-specific cytokeratins is shown. By phase
microscopy, two distinctly different cell types were present. One cell type was more
rounded, and it stained positive for cytokeratin, indicating its epithelial (E) origin.
Another cell type was more elongated and was cytokeratin negative, suggesting that
these were fibroblasts (F). Magnification: 300.
the two major types of cells that arise are either epithelial cells or fibroblasts,
or a combination of both. Morphologically, epithelial cells can be of almost
any size and shape, whereas fibroblasts are generally elongated, spindle-shaped,
and often grow in swirls. Cell morphology cannot definitively distinguish
cell type, so immunohistochemistry (IHC) for the presence of cytokeratins has
been used as a marker to identify epithelial cells, because fibroblasts do not
express the cytokeratins (see ref. 4 and Fig. 3).
3.4.1. IHC Staining Procedure (see Notes 810)
At the time of cell trypsinization for making drug plates (see Subheading
3.5. for details), an extra IHC plate is made by adding cells to wells of a
microtiter plate.
1. After 1 d in culture, rinse plates once in HBSS rapidly by flooding them.
2. Pour off the HBSS and then fix for 10 min with 20C 100% methanol by flooding the plates.
3. Individually pour off the methanol from each plate and immediately add TBS by
flooding the plate. After all of the plates have TBS on them, pour off the TBS and
then invert the plates and bang gently twice on an absorbent paper towel to
remove excess TBS from the wells.
4. Add 10 L of diluted primary antibody (monoclonal anticytokeratin) to each well
by forming a drop at the end of a pipet tip and then touching it to the side of the
well by holding the pipet at an angle of approx 45. Incubate for 60 min at room
temperature with diluted primary antibody.
The ChemoFx Assay
163
Table 2
IHC Decision-Making Algorithm
Cytokeratin
staining
Negative
Positive
Negative
Morphology
Fibroblast-like
Epithelial
Variable
Interpretationa
F
E
Not F, not E
F, fibroblast cells; E, epithelial cells.
5. After antibody incubation, rinse each plate with TBS by picking up the plate and
making two rapid rinses followed by a third 5-min soak. To prevent washing
away cells and disturbing the Vitrogen 100 coating (if present), do not add the
rinse solution directly over the wells containing antibodies. After all of the rinses
have been completed, pour off the TBS, and then invert the plates and bang gently
twice on an absorbent paper towel to remove excess TBS from the wells.
6. Incubate for 3060 min at room temperature with diluted secondary antibody conjugated to the fluorochrome Alexa 488. Add 10 L of diluted antibody to each
well by forming a drop at the end of the pipet tip and then touching it to the side
of the well by holding the pipet at an angle of approx 45.
7. Repeat step 5.
8. Stain for 5 min with DAPI diluted 1/1000 in TBS. Add 10 L to each well by
forming a drop at the end of the pipet tip and then touching it to the side of the
well by holding the pipet at an angle of approx 45.
9. Rinse rapidly two times with TBS, flood the plates with TBS, and then store at
4C until they are observed in the microscope. Bright cytoplasmic filament staining is indicative of the presence of cytokeratin in epithelial cells.
3.4.2. IHC Interpretation

1. By using phase and fluorescence, alone and in combination, estimate the percentage of cells that are positive for each antibody (see Fig. 3).
2. Observe the DAPI staining pattern to determine whether there is suspected
mycoplasma contamination. Normal DAPI staining is confined to the nucleus.
Cytoplasmic staining can be indicative of mycoplasma (see Note 11).
3.4.3. IHC Acceptance/Rejection Criteria (see Table 2)

1. If a culture contains 90% epithelial cells, then it is considered an E culture that
is suitable for testing.
2. If a culture contains 90% fibroblasts, then it is an F culture that cannot be
tested.
3. If a culture contains 1090% epithelial cells (or fibroblasts), then it is an E/F
culture that may need epithelial cell enrichment by differential trypsinization.
164
Ochs et al.
3.5. Plating Cells into Wells of Microtiter Plates

3.5.1. Routine Trypsinization
The following steps should be performed under a laminar flow hood using
proper aseptic technique unless otherwise stated:
1. Observe the explant/cell flask for growth (confluency) and the presence of debris/
contamination.
2. Pour off the growth medium and any remaining explants into a waste beaker.
Rinse the flask by pouring a sufficient amount of HBSS without calcium and
magnesium into the flask and swirling to rinse the cell monolayer and remove
any explant debris. Pour off the HBSS into a waste beaker.
3. Pipet an appropriate amount of 0.05, 0.25, or 0.025% trypsin into the flask using
the following guidelines:
a. Young cultures: 0.05% trypsin.
b. Densely confluent cultures, or older cultures: 0.25% trypsin.
c. Breast cell cultures: 0.025% trypsin.
4. Swirl to completely cover the cell monolayer.
5. Immediately observe the cells under an inverted microscope to determine whether
the cells are detaching. If cells are detaching, proceed to step 7 immediately. If
cells are not detaching, place the flask in a 37C incubator for 30 s to 1 min.
6. Remove the flask from the incubator and examine under the inverted microscope.
Gently strike the flask two to three times to dislodge the cells. If the cells are
detached, proceed to step 8. If the cells are not detached, incubate further at 37C,
observing the flask at 30-s to 1-min intervals, until the cells have detached. Then
proceed to step 7.
7. Add trypsin neutralizer solution to the flask. An excess amount that is at least
twice the amount of trypsin used will be required. Using a pipet, rinse the flask
several times, and transfer the cell suspension to a 15-mL conical centrifuge tube.
8. Centrifuge the cell suspension at 500800g for 3 min. Carefully decant or pipet
off the supernatant without disturbing the cell pellet.
9. Resuspend the cell pellet in 1 mL of appropriate growth medium.
3.5.2. Differential Trypsinization
If a culture is determined to have a fibroblast cell population, in addition to

the epithelial tumor cell population, it may be useful to perform a differential
trypsinization. Because fibroblasts have different adherent properties from
epithelial tumor cells, the fibroblasts can be differentially dissociated from the
tumor cell population.
1. Using 0.05% trypsin, pipet an appropriate amount of trypsin into the cell culture
flask.
2. Observe the flask immediately for cell detachment. Fibroblasts are not as adherent as tumor cells and will detach readily, leaving the tumor cells still adherent.
Gently strike the flask two to three times to dislodge the fibroblasts.
The ChemoFx Assay
165
3. Immediately add trypsin neutralizer solution to the flask. An excess amount that
is at least twice the amount of trypsin will be sufficient. Using a pipet, rinse the
flask several times, and transfer the fibroblast suspension to a 15-mL conical tube.
Discard the tube.
4. Rinse the flask with HBSS without calcium and magnesium, and follow steps
39 in Subheading 3.5.1.
3.5.3. Addition of Cells to Wells of Microtiter Plates

1. Dilute the cell suspension so that it contains 350500 cells in a 10-L volume.
2. Mix the cell suspension gently and thoroughly with a pipet in order to generate a
single cell suspension.
3. Pour the cell suspension into a sterile reagent reservoir.
4. Place the 60-well microtiter plates on the surface of the laminar flow hood and
remove the lids. There should be one plate for each drug. There should also be
one plate for IHC testing.
5. Using a multiwell pipettor, dispense 10 L of cell suspension into each well.
6. Transfer the plates into a 37C humidified incubator with 5% CO2. Incubated the
plated cells overnight to allow the cells to adhere to the bottom of the plate wells
before treatment with anticancer agent.
3.6. Drug Preparation, Application, and Removal of Drugs

3.6.1. Preparation of Anticancer Agents
Anticancer drugs come in various forms and with different types of packaging, necessitating different handling requirements. All drug procedures are performed in a Type IIB laminar flow hood that is vented to the outside, and
utilizing the proper safety precautions and sterile technique.
3.6.1.1. RECONSTITUTION
OF
LYOPHILIZED DRUG
1. Remove the plastic seal from the top of the vial and clean with an alcohol pad.
2. Insert a sterile 22-gauge needle into the rubber stopper to provide a vent for the
vacuum in the vial.
3. Clean the top of the diluent container with an alcohol pad. Appropriate diluent
may be sterile water, sterile saline (0.9%), or an alcohol diluent provided with
the drug vial by the manufacturer.
4. Using a sterile syringe, withdraw the appropriate amount of diluent and remove
any air bubbles from the syringe.
5. Dispense the diluent into a drug vial and vortex to mix completely.
6. Dispose of the needle into a sharps container.
3.6.1.2. RECONSTITUTION
OF
DRUG
IN
CAPSULE FORM
1. Carefully open the capsule with a slow twisting motion while holding the capsule
over a conical tube containing appropriate diluent. Avoid cracking the capsule. Dispense the contents of the capsule into the diluent and vortex to mix completely. Appropriate diluent may be propylene glycol, ethyl alcohol, dimethyl sulfoxide, or HBSS.
166
Ochs et al.
2. Discard the capsule into a chemotherapeutic drug waste container. Immediately

wipe up any powder residue from the hood, gloves, or other items with an alcohol pad. Change gloves if necessary.
3.6.1.3. WEIGHING DRUG
IN
POWDERED FORM
Some drugs are supplied in powder form and must be weighed to achieve the
appropriate concentration. Perform all weighing procedures in a vented hood
enclosure that contains an analytical balance.
1. Before proceeding, wipe the drug vial and sterile cryovial with an antistatic brush.
Removing static is crucial to an accurate measurement.
2. Weigh the empty cryovial and tare the analytical balance.
3. Using a clean spatula, withdraw an amount of drug from the vial and place it in
the cryovial.
4. Place the lid on the cryovial and weigh the filled container.
5. Adjust the amount of drug accordingly.
6. Add the weighed drug to an appropriate diluent and vortex to mix completely.
3.6.1.4. PREPARATION
OF
LIQUID DRUG
1. Remove the plastic seal from the top of the liquid drug vial, and then clean the
rubber stopper with an alcohol pad.
2. Using a sterile syringe, pull back on the plunger to fill the syringe with air. Insert
the needle into the top of the vial and dispense air into the vial to prevent backpressure. Withdraw an appropriate amount of drug concentrate from the vial and
dispense into a specified volume of HBSS.
3. Discard the needle into a sharps container.
3.6.2. Application of Anticancer Agents to Microtiter Plates

1. Remove plates to be treated from the incubator and place on an absorbent pad in
the laminar flow hood.
2. Using a multichannel pipettor, add 10 L of HBSS containing calcium and magnesium to each well of the control rows 5 and 8 (see Fig. 4). Then apply 10 L
of the appropriate drug dilution to each well in the appropriate row of the plate
(see Fig. 4).
3. Return the plates to the incubator and incubate for the appropriate time (see
Note 12).
3.6.3. Removal of Anticancer Agents from Microtiter Plates

1. Remove the drug-treated plates from the cell culture incubator and place them on
an absorbent pad.
2. Flood all of the plates with approx 15 mL of sterile HBSS with calcium and
magnesium.
3. Agitate the plates for 30 s by grabbing the pad at both ends and gently sliding
the pad back and forth with a left-to-right motion.
The ChemoFx Assay
167
Fig. 4. Schematic diagram of typical drug plate showing configuration that accomodates six replicates of six different drug doses and 12 wells of control cells. The
ChemoFx Assay tests clinically relevant drug doses. Doses 1 and 2 are subclinical,
doses 3 and 4 are in the range of the peak plasma levels found in vivo, and doses 5 and
6 are suprapharmacological.
4. Using a vacuum pump, suction off the HBSS with a sterile disposable pipet
attached to aspiration tubing. While suctioning, place the pipet in the corner of the
plate, avoiding the wells.
5. Repeat the rinse, agitation, and aspiration technique three additional times, for a
total of four rinses.
6. Pipet 2.53.0 mL of appropriate growth medium into each plate and agitate to
distribute the medium into each well. Make sure that there are no air bubbles in
any of the wells.
7. Rinse the plates four times and then place them back in the incubator. Incubate the
plates at 37C and 5% CO2 for 72 h, and then fix and stain with DAPI (see Subheading 3.7.) on the last day.
3.7. DAPI Staining

DAPI is a nucleic acid stain that stains nuclei specifically, with little or no
cytoplasmic labeling (see Fig. 5). Thus, counting of DAPI-stained nuclei can
serve as a surrogate for cell counting.
1. Make a DAPI working solution by dissolving the DAPI dilactate powder in distilled water to a final concentration of 0.013 g/mL or 400 nM.
2. Place the plates to be stained on an absorbent pad.
3. Pour the medium from all plates into a waste beaker.
4. Rinse each plate one time with HBSS to remove the remaining medium.
5. Fix the plates by flooding them with 95% ETOH and allow them to stand for at
least 5 min.
168
Ochs et al.
Fig. 5. Example of determination of CI by DAPI staining and counting of fluorescent nuclei. In this example, the control well contained 628 cells and the adjacent drugtreated well contained 72 cells. The CI for this drug dose was calculated as 89%.
Magnification: 100.
6. Pour off the ETOH into the waste beaker. Add bleach to the waste container and
pour down the sink and flush with water.
7. Flood each plate with the DAPI working solution, and allow to stand for a minimum of 10 min. Staining times up to 1520 min are acceptable.
8. Pour the DAPI stain from each plate into a container designated as DAPI Waste.
Do not dispose of DAPI solution down the sink drain.
9. Dip each plate into a beaker of running tap water, and flush the plate to remove
excess DAPI.
10. Allow the plates to drain off excess liquid and scan (count) them before they have
completely dried (see Subheading 3.8. for counting details).
11. Dispose of all articles contaminated with DAPI as chemotherapy waste.
3.8. Automated Cell Counting

3.8.1. Scanning of Drug-Treated Plates
Scan the DAPI-stained plates on an inverted fluorescence microscope, which

counts the number of live cells that remain in each well after drug treatment. The
cell counts from each well within a drug concentration row are averaged. The
standard deviation and coefficient of variation are calculated to give a variability flag that is used as a criterion for plate rejection (see Fig. 6 and Note 13).
The ChemoFx Assay
169
3.8.2. Calculation of CI (see Fig. 5)
The average cell counts from each drug concentration row are compared
with the average cell counts from the two control rows. The CI is calculated for
each drug concentration as a percentage of cells killed using the following
formula:
1 average number of cells counted in treated wells
CI = 100
average number of cells counted in control wells
From this information, the data are displayed as a dose-response graph,

which illustrates the percentage of cells killed at each drug concentration tested
(Fig. 6B).
4. Notes
1. Avoid submitting any specimens from patients with systemic infections or with
infectious foci contained within the tumor tissue.
2. Do not add fixative or preservative to any specimen.
3. Refrigerate specimens after surgery and prior to pickup. Do not freeze specimens.
4. If there is a potential for contamination, the specimen should be rinsed in antibiotic/antimycotic solution prior to initiating the explant procedure (see Note 5).
5. Specimens derived from some tumor sites (see Table 1) require treatment with a
special antibiotic/antimycotic rinse containing penicillin/streptomycin (300 U/mL),
ciprofloxacin (40 mg/mL), gentamycin (100 mg/mL), amphotericin B (Fungizone) (0.25 mg/mL), and nystatin (100 U/mL) to prevent possible contamination
of subsequent cultures. The method is as follows:
a. Place the tissue explant in a 50-mL conical centrifuge tube.
b. Add enough antibiotic/antimycotic wash solution to the tube to completely submerge the specimen. Rinse the solid tumor tissue for 5 min with gentle agitation.
c. Using a pipet, remove the wash solution and replace with fresh wash solution
as described in step b. Repeat this process for a total of three washes.
d. Rinse the tumor specimen for 5 min in HBSS (with calcium and magnesium)
without antibiotics.
e. Process the tumor specimen for explantation in the appropriate antibiotic/
antimycotic-containing growth medium.
f. Culture the tumor specimens in antibiotic/antimycotic growth medium until
there is no visible contamination when viewed under a microscope. Thereafter
make medium changes with complete growth medium without antibiotic/
antimycotic supplementation.
6. Collagenase treatment of tissue explant segments can be utilized to liberate cells
and stimulate growth. The procedure is as follows:
a. Remove explants and associated liquid medium from the flask and place in a
new flask of the same size.
170
Fig. 6. Automated cell counting and image analysis. (A) Screenshot illustrating the results of a typical scan of cell number vs
drug dose. The information that is automatically displayed includes the raw cell counts/well, the average cell number/drug dose, the
average cell number/control well, the standard deviation, and a dose-response graph (B) that represents the CI (% cell kill) at each
drug dose compared to the control wells.
The ChemoFx Assay
7.
8.
9.
10.
11.
12.
13.
171
b. Place new growth medium of the correct type into the old flask to keep any
cells that may be present fed and hydrated during the time that the explants are
being enzymatically treated.
c. Add one-tenth the culture volume of Type I collagenase stock to the new
flask containing the explants. The final collagenase concentration should be
100 U/mL. Incubate at 37C for 4 h without agitation.
d. Draw up the collagenase/medium suspension in a sterile pipet and place in a
conical tube.
e. Pellet the resulting cell suspension at 200g for 3 min.
f. Resuspend the pellet in HBSS with calcium and magnesium, recap the tube, and
agitate gently for 30 s. Centrifuge the cell suspension again at 200g for 3 min.
g. Resuspend the pellet in the appropriate growth medium, and add the cell suspension to the original flask and return to the incubator.
For optimal attachment and growth, some tumor types (see Table 1) require
growth on a Vitrogen 100-coated flask or plate as follows:
a. Dilute the vitrogen 100 stock solution (0.1 mL of Vitrogen 100/9 mL of water)
in sterile tissue culturegrade water.
b. Pipet the appropriate amount of diluted Vitrogen 100 solution into each flask,
or plate and swirl to ensure an even coating of the plastic surface.
c. Pour off excess liquid and allow the flasks or plates to air-dry overnight in a
laminar flow hood.
d. Once the flasks or plates are dry, store at room temperature in sterile sleeves.
Do not let the cells air-dry at any time (it sacrifices structure).
Do not touch the bottom of the wells with a pipet tip (it makes plastic shavings).
Do not create air bubbles in the wells (solutions will not reach the cells).
If mycoplasma contamination is suspected, the case is to be terminated, and all
plates or culture flasks associated with that patient are to be immediately disinfected with bleach and discarded in the biohazard waste.
The exposure time for each drug is dependent on the mechanism of action of that
drug. For example, for carmustine (BCNU) and lomustine (CCNU), it is 1 h;
for hydroxytamoxifen, it is 24 h, and for all other anticancer agents it is 2 h.
Reasons for failed plating or plate rejection include debri, fibroblast contamination, microbial contamination, too few cells, variable cell counts, excessive
trypsinization, and poor plating efficiency.
Acknowledgments
We wish to acknowledge the assistance of Linda Smith in the preparation of
the manuscript.
References
1. Ness, R. B., Wisniewski, S. R., Eng, H., and Christopherson, W. (2002) Cell viability assay for drug testing in ovarian cancer: in vitro kill versus clinical response.
Anticancer Res. 22, 11451149.
2. Kornblith, P., Wells, A., Gabrin, M. J., et al. (2003) In vitro responses of ovarian
cancer to platinums and taxanes. Anticancer Res. 23, 543548.
172
Ochs et al.
3. Kornblith, P., Wells, A., Gabrin, M. J., et al. (2003) Breast cancer-response rates to
chemotherapeutic agents studied in vitro. Anticancer Res. 23, 34053412.
4. Ochs, R. L., Fensterer, J., Ohori, N. P., et al. (2003) Evidence for the isolation,
growth, and characterization of malignant cells in primary cultures of human
tumors. In Vitro Cell Dev. Biol. Anim. 39, 6370.
5. Oberhammer, F., Bursch, W., Parzefall, W., et al. (1991) Effect of transforming
growth factor beta on cell death of cultured rat hepatocytes. Cancer Res. 51,
24782485.
6. Evans, D. L. and Dive, C. (1993) Effects of cisplatin on the induction of apoptosis in proliferating hepatoma cells and nonproliferating immature thymocytes.
Cancer Res. 53, 21332139.
7. Wilson, J. W., Wakeling, A. E., Morris, I. D., Hickman, J. A., and Dive, C. (1995)
MCF-7 human mammary adenocarcinoma cell death in vitro in response to hormone-withdrawal and DNA damage. Int. J. Cancer 61, 502508.
8. Iguchi, K., Hirano, K., Hamatake, M., and Ishida, R. (2001) Phosphatidylserine
induces apoptosis in adherent cells. Apoptosis 6, 263268.
9. Chu, E. and DeVita, V. T. Jr. (2001) In vitro drug response assays, in Cancer:
Principles and Practice of Oncology (DeVita, V. T. Jr., Hellman, S., and Rosenberg, S. A., eds.), Lippincott, Williams & Wilkins, Philadelphia, pp. 302304.
10. Fruehauf, J. P. and Bosanquet, A. G. (1993) In vitro determination of drug
response: a discussion of clinical applications. Principles Pract. Oncol. Updates 7,
116.
11. Fruehauf, J. P. (2002) In vitro assay-assisted treatment selection for women with
breast or ovarian cancer. Endocr. Relat. Cancer 9, 171182.
12. Andreotti, P. E., Cree, L. A., Kurbacher, K. M., et al. (1995) Chemosensitivity
testing of human tumors using a microplate adenosine triphosphate luminescence
assay: clinical correlation for cisplatin resistance of ovarian carcinoma. Cancer
Res. 55, 52765282.
13. OMeara, A. T. and Sevin, B.-U. (2001) Predictive value of the ATP chemosensitivity assay in epithelial ovarian cancer. Gynecol. Oncol. 83, 334342.
14. Singh, B., Li, R., Xu, L., et al. (2002) Prediction of survival in patients with head
and neck cancer using the histoculture drug response assay. Head Neck 24,
437442.
15. Tanigawa, N., Kitaoka, A., Yamakawa, M., Tanisakak, K., and Kobayashi, H.
(1996) In vitro chemosensitivity testing of human tumors by collagen gel droplet
culture and image analysis. Anticancer Res. 16, 19251930.
16. Nagourney, R. A., Brewer, C. A., Radecki, S., et al. (2003) Phase II trial of gemcitabine plus cisplatin repeating doublet therapy in previously treated, relapsed
ovarian cancer patients. Gynecol. Oncol. 88, 3539.
17. Moll, R., Franke, W. W., Schiller, D. L., Geiger, B., and Krepler, R. (1982) The
catalog of human cytokeratins: patterns of expression in normal epithelia, tumors
and cultured cells. Cell 31, 1124.
14
Evaluating Response to Antineoplastic Drug
Combinations in Tissue Culture Models
C. Patrick Reynolds and Barry J. Maurer
Summary
The mainstay of clinical antineoplastic chemotherapy is multiagent combinations, most of
which were developed empirically. Because of the desire to speed research and decrease costs,
there is increasing interest in moving new drugs into clinical trials in potentially active combinations based on preclinical testing data. Different mathematical models have been proposed for
evaluating drug interactions, which can be classified as synergistic (combinations demonstrating
greater than the additive activity expected from each agent alone), additive, or antagonistic (drugs
showing less activity in combination than expected from the sum of each agent alone). Here, we
briefly review some of the principles for testing cytotoxic drug interactions. We focus this review
on application of the Combination Index method (as developed by Chou and colleagues) in the
evaluation of drug interactions in cell culture assays.
Key Words
Drug synergism; drug antagonism; chemotherapy; combination index; isobologram.
1. Introduction
Virtually all curative chemotherapy regimens for cancer employ multiagent
drug combinations (1). Although ideal drug combinations would be those that
are synergistically active against malignant cells without increased systemic
toxicity, additive antitumor activity with a favorable toxicity profile can also be
clinically beneficial. The large majority of currently employed combination
chemotherapy regimens have been developed empirically. While patterns
of cross-resistance, degree of normal organ toxicity overlap, and mechanisms of
drug mechanisms are often considered in designing combination regimens, formal
preclinical testing of the combinations has played a minor role in clinical trials of
combination chemotherapy (2). Thus, whether clinically active multiagent comFrom: Methods in Molecular Medicine, vol. 110: Chemosensitivity: Vol. 1: In Vitro Assays
173
174
Reynolds and Maurer
binations are effective due to increased cytotoxicity for tumor cells that is additive or synergistic remains undefined for most combination chemotherapies. As
the number of drugs available for clinical trials continues to increase, the
number of possible combinations of agents increases almost exponentially. This
has enhanced interest in developing robust preclinical models to assess drug
combinations, which may identify clinically useful combinations, enabling
reduced development costs and more rapid clinical delivery (3). Of particular
importance is the ability to identify and preclude the testing of drug combinations that are likely to be antagonistic when tested in clinical trials.
Combination testing of drugs can be done directly (i.e., delivered concurrently), or sequentially, in time. Although the principles of defining interactions can be applied to either case, in general, most experimental studies
exploring synergy employ direct testing of drug combinations. Comparison of
the dose-response curves of two agents in combination (in fixed-ratio concentrations) to either agent alone may demonstrate apparent antagonism or apparent synergy by inspection alone when the skilled eye is employed. However,
formal quantitative analysis of experimental data is always preferred over
descriptive presentation, to preclude unintended observer bias, and also because
such analysis is generally required to distinguish additive from truly synergistic effects. Formal methods that have been employed to evaluate drug interactions include isobologram methodology (46), the Nonlinear Mixture (surface
response) model (5,79), and the combination index (CI) (6,1013). We focus
here on the adaptation of the CI method (based on the multiple drug-effect
equation of Chou-Talalay, originally derived for use in enzyme kinetic models)
to in vitro anticancer drug testing. CI analysis has been widely used for this
purpose, and its application is simplified by the availability of user-friendly
microcomputer software (14).
Drug interactions can also be studied in vivo, and elegant demonstrations of
synergy in mouse xenografts have been reported (6,1518). However, the numbers of variables that impact synergy studies in xenografts are substantial,
requiring large numbers of animals to achieve statistically valid results, with
consequent investment of time and cost. Formal models used for in vitro modeling of synergy are applicable to xenograft studies, but relatively few laboratories have the resources to conduct experiments of the size needed to truly
define synergy using either the isobologram or CI approaches. A model more
suitable for xenograft studies employing the F-test has been proposed, and
successfully applied, to the testing of drug combinations, but it still requires
a significant investment of resources (19). One alternative approach to large
xenograft experiments for defining synergistic drug interactions would be
first to define synergy in robust tissue culture models, and then to confirm a
Assessing Drug Combinations
175
beneficial interaction between the two drugs (i.e., increased antitumor response)
in more limited xenograft studies.
2. Materials
1. Data from a suitable cytotoxicity assay such as DIMSCAN (see Note 1).
2. Microsoft Excel (Microsoft Office). Data can be collected easily in the Excel program and manipulated, e.g., to calculate averages or log transformations, as
required.
3. SigmaPlot (Jandell, San Rafael, CA). Analyzed data can be copied from Excel to
SigmaPlot to create publication-quality graphics.
4. CalcuSyn (Biosoft, Cambridge, UK; www.biosoft.com). Data from both fixedratio and non-fixed-ratio drug combination experiments can be analyzed using
this Microsoft Windowsbased program. CalcuSyn analyzes drug interactions and
provides tables and graphics of a variety of values from application of medianeffect equations to the data. In addition to calculating the various tables, graphs,
and values for assessing drug interactions described below, CalcuSyn can determine the effective concentrations needed for each drug to produce a given dose
effect. In such calculations, the concentration of a drug needed to affect 90% of
cells is termed an EC90 (effective concentration for 90%). For cytostatic assessment, this is generally referred to as an IC90 (inhibitory concentration) and for
cytotoxic agents as an LC90 (lethal concentration).
3. Methods
3.1. CI Method
The CI method is based on the median-effect principle derived by Chou
(10,20,21). The median-effect equation correlates the drug dose and cytotoxicity or cytostatic effect in the following form:
fa / fu = (D/Dm)m or its alternative form, D = Dm[ fa / (1 fa)]1/m
in which D is the dose of the drug; Dm is the median-effect dose signifying the
potency, determined from the x-intercept of the median-effect plot; fa is the
fraction affected by the dose; fu is the fraction unaffected ( fu = 1 fa); and m
is an exponent that signifies the sigmoidicity (shape) of the dose-effect curve,
which is determined by the slope of the median-effect plot.
The median-effect equation is utilized to calculate Dx, which is the dose of
a drug that inhibits (or kills) x percent of cells. The CI is then calculated as
(D)1 + (D)2
CI = +
(Dx)1 + (Dx)2
The preceding equation is employed when mutually exclusive drugs are

used that have the same or similar modes of action (i.e., those drugs that are, by
176
Reynolds and Maurer
analogy to enzyme kinetics, competitive inhibitors of each other; e.g., two

agents that both inhibit the same active site of a kinase, or perhaps, two microtubule inhibitors) (10,21). For drugs with mutually nonexclusive mechanisms
of action (i.e., those drugs that are, by analogy, noncompetitive inhibitors of
each other; e.g., a DNA alkylator and a microtubule inhibitor, or a tyrosine
kinase inhibitor and a topoisomerase inhibitor), the following equation applies:
(D)1
(D)2
(D)1(D)2
CI = + +
(Dx)1
(Dx)2
(D)1(D)2
Owing to the complexity of whole-cell biological systems, the CalcuSyn

program automatically analyzes a data set using both the mutually exclusive
and mutually nonexclusive assumptions. The CI equation determines the
additive effect of drug combinations, such that synergism is defined as a
greater-than-the-expected-additive effect, and antagonism is defined as lessthan-an-expected-additive effect. Thus, CI = 1 indicates an additive effect, CI < 1
indicates a synergistic effect, and CI > 1 indicates antagonism. The precise biological significance of various degrees of synergism or antagonism remains to
be defined, but it has been proposed that CI values be interpreted as follows:
<0.1
0.10.3
0.30.7
0.70.9
0.91.1
1.11.45
1.453.3
>3.3
very strong synergism

strong synergism
synergism
moderate to slight synergism
nearly additive
slight to moderate antagonism
antagonism
strong to very strong antagonism
Because CI values may change with the fraction affected (Fa) in a nonlinear
manner, the CI should optimally be presented for each effective concentration
(EC) tested, or an overall CI value presented that is generally reflective of the
CI values calculated at the various ECs tested (e.g., EC50, EC90, and EC90 or
EC99).
3.2. Conducting Fixed-Ratio Analysis of Drug Interactions
A prerequisite for fixed-ratio calculations is the generation of accurate doseresponse curves for the agents tested, both alone and in combination. The accuracy of assessing drug interactions will depend directly on the accuracy of the
method used to assess their cytotoxicity or growth inhibition. In testing the
antitumor properties of antineoplastic drugs, the effect can be an inhibition of
growth or cytotoxicity (most assays measure a combination of both). The
177
dynamic range of the assay is important, because effective combinations should

achieve an inhibition of growth over time, or direct cytotoxicity, of >2 to 3
logs (the two-log rule) for an expectation of clinical response (based on current response definitions and calculated disease burdens in leukemia patients
and solid tumor masses) (22). Accuracy of the calculations is also impacted by
the number of different points taken along the dose-response curve.
To determine potency and shape in a dose-response relationship, three data
points are a minimum value, although CI can be calculated using any number
of data points. The linear correlation coefficient (r) of the median-effect plot
should be reasonably good (r > 0.9 is common for cell culture experiments).
To assess synergy, each drug should have some effect as a single agent. Both
the potency (Dm) and shape (m) parameters that are derived from dose-response
data are required for assessing synergism or antagonism. A CI can be derived
for a combination of a noneffective agent and an effective agent, and although
these can be referred to in terms of synergism or antagonism, such interactions
are more properly termed modulation (potentiation, augmentation, or inhibition) of the active agent by the inactive agent.
CI values for each data point of a nonconstant ratio experiment can also be
calculated, as long as m and Dm parameters are available for each single drug.
An example of such an experiment is the use of one drug at a fixed concentration while varying the concentration of the second drug. However, such an
experimental design does not allow calculation of an Fa-CI plot simulation
(though the points can still be placed on the Fa-CI plot), nor does it allow plotting of a classic isobologram.
The preferred experimental design is to employ each drug alone, and in combination, at a fixed ratio of concentrations (e.g., 21 or 13). Drugs with two
different units (micromolar vs micrograms/milliliter) can be analyzed directly
in fixed-ratio combinations. One approach to selecting values for the fixed ratio
combinations is to combine each drug at its equipotent ratio (i.e., the ratio of
the EC50 concentrations), and then to create a combination that is four- to eightfold higher than the EC50, and use serial dilutions of the highest concentration
combination to generate the dose-response curve (each single agent in the combination is tested alone in the same manner) (see Note 2).
3.3. Examples of Assessing Drug Interaction Data
The first example (Fig. 1) provided is from a published study that assessed
the synergistic interaction of fenretinide (a cytotoxic retinoid known to stimulate generation of ceramides in tumor cells in vitro) and safingol (a stereoisomer of sphinganine, a precursor of ceramide) (23). First, the data for Fig. 1
were obtained from the DIMSCAN assay of fixed-ratio exposures to each drug
178
Fig. 1. Effects of L-threo-dihydrosphingosine (safingol) on fenretinide (4-HPR) cytotoxicity in a neuroblastoma and a breast
cancer cell line (23). The cytotoxicity dose response of (A) the SK-N-RA neuroblastoma cell line and (B) the DoxR MCF-7 breast
cancer cell line to 4-HPR (H), safingol (S), and 4-HPR/safingol (31 ratio) (H+S) using a fluorescence-based assay employing digital imaging microscopy (DIMSCAN) is shown (24). Cell lines were exposed to drug(s) and responses were assayed at 4 d. () 4HPR; () safingol; () 4-HPR/safingol (31 ratio). Synergy was quantified by Combination Index (CIN) analysis and expressed as
log10 (CIN) vs fraction affected. By this method, log10 (CIN) < 0 indicates synergy; log10 (CIN) = 0 indicates an additive effect; and
log10 (CIN) > 0 indicates antagonism. Ninety-five percent confidence intervals are shown on CIN plots where calculable. Bars indicate 95% confidence intervals. Note that CIN is used as an abbreviation for Combination Index because the journal in which the data
were originally published used CI as an abbreviation for confidence interval.
179
alone and in combination, and the data were analyzed in CalcuSyn. Second,
values from the CI tables were then plotted using SigmaPlot to generate publication-quality graphics demonstrating the synergistic interaction between fenretinide and safingol.
The second example is a computer screen showing CalcuSyn after it has
generated the median-effect and CI plots that are used to analyze drug interactions for the combination of cyclophosphamide (as the active metabolite
4-hydroperoxycyclophosphamide [4-HC]) and etoposide (Fig. 2). The data, data
analysis, and methods of obtaining and analyzing the data associated with Fig. 2
are presented elsewhere in this volume (24) (see Note 3). A number of examples of the use of the CI approach to assess interactions of antineoplastic drugs
can be found in the literature (12,23,2529).
4. Notes
1. Determination of drug interactions for antineoplastic agents is ideally done using
an assay with a 3 log dynamic range, especially if each individual agent is capable of 1 to 2 logs of cytotoxicity. Some commonly used cytotoxicity assays (such
as 3-[4,5-dimethylthiazolyl-2]-2,5-diphenyltetrazolium bromide) cannot achieve
>2 logs of dynamic range with many cell types and, thus, should not be employed
for assessing synergy and antagonism. We find that the DIMSCAN assay, owing
to its 4 log dynamic range, is particularly suitable for drug combination testing
(23,25). DIMSCAN and other in vitro cytotoxicity assays are reviewed by
Keshelava et al. in Chapter 12 (24).
2. If the actual clinically achievable plasma level for each drug is known or suspected, then the ratio between the drugs should reflect this. The drug concentration ratios should also reflect the lower than maximal drug concentrations likely
to be achieved when using the combination owing to the additive systemic toxicities of each agent, and the possible lower drug levels achieved in tumor tissue relative to plasma. Otherwise, in vitro overmodeling may occur, which can diminish
the predictive value of the preclinical studies when the drug combination is tested
in clinical trials.
3. Details on the use of CalcuSyn are found in the software users manual (14).
Acknowledgments
We thank Dr. Nino Keshelava and Dr. Rita Grigoryan for the data used to
generate Fig. 2. This work was supported in part by the Neil Bogart Memorial
Laboratories of the T.J. Martell Foundation for Leukemia, Cancer, and AIDS
Research; and by National Cancer Institute grants CA82830 and CA81403.
180
181
References
1. Frei, E. and Antman, K. H. (1997) Combination chemotherapy, dose, and schedule,
in Cancer Medicine (Holland, J. F., Bast, R. C., Morton, D. L., Frei, E., Kufe, D. W.,
and Weichselbaum, R. R., eds.), Williams & Wilkins, Baltimore, pp. 817837.
2. Kaufman, D. C. and Chabner, B. A. (2001) Clinical strategies for cancer treatment: the role of drugs, in Cancer Chemotherapy & Biotherapy (Chabner, B. A.
and Longo, D. L., eds.), Lippincott Williams & Wilkins, Philadelphia, pp. 116.
3. Gitler, M. S., Monks, A., and Sausville, E. A. (2003) Preclinical models for defining efficacy of drug combinations: mapping the road to the clinic. Mol. Cancer
Ther. 2, 929932.
4. Tallarida, R. J. (2000) Drug Synergism and Dose-Effect Data Analysis, Chapman
& Hall/CRC, New York.
5. Tallarida, R. J. (2001) Drug synergism: its detection and applications. J. Pharmacol. Exp. Ther. 298, 865872.
6. Teicher, B. A. (2003) Assays for in vitro and in vivo synergy [review]. Methods
Mol. Med. 85, 297321.
7. White, D. B., Slocum, H. K., Brun, Y., Wrzosek, C., and Greco, W. R. (2003) A
new nonlinear mixture response surface paradigm for the study of synergism: a
three drug example. Curr. Drug Metab. 4, 399409.
8. Levasseur, L. M., Greco, W. R., Rustum, Y. M., and Slocum, H. K. (1997) Combined action of paclitaxel and cisplatin against wildtype and resistant human
ovarian carcinoma cells. Cancer Chemother. Pharmacol. 40, 495505.
9. Greco, W. R., Bravo, G., and Parsons, J. C. (1995) The search for synergy: a critical review from a response surface perspective [review]. Pharmacol. Rev. 47,
331385.
10. Chou, T. C. (1996) The median-effect principle and the combination index for
quantitation of synergism and antagonism, in Synergism and Antagonism in
Chemotherapy (Chou, T. C. and Rideout, D. C., eds.), Academic, San Diego,
pp. 61102.
11. Chou, T. C., Rideout, D., Chou, J., and Bertino, J. R. (1991) Chemotherapeutic
synergism, potentiation, and antagonism, in Encyclopedia of Human Biology (Dulbecco, R., ed.), Academic, San Diego, pp. 371379.
12. Chou, T. C., Motzer, R. J., Tong, Y., and Bosl, G. J. (1994) Computerized quantitation of synergism and antagonism of taxol, topotecan, and cisplatin against
human teratocarcinoma cell growth: a rational approach to clinical protocol design
[see comments]. J. Natl. Cancer Inst. 86, 15171524.
Fig. 2. (see opposite page) Screen captured from a computer analyzing data from a
fixed-ratio analysis of the combination of 4-HC + etoposide using a human neuroblastoma cell line. Dose-response curves and data tables from the same experiment as
shown in Fig. 2 can be found elsewhere in this volume (24). Shown are windows containing the dose-effect curves (upper left), the Fa-CI plot (upper right), the medianeffect plot (lower left), and a small portion of CI tables (lower right).
182
Reynolds and Maurer
13. Chou, T. C. (1998) Drug combinations: from laboratory to practice. J. Lab. Clin.
Med. 132, 68.
14. Chou, T. C. and Hayball, M. P. (1996) CalcuSyn Windows Software for Dose Effect
Analysis, Biosoft, Cambridge, MA.
15. Houghton, P. J., Stewart, C. F., Cheshire, P. J., et al. (2000) Antitumor activity of
temozolomide combined with irinotecan is partly independent of O6-methylguanineDNA methyltransferase and mismatch repair phenotypes in xenograft models. Clin.
Cancer Res. 6, 41104118.
16. Meco, D., Colombo, T., Ubezio, P., et al. (2003) Effective combination of ET-743
and doxorubicin in sarcoma: preclinical studies. Cancer Chemother. Pharmacol.
52, 131138.
17. Thompson, J., George, E. O., Poquette, C. A., et al. (1999) Synergy of topotecan
in combination with vincristine for treatment of pediatric solid tumor xenografts.
Clin. Cancer Res. 5, 36173631.
18. Coggins, C. A., Elion, G. B., Houghton, P. J., et al. (1998) Enhancement of irinotecan (CPT-11) activity against central nervous system tumor xenografts by alkylating agents. Cancer Chemother. Pharmacol. 41, 485490.
19. Tan, M., Fang, H. B., Tian, G. L., and Houghton, P. J. (2003) Experimental design
and sample size determination for testing synergism in drug combination studies
based on uniform measures. Stat. Med. 22, 20912100.
20. Chou, T.-C., Rideout, D., Chou, J., and Bertino, J. R. (1974) Relationships between
inhibition constants and fractional inhibitions in enzyme-catalyzed reactions with
different numbers of reactants, different reaction mechanisms, and different types
of mechanisms of inhibition. Mol. Pharmacol. 10, 235247.
21. Chou, T.-C. and Talalay, P. (1984) Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors. Adv. Enyzme
Regul. 22, 2755.
22. Harrison, S. (2002) Perspective on the history of tumor models, in Anticancer
Drug Development Guide (Teicher, B. A., ed.), Humana, Totowa, NJ, pp. 319.
23. Maurer, B. J., Melton, L., Billups, C., Cabot, M. C., and Reynolds, C. P. (2000) Synergistic cytotoxicity in solid tumor cell lines between N-(4-hydroxyphenyl)retinamide
and modulators of ceramide metabolism. J. Natl. Cancer Inst. 92, 18971909.
24. Keshelava, N., Frgala, T., Krejsa, J, Kalous, O, and Reynolds, C. P. (2005)
DIMSCAN: a microcomputer fluorescence-based cytotoxicity assay suitable for
pre-clinical testing of combination chemotherapy, in Chemosensitivity (Blumenthal, R. D., ed.), Humana Press, Totowa, NJ, 2005, vol. 1, chap. 12.
135140.
26. Xu, J. M., Azzariti, A., Colucci, G., and Paradiso, A. (2003) The effect of gefitinib
(Iressa, ZD1839) in combination with oxaliplatin is schedule-dependent in colon
cancer cell lines. Cancer Chemother. Pharmacol. 52, 442448.
183
27. Ricotti, L., Tesei, A., De Paola, F., et al. (2003) In vitro schedule-dependent interaction between docetaxel and gemcitabine in human gastric cancer cell lines. Clin.
Cancer Res. 9, 900905.
28. Takahashi, N., Li, W., Banerjee, D., et al. (2002) Sequence-dependent synergistic
cytotoxicity of ecteinascidin-743 and paclitaxel in human breast cancer cell lines
in vitro and in vivo. Cancer Res. 62, 69096915.
29. Topaly, J., Zeller, W. J., and Fruehauf, S. (2001) Synergistic activity of the new
ABL-specific tyrosine kinase inhibitor STI571 and chemotherapeutic drugs on
BCR-ABL-positive chronic myelogenous leukemia cells. Leukemia 15, 342347.
15
Image Analysis Using the Fluorochromasia Assay
to Quantify Tumor Drug Sensitivity
John F. Gibbs, Youcef M. Rustum, and Harry K. Slocum
Summary
A method of assessing chemosensitivity of tissue utilizing tissue fluorescence and image
analysis was implemented to provide a rapid and quantitative means of assessing the effect of
drugs on tissue metabolic activity and proliferative capacity. The fluorescent microscopic image
captured by a silicon-intensified target (low-light-detecting) camera and linked to an imageprocessing unit was measured for fluorescent brightness and tumor image area. An established
rodent model served to characterize the systems ability to measure serially the tumors metabolic
activity and growth. Further studies on fresh human tumors were conducted with a novel topoisomerase II inhibitor, NC-190. Tumor image area and fluorescent brightness were measured
24 h pretreatment, 48 h posttreatment, and 48 h postdrug removal. Fifty-five percent (28/51) of
fresh human tumors showed sensitivity to 48-h exposure to 10, 30, or 100 M NC-190. The
potential benefit of this technique is the ability to predict the response of tumors to chemotherapeutic agents as a laboratory tool for preclinical drug evaluation and clinically prior to the commencement of therapy.
Key Words
Chemosensitivity; image analysis; fluorescein; fluorochromasia assay; NC-190; topoisomerase
II inhibitor.
1. Introduction
An analytical image analysis system to assess drug responsiveness of solid
organ tumors was implemented to evaluate the activity of chemotherapeutic
agents utilizing the fluorochromasia assay. Fluorochromasia is the production of
a fluorescent molecule within a cell owing to metabolism of a nonfluorescent
precursor substrate. This provides the basis for a noninvasive assessment of
metabolic activity within a cell or tissue. The precursor substrate that we used
is fluorescein diacetate (FDA) (nonpolar), which diffuses across viable cellular
185
186
Gibbs et al.
Fig. 1. Fluorochromasia is a nondestructive process of intracellular accumulation of

fluorescein. FDA (nonpolar) diffuses across viable cellular membrane and is cleaved by
esterase(s) into the fluorescein ion (polar). The process is rapid at a concentration of
0.06 M for 10 min. Fluorescein diffuses out of the cell very slowly.
membrane and is cleaved by esterase(s) into the fluorescein ion (polar) (Fig. 1).
The process is rapid, and fluorescein ion is produced intracellularly (14). The
fluorescein emits green light under blue light illumination. Quantitative analysis within the assay would improve its general usefulness. A program was
developed to detect changes in fluorescent brightness and tumor area as an
index of drug responsiveness with computer-assisted imaging. The physicochemical components of the system were validated initially with a fluorescent
standard. Our fluorochromasia assay system deviated from that of the original
description by Rotman in the use of single instead of multiple tumor fragments.
Individual tumor fragments allow for more precise analysis. We initially evaluated the biological growth pattern and drug responsiveness characteristics with
a known fluoropyrimidine-sensitive rodent model, which had been well characterized in our laboratory (5). We then evaluated the activity of novel drugs on
fresh human tumor specimens. NC-190 is a novel benzophenazine topoisomerase II inhibitor that was selected as the most active analog of the original
compound NC-021 against the P388 murine leukemia cell line (6,7). The effect
of NC-190 on fresh human tissue is reported to illustrate the effectiveness of
the computer-assisted fluorochromasia assay.
2. Materials
1. RPMI-1640 medium, 10% fetal bovine serum, Dulbeccos solution, penicillin
streptomycin, and gentamicin (Gibco, Grand Island, NY).
2. FDA (mol wt = 416.39) (Aldrich, Milwaukee, WI) and sodium fluorescein (mol
wt = 376.28) (Alcon, Fort Worth, TX).
3. Purified collagen (Vitrogen 100; 3 mg/mL) (Celtrix, Santa Clara, CA).
4. 5-Fluorouracil (5-FU) (mol wt = 130.1) (Sigma, St. Louis, MO).
Quantitative Chemosensitivity Test
187
5. Non-heat-treated tea bag paper (cellulose matrix) (C.H. Dexter, Windsor

Locks, CT).
6. Twenty-mesh stainless-steel screen (Small Parts, Miami, FL).
7. Falcon 12-well plates.
8. Qualitative Whatman no. 1 filter paper.
9. Evans antistatic spray (Evans Specialty, Richmond, VA).
10. NC-190 (mol wt = 456.4). This was kindly supplied by Taisho. Drug solubility
and stability under culture conditions was determined by ultraviolet spectroscopy
using freshly diluted stock (5 mM) at pH 10.2 into pH 7.4 aqueous buffer before
and after 2 h incubation at 37C. In vitro drug concentrations up to 100 M could
be utilized without fear of precipitation or chemical breakdown owing to the physiological pH and temperature conditions employed.
3. Methods
3.1. Fluorescent Standard
A fluorescent standard is utilized to test the physical and chemical stability
of the system and to define its proper settings (see Note 1).
1. Place Whatman filter paper strips (2 2 mm) saturated with 1.5 L of linear dilutions of stock fluorescein solution on metal grids in Falcon multiwell plates.
2. Precondition plates with antistatic spray to avoid electrostatic effect on the fluorescent paper strips. This is only necessary for the fluorescent standards, not
for the papers bearing biological samples. The fluorescent standard is stable to
diminution in intensity from photobleaching over time.
3. Measure the integrated optical brightness and image area (8). The system is standardized for its ability to quantitate fluorescence and the area of the fluorescent
object.
3.2. Fluorescent Analytical Measurements

An image analysis system previously described was modified for this
assay (9).
1. Obtain a fluorescent image under an Olympus Highlight 2000 light source filtered to maximize blue light intensity, and view through a yellow barrier filter at
10 magnification with an Olympus stereomicroscope model SZH.
2. Connect the microscope to a Hamamatsu C2400 silicon-intensified target (SIT)
low-light-detecting camera at settings gain 0, offset 0, sensitivity minimal, and
shading off (Fig. 2).
3. Capture 32 frames (l frame = 1/32nd s) using an Argus 10 image processor
(Hamamatsu).
4. Maintain the background intensity threshold at 411 for the fluorescent standard
and 512 for biological tissue to enhance the border-vs-background delineation
(see Note 2).
188
Gibbs et al.
Fig. 2. Image capture of fluorescent object by blue light source connected via
fiberoptics to microscope. The magnified image is transferred to an image processor by
an SIT camera. After accumulation of the image and modification of the background,
the image is transferred to a Quantimet Q970, where the binary area and integrated
optical brightness are recorded. Data files are transferred to a PC/IBM computer for
analysis.
5. Transfer the integrated image to a Cambridge Quantimet Q970 (Leica, Deerfield,
IL) for programmed analytical measurements, which are recorded and stored on
removable disks (Iomega).
6. Transfer the data from the Cambridge Quantimet Q970 to a PC/IBM, format
through an Excel 4.0 file, and import into Sigmaplot 5.0 for graphic display. A
program is created to measure binary image area and to detect changes in fluorescent brightness. Figure 3 shows the pattern of recognition by the system. The
image is detected by the system as a series of measuring units termed pixels. A
background threshold is set, and each pixels fluorescent brightness is recorded, on
a scale of 256 levels of gray, and the sum of the brightness of all pixels in the
image area is calculated. This provides data on total integrated brightness (of all
pixels), and brightness per unit area, as the pixels per unit area is measured using
a sizing standard. Figure 4 shows the expected linear relationship between total
brightness and fluorescein concentration.
3.3. Fresh Rodent Specimens

The 1,2-dimethlylhydrazine-induced mucinous colon and jejunal adenocarcinoma described by Ward et al. (9) was passed by sc trocar technique into
Fischer rats. This model was selected for this study because of its fluoropyri-
189
Fig. 3. Computerized interpretation of fluorescent image. The letter A characterizes

the image capture. The area is measured by summating small processing units termed
pixels. The pixel is calibrated a priori to a known size. Each pixel additionally has an
associated fluorescent brightness unit. If a detector threshold is set to interpret pixel
brightness greater than a set value, the corresponding image is obtained.
midine sensitivity (1114). The transplanted tumor has a doubling time of

approx 4 d in the rodent after a latent period.
Fresh samples of cells were taken from a freezer for implantation every 3 to
4 mo so that the tumors did not undergo more than 10 serial passages in rats.
Therapeutic manipulation of the Ward rat tumor proceeded according to the
following scheme:
1.
2.
3.
4.
Measurement 24 h pretreatment measurement on explant d 7.

Drug treatment on explant d 8.
Evaluation for direct drug effect 48 h following drug removal on d 10.
Evaluation for tissue recovery 48 h postdrug treatment on d 12 (see Note 3).
Using these time points, the effect of 5-FU on the Ward rat tumor was ascertained using the imaging system. Figure 5 demonstrates the systems ability to
detect dose responsiveness in the rodent model. At the highest clinically achievable dose of 5-FU (500 M), there was a 39% decrease in area compared with
the nontreated group, 113% when compared with pretreatment measurements
immediately after drug treatment. There was an intermediate effect using
50 M 5-FU (84.5%). Dose responsiveness was not demonstrated between
50 and 150 M 5-FU. When evaluated for recovery 48 h following drug
removal, dose responsiveness was sustained; the tumor fragments did not
recover from the drug effect.
190
Gibbs et al.
Fig. 4. Changes in total brightness of the fluorescent image while measuring a constant area are detected by the image analysis system. The fluorescent standard reveals
a linear decrease in total brightness at a concentration between 24 and 37 M fluorescein. The total brightness units are arbitrarily set for the integrated optical brightness
above the preset threshold for a given pixel area.
3.4. Fresh Human Surgical Specimens

The effect of NC-190 on fresh human surgical tissues was evaluated. NC-190
had shown promising activity among a variety of cell lines in our laboratory.
There was a wide range of IC50 values among the cell lines evaluated even within
a given tumor type, suggesting intertumor heterogeneity. Fresh solid tumor specimens are sent to the laboratory on ice and mechanically disaggregated.
1. Slice the tumors into 0.5-mm slices with a Stadie-Riggs microtome and then
crosscut with scalpel blades. Prefabricate 24-well tissue culture plates (Falcon)
containing sterilized collagen-impregnated cellulose papers on metallic grids and
store at 4C for later use.
2. Following mechanical disaggregation, embed small tissue fragments in the collagen-impregnated cellulose mattress on top of the metallic grids.
3. Fill the wells to an air-fluid interface with culture medium, and return the plates
to a 95% air/5% CO2 incubator at 37C.
4. One day after implantation, incubate the samples in the presence of FDA (15 M)
for 30 min.
5. Assess fluorescent brightness and image area under blue light.
Fig. 5. Effect of 5-FU drug treatment on (A) tumor surface area and (B) total brightness of rodent colon adenocarcinoma. The drug treatment conditions were performed in
triplicate with an SD <20%. Pretreatment look was done on d 7. Subsequent repeat
imaging was done on d 10 for assessment of direct drug effect and on d 12 and 14 for
recovery effect. (From ref. 5 with permission.)
192
Gibbs et al.
Fig. 6. Fluorochromasia assay. Twenty-eight (55%) tumors displayed intermediate or

complete responsiveness to 100 M NC-190. Of the tumor types with adequate sample
sizes, lung and gastrointestinal (GI) malignancies had promising effects at all concentrations tested. Gynecological and breast malignancies also showed good responsiveness to drug treatment. Approximately 38% of surgical samples overexpressing
P-glycoprotein (P-GP) were responsive to NC-190 at 100 M. NC-190 sensitivity
against metastatic specimens ranged from 23% at 10 M to approx 62% at 100 M.
6. Magnify and capture the fluorescent microscopic image with an SIT (low-lightdetecting) camera linked to an image-processing unit. Print the enhanced image
with a digital Sony video printer.
7. On d 3 following fresh tumor implantation, start drug treatment with NC-190 at
concentrations of 10, 30, and 100 M.
8. After 48 h of exposure, wash the wells with phosphate-buffered saline, and replace
the culture medium.
9. On d 7 after implantation (48 h postexposure), reimage the samples after incubation with FDA. Thus, each culture serves as its own control, comparing drug
responsiveness by means of fluorochromasia before and after drug exposure.
Tumor responsiveness is graded as sensitive, intermediate, or resistant, depending
on the degree of image fluorescence or reduction in area.
193
10. Measure tumor image area and fluorescent brightness at 24 h pretreatment, 48 h

posttreatment, and 48 h postdrug removal. In our study, 55% (28/51) of fresh
human tumors showed sensitivity to 48 h of exposure to 10, 30, or 100 M
NC-190 (Fig. 6). Sensitivity is seen in 4/8 sarcomas and 23/40 epithelial tumors.
Lung malignancies appear to be most responsive to NC-190 treatment. Although
NC-190 did not compare in efficacy to proven tumor-responsive agents, it may be
most useful in combination with standard agents, or against tumors that have
failed first-line therapy. In this regard, it is notable that tumor specimens with
overexpression of P-glycoprotein showed an appreciable response to NC-190. The
response of NC-190 against human metastatic samples corresponding to the activity against the metastatic murine model appears promising for further testing
against metastatic tumors (see Note 4).
4. Notes
1. The fluorescent standard was necessary to control for instrument variations, which
otherwise could confound interpretation of results with variable biological samples. The analytical measurement program was tested for the ability to detect
decreasing fluorescent intensity while measuring a constant image area. Figure 4
shows the expected linear relationship between total brightness and fluorescein
concentration. An initial question was, could this image analysis system detect
preliminary changes in tumor fluorescence after drug treatment before a change in
tumor area could be seen? This was the primary rationale for incorporating the
integrated optical brightness program into the Quantimet measuring system after
initial pilot studies with the fluorescent standard. As seen in Fig. 4, the system was
capable of detecting small changes in fluorescence over a constant area in the
fluorescent standard. This may be unachievable in a transplanted model system
because in this tumor the growing fraction of cells is located at the periphery. A
decrease in brightness seems to be accompanied by a corresponding reduction in
area. Further evaluation to elucidate this possibility remains. The systems ability
to visualize the growing rim and necrotic center of tumor spheroids could lead to
studies with agents that work in a hypoxic environment.
2. Many inherent problems must be overcome in order to enable quantitation through
the use of an automated computerized image analysis system. Interpretation of
the image signal by the system must be validated and reproducible. Each part of
the hardware system was specifically tested through experiments designed to isolate and decouple the unit series imaging system for evaluation of the instrumentation prior to biological testing. The addition of the fluorescent standard proved
invaluable. An early problem detected through this systematic evaluation was the
presence of varied signal intensity. The original light source (Nikon) had an inadequate female adapter for the fiberoptic that created variability in motion from
the poor fit. Incorporation of the Olympus Highlight 2000 overcame this problem.
3. The variation in fluorescence between the fluorescent standard and biological
tissue was an additional problem. The standard represented an idealized situation.
It was a flat and dry object that did not disperse light and thus gave a precise
194
Gibbs et al.
image. By contrast, biological tissue must be kept in a moist, warm environment

and has a variable three-dimensional shape, thus dispersing light to various
degrees. Therefore, after reconfirmation of the system, it was found that imaging
of biological tissue required a higher light source setting than the fluorescent standard, and a 48-h recovery interval between observations. The growth characteristics of the Ward rat tumor obtained with the imaging system demonstrated a
pattern similar to the growth pattern seen in vivo. A latent period of tumor growth
was seen from 4 to 7 d. This period was followed by a rapid growth phase
between 7 and 15 d and succeeded by a plateau phase. Thus, the optimal time
point for drug testing would be d 715. During the rapid growth phase, there was
an increase in fluorescence around the rim of the spheroid-shaped tumor fragment. Tumors grew and remained viable up to 90 d postexplanation, at which
time the tumors sometimes had outgrown the matrix.
4. Despite the inability to determine linear dose responsiveness, the image system
was able to verify drug response in both animal and human tumor specimens with
5-FU and NC-190. Other tumor types and chemotherapeutic agents may show a
more consistent dose response.
References
1. Rotman, B., Teplitz, C., Dickinson, K., and Cozzolino, J. P. (1988) Individual
human tumors in short-term micro-organ cultures: chemosensitivity testing by
fluorescent cytoprinting. In Vitro Cell Dev. Biol. 24(11), 11371146.
2. Leone, L. A., Meitner, P. A., Myers, T. J., et al. (1991) Predictive value of the
fluorescent cytoprint assay (FCA): a retrospective correlation study of in vitro
chemosensitivity and individual responses to chemotherapy. Cancer Invest. 9(5),
491503.
3. Meitner, P. A. (1991) The fluorescent cytoprint assay: a new approach to in vitro
chemosensitivity testing. Oncology 5(9), 7588.
4. Rotman, B. and Papermaster, B. W. (1996) Membrane properties of living mammalian cells as studied by enzymatic hydrolysis of fluorogenic esters. Proc. Natl.
Acad. Sci. USA 55, 134141.
5. Gibbs, J. F., Slocum, H. K., Cao, S., and Rustum, Y. M. (1999) Image analysis for
quantitation of solid tumor drug sensitivity. Int. J. Surg. Invest. 1, 133138.
6. Nakaike, S., Yamagishi, T., Samata, K., et al. (1989) In vitro activity on murine
tumors of a novel antitumor compound, N--dimethylaminoethyl 9-carboxy5-hydroxy-10- methoxybenzo[a]phenazine-6-carboxamide sodium salt (NC-190).
Cancer Chemother. Pharmacol. 23, 135139.
7. Tsuruo, T., Naito, M., Takamori, R., et al. (1990) A benzophenazine derivative,
N--dimethylaminoethyl 9-carboxy-5-hydroxy-10-methoxybenzo [a]phenazine-6carboxamide, as a new antitumor agent against multidrug-resistant and sensitive
tumors. Cancer Chemother. Pharmacol. 26, 8387.
8. Slocum, H. K., Malmberg, M., Greco, W. R., Parsons, J. C., and Rustum, Y. M.
(1990) The determination of growth rates of individual colonies in agarose using
high-resolution automated image analysis. Cytometry 11, 793804.
195
9. Ward, J. M., Yamamoto, R. S., Weisburger, J. H., and Benjamin, T. (1973) Transplantation of chemically induced metastatic mucinous adenocarcinomas of the
jejunum and colon in rats. J. Natl. Cancer Inst. 51(6), 19931995.
10. Rustum, Y. M., Liu, L., and Zhang, Z. (1988) Role of dose, schedule, and route of
administration of 5-formyltetrahydrofolate: preclinical and clinical investigations,
in The Expanding Role of Folates and Fluoropyrimidines in Cancer Chemotherapy,
vol. 244 (Rustum, Y. and McGuire, J. eds.), Plenum, New York, pp. 3952.
11. Danhauser, L. L. and Rustum, Y. M. (1987) Potential for selective enhancement of
the in vivo metabolism of 1-B-D-arabinofuranosylcytosine in rats by thymidine
pretreatment. Cancer Res. 45, 20022007.
12. Danhauser, L. L. and Rustum, Y. M. (1984) Chemotherapeutic efficacy of 5-fluorouracil with concurrent thymidine infusion against transplantable colon tumors in
rodents. Cancer Drug Deliv. 1(4), 269282.
13. Lai-Sim Au, J., Walker, J. S., and Rustum, Y. (1983) Pharmacokinetic studies of
5-fluorouracil and 5-deoxy-5-fluorouridine in rats. J. Pharm. Exp. Ther. 227(1),
174180.
14. Gibbs, J. F., Slocum, H. K., Reska, N., Winslow, E., Frank, C. and Rustum, Y. M.
(1994) In vitro activity of a novel chemotherapeutic agent, N--dimethylaminoethyl-O-carboxy-5-hydroxy-10-methoxybenzo [a] phenazine-6-carboxamide
(NC-190), against human tumors. Proc. Am. Assoc. Cancer Res. 35, A2405.
16
Immunohistochemical Detection of Ornithine
Decarboxylase as a Measure of Chemosensitivity Testing
Uriel Bachrach
Summary
The development of reliable methods for the in vitro testing of sensitivity of cancer cells to
various drugs has been a longstanding objective in cancer treatment. The development of individualized chemotherapy could minimize undesired toxic side effects and increase the chance of
recovery. The known methods for in vitro chemosensitivity tests are mainly based on monitoring
the metabolic changes induced in cancer cells by the drugs. These experiments, as well as
attempts to cultivate isolated cancer cells, did not give reliable results. In this study, I used a
marker for proliferation to detect the effect of drugs on the potential of cancer cells to divide. An
ideal marker should be present in all cells, be expressed early in the cell cycle, and have a short
half-life. Ornithine decarboxylase (ODC), which catalyzes the conversion of ornithine to
putrescine, fulfilled these prerequisites. Because ODC has an extremely short half-life, it disappears when cellular proliferation is arrested. The decay of ODC was assayed both by determining its activity and by immunohistochemical analyses. This approach was successfully used to
determine the sensitivity of lymphocytes from hematological cancer patients to various drugs. It
is conceivable that this method could serve as an important tool to improve cancer chemotherapy.
Key Words
Chemosensitivity; cancer; in vitro; ornithine decarboxylase; polyamines; immunohistochemistry; chemotherapy.
1. Introduction
Despite many advances in understanding carcinogenesis, the mortality rates
from tumors remain stubbornly high. Surgery, radiotherapy, and chemotherapy
are currently the major means to treat cancer patients. Recent studies suggested
that two important weapons could be added to our arsenal to combat cancer.
One is the prevention of the disease. This can be achieved by improving the
environment and by selecting a proper diet. The second is to develop tailored
197
198
Bachrach
cancer therapy, optimized to individual patients. Oncology is currently an

empirical discipline in which all patients with a particular type of cancer are
treated as though they were the same. Tumors of the same type show heterogeneity of chemosensitivity, and patients with apparently identical tumor histologies do not always respond identically to the same drug regimen. As new
drugs come into the market, oncologists are faced with a bewildering array of
choices. As in the treatment of infectious diseases, an in vitro chemosensitivity
test for cancer could increase the chance of recovery; reduce undesired side
effects due to nonoptimal drugs; and minimize the emergence of multidrug
resistance (MDR) variants.
Early attempts to develop in vitro tailored chemosensitivity tests were
based on the growth of cancer cells in vitro and the formation of foci (1,2).
Soon it became apparent that this assay is not practical because of the low plating efficiency; only 1 to 2% of the cells grew in culture (3). Although individualization of chemotherapy is theoretically attractive, past attempts to provide
such information by using clonogenic assays have produced many papers and
little progress. Attempts to determine cellular growth by studying the incorporation of radioactive precursors into cellular macromolecules did not give reliable results (4,5). Subsequent attention was focused on nonclonogenic assays,
which were mainly based on the effect of drugs on the metabolic activity of
tumor cells. The 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide
assay was one of the earliest developed nonclonogenic assays (6,7). In this
assay, which requires relatively large numbers of cells, cell viability is detected
by the formation of fluorescent products from substrates added to the cells.
This assay, which is not widely used, was applied by Sargent et al. (8) to study
the chemosensitivity of hematological cancer patients. Another metabolic
chemosensitivity test is the adenosine triphosphatebased (ATP) tumor
chemosensitivity assay (ATP-TCA). This test is based on the loss of ATP from
dead cells (9). The ATP-TCA showed a good correlation with clinical outcome
of breast (10) and ovarian (11) cancers. The assay, which is mainly applied for
testing solid tumors, but not for hematological cancer, requires at least 20,000
cells per assay and lasts at least 6 d.
The present state of the art was discussed during a recent meeting of the
International Society for Chemosensitivity Testing in Oncology, which was held
in Homburg/Saar in 2002. The proceedings of the meeting were published in a
book (12).
The aforementioned nonclonogenic assays suffer from serious drawbacks;
they are based on studying the effect of drugs on the metabolic activities
of the cells, rather than studying their effect on cell proliferation. It is well
known that dead cells, or even cellular extracts, can be active metabolically.
An ideal in vitro chemosensitivity test should be based on the assay of a marker
Detection of Ornithine Decarboxylase
199
Fig. 1. Biosynthesis of polyamines.
for proliferation. Such a marker should be universal, found in almost all cells.
It should be expressed during the early stages of the cell cycle and should have
a short half-life, so that it will decay rapidly when cell proliferation is arrested.
Ornithine decarboxylase (ODC) (EC 4.1.1.17) can serve as a marker for proliferation (13,14). This enzyme (Fig. 1) catalyzes the conversion of ornithine
into the diamine putrescine, which is the precursor for the synthesis of the naturally occurring polyamines (13).
The polyamine spermidine and spermine play an essential role in growth and
proliferation processes. They accumulate in cancer cells and are found in high concentrations in the urine and blood of cancer patients (15). Moreover, polyamines
trigger the transformation of cultured NIH3T3 cultured fibroblasts (16).
Studies from my laboratory indicated that ODC could serve as a marker for
proliferation (17). Indeed, ODC was used to determine the chemosensitivity of
various cultured cells to different anticancer drugs (18,19). In those studies, the
in vitro chemosensitivity of the cultured cells was assessed by determining the
enzymatic activity of ODC. This assay required at least 106 cells and did not
200
Bachrach
Fig. 2. Detection of ODC by immunohistochemical methods in (A) drug-sensitive

and (B) -resistant cells.
Fig. 3. Immunohistochemical imaging in drug-resistant cells incubated for 24 h in

presence of 0.5 g/mL of vinblastne.
201
Fig. 4. Immunohistochemical imaging in drug-sensitive cells incubated for 24 h in

presence of 0.5 g/mL of vinblastne.
Fig. 5. Quantitative estimation of ODC by confocal laser scanning microscope.
202
Bachrach
Fig. 6. Detection of ODC in lymphocytes of normal individuals. Treatment:

, 2 g/mL PHA; , non-stimulated controls; + +, stimulated lymphocytes in the
presence of adriamycin; - - - - stimulated lymphocytes in the presence of methotrexate; , stimulated lymphocytes in the presence of vincristine.
permit the study of individual cells. To detect ODC in individual cells, an immunohistochemical staining method was developed (Figs. 24). This method permitted
the detection of ODC in individual cells and could be completed within 48 h (20).
The data presented in Figs. 24 provided a semiquantitative estimation of the
amounts of ODC in the cells. Definite quantitative results were obtained by
screening slides using a confocal laser scanning microscope (Fig. 5). This
approach was successfully used to test the in vitro chemosensitivity of hematological cancer patients (12,21,22).
My colleagues and I used the immunohistochemical ODC assay to test the
sensitivity of lymphocytes from 20 healthy individuals to various anticancer
drugs. As expected, cells from the healthy control subjects were sensitive to
all the drugs tested and ODC levels decreased in drug-treated cells (Fig. 6).
Lymphocytes from seven patients who did not respond to therapy and died
were MDR, and ODC was detected in drug-treated cells (Fig. 7).
These findings strongly suggest that MDR could be detected in specimens
taken from individual patients. Fifty patients who suffered from mild hematological cancers responded to therapy, parallel with the results of the ODC
Fig. 7. Detection of ODC in lymphocytes of deceased cancer patients. Treatment: ,

2 g/mL PHA; , nonstimulated controls; + +, stimulated lymphocytes in the
presence of adriamycin; - - - -, stimulated lymphocytes in the presence of
methotrexate; , stimulated lymphocytes in the presence of vincristine.
Patient no. 1 suffered from non-Hodgkins lymphoma with secondary CLL; no. 2 CLL
and no. 3 multiple myeloma. The in vitro test showed that all of them had MDR.
204
Bachrach
Fig. 8. Detection of ODC in lymphocytes of other cancer patients. Treatment: ,

2 g/mL PHA; , nonstimulated controls; + +, stimulated lymphocytes in the
presence of adriamycin; - - - -, stimulated lymphocytes in the presence of
methotrexate; , stimulated lymphocytes in the presence of vincristine.
Patient no. 1 suffered from hairy cell leukemia, no. 2 multiple myeloma and no. 3
essential thymocytosis. All of them responded to clinical treatment.
205
Fig. 9. Use of immunohistochemical assay to test sensitivity of leukemic blast cells

to polyphenol ()-epigallocatechin-3-gallate (EGCG) found in green tea.
chemosensitivity assay (Fig. 8). Part of the results was published (12,21,22). In
addition, the test was applied to bone marrow of a patient suffering from multiple myeloma (22). The ODC assay was also used to define the optimal dose
for treatment with an anticancer agent and was used for in vitro testing of new
drugs (Fig. 9) or their combination.
Materials
1. Escherichia coli JM109 cells carrying plasmid pGEM-1 containing a 1.8-kb
BamHI-EcoRI fragment of mouse ODC cDNA.
2. ODC antibodies.
3. M9 medium containing ampicillin (200 g/mL).
4. 1 mM Isopropyl--D-thiogalactopyranoside (IPTG).
5. Bacteria lysis buffer: 1% Nonident NP-40, 25 mM Tris-HCl (pH 7.4), 20 mM
MgCl2 + 1 mM phenylmethylsulfonyl flouride, 1 mM dithiothreitol, 1% aprotinin
(Sigma, St. Louis, MO).
6. Deoxycholic acid.
7. Deoxyribonuclease (Sigma).
8. Polyethylene glycol (PEG) 6000 (Merck, Darmstadt, Germany).
9. Freunds adjuvant (Sigma).
10. Dulbeccos modified Eagles medium (DMEM).
206
Bachrach
11. Ficoll-paque (Pharmacia Biotech, Uppsala, Sweden).

12. RPMI-1640 medium containing 50 U/mL of penicillin, 50 g/mL of streptomycin,
and 15% autologous plasma.
13. Phytohemagglutinin (PHA) (2 g/mL) (Sigma).
14. L-Polylysine (110) (Sigma).
15. 1% Bovine serum albumin (BSA) in phosphate-buffered saline (PBS).
16. Fluorescein isothiocyanate (FITC)conjugated Affinipure goat antirabbit IgG antibodies (Jackson, West Grove, PA).
17. Mounting solution: 90% glycerol, 10% PBS (pH 10.0), 0.1% sodium azide, and
5% 3,3-diaminobenzidine (DAB).
18. Streptavidin-conjugated goat antirabbit IgG antibodies (11500) (Jackson).
19. Biotin-conjugated peroxidase (1500) (Jackson).
20. 0.05% Diaminobenzidine, 0.01% hydrogen peroxide, and 0.02% dimethyl sulfoxide (DMSO) in PBS.
3. Methods
3.1. Preparation of ODC Antibodies (see Note 1)
1. Grow E. coli JM109 cells carrying the ODC cDNA fragment at 37C in
M9 medium containing ampicillin (200 g/mL) to an optical density of 0.6 at
550 nm.
2. After overnight induction of ODC, with 1 mM IPTG sediment the cells by
centrifugation.
3. Resuspend the pellet in bacteria lysis buffer at a ratio of 3 mL of buffer/g of
packed bacterial sediment.
4. Shake the suspension at room temperature for 20 min and then subject to three
cycles of freezing and thawing.
5. Add deoxycholic acid to the lysed bacteria (4 mg/3 mL of original suspension).
6. Shake the bacteria continuously until a viscous fluid results.
7. Add deoxyribonuclease (1 mg/3 mL of original suspension) to the extract, and
continue incubating at room temperature for another 30 min until the viscosity
disappears.
8. Centrifuge the suspensions and suspend the sediment obtained in lysis buffer supplemented with 1 M urea.
9. After shaking at 37C for 10 min, centrifuge the samples again, and suspend the
sediment in lysis buffer containing 2 M urea.
10. Continue the procedure with increasing urea concentrations to a total of 10 times
(at a final concentration of 10 M urea).
11. Analyze the supernatants from the 10 fractions by electrophoresis in polyacrylamide gels, and concentrate fractions rich in ODC (53-kDa band, usually fractions of 68 M urea) by placing them in dialysis tubing suspended in PEG.
12. Inject ODC-rich fractions intravenously and subcutaneously into rabbits (approx
500 g of ODC/rabbit) in the presence of complete Freunds adjuvant.
13. Repeat immunization twice by injecting the protein in the presence of complete
Freunds adjuvant in 3-wk intervals.
207
14. Bleed the rabbits and test sera for ODC neutralization activities.
3.2. Separation of Hematological Cancer Cells (see Note 2)

1. Mix heparinized fresh blood from hematological cancer patients with an equal
volume of DMEM.
2. Layer this suspension, which can be kept at room temperature for several hours,
on Ficoll-paque at a ratio of 41.
3. After centrifuging at 1200g for 25 min, collect the lymphocytes and wash twice
with DMEM by centrifuging at 700g for 10 min.
3.3. Incubation of Cells

1. Suspend lymphocytes in RPMI-1640 medium containing 50 U/mL of penicillin,
50 g/mL of streptomycin, and 15% autologous plasma.
2. To increase ODC activity, stimulate the cells by adding PHA at a concentration of
2 g/mL (23).
3. Incubate the stimulated cells at 37C in a 5% CO2 incubator with or without anticancer drugs at various concentrations.
4. After 24 h, wash the lymphocytes with PBS by centrifuging at 700g for
10 min.
3.4. Immunohistochemical Detection of ODC (see Note 3)

1. Prepare slides either by dropping the cells on slides covered with L-polylysine
(110 in double-distilled water) or by using a cytospin at 1200g for 5 min.
2. Dry the slides at room temperature for 2 h and fix cells with cold (20C)
methanol for 2.5 min.
3. Rinse the slides twice with PBS, and incubate at room temperature in a humidified box for 30 min with 1% BSA in PBS.
4. Incubate the slides overnight at 4C with anti-ODC antibodies (1400 in 1% BSA
in PBS).
3.4.1. Staining Method #1

1. After repeated rinsing, expose the slides for 1 h at 37C to FITC-conjugated
Affinipure goat antirabbit IgG antibodies (1100).
2. Wash the slides four times with PBS, and cover with a mounting solution (90%
glycerol; 10% PBS, pH 10.0; 0.1% sodium azide; and 5% DAB).
3. For quantitative analyses, examine the specimens using a confocal laser scanning
microscope, Photobus 1000 (Sarastro, Sweden) at 730 mV. The value of each section is the average fluorescent activities in the cells from two areas. Results are
expressed as arbitrary units per square millimeter area of scanned cells.
The proposed ODC chemosensitivity assay has the following advantages:

1. It is a marker for proliferation.
2. It can be applied to specimens from hematological and solid tumors.
208
Bachrach
Fig. 10. Immunohistochemical detection of ODC by staining lymphocytes exposed

to drugs.
3.
4.
5.
6.
It requires a small number of cells.

It allows the study of single cells.
It is fast and can be completed within 48 h.
It is relatively inexpensive.
Because a confocal laser scanning microscope is not available in many laboratories, the following simple modification has been used:
3.4.3. Staining Method #2
1. Separate, wash, and place the lymphocytes on slides as described above. After
fixation, treat the cells overnight with ODC antibodies, wash, and incubate for
30 min at 37C with streptavidin-conjugated goat antirabbit IgG antibodies (11500).
2. Wash the cells again and incubate at 37C for 30 min with biotin-peroxidase
(1500).
3. After washing, place the slides for 10 min in a solution containing 0.05% DAB,
0.01% hydrogen peroxide, and 0.02% DMSO in PBS.
4. After washing, dry the slides and stain with hematoxylin for 1 min.
209
Fig. 11. Immunohistochemical detection of ODC by staining cultured fibroblasts

exposed to drugs.
5. Examine the slides microscopically, and score the number of ODC-containing

cells (stained brown) (Figs. 10 and 11).
4. Notes
1. Instead of preparing ODC antibodies as described in Subheading 3.1., monoclonal antibodies (MAbs) cand be obtained from Sigma (cat. no. O-1136) or be
prepared as described in ref. 24 as follows:
a. Couple a hexadecapeptide (P16) representing the amino acid sequence
345360 of ODC to BSA and use for the production of MAbs.
References
1. Hamburger, A. W. and Salmon, S. E. (1977) Primary bioassay of human stem cells.
Science 194, 461463.
2. Ajani, J. A., Baker, F. L., Spitzer, G., et al. (1987) Comparison between clinical
response and in vitro drug sensitivity of primary human tumors in the adhesive
cell culture system. J. Clin. Oncol. 5, 19121921.
210
Bachrach
3. Salmon, S. E., Young, L., Scuderi, P., and Clark, B. (1987) Antineoplastic effects
of tumor necrosis factor alone and in combination with gamma-interferon on tumor
biopsies in clonogenic assay. J. Clin. Oncol. 5, 18161821.
4. Kern, D. H., Sondak, V. K., Morgan, C. R., and Hildebrand-Zanki, S. U. (1987)
Clinical application of the thymidine incorporation assay. Ann. Clin. Lab. Sci. 17,
383388.
5. Von-Hoff, D. D. and Weisenthal, L. (1980) In vitro methods to predict for patient
response to chemotherapy. Adv. Pharmacol. Chemother. 17, 133156.
6. Suto, A., Kubota, T., Shimoyama, Y., Ishibiki, K., and Abe, O. (1989) MTT assay
with reference to clinical effect of chemotherapy. J. Surg. Oncol. 42, 2832.
7. Xu, J. M., Song, S. T., Tang, Z. M., et al. (1999) Predictive chemotherapy of
advanced breast cancer directed by MTT assay in vitro. Breast Cancer Res. Treat.
53, 7785.
8. Sargent, J., Elgie, A., Williamson, C., and Taylor, C. (1997) The use of MTT assay
to study drug resistance in acute myeloid leukemiaan update. Adv. Blood Dis. 3,
3341.
9. Ahmann, F. R., Garewal, H. S., Schifman, R., Celniker, A., and Rodney, S. (1987)
Intracellular adenosine triphosphate as a measure of human cell viability and
drug-modulated growth. In Vitro Cell Dev. Biol. 23, 474480.
10. Cree, I. A., Kurbacher, C. M., Untch, M., et al. (1996) Correlation of the clinical
response to chemotherapy in breast cancer with ex vivo chemosensitivity. AntiCancer Drugs 7, 630635.
11. Konecny, G., Crohns, C., Pegram, M., et al. (2000) Correlation of drug response
with ATP tumor chemosensitivity assay in primary stage III ovarian cancer.
12. Bachrach, U. (2003) Recent chemosensitivity testing in oncology, in Recent Results
in Cancer Research, vol. 161 (Reinhold, U. and Tilgen, W., eds.), Springer Verlag,
Berlin, pp. 6270.
13. Pegg, A. E., Shantz, L. M., and Coleman, C. S. (1995) Ornithine decarboxylase as
a target for chemoprevention. J. Cell. Biochem. Suppl. 22, 132138.
14. Cohen, S. S. (1998) A Guide to the Polyamines, Oxford University Press, NY.
15. Bachrach, U. (1989) Polyamines as indicators of disease activity and response to
therapy in cancer, in The Physiology of Polyamines, vol. 2 (Bachrach, U. and
Heimer, Y. M., eds.), CRC Press, Boca Racon, FL, pp. 235250.
16. Tabib, A. and Bachrach, U. (1999) Role of polyamines in mediating malignant transformation and oncogene expression. Int. J. Biochem. Cell Biol. 31,
12891295.
17. Shayovits, A. and Bachrach, U. (1995) Ornithine decarboxylase: an indicator
for growth of NIH 3T3 fibroblasts and their c-Ha-ras transformants. Biochem. Biophys. Acta 1267, 107114.
18. Bachrach, U., Shayovitz, A., Marom, Y. Ramu, A., and Ramu, N. (1994) Ornithine
decarboxylasea predictor for tumor chemosensitivity. Cell Mol. Biol. 40,
957964.
211
19. Assaraf, Y. G., Drori, S., Bachrach, U., and Shaugan-Labay, V. (1994) Determination of multidrug resistance levels in cultured mammalian cells using ornithine
decarboxylase activity. Anal. Biochem. 216, 97109.
20. Shayovits, A. and Bachrach, U. (1994) Immunohistochemical detection of
ornithine decarboxylase in individual cells: potential application for in vitro
chemosensitivity assays. J. Histochem. Cytochem. 42, 607611.
21. Wang, Y., Ashkenazi, Y. J., and Bachrach, U. (1999) In vitro chemosensitivity
testing of hematological cancers: immunohistochemical detection of ornithine
decarboxylase. Anti-Cancer Drugs 10, 797805.
22. Wang, Y., Or, R., and Bachrach, U. (2000) Chemosensitivity testing of hematological cancers using ornithine decarboxylase as a marker. Int. J. Med. Biol. Environ. 28, 5156.
23. Faber, J., Menashe, M., Bachrach, U., and Desser, H. (1980) Formation of
putrescine and acetylspermidine from spermidine by cultured human lymphocytes.
FEBS Lett. 121, 165168.
24. Schipper, R. G. Romain, N., Otten, A. A., et al. (1999) Immunohistochemical
detection of ornithine decarboxylase. J. Histochem. Cytochem. 47, 13951404.
17
Immunohistochemistry of p53, Bcl-2, and Ki-67
as Predictors of Chemosensitivity
Mitsuyoshi Itaya, Jiro Yoshimoto, Kuniaki Kojima, and Seiji Kawasaki
Summary
Chemosensitivity is affected by molecular biological factors, including factors related to the
induction of apoptosis and the activity of proliferation. We analyzed immunohistochemically the
expression of p53, Bcl-2, and Ki-67 in various types of cancers and assessed the correlation
between this expression and chemosensitivity. Moreover, we investigated whether the expression of these factors could be a useful predictor for the clinical response to chemotherapy. Study
subjects comprised 63 preoperative patients with untreated malignant tumors (9 with esophageal
cancer, 12 with stomach cancer, 12 with colon cancer, 16 with liver cancer, and 14 with breast
cancer). Immunohistochemical staining (the labeled streptavidin biotin technique: LSAB method)
was used to assess expression of p53 protein, Bcl-2 protein, and Ki-67. A chemosensitivity test
was carried out with the histoculture drug response assay method using four drugs: mitomycin
C, 5-fluorouracil, doxorubicin hydrochloride (ADM), and cisplatin (CDDP). Immunohistochemical studies for p53 were found to be useful for predicting chemosensitivity.
Key Words
Immunohistochemical staining; LSAB, p53; Bcl-2; Ki-67; chemosensitivity; histoculture drug
response assay.
1. Introduction
It has been reported that the sensitivity of cancer cells to anticancer drugs is
affected by various factors, including factors related to the induction of apoptosis
(1). Furthermore, it has been shown that the apoptosis-associated p53 gene and
Bcl-2-related genes are involved in the induction of apoptosis (2,3) and thus affect
chemosensitivity. In addition, it has been reported that proliferation activityrelated
factors such as Ki-67 and proliferating cell nuclear antigen are also important in
determining chemosensitivity (4). Moreover, we reported that immunohistochemical studies for p53 were useful for predicting chemosensitivity (5).
213
214
Itaya et al.
Immunohistochemical staining techniques allow the visualization of tissue

(cell) antigens. These techniques are based on the immunoreactivity of antibodies and the chemical properties of enzymes or enzyme complexes, which
react with colorless substrate-chromogens to produce a colored end product.
The sensitivity of immunohistochemical stains was significantly improved
with the development of an indirect technique. In this two-step method,
enzyme-labeled secondary antibodies react with the antigen-bound primary
antibody. A further increase in sensitivity over the indirect technique was
achieved with the introduction of the peroxidase-antiperoxidase (PAP) enzyme
complex (6). In this method, the secondary antibody serves as a linking
antibody between the primary antibody and the PAP complex (6). Subsequent
developments in immunohistochemistry (IHC) exploited the strong affinity of
avidin for biotin and resulted in the avidin-biotin complex (ABC) method (7).
This technique employs an enzyme-labeled avidin-biotin complex, which is
mixed prior to use and forms a complex with a biotinylated secondary antibody. The ABC method increased reagent sensitivity when compared with the
PAP method.
The LSAB method that we used is based on a modified labeled avidin-biotin
(LAB) technique in which a biotinylated secondary antibody forms a complex
with peroxidase-conjugated streptavidin molecules (8). In comparison with the
ABC method, the LAB/LSAB method has been reported to be four to eight
times more sensitive (9).
The LSAB method is a sensitive and versatile IHC procedure that permits
the simultaneous processing of numerous specimens with rabbit or mouse primary antibodies in a few hours. In this chapter, we describe the technique and
procedure for immunohistochemical staining (LSAB method) of p53 protein,
Bcl-2 protein, and Ki-67 as predictors of chemosensitivity.
2. Materials
1.
2.
3.
4.
5.
6.
7.
8.
9.
Graded series of alcohol.

Xylene.
Paraffin wax.
Target retrieval buffer: 0.01 M citrate buffer, pH 6.0.
0.05 M Tris-buffered saline (TBS) or 0.02 M phosphate-buffered saline (PBS)
buffer (pH 7.27.6) without sodium azide.
Dako LSAB +/HRP kit.
Hydrogen peroxide.
Normal goat serum (NGS) or bovine serum albumin.
Antibodies: monoclonal mouse antihuman p53 protein (clone DO-7; Dako), monoclonal mouse antihuman BCL2 oncoprotein (clone 124; Dako), monoclonal
mouse antihuman Ki-67 antigen (clone MIB-1; Dako).
Predictors of Chemosensitivity
215
Table 1
Immunohistochemical Staining Procedure (LSAB Method)
1. Specimen preparation (fixation, paraffin embedding, sectioning) (see
Subheading 3.1.).
2. Deparaffinize sections and rehydrate in distilled water (see Subheading 3.2.).
3. Heat the sections three times for 3 min each with 0.01 M citrate buffer
(pH 6.0) in a microwave oven.
4. Leave the sections in citrate buffer at room temperature (2025C) for
1520 min to cool.
5. Wash the sections in PBS buffer (pH 7.6) once for 5 min.
6. Cover the sections with 3% hydrogen peroxide and incubate for 5 min (see
Subheading 3.4.1.).
7. Rinse the sections gently with distilled water and place in fresh PBS buffer for
5 min.
8. Cover the sections with NGS and incubate for 5 min (see Subheading 3.4.2.).
9. Cover the sections with primary antibody (see Subheading 3.4.3.).
20 min.
11. Cover the sections with link antibody and incubate for 30 min (see
Subheading 3.4.4.).
12. Rinse the sections gently with distilled water and place in fresh PBS buffer
for 20 min.
13. Cover the sections with streptavidin-HRP and incubate for 15 min (see
Subheading 3.4.5.).
20 min.
15. Cover the sections with DAB solution and incubate for 510 min (see
Subheading 3.4.6.)
16. Wash the sections in distilled water for 5 min.
17. Counterstain with methylgreen (if required), dehydrate, coverslip, and mount
(see Subheadings 3.4.7. and 3.4.8.).
10. Biotin-labeled affinity-isolated goat antirabbit and goat antimouse immunoglobulin (Ig) in 0.01 M PBS (1200 dilution).
11. Horseradish peroxidase (HRP)-conjugated streptavidin: 200 L of 1.25 M TrisHCl buffer (pH 7.6) + 4 mL of distilled water + 40 L of peroxidase streptavidin.
12. Diaminobenzidine (DAB) solution: 1 mL of DAB in 0.05 M Tris-HCl buffer,
pH 7.6, 1501100 dilution, containing 17 L of 30% hydrogen peroxide. Mix
well before incubating.
13. 3-Amino-9-ethylcarbazole (AEC) in N,N-dimethylformamide and acetate buffer,
pH 5.0, containing hydrogen peroxide.
14. Methyl green.
216
Itaya et al.
3. Methods
3.1. Preparation of Specimen (see Table 1, step 1)
Prior to immunohistochemical staining, tissues must be fixed and processed.
Fixation prevents autolysis and putrefaction of excised tissues, preserves antigenicity, enhances the refractive index of tissue constituents, and increases the
resistance of cellular elements to tissue processing. Tissue processing includes
dehydration, cleaning of dehydrating agents, infiltration of embedding media,
embedding, and sectioning of tissues.
3.1.1. Fixation (see Note 1)
Survival of tissue antigens for immunological staining may depend on the

type and concentration of fixative, the fixation time, and the size of the tissue
specimen to be fixed (10,11). Most formaldehyde-based fixatives contain 10%
formalin, a neutral salt to maintain tonicity, and a buffering system to maintain
pH. These fixatives are well tolerated by tissues, exhibit good histological penetration, and are well suited for labeled streptavidin-biotin immunostains.
1. Fix small blocks of tissue (10 10 3 mm) in 510 mL of neutral buffered formalin per block for up to 24 h.
3.1.2. Processing and Paraffin Embedding (see Note 2)
After fixation, processing may be completed using an automatic tissue

processor.
1. Dehydrate tissues using graded alcohols, clear with xylene or xylene substitute,
and infiltrate with paraffin wax.
2. Embed the tissue with paraffin wax in molds or cassettes, which facilitate tissue sectioning. Tissue block may be stored or sectioned on completion of embedding. Properly fixed and paraffin-embedded tissue will keep indefinitely if stored at 225C.
3.1.3. Sectioning (see Note 3)

1. Collect sectioned tissues from paraffin-embedded blocks on clean glass slides.
2. Cut samples into slices 4 m thick to make slide preparations.
3. Dehydrate in an oven for 1 to 2 h at 60C or less. Slides with paraffin-embedded
tissue sections can be kept indefinitely if stored at 225C.
3.2. Deparaffinization (see Table 1, step 2 and Note 4)

Prior to staining, tissue slides must be deparaffinized to remove embedding
media and then rehydrated. Avoid incomplete removal of paraffin. Residual
embedding media will result in increased nonspecific staining.
1. Place slides in a xylene bath and incubate for 5 min. Change baths and repeat
twice.
2.
3.
4.
5.
6.
217
Tap off excess liquid and place slides in 100% ethanol for 1 min.
Tap off excess liquid and place slides in distilled or deionized water for 2 min.
3.3. Antigen Target Retrieval (see Table 1, steps 35 and Note 5)

1. Place slides in a slide holder and fill any empty positions with blank slides. This
will ensure that the same amount of glass is heated on each occasion in order to
ensure reproducibility.
2. Fill an incubation container with target retrieval buffer and insert a plastic slide
holder. Cover the container loosely with a lid to minimize evaporation.
3. Place the container in the middle of a microwave oven. Heat the sections three times
for 3 min each time. Refill the container with distilled water during the treatments.
The tissue sections must under no circumstance dry out during the heating periods.
4. After the third treatment, remove the container from the microwave oven, and leave
the sections in the target retrieval buffer at room temperature for 1520 min to cool.
5. Wash the sections once in TBS or PBS buffer (pH 7.6) for 5 min.
3.4. Immunohistochemical Staining Procedure (see Note 6)

The labeled streptavidin biotin technique, HRP (LSAB method; Dako LSAB +/
HRP kit) is used for immunohistochemical staining.
3.4.1 Peroxidase Blocking Step (see Table 1, steps 6 and 7)
1. Tap off excess liquid from slide. Using a lintless tissue (such as a Kimwipe or
gauze pad), carefully wipe around the specimen to remove any remaining liquid
and to keep reagents within the prescribed area.
2. Apply enough 3% hydrogen peroxide to cover the specimen in order to block
endogenous peroxidase activity (see Note 7).
3. Incubate for 5 min in a humid chamber at room temperature (2025C). (see
Note 16).
4. Rinse gently with distilled water from a wash bottle; do not focus flow directly on
the tissue.
5. Place sections in a fresh TBS or PBS buffer bath for 5 min (see Note 8).
3.4.2. Blocking of Nonspecific Binding (see Table 1, step 8 and Note 9)

1. Tap off excess buffer and wipe slides as in Subheading 3.4.1., step 1.
2. Apply enough NGS solution (or BSA) to cover the specimen.
3. Incubate for 20 min in a humid chamber at room temperature (2025C). (see
Note 16).
3.4.3. Application of Primary Antibody (see Table 1, steps 9 and 10)

1. Tap off excess liquid and wipe slides as in Subheading 3.4.1., step 1.
2. Apply enough primary antibody or negative control reagent to cover the specimen.
218
Itaya et al.
Primary antibodies include the following:

a. Monoclonal mouse antihuman p53 protein (clone DO-7; Dako). Dilution
range: 125150; incubation: 3060 min at room temperature (2025C) (see
Note 10 and 16).
b. Monoclonal mouse antihuman BCL2 oncoprotein (clone 124; Dako). Dilution
range: 1501100; incubation: 3060 min at room temperature (2025C)
(see Note 11 and 16).
c. Monoclonal mouse antihuman Ki-67 antigen (clone MIB-1; Dako). Dilution
range: 1751150; incubation: 3060 min at room temperature (2025C)
(we incubated 60 min regularly) (see Note 12 and 16).
3. Incubate for 30 min in a humid chamber at room temperature (2025C) (see
Note 16).
4. Rinse gently with distilled water as in Subheading 3.4.1., step 4.
5. Place the sections in a fresh TBS or PBS buffer bath for 20 min (see Note 13).
3.4.4. Biotinylated Second Antibody Step (see Table 1, steps 11 and 12

and Note 14)
1. Immediately tap off excess buffer and wipe slides as in Subheading 3.4.1., step 1.
2. Apply enough link antibody (biotin-labeled affinity-isolated goat antirabbit and
goat antimouse Ig in 0.01 M PBS [1200 dilution]) to cover specimen.
Note 16).
4. Rinse gently with distilled water as in Subheading 3.4.1., step 4. (see
Note 15).
5. Place the sections in a fresh TBS or PBS buffer bath for 20 min.
3.4.5. Streptavidin-HRP Step (see Table 1, steps 13 and 14)

1. Wipe slides as in Subheading 3.4.1., step 1.
2. Apply enough HRP-conjugated streptavidin (mix well before using) to cover the
specimen.
Note 16).
4. Rinse gently with distilled water as in Subheading 3.4.1., step 4.
5. Place the sections in a fresh TBS or PBS buffer bath for 20 min.
3.4.6. DAB (or AEC) Substrate-Chromogen Step (see Table 1, steps 15

and 16 and Note 17)
1. Wipe slides as in Subheading 3.4.1., step 1.
2. Apply enough DAB solution to cover the specimen.
3. Incubate for 510 min (10 min for AEC) at room temperature (2025C) (see
Note 16).
4. Rinse gently with distilled water as in Subheading 3.4.1., step 4. Collect DAB
chromogen waste in a hazardous materials container for proper disposal.
219
Fig. 1. Immunohistochemical staining for (A) P53 AND (B) Bcl-2 in patient with
colon cancer. (A) More than 60% of the nuclei of tumor cells were stained with DO-7
(anti-p53 monoclonal antibody [MAb]) (100). (B) The photomicrograph shows positive immunostaining with anti-Bcl-2 MAb (100). This patient with p53 overexpression
and Bcl-2-positive expression showed low chemosensitivity and no response to adjuvant
chemotherapy.
3.4.7. Methylgreen Counterstain (see Table 1, step 17 and Note 18)

1. Wash the sections in distilled water for 5 min.
2. Immerse the slides in a bath of methylgreen and incubate for 3060 min at room
temperature (2025C), depending on the strength of methylgreen used.
3. Place the slides in a propanol bath for two times for 30 s.
4. Place the slides in a xylene bath for four times for 30 s.
3.4.8. Mounting (see Table 1, step 17 and Note 19)
Specimens may be mounted and coverslipped with an aqueous-based mounting medium (such as Dako Faramount or Glycergel).
220
Itaya et al.
Table 2
Results if Immunohistochemical Staining for p53, Bcl-2, and Ki-67
p53 overexpression
Type of cancer
Esophageal
Stomach
Colon
Liver
Breast
Total
Negative
2/9 (22.2)
5/10 (50.0)
5/11 (45.5)
10/15 (66.7)
8/14 (57.1)
30/59 (50.8)
Positive
Bcl-2 expresssion
Negative
Positive
7/9 (77.8)
6/9 (66.7)
3/9 (33.3)
5/10 (50.0) 5/10 (50.0) 5/10 (50.0)
6/11 (54.5) 6/11 (54.5) 5/11 (45.5)
5/15 (33.3) 15/15 (100)
0/15 (0.0)
6/14 (42.9) 5/14 (35.7) 9/14 (64.3)
29/59 (49.2) 37/59 (62.7) 22/39 (37.3)
Ki-67 (LI)a
15.9 3.7
15.5 4.2
23.1 5.6b
9.2 3.8
6.4 1.7
13.5 1.9
Data represent the mean SE.

p < 0.05, compared with liver and breast cancer. Data represent the number of negative or
positive cases/number of evaluable cases (%).
b
3.5. Evaluation of the Immunohistochemical Staining:

p53 and Ki-67 Staining (see Note 20)
For p53 staining, patients having a labeling index (LI) of 10% or more are
assessed as p53 protein overexpression cases (Fig. 1A). For Bcl-2 staining, a
case staining Bcl-2 positive is defined as a patient whose cancer cells are
stained more intensely than the control cells (Fig. 1B). All labeled nuclei with
brown staining are regarded as immunoreactive regardless of the staining intensity or pattern. More than 2000 cells are counted at a final magnification of
~400 in each section to determine the LI, which is the number (percentage) of
positive cells (5) (Table 2).
3.6. Chemosensitivity Assay (HDRA With the MTT Assay)
For culturing tumor tissues, the histoculture technique reported by Hoffman,
et al. was used (12), and, for assessing the effectiveness of drugs, the MTT
assay was applied (13). The assay system combining these two techniques is
referred to as HDRA. Four anticancer drugs were used in the present investigation: MMC, 5-Fu, ADM (Kyowa Hakko Kogyo Co., Ltd., Tokyo, Japan) and
CDDP (Bristol-Myers Squibb K.K., Tokyo, Japan).
4. Notes
1. It is important to maintain optimal, standardized fixation conditions whenever
possible in order to obtain reproducible staining. When possible, use of thinner
specimens coupled with shorter fixation times is recommended. Prolonged expo-
2.
3.
4.
5.
6.
221
sure to fixatives may result in the masking of antigens and contributes to reduced
staining. However, shrinkage or distortions may occur in poorly fixed and embedded tissue specimens. Bouins, B-5, and Zenkers fluid are alternative fixatives
for the preservation of tissue antigens sensitive to routine formalin fixation.
It is necessary to dehydrate tissues completely in order to penetrate paraffin wax
into specimen, so the tissues are dehydrated using graded alcohols for up to 12 h.
To minimize denaturing of antigen, do not expose tissues to temperatures in
excess of 60C during processing.
For increased adhesion of tissue sections during the immunohistochemical staining
procedure, the use of poly-L-lysine-coated slides or aminopropyltriethoxysilanecoated slides (Dako Silanized Glass Slides) is suggested. Coated slides are
recommended for staining procedures requiring proteolytic digestion or target
retrieval.
Xylene and alcohol solutions should be changed after 40 slides. Toluene or xylene
substitutes such as Histoclear may be used in place of xylene. If necessary, rehydrated tissues may be kept in buffer solution at 28C for up to 18 h prior to use.
Allow tissues to come to room temperature (2025C) before staining.
Tissue specimens are in the majority of cases fixed in formalin, followed by paraffin embedding, for microscopic assessment of tissue morphology. The threedimensional structure of protein is altered to a variable degree during this process.
Whereas some antibodies readily react with antigens resistant to formalin fixation and paraffin embedding, others are directed against antigens that lose their
immunological reactivity after routine processing. However, the introduction
of heat-based pretreatment methods, such as microwave oven heating, pressure
cooking, or autoclaving, has brought about a further strong improvement and has
broadened the use of IHC as a very important tool in routine pathology. Microwave oven heating is a highly efficient method of heating aqueous solutions
(1420). Make sure that the microwave oven is 750800 W. The use of microwave
ovens with turntables gives a more uniform heating and should be preferred for
consistent results. In order to obtain optimal results, the microwave treatment time
must be decided based on the choice of target retrieval buffer, the target retrieval
buffer volume, the number of slides, the exposure time, the cooling time, and the
antigen to be detected (15). Silane-coated tissue slides or slides coated with other
suitable adhesive must be used. Metallic material must under no circumstances be
present in the microwave oven during use.
During the immunohistochemical staining procedure, wear appropriate personal
protective equipment to avoid contact with the eyes and skin. All reagents should
be equilibrated to room temperature (2025C) prior to immunostaining. Do not
store these components or perform staining in strong light, such as direct sunlight. Enzyme and chromogens may be affected adversely if exposed to excessive light levels. Never pipet reagents by mouth and avoid contacting the skin and
mucous membranes with reagents and specimens. If reagents or specimens come
in contact with sensitive areas, wash with a copious amount of water.
222
Itaya et al.
7. Endogenous peroxidase activity can be found in hemoprotein such as hemoglobin,

myoglobin, cytochrome, and catalase as well as in eosinophils (21,22). In formalinfixed tissue, this activity can be inhibited by incubating specimens with 3% hydrogen peroxide for 5 min prior to application of the primary antibody. However,
this procedure does not abolish the reddish-brown pigment of hemoproteins.
Alternatively, a solution of methanolhydrogen peroxide can be used. Tissue from
persons infected with hepatitis B virus and containing hepatitis B surface antigen
may exhibit nonspecific staining with HRP. Some antigen may become denatured
with this procedure.
8. The buffered saline washes should be without sodium azide, which inactivates
HRP and results in negative staining.
9. NGS acts as a protein block reagent and reduces excessive nonspecific staining.
NGS solution consists of normal goat serum in 0.01 M PBS (120 dilution).
10. Monoclonal mouse antihuman p53 protein is intended for use in IHC. The antibody labels wild-type (WT) and mutant-type p53 protein and is a useful tool for
the identification of p53 accumulation in human neoplasias (2325). p53 is a
nuclear phosphoprotein with a molecular mass of 53 kDa. WT p53 protein is present in a wide variety of normal cells, but the protein has a very short half-life and
thus is present in only minute amounts (23), generally in the detection level of the
immunohistochemical method (26). In addition, it has been indicated that positive
immunohistochemical staining can be understood to represent an expression
of mutant-type p53 protein (27). Somatic mutation of the p53 gene is a very
frequent event in the development of human neoplasia, and because mutant p53
protein is often much more stable than WT p53 protein, the mutant p53 protein
accumulates to a high level (23). For example, p53 protein accumulation was
observed in 76% of 212 human malignant lesions, including breast, colon, bladder, and uterine carcinoma and soft-tissue sarcomas (28). WT p53 protein functions as a transcription factor; that is, as a modulator that can turn crutial genes
either on or off, it also inhibits DNA replication and is a checkpoint control molecule for progression of the cell cycle. Furthermore, p53 protein is involved in the
regulation of apoptosis (24). In transfection assays, WT p53 behaves as a tumor
suppressor, and mutant p53 behaves as a dominant transforming oncogene (23).
However, the optimal conditions (dilution of the antibody, incubation time, and
temperature) may vary depending on the specimen and preparation method and
should be determined by each individual laboratory. Dilutions for this and other
antibodies should be prepared using antibody diluents that contain 0.01 M PBS,
pH 7.27.6, and 1% BSA. The BSA acts as a protein block, so diluents without
protein carrier are not recommended.
11. Bcl-2 oncoprotein is a blocker of apoptosis cell death. Gene transfer experiments
have shown that elevated levels of this protein can protect a wide variety of cells
from diverse cell death stimuli ranging from growth factor withdrawal and cytotoxic lymphokines to viral infection and DNA-damaging, anticancer drugs and
radiation (29,30). Bcl-2 oncoprotein resides on the cytoplasmic side of the mitochondrial outer membrane, endoplasmic reticulum, and nuclear envelope (29,31)
12.
13.
14.
15.
16.
17.
18.
223
and has a molecular mass of 26 kDa (30). The Bcl-2 gene is involved in the
t(1418) chromosomal translocation found in 85% of human follicular lymphomas
and 20% of diffuse B-cell lymphomas (31). In this translocation, the Bcl-2 gene
at chromosome segment 18q21 is juxtaposed with the Ig heavy chain locus at
14q32, resulting in deregulated expression of Bcl-2 oncoprotein (31).
The MIB-1 antibody has now been established as a reference monoclonal mouse
antibody for demonstration of the Ki-67 antigen in formalin-fixed, paraffinembedded specimens. The Ki-67 antigen is a nuclear protein, which is defined by
its reactivity with MAb from the Ki-67 clone (32). Two isoforms of 345 and
395 kDa have been identified (33). The Ki-67 antigen is preferentially expressed
during all active phases of the cell cycle (G1, S, G2, and M phases), but it is absent
in resting cells (G0 phase) (32). During interphase, the antigen can be exclusively
detected within the nucleus, whereas in mitosis most of the protein is relocated to
the surface of the chromosomes. The antigen is rapidly degraded as the cell enters
the nonproliferative state (30), and there appears to be no expression of Ki-67
during DNA repair processes (34).
To reduce excessive nonspecific staining, TBS with Tween-20 (TBST; Dako) or
0.05 M Tris containing 0.3 M NaCl and 0.1% Tween, pH 7.6, is recommended for
use as a wash buffer.
If the staining protocol must be interrupted, slides may be kept in a buffer bath
following incubation of the link antibody for up to 1 h at room temperature
(2025C) without affecting staining performance.
Distilled water may be used for rinsing the hydrogen peroxide, substratechromogen solution, and counterstain.
All incubations should be performed at room temperature. Do not allow tissue
sections to dry during the immunohistochemical staining procedure. Dried tissue
sections may display increased nonspecific staining, so we strongly recommend
that the tissue sections always be incubated in a humid chamber during the treatment or staining procedure.
DAB may be harmful if inhaled, contacted with skin, or if swallowed. The material is irritating to the eyes and skin. If skin contact should occur, flush the
affected area with soap and water. Although DAB is structurally related to benzidine, there is no evidence of the carcinogenicity of DAB. The DAB substratechromogen is sensitive to contamination from a variety of oxidizing agents such
as metals, bacteria, dust, and commonly used laboratory glassware. To avoid contamination and premature expiration, avoid exposing the DAB solution to any
potential source of contamination and never pipet directly from the bottle. Use the
provided graduated test tube to measure the amount of buffered substrate needed.
Mix well and apply the solution using the provided transfer pipet. After use, rinse
the graduated test tube and pipet thoroughly with distilled water. Do not return
excess DAB solution to the primary storage container.
Depending on the length of incubation and potency of the methylgreen used,
counterstaining will result in a pale to dark coloration of cell nuclei. Excessive or
incomplete counterstaining may compromise proper interpretation of results. Pre-
224
Itaya et al.
cipitates may form if specimens are allowed to dry during the staining procedure.
This may be apparent at the edge of the specimen. Scan tissues at 40 magnification to avoid misinterpretation of results.
19. Slides may be read when convenient. However, some fading may occur if slides
are exposed to strong light over a period of 1 wk. To minimize fading, store slides
in the dark at room temperature (2025C).
20. If ther is no staining of any slides, these are the possible causes and their solutions:
a. Cause 1: Reagents were not used in the proper order. Solution: Review application of the reagents.
b. Cause 2: Sodium azide was in the buffer bath. Solution: Use fresh azide-free
buffer.
c. Cause 3: Substrate-chromogen reagent was mixed incorrectly or not working.
Solution: Test the prepared substrate-chromogen with a drop of the streptavidin
solution. If no color reaction occurs, make a fresh substrate-chromogen solution.
(see Table 1, step 15).
If there is weak staining of all slides, these are the possible causes and their solutions:
a. Cause 1: Sections retained too much solution after the wash bath. Solution:
Gently tap off excess solution before wiping around the section. (see Subheading 3.4.1., step 1).
b. Cause 2: Slides were not incubated long enough with antibodies or substrate
mixture. Solution: Review the recommended incubation times, (see Table 1,
steps 9, 10, and 11).
c. Cause 3: Incompatible counterstain or mounting media dissolved the reaction
product. Solution: For AEC immunostained slides, use only aqueous-based
counterstains and mounting media, (see Table 1, step 17).
If there is excessive background staining in all slides, these are the possible causes
and their solutions:
a. Cause 1: Specimens contained high endogenous peroxidase activity. Solution:
Incubate the slides with fresh hydrogen peroxide, (see Table 1, step 6).
b. Cause 2: Paraffin was incompletely removed. Solutions: Use fresh xylene or
toluene baths; if several slides are stained simultaneously, the second xylene
bath should contain fresh xylene, (see Table 1, step 2).
c. Cause 3: Slides were not properly rinsed. Solution: Use fresh solutions in
buffer baths and wash bottles.
d. Cause 4: There was a faster than normal substrate reaction owing to excessive
room temperature. Solution: Use a shorter incubation time with substratechromogen solution.
e. Cause 5: Sections dried during the staining procedure. Solutions: Use a humidity chamber; wipe only three to four slides at a time before applying reagent.
f. Cause 6: The primary antibody was too concentrated. Solution: Use a higher
dilution of the primary antibody, (see Table 1, step 9).
If tissue sections are detaching from the slides, these are the possible cause and
their solutions:
225
a. Cause 1: Incorrect slides were used. Solutions: Use poly-L-lysine slides for
most staining; for primary antibodies that require target retrieval techniques,
use silanized slides, (see Table 1, step 1).
b. Cause 2: Slides were not properly prepared prior to tissue mounting. Solution:
Try a different lot number of the silanized slides, (see Table 1, step 1).
If there is excessively strong specific staining, these are the possible causes and
their solutions:
a. Cause 1: Too much antibody was used. Solution: Do serial dilutions of primary antibody to determine optimum dilution, (see Table 1, step 9).
b. Cause 2: The incubation for primary antibody, biotinylated link, or streptavidin-HRP was too long. Solutions: Determine the appropriate staining protocol for the antibody; length of incubation for primary antibody, biotinylated
link, and streptavidin-HRP (see Table 1, steps 9, 10, 11, 12 and 13).
References
1. Barry, M. A., Behnke, C. A., Eastman, A. (1990) Activation of programmed cell
death (apoptosis) by cisplatin, other anticancer drugs, toxins and hyperthermia.
Biochem Pharmacol. 40, 23532362.
2. Lowe, S. W., Bodis, S., McClatchey, A., et al. (1994) p53 status and the efficacy
of cancer therapy in vivo. Science 266, 807810.
3. Vaux, D. L., Agulia, H. L., and Weissman, I. L. (1992) Bcl-2 prevents death of
factor-deprived cells but fails to prevent apoptosis in targets of cell mediated
killing. Int. Immunol. 4, 821824.
4. Saito, A., Korenaga, D., Maehara, Y., Baba, H., and Okamura, T. (1992) In vitro
succinate dehydrogenase chemosensitivity of gastric carcinoma: relationship
between the DNA content. Cancer Chemother. Pharmacol. 29, 185189.
5. Itaya, M., Yoshimoto, J., Kojima, K., and Futagawa, S. (1999) Usefulness of p53
protein, Bcl-2 protein and Ki-67 as predictors of chemosensitivity of malignant
tumors. Oncol Rep. 6, 675682.
6. Farr, A. G. and Nakane, P. K. (1981) Immunohistochemistry with enzyme labeled
antibodies. J. Immunol. Meth. 47, 129144.
7. Hsu, S. M., Raine, L., and Farger, H., et al. (1981) Use of avidin-biotin-peroxidase
complex (ABC) in immunoperoxidase techniques: a comparison between ABC
and unlabeled antibody (PAP) procedures. J. Histochem. Cytochem. 29, 577.
8. Guesdon, J. L., Ternyck, T., and Avrameas, S. (1979) The use of avidin-biotin
interaction in immunoenzymatic techniques. J. Histochem. Cytochem. 27,
11311139.
9. Giorno, R. (1984) A comparison of two immunoperoxidase staining methods based
on the avidin-biotin interaction. Diagn. Immunol. 2, 161166.
10. Larsson, L.-I. (1993) Tissue preparation methods for light microscopic immunohistochemistry. Appl. Immunohistochem. 1, 216.
11. Nadji, M. and Morales, A. R. (1983) Immunoperoxidase: 1. The technique and its
pitfalls. Lab. Med. 14, 767.
226
Itaya et al.
12. Suto, A., Kubota, T., Shiroyama, Y., Ishibiki, K., and Abe, O. (1989) MTT assay
with reference to the clinical effect of chemotherapy. J. Surg. Oncol. 42, 2835.
13. Furukawa, T., Kubota, T., Watanabe, M., et al. (1992) High in vitroin vivo correlation of drug response using spongegel-supported three-dimensional histoculture
and the MTT end point. Int. J. Cancer 51, 489498.
14. Shi, S. R., Imam, S. A., Young, L., Cote, R. J., and Taylor, C. R. (1995) Antigen
retrieval immunohistochemistry under the influence of pH using monoclonal antibodies. J. Histochem. Cytochem. 43, 193201.
15. Shi, S. R., Key, M. E., and Kalra, K. L. (1991) Antigen retrieval in formalin-fixed,
paraffin embedded tissues: an enhancement method for immunohistochemical
staining based on microwave oven heating of tissue sections. J. Histochem.
Cytochem. 39, 741748.
16. Leong, A. S.-Y. and Milios, J. (1993) An assessment of the efficacy of the
microwave antigen-retrieval procedure on a range of tissue antigens. Appl.
Immunohistochem. 1, 267274.
17. Taylor, C. R., Shi, S. R., and Cote, R. J. (1996) Antigen retrieval for immunohistochemistry: status and need for greater standardization. Appl. Immunohistochem.
4, 144166.
18. Werner, M., von Wasielewske, R., and Komminoth, P. (1996) Antigen retrieval,
signal amplification and intensification in immunohistochemistry. Histochem. Cell.
Biol. 105, 253260.
19. Gown, A. M., de Wever, N., and Battifora, H. (1993) Microwave-based antigen
unmasking: a revolutionary new technique for routine immunohistochemistry.
Appl. Immunohistochem. 1, 256266.
20. Caitoretti, G., Piferi, S., Parravicini, C., et al. (1993) Antigen unmasking on
formalin-fixed, paraffin-embedded tissue sections. J. Pathol. 171, 8398.
21. Banerjee, D. and Pettit, S. (1984) Endogenous avidin-binding activity in human
lymphoid tissue. J. Clin. Pathol. 37, 223225.
22. Elias, J. M. (1990) Immunohistopathology: a practical approach to diagnosis,
American Society of Clinical Pathologists Press, Chicago.
23. Vojtesek, B., Bartek, J., Midgley, C. A., and Lane, D. P. (1992) An immunochemical analysis of the human nuclear phosphoprotein p53: new monoclonal antibodies
and epitope mapping using recombinant p53. J. Immunol. Methods. 151, 237244.
24. Nieder, C., Petersen, S., Petersen, C., and Thames, H. D. (2001) The challenge of
p53 as prognostic and predictive factor in Hodgkins or non-Hodgkins lymphoma
[review]. Ann. Hematol. 80, 28.
25. Ramael, M., Lemmens, G., Eerdekens, C., et al. (1992) Immunoreactivity for p53
protein in malignant mesothelioma and non-neoplastic mesothelium. J. Pathol.
168, 371375.
26. Cooper, K. and Haffajes, Z. (1997) bcl-2 and p53 protein expression in follicular
lymphoma. J. Pathol. 182, 307310.
27. Fujiwara, T., Cai, D. W., Georges, R. N., Mukhopadhyay, T., Grimm, E. A., and
Roth, J. A. (1994) Therapeutic effect of a retroviral wild-type p53 expression
vector in an orthotopic lung cancer model. J. Natl. Cancer Inst. 86, 14581462.
227
28. Bartek, J., Bartkova, J., Vojtesek, B., et al. (1991) Aberrant expression of the p53
oncoprotein is a common feature of a wide spectrum of human malignancies.
Oncogene 6, 16991703.
29. Adams, J. M. and Cory, S. (1998) The bcl-2 protein family: arbiters of cell survival. Science 281, 13221326.
30. Kusenda, J. (1998) Bcl-2 family proteins and leukeamia [minireview]. Neoplasma
45, 117122.
31. Yang, E. and Korsrheyer, S. J. (1996) Molecular thanatopsis: a discourse on the
BCL2 family and cell death [review]. Blood 88, 386401.
32. Gardes, J., Lemke, H., Baisch, H., Wacker, H.-H., Schwab, U., and Stein, H.
(1984) Cell cycle analysis of a cell proliferationassociated human nuclear antigen
defined by the monoclonal antibody Ki-67. J. Immunol. 133, 17101715.
33. Gardes, J., Li, L., Schlueter, C., et al. (1991) Immunobiochemical and molecular
biologic characterization of the cell proliferation-associated nuclear antigen that
is defined by monoclonal antibody Ki-67. Am. J. Pathol. 138, 867873.
34. Key, G., Kubbutat, M. H., and Gerdes, J. (1994) Assessment of cell proliferation
by means of an enzyme-linked immunosorbent assay based on the detection of
the Ki-67 protein. J. Immunol. Methods 177, 113117.
Index
229
Index
3H-uridine, 140
cytotoxic index, 169

cytotoxicity assay, 29, 42, 87, 121, 139
A
accuracy, 10
antagonism, 173
antibody, 206, 214
antigen target retrieval, 217
apoptosis, 127
ATP tumor chemosensitivity assay
(TCA), 8, 101, 140
attached (adherent) cells, 24, 73
automated cell counting, 168
B
bcl-2, 11, 213
C
CalcuSyn, 175
cell culture, 23, 155
cell dissociation, 61
cell survival, 24
cell-counting, 31
ceramide, 11
ChemoFx assay, 155
clonogenic assay, 6, 21
collagen gel droplet, 7, 59, 140
colony-forming, 6
colorimetric, 69
combination index (CI), 175
confocal laser scanning microscope,
201
Coulter counter, 22
cytokinesis block micronucleus assay, 9
D
DAPI staining, 167
diaminobenzidine (DAB), 218
differential staining cytotoxicity (DiSc)
assay, 7, 49
digital image microscope, 87, 139
DIMSCAN139
dose-response curve, 5, 149
drug doses, 42, 109
drug interaction, drug combinations,
114, 173
drug resistance, 49, 87
dye exclusion, 49
dye exclusion, 140
dynamic range, 140
E
ex vivo assay, 155
extreme drug resistance assay, 9
F
feeder cells, 34
firefly luminescence reaction, 111
fixation, 216
fixed-ratio analysis, 176
flow cytometry, 123
fluorescein diacetate (FDA), 91, 142,
185
fluorescence plate reader, 132
fluorescence-based assay, 139
229
230
fluorescent cytoprint assay, 9
fluorochromasia assay, 185
fluorometric microculture cytotoxicity
assay (FMCA), 140
formazan, 74
Index
microtiter plate, 41, 110, 164
monolayer cultures, 24
MTS, 71
MTT, 69, 81, 140
N
G
gelatin sponge, 80
genomic, 11
gentian violet, 27
green fluorescent protein (GFP), 121
growth conditions, 160
growth kinetics, 41
H
hemocytometer, 22
high-throughput, 131
histoculture drug response assay
(HDRA), 9, 79
hypoxia, 87
I
image analysis, 64, 185, 200
immunohistochemistry (IHC), 11, 161,
197, 213
inhibitory concentration (IC50), 5, 46, 83
isobologram, 173
K
kern assay, 7
Ki-67, 213
L
labeled streptavidin biotin technique
(LSAB), 213
leukemic cells, 51
limiting dilution, 32
luciferin-luciferase, 105
luminescence, 101
lymphoprep, 50
M
MDR, 11
microcomputer, 139
nonadherent cells, 29, 74

nonspecific binding, 217
O
O2- controlled chamber, 95
optical density, 75
ornithine decarboxylase (ODC),
197
P
p53, 11, 213
paraffin embedding, 216
p-glycoprotein, 11
plating efficiency, 24
polyamines. 197
primary cancer cells, 61
proteomic, 11
Q
quality control, 114
R
radioactive isotopes, 140
S
sensitivity, 10
serial dilutions, 25
serial dilutions, 73
specificity, 10
standard curve, 113
Sulforhodamine B (SRB) assay,
8, 39, 140
surgical specimen, 157, 190
survival curve, 26, 36, 104
synergism, 173
Index
T
telomerase, 11
tetrazolium dye, 69
tissue explant, 80, 158
231
transfection, 124
trypan blue, 31
trypsinization, 164
two-drug combination, 114

Chemosensitivity Vol 1

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chemosensitivity Vol 1

Uploaded by

Copyright:

Available Formats

M E T H O D S I N M O L E C U L A R M E D I C I N E TM

John M. Walker, SERIES EDITOR

99. Pain Research: Methods and Protocols,

2005 Humana Press Inc.

II, Part II, Genomics, Proteomics, and Chemosensitivity, addresses DNA

PART II. IN VITRO MEASURES

2 Clonogenic Cell Survival Assay

12 DIMSCAN: A Microcomputer Fluorescence-Based Cytotoxicity

KUNIAKI KOJIMA Department of Breast and Endocrine Surgery, Juntendo

PART I. MEASUREMENTS OF DNA DAMAGE, CELL DEATH,

12 Flow Cytometric Measurement of Functional

PART II. GENOMICS, PROTEOMICS,

14 Gene Expression Profiling to Characterize Anticancer

PART III. IN VIVO ANIMAL MODELING

19 Clinically Relevant Metastatic Breast Cancer

23 Evaluation of Chemosensitivity of Micrometastatses

In addition to the standard, DNA-damaging and metabolic inhibiting agents

Fig. 1. Theoretical dose response curve.

3. Cell Metabolic Activity Assays

established tumor cell lines than in evaluating chemosensitivity of primary

5. Molecular Assays of Chemosensitivity

DNA microarrays permits large-scale screening of genes and the simultaneous

tion related to chemosensitivity testing are worthy of consideration. First, work

produce a large number of progeny is considered dead. This is very commonly

Clonogenic Cell Survival Assay

2.2 Staining of Plates

3.3. Irradiation of Flasks and Performance of Plating Experiment

Clonogenic Cell Survival Assay

be counted. Label the first set of triplicate dishes as 0 Gy A and 0 Gy B (A and

Clonogenic Cell Survival Assay

3.4. Staining of Plates

3.5. Counting of Colonies

High-Sensitivity Cytotoxicity Assays

3.3. Limiting Dilution Clonogenic Assay (2) (see Note 3)

High-Sensitivity Cytotoxicity Assays

3.4. Preparation of Feeder Cells for Limiting Dilution Assay

High-Sensitivity Cytotoxicity Assays

clustering occurs; although it is simple to count small clusters (of 5 cells) by

High-Sensitivity Cytotoxicity Assays

in which T is the OD after exposure to a certain concentration of a drug, T0 is

Sulforhodamine B Assay and Chemosensitivity

2.2. SRB Procedure

10% (w/v) TCA.

Sulforhodamine B Assay and Chemosensitivity

In contrast to the original method as described by Skehan and colleagues, we

Sulforhodamine B Assay and Chemosensitivity

3.4. Data Processing and Interpretation of Results

Sulforhodamine B Assay and Chemosensitivity

dramatically. Sometimes, it might be beneficial to store plates on a warming

assays, as reviewed by several researchers (1,3,68). Clonogenic assays have

DiSC Assay for Chemosensitivity Testing

3. Immunomagnetic Dynabeads M-450 (Dynal, Norway) (18) to remove contaminating lymphocytes.

DiSC Assay for Chemosensitivity Testing

3.2. DiSC Assay (see Notes 5 and 6)

5. Have an experienced technician determine the ratio of living (MGG-stained)

DiSC Assay for Chemosensitivity Testing

by reducing the number of analyses, such as by calculating the number of living

DiSC Assay for Chemosensitivity Testing

Fig. 1. Process of CD-DST method.

3.1.2. Dissociation of Tumor Cells

3.1.3. Preculture Using CG Flask

2. Inoculate 5 mL of the suspension into a 25-cm2 CG flask, and preculture for

3.1.4. Collection of Viable Cells

3.2. Collagen Gel Droplet Embedded Culture Drug Sensitivity Test