ANATOMIC PATHOLOGY
Original Article
PCR Amplification from
Paraffin-Embedded Tissues
Effects of Fixative and Fixation Time
CATHERINE E. GREER, B.S., SANDRA L. PETERSON, H.T., NANCY B. KIVIAT, M.D.,
AND M. MICHELE MANOS, PH.D.
The polymerase chain reaction (PCR) DNA amplification method
is a powerful new tool for the retrospective analysis of paraffinembedded tissue (PET). The technique has afforded the sensitive
and specific detection of nucleic acid sequences associated with
genetic and infectious diseases. However, PET processing conditions vary in their suitability for amplification. The authors
have examined the effects of 11 fixatives at threefixationtimes.
The effect of fixation was measured by the ability of the DNA
in a treated tissue to serve as a template for the amplification
of DNA fragments that ranged from 110 to 1,327 base pairs in
The polymerase chain reaction (PCR) DNA amplification
method' 2 has been used in a number of disciplines to
retrospectively analyze paraffin-embedded tissue (PET).
Geneticists have begun to use archival clinical specimens
to investigate heritable diseases. Cancer researchers have
analyzed the relationship between oncogene mutations
and tumorigenesis,3'4and have identified subsequent metastases of primary tumors. 5 Furthermore, the technique
has allowed pathologists to better correlate histopathologic changes in tissues with infectious agents such as viruses.6"10
However, just as researchers have had variable success using PET DNA for Southern"" 13 blot and in situ
hybridization 1415 studies, so have researchers using PET
for PCR amplification analyses. 31617 Investigators using
Southern blot analysis of PET identified the fixation process as the primary source of the reduction in size of extractable DNA. 3 ' 3 1 6 Goelz and associates13 reported that
25% of the samples studied were not useful for Southern
length. Specimens fixed in acetone or 10% buffered neutral formalin were found to be best suited for subsequent analysis by
PCR. A second group of fixatives, including Zamboni's, Clarke's,
paraformaldehyde, formalin-alcohol-acetic acid, and methacarn,
compromised amplification efficiency. Tissues treated with Carnoy's, Zenker's, or Bouin's, respectively, were even less desirable
for amplification analysis. (Key words: PCR; Amplification; Fixation; Paraffin-embedded tissue) Am J Clin Pathol 1991;95:117124
blot analysis because of background problems caused by
degraded DNA. The DNA extracted from fixed tissues
has been reported to range in size from 10 to 24,000 base
pair (bp). In two studies, the size of DNA fragments from
tissues fixed in 10% buffered neutral formalin (BNF)
ranged from 100 to 1,500 bp 13 and 4,000-9,000 bp. 16 In
some cases, DNA that is too degraded for Southern or
other hybridization analyses may be suitable for PCRbased methods. In studies of ancient tissue specimens,
Paabo and associates18 examined the effect of DNA damage on amplification and found that, in general, amplification fragments larger than 200 bp could not be efficiently generated.
It is of value to researchers to know which tissue processing conditions are most conducive to subsequent PCR.
This knowledge can be incorporated into protocols for
prospective studies or for identification of appropriate
samples for retrospective analysis. Although our interest
is mainly in retrospective medical studies, determining
the best methods for preparing and storing specimens for
DNA analysis is an urgent need for scientists in other
disciplines as well. Zoologists, evolutionary biologists,
From the Department of Infectious Diseases. Cetus Corporation,
museum
Emeryville, California, and the Department of Pathology. University
of curators, and scientists interested in biologic diversity require fixation methods that will preserve invaluWashington, Seattle, Washington.
able specimens for future study.19
Received April 17, 1990; accepted for publication May 28, 1990.
In this study, we examined the effects of a broad specAddress reprint requests to Dr. Manos: Department of Infectious Distrum of fixatives and three fixation times. The effect of
eases, Cetus Corporation, Emeryville, California 94608.
117
118
ANATOMIC PATHOLOGY
Article
fixation was measured by the ability of a treated tissue to
serve as template for the amplification of DNA fragments
that ranged from 110 to 1,327 bp in length.
(72% ethanol, 10% [v/v] formalin, 5% [v/v] acetic acid,
pH 2.9). Treatment after fixation was a one-hour wash
in 70% (v/v) alcohol.
Preparation of Samples for PCR
MATERIALS AND METHODS
Fixatives and Tissue Processing
Two fresh cervical samples without signs of autolysis,
taken from separate hysterectomies, were cut into sections
measuring approximately 3 X 2 X 5 mm.
Eleven fixation methods (detailed below) were evaluated. For ease of processing they were randomly divided
into two groups. Each group included 10% (volume/volume [v/v]) BNF as a standard. All fixatives were freshly
prepared. Tissue sections were processed in each fixative
for 1, 4, and 24 hours and, depending on fixative type,
received additional processing. After fixation, samples
were stored in 70% (v/v) isopropyl alcohol until they were
processed for paraffin embedding. Preparation for paraffin
embedding included one 80% (v/v), two 95% (v/v), and
three 100% isopropyl alcohol washes, followed by two xylene and paraffin changes. Samples were embedded with
an Autotechnicon* (Technicon, Tarrytown, NY) paraffin
embedder. Fixation methods were as follows:
Set 1 (cervical sample 1): (a) 10% BNF (10% [v/v] formalin, 75 mmol/L sodium phosphate, pH 6.7); (b) alcoholic formalin (10% [v/v] formalin, 85.5% [v/v]
ethanol); (c) acetone (100% acetone); (d) Zenker's fixative (5% [v/v] acetic acid, 175 mmol/L mercuric chloride, 81 mmol/L potassium dichromate, and 31 mmol/
L sodium sulfate, pH 2.3). Treatment after fixation was
a 12-hour running-water wash: (e) Carnoy's fixative
(60% ethanol [v/v], 30% [v/v] chloroform, 10% [v/v]
acetic acid, pH 3.5). Treatment after fixation was a onehour wash in 70% ethanol: (f) methacarn (60% [v/v]
methanol, 30% [v/v] chloroform, 10% [v/v] acetic acid,
pH 3.0). Treatment after fixation was a one-hour wash
in 70% (v/v) ethanol.
Set 2 (cervical sample 2): (g) 10% BNF; (h) Bouin's fluid
(71% [v/v] saturated picric acid, 24% (v/v) formalin,
5% (v/v) acetic acid, pH 1.39). Treatment after fixation
consisted of three two-hour 50% (v/v) alcohol washes:
(i) Zamboni's fixative (2% [v/v] paraformaldehyde, 15%
[v/v] saturated picric acid, 24 mmol/L sodium phosphate, pH 4.0). Treatment after fixation consisted of
three two-hour 50% (v/v) alcohol washes: (j) paraformaldehyde (4% [w/v] paraformaldehyde, 66 mmol/L
S-collidine, 5 mmol/L CaCl2, pH 7.4); (k) Clarke's fixative (75% [v/v] ethanol, 25% [v/v] acetic acid, pH
1.76). Treatment after fixation was a one-hour wash in
70% (v/v) ethanol; (1) formalin-alcohol-acetic acid
A.J.C.P. •
Dry sections, 5 nm thick, were sliced from each sample
and placed in a 1.5-mL microfuge tube with sterile toothpicks. The microtome and blade were carefully cleaned
with xylene between each block to prevent sample-tosample contamination. We followed published protocols
to prevent sample contamination and PCR "carry over." 172021
Deparaffinization of sections was performed with the
procedure of Shibata and colleagues10 as modified by
Wright and Manos.17 Briefly, each section was extracted
twice with octane, followed by two 100% (v/v) ethanol
rinses. (Although octane has been routinely used in our
laboratory for deparaffinization, commercially available
solvents such as Americlear® [Baxter Scientific] or HemoDe* [Fisher Scientific] may be substituted.) Tissues were
pelleted between each extraction. Two drops of HPLCgrade acetone were added, and the tissue was dried in a
55 °C heat block.
Tissues were resuspended in 0.2 mL digestion buffer
(200 Mg/mL proteinase K in 50 mmol/L TRIS-HC1 [pH
8.5], 1 mmol/L EDTA, and 1% [w/w] Laureth-12* [Mazer
Chemicals, Gurnee, IL]). Samples were incubated for three
hours at 55 °C. The proteinase K was heat inactivated at
95 °C for 10 minutes.
To determine the size of the DNA, aliquots of prepared
samples were electrophoresed on 1.5% (w/v) agarose gels,
stained with ethidium bromide, and photographed under
ultraviolet (UV) light. Average sizes were determined after
the photographic negative image was captured by video
camera (model CCD°; Polaroid Corporation, Cambridge,
MA) and the data digitized with the use of the JAVA2*
(Jandel Scientific, Sausalito, CA) computer software.
Graphs were generated that plotted the intensity of the
ethidium bromide-stained sample DNA versus migration
(relative to molecular weight standards). Values for both
the full range of DNA fragment size and the range of most
of the fragments were obtained for each sample.
Phenol/Chloroform
Extraction
In experiments to evaluate the effects of additional purification, a phenol/chloroform extraction followed by a
chloroform extraction and 100% (v/v) ethanol, 2 mol/L
ammonium acetate precipitation was performed on sections of the 24-hour-fixed tissues of selected fixatives. The
precipitate was suspended in 10 mmol/L Tris-HCl, pH
8.0; 1 mmol/L EDTA (TE).
1991
119
GREER ET AL.
PCR Amplification from raffin-Embedded Tissues
Primer Selection
Five primer pairs for the human /3-globin gene (described in Table 1) provided amplification products of
110, 268, 536, 989, and 1,327 bp.
Polymerase Chain Reaction
Aliquots of 1 and 10 ^L from prepared sections of each
fixative time point were amplified in duplicate with each
of the five primer pairs, separately. Each reaction contained (in a total of 100 /nL) 1 or 10 nL of sample DNA;
100 nmol/L of each primer; 200 /imol/L of each dNTP
(dATP, dGTP, dCTP, dTTP); 2.5 units of Taq DNA
polymerase (Amplitaq®; Perkin-Elmer Cetus, Norwalk,
CT); 50 mmol/L KC1, 4 mmol/L MgCl2, and 10 mmol/
L TRIS-HC1 (pH 8.5). All reaction tubes included a 100nL mineral oil overlay. Cycling parameters were 1 minute
at 95 °C, 1 minute at 55 °C, and 2 minutes at 72 °C for
40 cycles, followed by an additional 5 minutes at 72 °C,
in a DNA Thermal Cycler* (Perkin Elmer Cetus). Purified
DNA from the K562 cell line (ATCC no. CCL243) was
used as a human DNA control, and sterile, distilled water
was used as a negative control.
PCR Inhibition Studies
Previously described methods 22 were used for amplifications in the human papillomavirus (HPV) recovery experiment. Briefly, the HPV LI consensus primers MY11
and MY09 (500 nmol/L each) were used to amplify a
fragment of approximately 450 bp. The buffer and
amounts of dNTPs and Amplitaq® were identical to those
in the /3-globin amplifications. A reaction master mix was
made and split, with half of the mix receiving approxi-
Primers
Primer Pair
PC03/PC04
GH20/PC04
RS42/K.M29
RS80/RS40
K.M29/RS80
Analysis of Amplified DNA
Aliquots representing 1/20 of each amplification were
resolved on 7% (w/v) polyacrylamide gels, stained with
ethidium bromide, and photographed under UV light.
RESULTS
Preparation of Paraffin-Embedded
Tissues
To focus on the effects of different fixation protocols,
several variables routinely encountered in clinical settings
were eliminated. Factors such as the type and amount of
tissue and the time between tissue sampling and processing
were standardized to ensure that the amount and integrity
of prefixation DNA were consistent for all fixation conditions. Consequently, fresh tissue was obtained and sectioned into approximately equal pieces that were processed
in each of 11 different fixatives for 1, 4, and 24 hours.
The tissues were then prepared for paraffin embedding
and the resulting blocks were sliced, yielding 5-fim dry
sections. The range in amount of tissue per slice was 0.040.13 mm 3 , with a mean of 0.1 mm 3 . Sections from each
fixative time point were prepared by deparaffinization and
protease digestion as described.17
PCR Strategy
We suspected that different fixation protocols were the
primary cause of variable success in amplifying DNA sequences from fixed tissues. To measure this effect, we used
treated sample DNA as a template for the amplification
of five different DNA fragment lengths that ranged from
110 to 1,327 bp. We expected that the less damaging a
fixation method was to sample DNA, the longer the resulting amplification products could be.
The oligonucleotide primers used to amplify the five
distinct fragments are clustered in the human /3-globin
gene and were chosen specifically for their high efficiency
in amplifications (Table 1). Optimal conditions, including
primer and MgCl2 concentrations, were determined for
each primer pair.
TABLE I. SEQUENCES OF H U M A N 0-GLOBIN
AMPLIFICATION PRIMERS
PC03
PC04
GH20
KM29
RS40
RS42
RS80
mately 0.75 ng (per reaction) of purified DNA from the
HeLa cell line (ATCC no. CCL2), which contains 10-50
integrated copies of HPV 18 per cell. Both mixes were
run in duplicate with aliquots of 10 /xL from the prepared
samples of the 24-hour time point of the following fixatives: 10% (v/v) BNF, Zenker's, Carnoy's, and Bouin's.
5'ACACAACTGTGTTCACTAGC 3'
CAACTTCATCCACGTTCACC
GAAGAGCCAAGGACAGGTAC
GGTTGGCCAATCTACTCCCAGG
ATTTTCCCACCCTTAGGCTG
GCTCACTCAGTGTGGCAAAG
TGGTAGCTGGATTGTAGCTG
Predicted Product Size
110 bp
268 bp
536 bp
989 bp
1,327 bp
Effect of Fixation on PCR
The predicted DNA fragment sizes for each of the five primer pairs used in this study also are
listed.
Examples of results from amplification experiments are
shown in Figure 1. Amplified DNA from the 10% (v/v)
Vol. 95 •No. 2
120
ANATOMIC PATHOLOGY
Original Article
.1327bp_
— 989 —
— 536—
— 268-^
—110 —
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
1 2 3 4 5 6 7 8 9 10 1112 13 1415
I—1327—I
I—989 — I
I—536I—268 — I
1—110bp—I
D
FlG. 1. Examples of PCR results. Amplification products ('/20 of reaction) were electrophoresed on 7% acrylamide, stained with ethidium bromide,
and photographed with UV light. A. Amplification from samples fixed in 10% BNF; (B) acetone; (C) Carnoy's; and (D) Bouin's. Sizes of PCR
products are noted at center. Far left unmarked lanes show a 123-bp ladder (BRL), and the far right lanes show 0X174 DNA cleaved with Hae III
(BRL) as size markers. Lanes 1, 4, 7, 10, and 13 are amplifications from tissuesfixedin the corresponding reagent for 1 hour; lanes, 2, 5, 8, 11, and
14 from thosefixedfor 4 hours; lanes 3, 6. 9, 12. and 15 from thosefixedfor 24 hours.
BNF-fixed tissue are shown in Gel A. There were abundant amplification products from most fixation time
points until the 24-hour time point/989 bp fragment,
where a reduction in the amount of product was evident.
Although not clearly visible in the photograph in Figure
1, the 1,327-bp product for the 1-, 4-, and 24-hour samples
was present, but at a reduced level. Amplifications from
the acetone-treated samples, displayed in Gel B, produced
abundant fragments of all sizes, from all time points. The
effect of fixation time is clearly demonstrated in Gel C.
Tissues fixed in Carnoy's showed reduction in amplification efficiency starting at the 24-hour time point of the
110-bp fragment. For each subsequent increase in fragment size, there was a clear reduction in product yield
ending with the 989-bp product of the four-hour time
point. Gel D represents the limited ability to amplify DNA
A.J.C.P. • February 1991
GREER ET AL.
PCR Amplification from Paraffin-Embedded Tissues
from tissues fixed in Bouin's. The only amplification
product detected was that of the 110-bp primer pair at
the one-hour fixation.
Table 2 summarizes the results for the 10-/tL sample
amplification from the 1-, 4-, and 24-hour time points
with the five primer pairs. In all cases, the 10-juL DNA
sample volume provided equal or better amplification
than the 1 -/xL sample volume. Results were scored as a
plus or minus, based on the presence or absence of an
ethidium-bromide-stained band of the appropriate molecular weight as resolved on a 7% (w/v) acrylamide gel.
In the bar graph displayed in Figure 2, results are com-
1400 bp
121
Alcoholic Formalin
Methacarn
Zamboni's
Paraformaldehyde
Clarke's
Formalin - Alcohol - Acetic Acid
10% BNF
Acetone
1200
a>
N
W
u
3
13
S
Q.
CC
O
Q.
1000
800
•
1 Hour
•
24 Hours
600
TABLE 2. SUMMARY OF AMPLIFICATION RESULTS
200
Fixation
Fixation
Time
(hr)
110
bp
268
bp
536
bp
989
bp
1,327
bp
Group: 1
Tissue Set 1
10% BNF
Alcoholic
formalin
Acetone
Zenker's
Carnoy's
Methacarn
Tissue Set 2
10% BNF
Bouin's
Zamboni's
Paraformaldehyde
Clarke's
Formalinalcoholacetic acid
1
4
24
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
1
4
24
1
4
24
1
4
24
1
4
24
1
4
24
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-
1
4
24
1
4
24
1
4
24
1
4
24
1
4
24
+
+
+
+
+
+
+
+
+
+
—
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-+
+
+
+
+
+
+
+
+
+
-
1
4
24
+
+
+
+
+
+
+
+
+
+
+
+
+
-
4-
+
+
+
+
+
+
+
+
+
—
Successful amplification of a particular size fragment is denoted by +. This reflects the visualization
of a band of the expected migration on an ethidium bromide-stained polyacrylamide gel (see
Fig. I).
— denotes that no amplification product was visible.
bp = basepairs.
2
3
4
5
FIG. 2. Effect of fixation on PCR. This graph illustrates results from
tissues fixed 1 or 24 hours, as detailed in Table 1. Fixatives included in
Groups 1-5 (x-axis) are noted at the top of the bars. The y-axis reflects
PCR product size (basepairs). Bars represent the maximum size PCR
product attained from each sample.
pared for all of the 1- and 24-hour fixations. Because the
results from the one- and four-hour fixations were similar,
they were not included in this graphic representation of
the results.
Although the use of a hybridization detection system
may have provided a more sensitive detection limit, it
would not have changed the overall pattern of results provided by the stained-gel system. From the results, it is
clear that acetone and 10% (v/v) BNF fixation permitted
amplification of larger products than the other fixatives
tested. A second group, including Zamboni's, Clarke's,
paraformaldehyde, formalin-alcohol-acetic acid, and
methacarn, appeared compatible with amplification
methods, but to a lesser extent than acetone and 10%
(v/v) BNF. Fixation with Carnoy's gave less efficient amplification, followed by Zenker's, and, lastly, Bouin's.
Additional Purification of Samples
An experiment was conducted to determine whether
additional purification of the DNA would improve the
amplification efficiency in samples that previously yielded
moderate or poor amplification. Phenol/chloroform extraction and ethanol precipitation were performed on the
prepared sample of the 24-hour time point of each of the
following fixatives: 10% (v/v) BNF, acetone, Zenker's,
Carnoy's, methacarn, Bouin's, Zamboni's, paraformal-
Vol. 95 • No. 2
ANATOMIC PATHOLOGY
122
Original Article
dehyde, and formalin-alcohol-acetic acid. The purified
DNA was suspended in TE buffer at a volume equal to
that of the prepurified sample. Amplification was performed as with the crudely prepared tissues. The results
(data not shown) indicated that there was little or no significant improvement in the amplification after purification.
cation of the added DNA was observed, indicating that
an inhibitor was not present in any of the sample solutions.
Size of DNA in PET Samples
Finally, to determine the size of the DNA that remained
after the tissue-fixation processes, all prepared samples
were analyzed by electrophoresis on 1.5% (w/v) agarose
gels. The results, summarized in Table 3, revealed that
the fixation procedures had a broad spectrum of effects
on DNA size. Many of the samples appeared as diffuse
smears, the size ranges of which are noted. In general,
DNA size was not significantly affected by fixation time;
tissues fixed in Zamboni's were the only samples that
showed a significant effect of fixation time and size of
recoverable DNA.
Addition-Recovery
The results from the phenol/chloroform extracted
samples suggested that additional purification of the DNA
was not advantageous but did not address whether the
DNA was suitable for amplification or whether a PCR
inhibitor was remaining from fixation. Therefore, an addition-recovery experiment, to identify such inhibitors,
was performed on the prepared sample of the 24-hour
time point of four fixatives: 10% (v/v) BNF, Zenker's,
Carnoy's, and Bouin's. Approximately 0.75 ng of purified
DNA from the HeLa cell line (representing approximately
1,000 copies of HPV18 DNA) was mixed with 10 nL of
each prepared sample DNA. A sample control (without
added HeLa DNA) was also tested. The samples were
subsequently amplified with an HPV-specific primer pair
that produces an amplification product of approximately
450 bp. 9
As can be seen in Figure 3, no reduction in amplifi-
DISCUSSION
Our goal was to establish, under optimal conditions, a
guide for tissue fixation procedures that would be most
conducive to subsequent use in PCR. Clearly, acetone
and 10% (v/v) BNF were best suited, under these experimental conditions. A second group of tissues, fixed in
Zamboni's, Clarke's, paraformaldehyde, formalin-alcohol-acetic acid, and methacarn, followed in efficiency of
CD
\_
Q)
J*
c
CD
m
O
i _
N
+
V)
o
CO
w
"><
o
c
CO
o
CDT3
>< CO
L_<
OZ
Z Q
n _ +M - +M - +ir
492bp—
- 450bp
123bp-
A.J.C.P. • February 1991
FIG. 3. Test for PCR inhibitors. Amplification reactions containing lO/jLofthe
prepared PET (24-hour fixation with reagent noted)
were carried out with HPVspecific primers, with (+) and
without (—) the addition of
purified HPV DNA. Aliquots
(V20) of the reaction were analyzed by 7% polyacrylamide
gel electrophoresis, ethidium
bromide staining, and photography under UV light.
The 450-bp fragment is the
HPV PCR product. The final
two lanes are control reactions containing no prepared
PET sample. The first lane
shows a 123-bp ladder (BRL).
GREER ET AL.
PCR Amplification from 'araffin-Embedded Tissues
TABLE 3. SIZE OF DNA FRAGMENTS RECOVERED
FROM TISSUES TREATED
Fixative
Size Range
(kb)
Average
Range (kb)
10% BNF
Alcoholic formalin
Acetone
Zenker's
Carnoy's
Methacarn
Bouin's
Zamboni's 1, 4 hr
24 hr
Paraformaldehyde
Clarke's
Formalin-alcohol-acetic acid
0.25->24
0.25-12
0.25->24
0.1-2
0.1-1.5
0.1-1.5
<0.123
0.1-7
<0.1-2
0.7->24
0.2-2.5
0.25-4.5
2-5
1-4
2-5
1-2
0.7-1.5
0.7-1.5
<0.1
1-2
0.2-0.5
0.2-5.0
0.8-1.0
1.0-3.0
DNA samples were electrophoresed on 1.5% agarose gels, ethidium bromide-stained, and photographed under UV light. The photographic negative was captured by video camera and the
data digitized using computer software.
Size range = full range of DNA fragment sizes observed.
Average range = size range of the majority of sample DNA fragments.
kb = kilobases.
amplification. Carnoy's, Zenker's, and Bouin's increasingly compromised amplification. Also, results indicated
that, in general, the longer the fixation time, the less efficient the subsequent amplification. Furthermore, tissues
processed by the less desirable procedures were particularly
poor templates for the amplification of larger DNA fragments.
In choosing optimal conditions for this study, we attempted to eliminate variables that might obscure the effect of the fixatives under consideration. For example, we
used approximately equal amounts of freshly fixed tissue.
In practice, tissue amount and the time between sampling
and processing are highly variable. These variables all affect the condition of the DNA as shown by Dubeau and
associates,12 who determined that large molecular weight
DNA could not be recovered from autolyzed or inadequately fixed tissues and that tissue necrosis resulted in
small molecular weight DNA.
Some generalizations can be made concerning the type
of fixative used and the resulting size of the DNA. The
tissues processed with acetone or fixatives that contained
buffered cross-linking agents (e.g., 10% [v/v] BNF, paraformaldehyde) tended to have larger DNA fragments than
the tissues treated with acid fixatives (e.g., Bouin's, Zenker's). Intermediate to these was Zamboni's, which contained both acid and cross-linking agents. Carnoy's,
methacarn, and Clarke's are acid-containing alcohol fixatives and resulted in smaller DNA fragments. However,
the size of the fragmented DNA does not appear to be
the sole predictor of successful PCR amplification. Although the fixatives contained in Group 2 (see Fig. 2)
afforded similar amplification results, they rendered tissue
DNA fragments of a variety of size ranges (see Table 3).
123
This suggests there are other factors that limit the size of
the amplification product, perhaps the frequency of crosslinking moieties present in the DNA.
The results suggested that there was no advantage provided by additional purification of the fixed samples.
However, the high efficiency of the primers used in this
study may have precluded an observation of improvement
in amplification. Amplification with primers that are degenerate or less efficient may be enhanced by additional
purification of the target DNA. Data from our laboratory
and those of other investigators (C. Wheeler, personal
communication) suggest that some 10% BNF fixed samples, incapable of efficient amplification, were rescued after
phenol/chloroform extraction. Unfortunately, the extra
manipulation greatly increases the potential for carryover
contamination.
It is encouraging to observe that fixation procedures
that are desirable for DNA in situ hybridization 1415 also
are appropriate for PCR amplification, thus allowing direct comparison of results. Nuovo and Richart' 4 evaluated
the effects of 10% (v/v) BNF, Bouin's fixative, and Hartman's fixative on DNA in situ hybridization. Their results
strongly suggest that 10% (v/v) BNF is the best fixative
for a high correlation between cytologic and in situ results
and that Hartman's and Bouin's are not desirable. We
have shown that Bouin's was not compatible with PCR
methods, and Hartman's may also be undesirable (K.
Shah and R. Kurman, personal communication).
There are many factors that affect the overall efficiency
of the amplification that should be considered when
studying PETs. For example, optimization of the proteinase K. digestion volume is important. Tissues should
be fully digested, with little or no remaining precipitate,
to efficiently liberate DNA. The amount of prepared sample used can also affect the efficiency of amplification. In
this study, a tenfold range in the amount of sample DNA
was used without impeding amplification. Work in our
laboratory suggests that a titration of each sample volume
is desirable for each specimen.17 Dilution of a prepared
sample may be necessary for successful amplification,
particularly with sections containing large amounts of
tissue.
The age of the PETs was uniform in this study and was
not considered a factor affecting PCR. However, age of
the sample is an important consideration for investigators
interested in archival samples. Successful amplification
from 40-year-old PETs has been reported.23 However,
Goelz and associates'3 reported that the size of DNA fragments prepared from samples four to six years old was
often smaller than the size of samples less than two years
of age. Researchers in both our laboratory and at the New
Mexico Tumor Registry (C. Wheeler, personal communication) have reported a decline in ability to amplify
Vol. 95 •No. 2
124
ANATOMIC PATHOLOGY
Original Article
from 10% (v/v) BNF-fixed tissues that are five years old
or older.
We strongly urge a thorough review of the documentation regarding the processing of each sample when designing large retrospective studies. Simply relying on
someone's recollection of how samples were processed in
the past is not sufficient and can lead to wasted time and
resources. Also, modifications offixationprocedures may
lead to interlaboratory variation in subsequent amplifications. Differences in the quality or age of the chemicals
used in sample processing may also affect amplification.
Samples from resource-poor environments, which are often fixed in lower-grade chemicals, may benefit from additional purification before amplification (L. Villa, personal communication).
Clearly, DNA amplification methods will be valuable
tools for the analysis of fixed tissue specimens. The information presented here may serve as a guideline for the
fixation of current specimens, as well as for choosing appropriate archival material for study. Older specimens,
or those fixed under suboptimal conditions, may require
the use of primer pairs that generate smaller amplification
products. Considering the many variables involved in PET
processing and the differing amplification efficiencies of
primer pairs used in PCR, it is highly recommended that,
before a large retrospective or prospective study is started,
a comprehensive pilot study be performed.
Note added in proof: Preliminary studies suggest that OmniFix*
(American Histology Reagent Co., Stockton, CA) may be superior to
10% BNF for subsequent amplification of PETs.
Acknowledgments. The authors thank the following Cetus employees
for their generous assistance: K. Lee for his excellent technical assistance,
R. Saiki for his recommendations concerning beta-globin primers, R.
Watson for his guidance in evaluating the size ranges of the sample DNA
after fixation, E. Ladner and S. Nilson for help with the figures, and R.
Kurka for manuscript preparation. They also thank C. Wheeler, D.
Mones, L. Villa, K. Shah, and R. Kurman for helpful discussions about
paraffin-embedded tissues, and J. Sninsky, T. White, H. Bauer, and M.
McClelland for review of the manuscript.
REFERENCES
1. Faloona F, Mullis K. Specific synthesis of DNA in vitro via a polymerase catalyzed chain reaction. Methods Enzymol 1987; 155:
335.
2. Saiki RK, Scharf S, Faloona F, et al. Enzymatic amplification of [)globin genomic sequences and restriction site analysis for diagnosis
of sickle cell anemia. Science 1985;230:1350.
3. Burmer GC, Rabinovitch PS, Loeb LA. Analysis of c-Ki-ras mutations in human colon carcinoma by cell sorting, polymerase
chain reaction, and DNA sequencing. Cancer Res 1989:49:21412146.
4. Frye RA, Benz CC, Liu E. Detection of amplified oncogenes by
differential polymerase chain reaction. Oncogene 1989;4:11531157.
5. Shibata D, Cosgrove M, Arnheim N, Martin WJ, Martin SE. Detection of human papillomavirus DNA in fine-needle aspirations
of metastatic squamous-cell carcinoma of the uterine cervix using
the polymerase chain reaction. Diagn Cytopathol 1989;5:40-43.
6. Brandsma JL, Lewis AJ, Abramson A, Manos MM. Detection and
typing of papillomavirus DNA in formalin-fixed paraffin-embedded tissue. Arch Otolaryngol Head Neck Surg 1990; 116:844-848.
7. Cao M, Xiao X, Egbert B, Darragh TM, Yen TSB. Rapid detection
of cutaneous herpes simplex virus infection with the polymerase
chain reaction. J Invest Dermatol 1989;82:391-392.
8. Levi JE, Delcelo R, Alberti VN, Torloni H, Villa LL. Human papillomavirus DNA in respiratory papillomatosis detected by in situ
hybridization and the polymerase chain reaction. Am J Pathol
1989;135:1179-1184.
9. Resnick RM, Cornelissen MTE, Wright DK, et al. Detection and
typing of human papillomavirus in archival cervical cancer specimens by DNA amplification with consensus primers. J Natl
Cancer Inst 1990;82:1477-1484.
10. Shibata DK, Arnheim N, Martin WJ. Detection of human papilloma
virus in paraffin-embedded tissue using the polymerase chain reaction. J Exp Med 1988;167:225-230.
11. Bramwell NH, Burns BF. The effects of fixative type and fixation
time on the quantity and quality of extractable DNA for hybridization studies on lymphoid tissue. Exp Hematol 1988;16:730732.
12. Dubeau L, Chandler LA, Gralow JR, Nichols PW, Jones PA. Southern blot analysis of DNA extracted from formalin-fixed pathology
specimens. Cancer Res 1986;46:2964-2969.
13. Goelz SE, Hamilton SR, Vogelstein B. Purification of DNA from
formaldehyde-fixed and paraffin embedded human tissue.
Biochem Biophys ResCommun 1985;130:118-126.
14. Nuovo GJ, Richart RM. Buffered formalin is the superior fixative
for the detection of HPV DNA by in situ hybridization analysis.
Am J Pathol 1989;134:837-842.
15. Nuovo GJ, Silverstein SJ. Methods in laboratory investigation:
comparison of formalin, buffered formalin, and Bouin's fixation
on the detection of human papillomavirus deoxyribonucleic acid
from genital lesions. Lab Invest 1988;59:720-724.
16. Impraim CC, Saiki RK, Erlich HA, Teplitz RL. Analysis of DNA
extracted from formalin-fixed, paraffin-embedded tissues by enzymatic amplification and hybridization with sequence-specific
oligonucleotides. Biochem Biophys Res Comm 1987; 142:710716.
17. Wright DK, Manos MM. Sample preparation from paraffin-embedded tissues. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ,
eds. PCR protocols: a guide to methods and applications. Berkeley:
Academic Press, 1990:153-158.
18. Paabo S. Ancient DNA: extraction, characterization, molecular
cloning, and enzymatic amplification. Proc Natl Acad Sci USA
1989;86:1939-1943.
19. Arnheim N, White T, Rainey WE. Application of PCR: organismal
and population biology. Bioscience 1990;40:174-182.
20. Kwok S. Procedures to minimize PCR-product carry-over. In: Innis
MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR protocols: a
guide to methods and applications. Berkeley: Academic Press,
1990:142-145.
21. Kwok S, Higuchi R. Avoiding false positives with PCR. Nature
1989;339:237-238.
22. Ting Y, Manos MM. Detection and typing of genital human papillomaviruses. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ,
eds. PCR protocols: a guide to methods and applications. Berkeley:
Academic Press, 1990:356-367.
23. Shibata D, Martin WJ, Arnheim N. Analysis of DNA sequences in
forty-year-old paraffin-embedded thin-tissue sections: a bridge
between molecular biology and classical histology. Cancer Res
1988;48:4564-4566.
A.J.C.P. • February 1991