Matrix Manipulation Affects Attachment and Growth of Breast Cancer Cells in a Bone-like Microenvironment in Vitro

Shelby J. Foster, Donna M. Sosnoski, Andrea M. Mastro

Penn State University, Biochemistry and Molecular Biology, University Park, PA
Experimental Design

Abstract
Breast cancer notoriously metastasizes to the bone, but it is unknown why breast cancer cells colonize there. Breast cancer cells may
become dormant in the skeleton until stimulated to grow. The aim of this project is to better understand the role of the bone matrix in
determining dormancy or growth. Breast cancer cells were co-cultured with MC3T3-E1 mouse osteoblasts that formed a collagenous
matrix. The matrix was manipulated by fixation, decellularization, and by varying the medium estrogen levels as the osteoblasts grew.
MDA-MB-231-BRMS1GFP (BRMS), a metastatic suppressed variant of the human MDA-MB-231 breast cancer cell line, was used
because these cells do not grow in the bone, allowing for observation of growth or dormancy. BRMS were added to a matrix produced
by differentiated MC3T3-E1 cells.
BRMS proliferated more on a culture of osteoblasts and their matrix when fixed with 1% paraformaldehyde than on a live osteoblasts
matrix. Why do the cells proliferate more on the fixed matrix? We ruled out the possibility that live osteoblasts and cancer cells
competed for growth medium. Another possibility was that BRMS attach differently to the matrices with live or fixed osteoblasts. We
compared attachment of BRMS to live, fixed, and decellularized matrices and to tissue culture plastic, and after 60 minutes it was seen
that more BRMS cells attached to a fixed matrix compared to the other conditions. Fluorescent microscopy revealed that BRMS had
many attachment points on a fixed matrix. On the live osteoblasts matrix, the cancer cells remained attached to one another instead of
attaching to the matrix. Experiments will determine if the attachment explains the effect on proliferation by testing for attachment factors
such as -actin and focal adhesion kinase.
Breast cancer is more prevalent in post-menopausal women. Estrogen production decreases with menopause, leading to the
hypothesis that the effects of estrogen deprivation on bone cause metastasis. Osteoblasts were grown in the presence of ICI, an
estrogen inhibitor preventing estrogen from binding. Osteoblasts grown with and without ICI were assayed for alkaline phosphatase
(AP), stained with von Kossa (mineralization), and with ChondrexTM (collagenous and noncollagenous proteins). While AP and
mineralization levels were unaffected by the estrogen inhibition, collagenous proteins production decreased. We will study breast
cancer cell attachment to matrices grown without estrogen. It is hypothesized that proliferation will increase on a matrix with less
collagen, similar to the increased proliferation seen on a fixed matrix.

Introduction
The attachment and growth of metastatic breast cancer cells on a bone matrix are dependent on
the structure of the matrix surface. Initially, it was observed that breast cancer cells proliferated
more on a fixed osteoblast matrix than a live osteoblast matrix. Further experimentation showed
that differences in growth were caused by differences in initial attachment to the matrix. We
hypothesized that breast cancer cells will attach more readily to an altered matrix structure.
Metastasis to bone occurs most often in post-menopausal women when estrogen levels are
lower. Experiments were performed in order to determine the effects of estrogen on the bone
matrix development and structure. While osteoblast differentiation was unchanged in response to
estrogen levels, the structure of the developed matrix varies. Estrogen deprivation changed the
apparent structure of the matrix and decreased its protein content. We hypothesize that estrogen
deprivation during bone matrix production leads to a structure that is conducive to colonization of
breast cancer cells in the bone.
Experimental Design
Cell Lines
MC3T3-E1 murine calvarial osteoblasts
MDA-MB-231GFP human metastatic breast cancer (231)

Experimental Design

Osteoblasts grown in plates for 3-4 weeks in

control, estrogen inhibited, or estrogen
supplemented media

Stain for alkaline phosphatase

or von Kossa

Matrix stained with CNA-35

collagen binding protein
labeled with Alexafluor 568

Osteoblasts grown in chamber slides for 3 weeks

in control, estrogen inhibited, or estrogen
supplemented media

MDA-MB-231 GFP and

BRMS GFP were added to
cell matrix for 6 hours,
matrices washed with PBS,
fixed

Imaged with
fluorescent confocal
microscopy

Bone matrices decellularized using

deoxycholate solution

Results
Light microscopy/
extract dye to
quantify protein

Chondrex
stain

a.

Estrogen inhibition did not affect osteoblast differentiation,

matrix mineralization

a. normal

b. estrogen inhibited

a. normal

Figure 2. Four-week osteoblast matrices stained

for alkaline phosphatase. Osteoblasts were
differentiated in a 24-well plate for 4 weeks in
normal medium (a) or medium with ICI (b) before
fixation and staining for alkaline phosphatase. The
substrate of the alkaline phosphatase enzyme, a
marker of cell differentiation, stains a blue-purple
color. Light microscope images were taken after
allowing the stain to dry. Scale bar = 100 m.

b.

c.

Figure 6. Estrogen levels affected matrix structure. Three week old osteoblast
matrices generated in (a) normal medium, (b) medium with estrogen inhibitor ICI
182,789, and (c) estrogen supplemented medium and stained with CNA35

b. estrogen inhibited

Figure 3. Four-week control osteoblast matrix

stained with von Kossa. Osteoblasts were
differentiated in a 24-well plate for 4 weeks in normal
medium (a) or ICI medium (b) before fixation and
staining with the von Kossa assay for mineralization.
Calcium mineralization is shown by the dark brown
specks. Light microscope images were taken after
allowing the stain to dry. Scale bar = 100 m.

Figure 7. Cancer cell attachment is affected by matrix surface. BRMS and 231 cells were
applied to normal, estrogen inhibited, and estrogen supplemented matrices and allowed to
attach for 6 hours. Adherent cancer cells were then enumerated in 10 random fields. *p<0.05
***p<0.001

MDA-MB-231

MDA-MB-231BRMS1GFP metastasis-suppressed variant of MDA-MB-231 (BRMS)

Conditions
Normal osteoblasts grown in differentiation media (alpha-MEM, 10% FBS, 1% PS, 10 mM Bglycerophosphate, 50 ug/mL ascorbic acid) Estradiol concentration 52 pg/mL
Estrogen inhibited (ICI) differentiation media with 1 mM estrogen receptor inhibitor ICI182,789
Estrogen supplemented (E2) differentiation media with 200 pg/mL estradiol

Estrogen inhibition decreased protein composition of bone matrix

Matrices unaltered (normal), fixed with 4%

paraformaldehyde, or decellularized with
deoxycholate

Osteoblasts grown in plates for 4 weeks

Addition of GFP cancer cells to matrices and

bare plastic control; cells allowed to proliferate or
attach for specified time periods

Attachment of BRMS
2.010 5

b.

c.

d.

e.

f.

MDA-MB-231BRMS
a. normal

Lysis of cancer cells at specified times;

fluorescence of lysates measured in order to
determine cell counts

a.

b. estrogen inhibited

Figure 4. Four-week control osteoblast matrix stained

with Chondrex. Osteoblasts were differentiated in a 24well plate for 4 weeks in normal media (a) or ICI media (b)
before decellularization of the matrix and staining with the
Chondrex kit for collagenous and non-collagenous. After
washing away excess dye, samples were allowed to dry
and were photographed using light microscopy.

Figure 8. Matrix structure affected actin filament arrangement in breast cancer cells. (a) 231
on normal matrix, (b) 231 on estrogen inhibited matrix, (c) 231 on estrogen supplemented matrix,
(d) BRMS on normal matrix, (e) BRMS on estrogen inhibited matrix, (f) BRMS on estrogen
supplemented matrix.

cell count

1.510 5
1.010 5

E2 3 wk

5.010 4

ICI 3 wk

11.27

4.79

plastic

a.

decell

25.78

8.72

normal 3 wk

live

41.86

40.64

fixed

b.

collagenous

c.

non-collagenous
E2 4 wk

Figure 1. Breast cancer cells attach and proliferated more on a fixed osteoblast matrix. (a.) In a 6 hours, more cancer
cells attached to fixed and decellularized matrices in comparison to a live matrix. (b.) MDA-MB-231 BRMS cells proliferated
more on a fixed matrix than on a live matrix. (c.) MDA-MB-231 cells proliferated more on a fixed osteoblast matrix than on a live
osteoblast matrix.
Acknowledgements
This work was supported by the U.S. Army Research and Materiel Command Breast Cancer Program, W81XWH-12-1-0127, Eberly College of
Science, Penn State Biochemistry and Molecular Biology Department, and the Schreyer Honors College
References
Schairer C, Lubin J, Troisi R, Sturgeon S, Brinton L, Hoover R. Menopausal Estrogen and Estrogen-Progestin Replacement Therapy and Breast Cancer Risk.
JAMA. 2000;283(4):485-491. doi:10.1001/jama.283.4.485.
Mastro, Andrea M., and Erwin A. Vogler. "A Three-Dimensional Osteogenic Tissue Model for the Study of Metastatic Tumor Cell Interactions with Bone." Cancer
Research 69 (2009):.
Norgard, Robert J., Donna Sosnoski, and Andrea M. Mastro. 2014 The Effects of Breast Cancer Cells on Collagen Production and Architecture in an in Vitro
Bone-like Matrix. Pennsylvania State University, n.d. Print. Abstract for Undergraduate Research Presentation.

11.27

ICI4 wk

41.86

4.77

normal 4 wk

27.5

10.23
0

40.00
10

20

30

40

50

60

70

protein concentra+on (ug)

Figure 5. Four-week control osteoblast matrix stained

with Chondrex. Bound Chondrex dye was eluted from
matrix sections. The OD of the eluted dye at 540 nm and
605 nm was used to calculate collagenous and noncollagenous proteins respectively.

a.

b.

Figure 9. Matrix structure did not affect proliferation of breast cancer cells. (a) BRMS growth on
estrogen inhibited and supplemented matrices, (b) 231 growth on estrogen inhibited and
supplemented matrices

Conclusions
While estrogen deprivation did not affect osteoblast differentiation or matrix mineralization, matrix
structure and protein content are affected by estrogen levels.
Bone matrix alterations affected breast cancer attachment and proliferation.
Attachment of cancer cells is increased on an estrogen altered matrix.
Future Directions
Stain for markers of attachment and adhesion (focal adhesion kinase, integrins).
Analyze matrix composition, thickness, topography, adhesion, and elasticity via atomic force
microscopy.