Journal of Pathology

J Pathol 2001; 195: 6671.

DOI: 10.1002 / path.921

Review Article

Unlocking the archive gene expression in parafnembedded tissue

F. Lewis1*, N. J. Maughan1, V. Smith2, K. Hillan2 and P. Quirke3
1

Histopathology, The Leeds Teaching Hospitals NHS Trust, Leeds, UK

Genentech Inc., South San Francisco, USA
3
Academic Unit of Pathology, University of Leeds, Leeds, UK
2

* Correspondence to:
F. Lewis, Histopathology, The
Leeds Teaching Hospitals NHS
Trust, Leeds, UK.
E-mail: fraserl@pathology.leeds.
ac.uk

Abstract
The histopathology archive represents a vast, well-characterized source of specimens covering
virtually every disease and is available for molecular biological investigation. The archive has in
recent years become widely used for molecular genetic analysis and DNA can be routinely
extracted from formalin-xed, parafn-embedded tissue. More recently, archival specimens have
become a source of material for extensive analysis of mRNA expression utilizing DNA
microarrays, real-time quantitative reverse transcriptase polymerase chain reaction (PCR), and
in situ hybridization and amplication techniques. These techniques will enable a greater
understanding of the changes that occur in gene function during every stage of the development of
disease and will lead to better diagnosis, better evaluation of prognosis, and better treatment
through targeted therapeutic regimes. Copyright # 2001 John Wiley & Sons, Ltd.
Keywords: archive; formalin-xed; parafn-embedded tissue; RNA extraction; reverse
transcriptase PCR; real-time quantitative PCR; in situ hybridisation; in situ amplication;
expression arrays

Introduction

Extraction of RNA from archival specimens

The importance of the histopathology archive cannot

be underestimated. Tissue specimens have been preserved as parafn blocks for over a century and this
archive represents a historical collection of virtually
every disease. With appropriate consent, these specimens are available for investigation as new technologies develop for their elucidation and diagnosis. This
archive has already proved invaluable for the development of many immunohistochemical assays now
utilized in routine diagnostic procedures.
Over recent years the archive has proved an
invaluable source of DNA for molecular genetic
analysis and the identication of infectious agents.
Extraction of DNA from formalin-xed, parafnembedded tissue is well documented and now a routine
diagnostic process [1].
More recently, it has become desirable to utilize
archival specimens as a source of RNA in order to
investigate the changes in expression of mRNA that
occur during a disease process. There is a requirement
to analyse expressed mRNA using a number of
emerging techniques including the identication of
up- and down-regulated genes using expression
arrays, quantitative analysis by real-time quantitative
reverse transcriptase polymerase chain reaction, and
morphological localization using in situ technologies.
Most of these procedures require a reliable recovery of
RNA from the preserved tissue and new extraction
methodologies have become available to accomplish
this.

Extraction of RNA from parafn sections has proved

to be problematic. Over recent years, researchers have
attempted to overcome many of these problems and
ampliable RNA can now be extracted successfully
from archival specimens.
RNA is readily extracted from fresh clinical specimens using a guanidinium thiocyanatecaesium
chloride gradient [2] or a combined guanidinium
thiocyanate acidphenol chloroform procedure [3].
These methods produce high yields of high quality
total RNA (Figure 1) that can be utilized for further
investigation. The methods can be time-consuming, so
many simplied methods for RNA extraction have
become available from numerous commercial suppliers
to speed up the process without loss of quality of the
extracted RNA. Methods such as RNeasy (Qiagen),
Purescript (Gentra), Trizol (Life Technologies), and
Totally RNA (Ambion) can be adapted for the
extraction of RNA from most fresh clinical specimens
to yield reasonable quantities of RNA at a quality
suitable for reverse transcriptase PCR amplication
procedures.
The application of these methods to formalin-xed,
parafn-embedded tissue, however, generally results in
failure to extract RNA in sufcient quantity to enable
further investigation. Some success has been achieved
by utilizing a guanidinium thiocyanateacid phenol
chloroform method for the extraction of hepatitis
C RNA from stained sections of liver tissue [4] and
the commercial version of this method (Trizol) has

Copyright # 2001 John Wiley & Sons, Ltd.

Unlocking the archive gene expression in parafn-embedded tissue

Figure 1. Example of total RNA extracts electrophoresed on a

2% agarose gel showing the quality of RNA from a fresh normal
and a tumour specimen (lanes 3 and 4) and from two formalinxed, parafn-embedded normal and tumour specimens (lanes 6
and 7, and 9 and 10). Lane 1 shows a 250 bp size standard

similarly been successfully employed for the extraction

of this RNA virus [5]. However, in order to improve
the yield of RNA from formalin-xed, parafnembedded tissue a number of procedures have been
proposed.
A method employing a thermal cycler and Chelex100 extraction has been shown to be successful in
extracting RNA from around 84% of sections tested [6]
and the use of a procedure utilizing sonication and
oligo(dT)25 paramagnetic beads also resulted in
successful RT-PCR [7]. Another method relies on the
binding of RNA to acid-treated glass beads in the
presence of high-molarity guanidinium salt [8]. This
method takes just 1 h to complete and tissue sections
do not need to be dewaxed prior to extraction. One
drawback of this approach is that whilst it is highly
successful for acetone-xed specimens, a lower success
rate is observed from formalin-xed specimens, a
xation procedure that will have been applied to the
majority of specimens in a histology archive.
By far the most successful method for the recovery
of total RNA from formalin-xed, parafn-embedded
tissue utilizes a proteinase K digestion prior to acidphenol chloroform extraction and carrier precipitation
[1,9,10]. This method is also employed in a recently
introduced commercial kit the Parafn Block RNA
Isolation Kit (Ambion). Whereas chatropic agents,
such as guanidinium hydrochloride, fail to solubilize
dewaxed sections from formalin-xed, parafnembedded tissue, these same sections are completely
solubilized in proteinase K [9]. This enzyme readily
destroys proteins and this property appears to be little
affected by the highly cross-linked nature of the
Copyright # 2001 John Wiley & Sons, Ltd.

67

proteins following formalin xation. The result of

incubation of the sections in proteinase K is the release
of RNA from the cross-linked matrix, enabling its
purication by acid-phenol chloroform extraction. The
low yields of RNA obtained result in a very dilute
aqueous solution that requires the addition of a
precipitant carrier such as glycogen or linearized
acrylamide to ensure its precipitation.
The effect of formalin xation on the structure of
RNA has indicated that all four bases show the
addition of mono-methylol [CH(2)OH] groups at
various rates with the additional dimerization of
adenine groups by methylene bridging [9]. The majority of the methylol groups can be removed from the
bases by simply elevating the temperature in a buffer
solution and the successful isolation of total RNA
following extended proteinase K digestion for up to 5
days [1] may be attributable to the continuing removal
of these groups over the time period to restore the
template activity of the RNA.
There is no doubt, however, that the RNA extracted
from formalin-xed, parafn-embedded tissue is
signicantly degraded (Figure 1) [10,11] and that
attempted amplication of long fragments should be
avoided. Attempts to amplify fragments longer than
200 bp are usually unsuccessful and in our experience,
amplication of fragments in a range 60120 bp
generally results in a very high success rate approaching 100%.
Another problem that affects the ability to produce
an ampliable cDNA from RNA from archival tissue
is the effective poisoning of RNA extracts by the use of
RNase destroyers and inhibitors during the extraction
process. The fear that the RNA in a formalin-xed
tissue will be destroyed by RNases during extraction is
mainly unfounded. Although RNases are very stable
and highly active and do not require cofactors to
function, endogenous RNases will have been totally
inactivated by cross-linking during the formalin xation process. Extraction with proteinase K ensures that
this cross-linked enzyme is totally destroyed, thus
avoiding any potential reactivation during reversal of
the xation in aqueous buffers. Therefore, the only
potential source of RNase is from glassware and
buffers used in and human contact during the extraction procedure. Provided that high quality reagents are
used, buffers are made with RNase-free water, and all
glassware is kept meticulously clean, this ubiquitous
source of RNase is not sufcient to cause a problem.
The tendency, however, to treat all glassware and
reagents with diethyl pyrocarbonate (DEPC), as a
precaution against potential RNase activity, is likely to
have a detrimental rather than a helpful effect on the
ability to amplify the extracted RNA. This is because
just trace amounts of DEPC will modify purine
residues in RNA by carboxymethylation and this
carboxymethylated RNA will translate with extremely
low efciency, producing very low yields of cDNA.
Success or failure to extract RNA from archival
tissue also depends on the xation procedure used to
J Pathol 2001; 195: 6671.

F. Lewis et al.

68

x the specimens. The chemical nature of the xatives

and the xation time can have a signicant effect on
the ability to extract macromolecules from processed
tissues [12,13]. Successful extraction can be accomplished from tissues xed in precipitating xatives such
as acetone, Clarkes, and Carnoys, and from crosslinking xatives such as formalin, neutral buffered
formalin, and paraformaldehyde. However, neither
DNA nor RNA can be successfully extracted from
highly cross-linking xatives such as glutaraldehyde,
modied formalins containing mercuric chloride, and
Bouins xative. It is probable that the histological
specimens stored in the archive will have been xed in
formalin although the length of time of xation will
not be known. It is likely that those specimens that
occasionally fail to yield transcribable RNA will have
been xed for longer than the acceptable time that
ensures successful extraction of total RNA.
Even though successful extraction of total RNA
from archival specimens results in low yields of a
highly degraded product, it can be utilized as a
template for a large number of applications. Careful
extraction of total RNA from just 10r5 mm sections
of formalin-xed, parafn-embedded tissue can be
reverse-transcribed to yield sufcient cDNA to perform around 100 PCR assays [14].

Utilization of total RNA extracted from

formalin-xed, parafn-embedded tissue
Reverse transcriptase PCR
Reverse transcriptase PCR assays should be designed
with the knowledge that the total RNA extracted from
formalin-xed, parafn-embedded tissue is signicantly
degraded (Figure 1) [10,11]. Provided that small
regions of the target mRNA are selected, then specic
amplicons will be produced. As a rule, the smaller the
amplicon size, the more chance of success, and
amplication of fragments below 200 bp should be
achievable.
The degradation of the RNA can, however, have an
inuence on the ability to perform the reverse
transcription step. Success at this stage is essential
and failure of the PCR is usually due to failure to
produce sufcient or any cDNA. Traditionally, the
reverse transcription step is primed with a polyT
oligomer. It is likely, however, that in its degraded
state, the mRNA in the total RNA extract will have
lost some or all of its 3k polyA tails. Loss of this
priming site for polyT is the main cause of failure of
the reverse transcription step [14]. It is recommended,
therefore, that the reverse transcription step is primed
with random hexamers, or with the specic antisense
primer that will be used latterly in the PCR. Avoiding
inhibitory contaminants such as DEPC is also important to ensure successful reverse transcription and
successful amplication of cDNA.
Standard reverse transcription PCR performed on
archival specimens has found most application for the
Copyright # 2001 John Wiley & Sons, Ltd.

identication of RNA virus (Figure 2) or expressed

gene fusion products associated with chromosomal
translocations. Whilst it is possible to identify
expressed genes in these specimens, no quantitative
data can be elucidated using this method. In order to
identify changes in gene expression, newer emerging
technologies such as expression arrays and reverse
transcriptase real-time quantitative PCR have to be
utilized.

Expression arrays
DNA expression arrays can be used to identify
qualitatively up- and down-regulation of genes at
certain stages of a disease. Most array technology
requires the use of large quantities of high quality
mRNA which has to be recovered from fresh clinical
specimens. Where prospective collection of specimens
is problematic due to the rarity of the disease, it has
been proposed that RNA extracted from archival
specimens could be utilized on custom expression
arrays. Despite the very low molecular weight of the
RNA typically isolated from parafn-embedded material, cDNA probes can be generated using sensitive
labelling strategies, without requiring any additional
nucleic acid amplication. High-quality data can be
obtained on hybridization to a microarray: preliminary
studies using tumour samples suggest that up to 80%
of genes can be quantitatively detected in a parafnembedded sample, compared with a fresh-frozen
sample from the same patient (Genentech, unpublished
data). Factors such as sample collection and the degree
of xation will probably affect the extent of applicability of microarray analysis to parafn-embedded
material, but the potential is exciting.

Validation of the results from expression arrays

The vast amount of data collated from expression
arrays gives an indication of the changes that are
occurring in the disease under investigation. The data
from the arrays are neither absolute nor quantitative,

Figure 2. Reverse transcriptase PCR amplication of measles

virus from total RNA extracted from formalin-xed, parafnembedded tissue (lanes 1 and 2), negative extraction control
(lane 3), negative PCR controls (lanes 5 and 6), and positive
control (lane 8)
J Pathol 2001; 195: 6671.

Unlocking the archive gene expression in parafn-embedded tissue

but merely a snapshot of the changes that may be

taking place. Thus, the data have to be validated by
other means in order to establish the signicance of the
differences occurring between the normal and disease
condition under investigation. The application of realtime quantitative PCR to conrm or refute the changes
identied in a large series of cases would validate the
results obtained from an expression array. A method
of in situ detection of the expression that would
morphologically identify the cells that are subject to
the changes taking place is also necessary. To achieve
validation of the results from the arrays using
prospectively collected fresh specimens would take a
considerable length of time, especially if the disease
under investigation is relatively rare. Therefore validation of the results is best performed on archival
specimens.

Sequence detection methods real-time

quantitative reverse transcriptase PCR
Real-time quantitative PCR methods utilising TaqMan
probes [15] or molecular beacons [16] have become
standard methods for the measurement of expression
levels in cells. These assays involve the use of
uorogenic probes, which release a uorescent reporter
during every cycle of PCR, enabling the accumulation
of uorescence to be directly related to the quantity of
RNA present in the original extract. For example, the
TaqMan assay is a PCR method that exploits the
5k nuclease activity of Taq DNA polymerase to cleave
a probe during amplication. The system utilizes a
uorogenic probe the TaqMan probe which is
labelled with a reporter dye at the 5k end and a
quencher dye at the 3k end. When the probe is intact,
the proximity of the reporter dye to the quencher dye
results in suppression of the reporter uorescence
primarily by Forster-type energy transfer (FRET)
[17,18]. During PCR, in the presence of the TaqMan
probe, if the target of interest is present, the probe
specically anneals between the forward and reverse
primer sites. The 5k-3k nucleolytic activity of the Taq
polymerase cleaves the probe between the reporter and
quencher only if this annealing to the target has
occurred. The cleavage of the probe separates the
reporter and quencher dyes, resulting in an increase in
the uorescence of the reporter dye which can be
detected by monitoring in a sequence detection instrument such as the Applied Biosystems 7700 sequence
detection system. As the probe fragments are displaced
from the target, the polymerization of the strand
continues.
The process occurs during every cycle and does not
interfere with the exponential accumulation of product.
The accumulation of product is directly proportional
to the accumulation of the uorescence being detected
and this increase in uorescence is only detected if the
target sequence is complementary to the probe and
is amplied during PCR. Thus, any non-specic
Copyright # 2001 John Wiley & Sons, Ltd.

69

amplication is not detected [15,19,20] Hence highly

specic detection and quantitation of the target
sequence occurs.
A major advantage of this technology is its ability to
utilize very small target sizes for amplication
(50150 bp) and it is thus well suited to amplify
cDNAs transcribed from the highly degraded total
RNA extracted from archival specimens. As degradation of the RNA is universal and as the quantication
of a target mRNA is measured relative to that of a
standard housekeeping mRNA, then accuracy and
specicity should be maintained despite the degraded
nature of the RNA extract.
Another advantage of the assay is its ability to
distinguish between RNA and contaminating DNA in
the extract. This is achieved by designing the primers
so that they span an exonexon boundary, ensuring
that the spanned intron is large enough to fail to
amplify during the PCR process. However, where large
introns are not present in the gene sequence, or where
an associated pseudogene has been identied, care has
to be exercised to validate the results. Pseudogenes are
a particular problem in real-time quantitative PCR
assays. They are found throughout the human genome
and have nucleotide sequences that have similarities to
functional genes but are themselves non-functional.
They are commonly associated with the genes used
extensively as housekeeping genes, e.g. b-actin,
GAPDH, and b-globin, in real-time quantitative PCR
assays. Extensive DNase digestion may have to be
performed on total RNA extracts to ensure that
amplication of contaminating DNA does not occur.
Recently, a reliable extraction method for total RNA
from formalin-xed, parafn-embedded tissue has been
utilized which yields extracts free from contaminating
DNA (Figure 3) [14].

In situ hybridisation and in situ

amplication
In order to understand, morphologically, the changes
taking place in cells undergoing changes in expression,
the mRNA has to be visualized microscopically in nondisrupted cells and tissue sections. This can be
accomplished by utilising an in situ labelling technique.
The technique of in situ hybridization on parafn
sections is well documented [21,22], but there is a
consensus that radiolabelled probes have to be utilized
for the detection of rare mRNA as hapten labels and
reporters visualized by light microscopy lack the
necessary sensitivity. This has practical implications
for most histopathology laboratories where the use of
radioactive materials is not possible. Increasing the
sensitivity of detection of mRNA using hapten labels
can be accomplished using in situ reverse transcriptase
PCR amplication of the target under investigation
and a number of methods have been proposed. This
method, however, is difcult to perform reliably and
reproducibly and remains exclusive to the laboratories
J Pathol 2001; 195: 6671.

F. Lewis et al.

70

Figure 3. DNA contamination of total RNA extracts from archival specimens was identied by performing PCR directly on RNA
extracts without a reverse transcriptase step. No amplication of thymidylate synthase was observed in 45 total RNA extracts.
Three out of 45 of the same RNA samples showed amplication of GAPDH (circles) as a result of amplication of the GAPDH
pseudogene, indicating that 3/45 samples were contaminated with DNA. The squares show amplication from cDNA for TS and
GAPDH

that have developed the techniques. As there is a

tendency to overamplify the target, interpretation of its
localization is sometimes difcult to achieve. Using
standard in situ hybridization techniques, a system of
signal amplication has been applied to increase the
sensitivity of detection of hapten labels and this has
been used successfully to localize rare mRNA in nondisrupted cells and tissues [23,24].
More recently, there has been a resurgence in the use
of dinitrophenol as a hapten of choice. This hapten has
been shown to be more sensitive than digoxigenin,
biotin, and uorescein, and is less prone to problem
background staining [25]. A recently introduced tyramine signal amplication for dinitrophenol (TSA
Plus, NEN Life Sciences) increases the sensitivity of
detection to levels expected with radiolabels, so
application of non-radioisotopic in situ hybridization
for the localization of rare mRNA in parafn sections
should now be possible as a routine procedure.

Conclusion
The prospective collection of specimens can take a
considerable length of time, particularly for rarer
diseases. The archive can provide a valuable source of
material from which expressed genes can be investigated at virtually every stage of development of a
disease. As better techniques evolve for the extraction,
quantitation, and localization of RNA from formalinxed, parafn-embedded tissue, the value of the
archive will grow even stronger. The ultimate goal is
to develop a greater understanding of the changes that
occur in gene function during the progression of a
disease. The knowledge acquired will enable a better
diagnosis and a better evaluation of the prognosis
tailored to the disease occurring within an individual
Copyright # 2001 John Wiley & Sons, Ltd.

and will lead to better treatment through targeted

therapeutic regimes.

References
1. Jackson DP, Lewis FA, Taylor GR, Boylston AW, Quirke P.
Tissue extraction of DNA and RNA and analysis by the
polymerase chain reaction. J Clin Pathol 1990; 43: 499504.
2. Mehra M. RNA isolation from cells and tissue. In Laboratory
Guide to RNA: Isolation, Analysis and Synthesis, Krieg PAA
(ed.). Wiley-Liss, New York, 1996; 120.
3. Chomczynski P, Sacchi N. Single-step method of RNA isolation
by acid guanidinium thiocyanatephenolchloroform extraction.
Anal Biochem 1987; 162: 156159.
4. Saito K. Morphology and molecular pathology: detection of
hepatitis C virus RNA sequences in stained sections by
microscopy-directed selective extraction. Rinsho Byori 1998; 46:
4955.
5. Greenson JK, Svoboda-Newman SM, Merion RM, Frank TS.
The histologic progression of recurrent hepatitis C in liver
transplant allografts. Am J Sur Pathol 1996; 20: 731738.
6. Coombs NJ, Gough AC, Primrose JN. Optimisation of DNA
and RNA extraction from archival formalin-xed tissue. Nucleic
Acids Res 1999; 27: e12.
7. Houze TA, Gustavsson B. Sonication as a means of enhancing
the detection of gene expression levels from formalin-xed,
parafn-embedded biopsies. Biotechniques 1996; 21: 10741078,
1080, 1082.
8. Koopmans M, Monroe SS, Cofeld LM, Zaki SR. Optimization
of extraction and PCR amplication of RNA extracts from
parafn-embedded tissue in different xatives. J Virol Methods
1993; 43: 189204.
9. Masuda N, Ohnishi T, Kawamoto S, Monden M, Okubo K.
Analysis of chemical modication of RNA from formalin-xed
samples and optimization of molecular biology applications for
such samples. Nucleic Acids Res 1999; 27: 44364443.
10. Krafft AE, Duncan BW, Bijwaard KE, Taubenberger JK, Lichy
JH. Optimization of the isolation and amplication of RNA
from formalin-xed, parafn-embedded tissue: the Armed Forces
Institute of Pathology Experience and Literature Review. Mol
Diagn 1997; 2: 217230.
11. Stanta G, Schneider C. RNA extracted from parafn-embedded
J Pathol 2001; 195: 6671.

Unlocking the archive gene expression in parafn-embedded tissue

12.

13.

14.

15.

16.
17.
18.
19.

human tissues is amenable to analysis by PCR amplication.

Biotechniques 1991; 11: 304, 306, 308.
OLeary JJ, Browne G, Landers RJ, et al. The importance of
xation procedures on DNA template and its suitability for
solution-phase polymerase chain reaction and PCR in situ
hybridization. Histochem J 1994; 26: 337346.
Goldsworthy SM, Stockton PS, Trempus CS, Foley JF,
Maronpot RR. Effects of xation on RNA extraction and
amplication from laser capture microdissected tissue. Mol
Carcinog 1999; 25: 8691.
Lewis FA, Rea N, Bell SM, Cross D, Maughan N, Quirke P.
Extraction of total RNA from formalin xed parafn embedded
tissue for application in real time quantitative PCR methods. In
preparation.
Livak KJ, Flood SAJ, Marmaro J, Giusti W, Deetz K.
Oligonucleotides with uorescent dyes at opposite ends provide
a quenched probe system useful for detecting PCR product and
nucleic acid hybridisation. PCR Methods Appl 1995; 4: 357362.
Tyagi S, Kramer FR. Molecular beacons: probes that uoresce
upon hybridization. Nat Biotechnol 1996; 14: 303308.
Forster VT. Zwischenmolekulare Energiewanderung Fluoreszenz. Ann Phys (Leipzig) 1948; 2: 5575.
Lakowicz JR. Energy transfer. In Principles of Fluorescent
Spectroscopy. Plenum Press: New York, 1983; 303339.
Holland PM, Abramson RD, Watson R, Gelfand DH. Detection
of specic polymerase chain reaction products by utilizing the

Copyright # 2001 John Wiley & Sons, Ltd.

20.

21.

22.

23.

24.

25.

71

5k-3k exonuclease activity of Thermus aquaticus DNA polymerase.

Proc Natl Acad Sci U S A 1991; 88: 72767280.
Holland PM, Abramson RD, Watson R, Wili S, Saiki RH,
Gelfand DH. Detection of specic polymerase chain reaction
product by utilising the 5k-3k exonuclease activity of Thermus
Aquaticus DNA polymerase. Clin Chem 1991; 38: 462463.
Poulsom R, Longcroft JM, Jeffery RE, Rogers LA, Steel JH. A
robust method for isotopic riboprobe in situ hybridisation to
localise mRNAs in routine pathology specimens. Eur J Histochem
1998; 42: 121132.
Wilson KH, Schambra UB, Smith MS, et al. In situ hybridization: identication of rare mRNAs in human tissues. Brain Res
Brain Res Protoc 1997; 1: 175185.
Speel EJ, Saremaslani P, Roth J, Hopman AH, Komminoth P.
Improved mRNA in situ hybridization on formaldehyde-xed
and parafn-embedded tissue using signal amplication with
different haptenized tyramides. Histochem Cell Biol 1998; 11:
571577.
Yang H, Wanner IB, Roper SD, Chaudhari N. An optimized
method for in situ hybridization with signal amplication that
allows the detection of rare mRNAs. J Histochem Cytochem
1999; 47: 431446.
Harper SJ, Bailey E, McKeen CM, et al. A comparative study of
digoxigenin, 2,4-dinitrophenol and alkaline phosphatase as
deoxyoligonucleotide labels in non-radioactive in situ hybridisation. J Clin Pathol 1997: 50; 680690.

J Pathol 2001; 195: 6671.