Southern Blot Analysis of DNA Exfracted from Formalin...
7
(CANCERRESEARCH 46, 2964-2969, June l986J ABSTRACF We have developeda method for the extraction ofDNA fromformalin fixed, paraffin-embedded pathology specimens. High-molecular-weight DNA was recovered from well-fixed nonautolyzed samples of viable tissue. DNA recovered from samples exposed to picric acid or mercuric chloride containing fixatives was not intact. Increasing the formalin fixation time decreased the amount of intact DNA available. When these limitations were taken into consideration, the procedure allowed for the removal ofdegraded and chemically modified DNA from the preparation, and the final product was suitable for quantitative and qualitativeanalysis by Southern or dot blotting techniques. Digestion with methylation sensitive restriction endonucleases showed that DNA methylation pat term were not altered after formalin fixation. INTRODUCFION The field of molecular biology has expanded rapidly over the last few years, and a large amount of new information has accumulated on the organization of the human genome. The genome is not as static as previously thought, and gene rear rangement (1), amplification (2), deletion (3), mutation (4), and altered DNA methylation patterns (5, 6) are examples of deter minants now thought to be important for the regulation of gene expression. Similar modifications are also thought to take place in neoplastic processes and may have prognostic significance. For example, gene deletions (3), amplifications (7), and the presence of some restriction fragment length polymorphisms (8) have been associated with a poor prognosis for specific tumors. The useof this new knowledge for the study of tumor biology and also as a clinical tool has been delayed becauseof the need to use fresh or frozen tissue as a source of DNA, making it difficult to conduct retrospective studies on a large number of patients. These restrictions would be overcome if DNA was obtainable from the vast archives of material available in pa thology departments throughout the world. Since clonable and restrictable DNA has been obtained from sources as ancient as Egyptian mummies (9), we were encouraged to develop an extraction method for DNA from formalin-fixed, paraffin-em bedded tissue. We are particularly interested in analyzing gene rearrangements, methylation patterns, and amplification using Southern blotting and dot blotting techniques. Formylation of nucleic acids produces Schiff baseson free amino groups of nucleotides (10.-i 2), and exposure of nucleo proteins to formaldehyde results in the formation of cross-links between proteins and DNA (13). The fact that both of these processesare reversible in aqueoussolutions (12—14) suggested to us that DNA could be recovered from formalin-fixed tissue. Indeed, Goelz et a!. (15) have recently shown that DNA which can be cut by restriction endonucleasesand hybridized to DNA Received 10/17/85; revised 1/21/86; accepted 3/5/86. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734solelyto indicatethis fact. I This work was supported by Grant CA40422 from the National Cancer Institute and by the Dorothy and Hugh Edmondson Research Fund. 2To whom requestsfor reprints should be addressed. probes can be obtained from embedded specimens. In the pres ent paper, we have characterized the effect of formalin fixation and paraffin embedding on the availability ofDNA for Southern blot analysis and have developed an extraction method which allows for the removal of degraded and irreversibly modified DNA from the preparation. The purified DNA is suitable for quantitative as well as qualitative Southern blot analyses. MATERIALS AND METHODS Source and Handling of Tissue. Fresh tissue was obtained from the Department of Pathology of the Kenneth Norris Hospital in Los Angeles. Unfixed tissue samples were frozen in isopentane and kept at —70T until DNA was extracted using the method previously described by Jones and Taylor (16). Fixed samples were sliced, washed several times with 10% buffered formalin, fixed in an excess volume of for main, and processed for paraffinembeddingin parallelwith routine surgical pathology specimens. Old blocks of paraffin-embedded, for main-fixed tissue were obtained from the Kenneth Norris Hospital or from the archives of the Los Angeles County Hospital. Digestion with Restriction Endonucleases.DNA wasincubated over night at 37C using 4 units ofenzyme per @g of DNA as recommended bythe manufacturer (New England Biolabs, Beverly,MA). The follow ing morning one additional unit of enzyme was added per @gofDNA, and incubation was continued for 8 h or until completeness of digestion was demonstrated by running aliquots of the digest on a minigel. Gel ElectrophoresisandSouthernBlotting.Aliquots of restriction endonucleasedigests were electrophoresed on 1% agarosegels in 40 mM Tris-acetate-l mM EDTA, pH 7.4. After electrophoresis the DNA was stained with ethidium bromide and visualized by UV illumination. The DNA was then transferred to nitrocellulose filters (Millipore, Bedford,MA) usingthe procedureof Southern(17).The filters were air dried and baked between 2 sheets of Whatman No. 3MM paper undervacuumfor 3 h at 80C. Preparation of 32P-labeled DNA Probes. The c-Ha-ms probe (clone pT24-C3) was obtained from Dr. Mariano Barbacid from the National Cancer Institute. The probe was labeled with [32PIdCTPto specific activities >i0@ cpm/@@gusing the procedure of Feinberg and Vogelstein (18). DNA Hybridization Conditions. Nitrocellulose filters were prehy bridized overnight at 68T in a solution containing 6x SSC,3 0.5% SDS, 5x Denhardt's solution (19), and denatured salmon sperm DNA (200 @@g/ml) (Sigma). The prehybridization buffer was then discarded, and the 32P-labeledprobes were hybridizedto the filters for 60—65 h at 68T in 6x SSC-0.01 M EDTA-5x Denhardt's solution-0.5% SDS denatured salmon sperm DNA (100 @ig/ml). Hybridization mixes were filtered through Millex-HA 0.45-sm filter units (Millipore) as recom mendedby Meinkoth andWahl (20)prior to addition to the nitrocel lulose filter to aid in the reduction of background. Following hybridi zation, the filters were washed at room temperature once in 2x SSC 0.5% SDS for 5 mm, once in 2x SSC-0.1 % SDS for 15 mm, and then for 1—3 h at 55C in 0.lx SSC-0.5% SDS with multiple buffer changes. The filters were exposed to Kodak XAR-5 X-ray film with 2 DuPont Lightning-Plus intensifying screens for 10—14days at —80'C. Dot Blotting. Ten-pg samples of formalin-fixed DNA were precipi tated in ethanol and resuspendedin 9 @zl of 0.lx SSC. One @tl of 2 N NaOH was added, and the samples were boiled for 5 min. Twenty j@l 3The abbreviations used are: SSC, 0.15 M NaCl-0.0l5 M sodium citrate; SDS, sodium dodecyl sulfate. 2964 Southern Blot Analysis of DNA Exfracted from Formalin-fixed Pathology Specimens' Louis Dubeau,2 Lois A. Chandler, Julie R. Gralow, Peter W. Nichols, and Peter A. Jones Urological Cancer Research Laboratory, USC Comprehensive Cancer Center IL D., L A. C., I. R. G., P. A. I.], and Department ofPathology, Kenneth Norris, Jr., Hospital(P. W. N.], University ofSouthern California, LosAngeles, California 90033 Research. on November 19, 2018. © 1986 American Association for Cancer cancerres.aacrjournals.org Downloaded from
Transcript of Southern Blot Analysis of DNA Exfracted from Formalin...
(CANCERRESEARCH46, 2964-2969, June l986J
ABSTRACF
We havedevelopeda methodfor the extraction ofDNA fromformalinfixed, paraffin-embedded pathology specimens. High-molecular-weightDNA was recovered from well-fixed nonautolyzed samples of viabletissue. DNA recoveredfrom samples exposed to picric acid or mercuricchloride containing fixatives was not intact. Increasing the formalinfixation time decreased the amount of intact DNA available. When theselimitations were taken into consideration, the procedure allowed for theremoval ofdegraded and chemically modified DNA from the preparation,and the final productwassuitable for quantitativeand qualitativeanalysisby Southern or dot blotting techniques. Digestion with methylationsensitive restriction endonucleases showed that DNA methylation patterm were not altered after formalin fixation.
INTRODUCFION
The field of molecular biology has expanded rapidly over thelast few years, and a large amount of new information hasaccumulated on the organization of the human genome. Thegenome is not as static as previously thought, and gene rearrangement (1), amplification (2), deletion (3), mutation (4), andaltered DNA methylation patterns (5, 6) are examples of determinants now thought to be important for the regulation of geneexpression. Similar modifications are also thought to take placein neoplastic processes and may have prognostic significance.
For example, gene deletions (3), amplifications (7), and thepresence of some restriction fragment length polymorphisms(8) have been associated with a poor prognosis for specifictumors.
The useof this new knowledge for the study of tumor biologyand also as a clinical tool has beendelayed becauseof the needto use fresh or frozen tissue as a source of DNA, making itdifficult to conduct retrospective studies on a large number ofpatients. These restrictions would be overcome if DNA wasobtainable from the vast archives of material available in pathology departments throughout the world. Since clonable andrestrictable DNA has been obtained from sourcesas ancient asEgyptian mummies (9), we were encouraged to develop anextraction method for DNA from formalin-fixed, paraffin-embedded tissue. We are particularly interested in analyzing generearrangements, methylation patterns, and amplification usingSouthern blotting and dot blotting techniques.
Formylation of nucleic acids produces Schiff baseson freeamino groups of nucleotides (10.-i 2), and exposure of nucleoproteins to formaldehyde results in the formation of cross-linksbetween proteins and DNA (13). The fact that both of theseprocessesare reversible in aqueoussolutions (12—14)suggestedto us that DNA could be recovered from formalin-fixed tissue.Indeed, Goelz et a!. (15) have recently shown that DNA whichcan be cut by restriction endonucleasesand hybridized to DNA
Received 10/17/85; revised 1/21/86; accepted 3/5/86.The costs of publication of this article were defrayedin part by the payment
of page charges. This article must therefore be hereby marked advertisement inaccordancewith 18 U.S.C. Section 1734solelyto indicatethis fact.
I This work was supported by Grant CA40422 from the National CancerInstitute and by the Dorothy and Hugh Edmondson ResearchFund.
2To whom requestsfor reprints should be addressed.
probes can be obtained from embeddedspecimens.In the present paper, we have characterized the effect of formalin fixationand paraffin embedding on the availability ofDNA for Southernblot analysis and have developed an extraction method whichallows for the removal of degraded and irreversibly modifiedDNA from the preparation. The purified DNA is suitable forquantitative as well as qualitative Southern blot analyses.
MATERIALS AND METHODS
Source and Handling of Tissue. Fresh tissue was obtained from theDepartment of Pathology of the Kenneth Norris Hospital in LosAngeles. Unfixed tissue samples were frozen in isopentane and kept at—70T until DNA was extracted using the method previously describedby Jones and Taylor (16). Fixed samples were sliced, washed severaltimes with 10% buffered formalin, fixed in an excess volume of formain, and processedfor paraffinembeddingin parallelwith routinesurgical pathology specimens.Old blocks of paraffin-embedded, formain-fixed tissue were obtained from the Kenneth Norris Hospital orfrom the archives of the Los Angeles County Hospital.
Digestion with Restriction Endonucleases.DNA was incubated overnight at 37C using 4 units ofenzyme per @gof DNA as recommendedbythe manufacturer (New England Biolabs, Beverly,MA). The following morning one additional unit of enzyme was added per @gof DNA,and incubation was continued for 8 h or until completeness of digestionwas demonstrated by running aliquots of the digest on a minigel.
Gel Electrophoresisand SouthernBlotting. Aliquots of restrictionendonucleasedigests were electrophoresed on 1% agarosegels in 40mM Tris-acetate-l mM EDTA, pH 7.4. After electrophoresis the DNAwas stained with ethidium bromide and visualized by UV illumination.The DNA was then transferred to nitrocellulose filters (Millipore,Bedford,MA) usingthe procedureof Southern(17).The filters wereair dried and baked between 2 sheets of Whatman No. 3MM paperundervacuumfor 3 h at 80C.
Preparation of 32P-labeled DNA Probes. The c-Ha-ms probe (clonepT24-C3) was obtained from Dr. Mariano Barbacid from the NationalCancer Institute. The probe was labeled with [32PIdCTPto specificactivities >i0@ cpm/@@gusing the procedure of Feinberg and Vogelstein(18).
DNA Hybridization Conditions. Nitrocellulose filters were prehybridized overnight at 68T in a solution containing 6x SSC,3 0.5%SDS, 5x Denhardt's solution (19), and denatured salmon sperm DNA(200 @@g/ml)(Sigma). The prehybridization buffer was then discarded,and the 32P-labeledprobes were hybridized to the filters for 60—65h at68T in 6x SSC-0.01 M EDTA-5x Denhardt's solution-0.5% SDSdenaturedsalmon sperm DNA (100 @ig/ml).Hybridization mixes werefiltered through Millex-HA 0.45-sm filter units (Millipore) as recommendedby Meinkoth andWahl (20)prior to addition to the nitrocellulose filter to aid in the reduction of background. Following hybridization, the filters were washed at room temperature once in 2x SSC0.5% SDS for 5 mm, once in 2x SSC-0.1 % SDS for 15 mm, and thenfor 1—3h at 55C in 0.lx SSC-0.5% SDS with multiple bufferchanges.The filters were exposed to Kodak XAR-5 X-ray film with 2 DuPontLightning-Plus intensifying screens for 10—14days at —80'C.
Dot Blotting. Ten-pg samples of formalin-fixed DNA were precipitated in ethanol and resuspendedin 9 @zlof 0.lx SSC. One @tlof 2 NNaOH was added, and the samples were boiled for 5 min. Twenty j@l
3The abbreviations usedare: SSC, 0.15 M NaCl-0.0l5 M sodium citrate; SDS,sodium dodecyl sulfate.
2964
Southern Blot Analysis of DNA Exfracted from Formalin-fixed Pathology
Specimens'
Louis Dubeau,2 Lois A. Chandler, Julie R. Gralow, Peter W. Nichols, and Peter A. Jones
Urological Cancer Research Laboratory, USC Comprehensive Cancer Center IL D., L A. C., I. R. G., P. A. I.], and Department ofPathology, Kenneth Norris, Jr.,Hospital(P. W. N.], University ofSouthern California, LosAngeles, California 90033
Research. on November 19, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from
DNA FROM FORMALIN-FIXED TISSUE
of 3 M NaC1 were added, and the tubes were put on ice. Each samplewas spotted onto Millipore (Bedford, MA) type HA, 0.45-@tmnitrocellulose paper. The paper was dried on a piece of Whatman No. 3MMfilter paper for 30 mm, washed twice for 15 mm in 6x SSC, and driedfor an additional 30 mm. The nitrocellulose paper was baked at 80Cunder vacuum for 3 h between 2 sheets of Whatman No. 3MM paper.Prehybridization and hybridization with a 32P-labeledc-Ha-ras probewere done as for Southern blots.
Extraction of DNA from Formalin-fixed, Paraffin-embedded Specimens.A flowdiagram of the various steps used for the DNA extractionprocedureis shownin Fig. 1. Commentsabouteachof thesestepsaregiven in the rest of this section as well as in “Results.―
Selection of Paraffin Blocks. The single most important cause offailure to recover intact high-molecular-weight DNA from formalinfixed pathology specimens is the use of inadequately fixed tissue.Histological sections of paraffin blocks should therefore be examinedprior to DNA extraction, and the blocks should not be usedif autolysisis evident. Examination of histological sections may also reveal thepresence of tissue necrosis. Such areas should be removed with a scalpelblade before starting the extraction. Intact DNA could not be recoveredfrom tissue that had been exposed to picric acid (Bouin) or mercuricchloride (Zenker, B5) containing fixatives. The amount of intact DNAavailable from the tissue decreased as the length of fixation in formalinwas increased.
Tissue Homogenization. One effect of formalin fixation is to producea hard rubbery tissue difficult to homogenize. We have taken advantageof the fact that the fixed tissue was embedded in paraffin blocks to cutit into thin slicesusingthe microtome.Five-tim-thicksliceswerecutuntil the whole paraffin block was exhausted or enough tissue wasobtained,andthesliceswerecollectedinto acentrifugetube.Practicallyevery cell present in the tissue was cut at this setting, thus providing afirst homogenization step.
Solubilization of DNA. The thin tissue slices were deparaffinized andrehydrated by resuspension in 50 ml ofxylene(J. T. Baker, Phillipsburg,NJ) and washing 3 times by centrifugation in xylene, absolute ethanol,
CUT EMBEDDED TISSUE WITH MICROTOME
@1@
DEPARAFFINIZE AND REHYDRATE
FIRST INCUBATION IN SOLUTION A(discard supernatant)
¶1?SECOND INCUBATION @NSOLUTION A
(save supernatant)
EXTRACT WITH CHLOROFORM - ISOAMYL ALCOHOL
DIGEST WITH RNase AND PRONASE
EXTRACT WITH PHENOL
SPOOL(discard unspoolable DNA)
DIALYZE
Fig. 1. Flow diagram for extraction of DNA from formalin-fixed, paraffinembedded tissues.
and SSC. The final pellet was resuspended in 5 ml of SSC containing1% SDS and 100 @Lgof proteinase K per ml (E. Merck, Darmstadt,Germany) (Solution A). If the block contained only small amounts oftissue (less than 25% of the volume of the block), it was resuspendedin 2.5 ml of Solution A insteadof 5 ml. The preparationwas thenincubated at 37C in a plastic centrifuge tube that was taped inside an850-cm2 tissue culture roller bottle (Falcon, Oxnard, CA) and rotatedat a speed of one revolution every 2.5 min. All of the DNA present inthe tissue was slowly released in a soluble form during this process(Fig. 2). The rate of release was not the same for different types oftissue and depended on the density and abundance of extracellularstroma. For highly cellular tissue containing little stroma, such asspleen (Fig. 3a), over 80% of the DNA was released within 24 h (Fig.2). However, it took 6 days before a similar percentage of the totalDNA could be recovered from a uterine leiomyoma containing largequantities of dense stroma (Figs. 2 and 3). DNA from a metastaticteratoma of moderate cellularity containing areas of loose stroma wassolubilized at an intermediate rate (Fig. 2). The rate of solubilizationand the amount of total DNA recovered were not altered when thelength oftime the tissue had been exposed to formalin was varied from12 h to 5 days (results not shown).
Incubation in Solution A was done with 2 consecutive extractions inorder to remove low-molecular weight DNA from the preparation.After 6 to 12 h of incubation (24 h for very fibrous tissue),thepreparation was centrifuged, and the supernatant was discarded. Thepelletwasthenresuspendedin freshSolutionA andput backat 37T.This step also allowed removal of pigments such as hemoglobin. Thesecond incubation was stopped after 2 to 6 days, depending on thedensity of the tissue stroma.
Extraction with Chloroform-Isoamyl Alcohol. The preparation wascentrifuged to remove insoluble debris after the second incubation inSolutionA, andthe supernatantwastransferredto a round-bottomedglass flask. An equal volume of a mixture of 24 parts of chloroformand one part of isoamyl alcohol was added, and the flask was shakenmanually for 15 mm. The suspension was then transferred to a Corextube (Corning, Palo Alto, CA), and the phases were separated bycentrifugation at 7000 rpm for 5 mm in the 55-34 rotor of a Sorvallcentrifuge. The top phase was removed using a bent-tip Pasteur pipet,transferredto anotherCorex tube,and precipitatedovernightwith 2volumes of absolute ethanol at —20°C.The precipitate was recoveredby centrifugation for 30 mm at 7000 rpm in a Sorvall centrifuge andresuspended in 0.5 to 1.0 ml of SSC for each paraffin block originallyused. Larger volumes did not allow for spooling of the DNA after thephenol extraction (see later).
DAYSFig. 2. Effect of stromal density on the rate of extraction of DNA from
formalin-fixed tissues. Paraffin blocks from the spleen (•),teratoma (A), anduterine leiomyoma (U) shown in Fig. 3 were cut, and the tissue slices wereincubated at 37C in Solution A after rehydration. The samples were centrifugedat the indicated time intervals, the supernatants were saved, and the pellets wereresuspendedand reincubatedin fresh Solution A until the next time point. Theamount of total DNA present in each supernatant as well as in the insoluble pelletremaining after 10 days was determined using the diphenylamine method (29).Total DNA originally present in each sample wascalculated by adding the amountof DNA in each supernatant to the DNA left in the residual pellet after 10 days.
0‘UN-J
U)
z0
.-J
0
2965
Research. on November 19, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from
DNA FROM FORMALIN-FIXED TISSUE
heat-treated RNase and self-digested Pronase was done as for DNAextraction from unfixed tissue.
Extraction with Phenol. After digestion with Pronase, the preparationwas extracted with phenol as for DNA from fresh tissue. One volumeof ethanol was added to the aqueous phase of the phenol extract, andthe DNA was spooled with a glass rod, washed once with 70% ethanol,dissolved in 0.lx SSC, and dialyzed extensively against this bufferunder sterile conditions. Variable quantities of unspoolable DNA werealways seen in the phenol extract. This unspoolable material was notsuitablefor further analysisandwasdiscarded.
RESULTS
- .‘ ,. ,, .@ ,‘.1.@.;,. ‘,@‘@@ @@;@::‘“%!@
@;@@@ ‘..:@@ -‘@:@ ..‘@.‘... ,@, .,
@....@ ..@@@ ‘@ y
It should be stressed that intact high-molecular-weight DNAcould not be recovered from autolyzed and/or inadequatelyfixed tissues, and any area of tissue necrosis in the originalsample generated increased amounts of degraded low-molecular-weight DNA in the preparation. When care was taken toselect well-fixed viable tissue specimens, DNA preparationscould be obtained that had UV absorbancespectra very similarto those obtained from unfixed DNA samples (results notshown). Thus formalin fixation and processing for histologydid not damage the DNA so that its UV absorbance propertieswere appreciably altered.
Effect of Extraction Time. Because variable amounts of degraded DNA were always present in our preparations, wewanted to see if this degraded material was solubilized at thesame rate as high-molecular-weight DNA during the incubationstep in Solution A and to determine if this step could thereforebe used to remove the low-molecular-weight material. Paraffinblocks from 3 different inadequately fixed breast tumors wereselected in order to obtain DNA preparations with increasedamounts of degraded low-molecular-weight material. Theblocks were sectioned with the microtome, rehydrated, andincubated in Solution A for 24 h. The preparations were thencentrifuged, and the pellets were resuspended in fresh SolutionA. The supernatantsand the resuspendedpelletswereput backat 37C for an additional 60 h. DNAs from each tube werepurified using the method described in “Materialsand Methods―except that the purified samples were collected by centrif
@ -@ ugation instead of spooling after the phenol extraction step.— DNA from each purified sample was electrophoresed on a 1%
agarose gel (Fig. 4). The results clearly show that, for all 3cases, low-molecular-weight DNA with high electrophoreticmobility was recovered during the first 24 h, but not during thesecond incubation period. Thus, low-molecular-weight DNA
. . unsuitable for Southern blot analysis can be removed by performing 2 consecutive incubations in Solution A and discardingthe supernatant obtained after the first incubation. It should be
: emphasized that poorly fixed autolyzed material was used for. @: this experiment and that most of the DNA recovered from
.. : adequately fixed tissues would not have penetrated I % agarose
gels unless it had first been digested by restriction endonucleases(results not shown).
Effect of Length of Fixation in Formalin. When tissue specimens are processed in surgical pathology laboratories, formalinfixation times often vary from a few hours up to 5 days. Certaintypes of tissue such as lymph nodes are routinely fixed for atleast 48 h in some centers. We determined the effect of fixationtime on DNA extractability, because it seemed likely that thelength of fixation would be an important determinant of thequality of DNA present in a specimen. Tissue samples from ametastatic ovarian carcinoma were fixed for 12, 36, 60, and110 h. The corresponding DNAs were spooled on a glass rodafter phenol extraction. After spooling, variable quantities ofprecipitate were seenremaining in each sample. These precip
;- - -:l.@'@@
‘:@•@:@
@ ; fl@ :;.s. -@@ :@:@ .@‘.: .@ - @.
. ‘:â€â€¢@ @. .:@
@ ...“,‘ •r@@'
@ . ,4_@@
..‘. ,@ ,@ @.: -: • @•@@
@@ .,@ - . @“4@1.' @, •.. ;:@@@@@
\, S@ , % IiiI@ @,, j!
@ ‘ “@ - @. .-.@. “••4@@•@ (I
.@@@
:@ •@ @, I@@ : , .@
Fig. 3. Histological sections of tissues with different amounts and quality ofstroma. Histologicalsectionsof the 3 differentorgans used for Fig. 2 are shown.a, Normal spleen removed for staging of Hodgkin's disease; the section showsportions of the splenic white and red pulps, note the degree of cellularity andrelative paucity of stroma b, metastatic teratoma proliferating glandular structures and mesenchymalcell types are seen;extracellularstroma is abundant, butits density is less than in c; c, uterine leiomyoma showing proliferating smoothmuscle cells in a dense stroma. x570.
Digestion with RNase and Pronase. The resuspended pellet wasvortexed gently and allowed to dissolve for 30 to 60 min at roomtemperature. After this period of time most of the precipitate was stillin an insoluble form. The suspension was transferred to a 37C waterbath, and RNase digestion was started immediately. Digestion with
2966
Research. on November 19, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from
Table 1 Effectoftime offixation on the ability to spoolDNADNAwas extracted from paraffin blocks of a metastatic ovarian adenocarci
noma that had been fixed for 12, 36, 60, and 110 h. After the phenolextractionstep,DNAs were spooled, and the unspoolable material left behind in eachsamplewas
precipitated overnight at —20C in 2 volumes ofabsolute ethanol. TotalDNAwasquantitated in each sample by measuring uv absorbance at 260nm.TotalLength
of spoolable Total un- % oftotalfixationDNA spoolableDNA(h)(,ig) DNA (jig)spoolable12
829 6195736474 8893560211 6962311049 568 8
DNA FROM FORMALIN-FIXED TISSUE
b12
a12
b C d.
@. h.
•a@ö
e
C12
f
kb
23.1 -
6.6 -
2.3-
0.9-
0.6-
0.2-
Fig. 5. Dot blots ofspoolable and unspoolable DNA samples. DNA from eachsample of spoolable and unspoolable material in Table 1 was dot blotted ontonitrocellulose, hybridized to 3@P-labeledc-Ha-re., probe, and autoradiographed.Dots a, b, C,and d represent spooled samples fixed for 12, 36, 60, and 110 h,respectively. Dots e,f, g, and h are the corresponding unspooled samples.
Ha-ras probe (Fig. 5). All spoolable DNAs showed good hybridization, although dot intensity of the sample fixed for 110h wasslightly decreased.Hybridization intensities from samplesof unspoolable DNA were reduced but showed a gradual increasewith increasing fixation time. DNA Samples a, b, c, d,and h were digested with MspI, electrophoresed on an agarosegel, stained with ethidium bromide, and visualized under UVlight (Fig. 6). Although the staining intensity of the DNAentering the gel was similar for each sample of spooled DNA,most ofthe DNA in Sample h, which is the unspoolable samplethat showed the highest hybridization intensity in Fig. 5, couldnot penetrate the gel. When similar experiments were doneusing the other unspoolable samples, the data were eithersimilar or showed mainly low-molecular-weight DNA (resultsnot shown). Thus, it was important that samples be kept concentrated enough during the extraction procedure to allow forspooling of the DNA and that any unspoolable material bediscarded.
Digestion with Restriction Endonucleases. Fig. 7 is a stainedgel of DNA extracted from unfixed and fixed samples of ametastatic ovarian adenocarcinoma. A certain amount of incompletely digested material was present in the fixed sampleeven after taking the precautions mentioned above. Althoughthe nature of this material is not clear, it was observed insamples from paraffin-embedded tissue that had been fixed inethanol without exposure to formalin (results not shown). Someof this undigestible material may therefore represent singlestranded DNA generated by exposure to elevated temperaturesduring the embedding process. Increasing either the amount ofenzyme or the length of digestion did not improve the completenessof the reaction. The amount of DNA loaded on a gelshould be 2 to 3 times the amount normally used for DNAfrom unfixed samples to account for the presence of this unrestrictable material. Most of this material did not transfer tonitrocellulose filters during Southern blotting procedures (results not shown).
Southern blots ofDNA obtained from fresh and fixed tissues,digested with either MspI or HpaII endonucleases and hybridized to 32P-labeledc-Ha-ras gene probe, were also compared(Fig. 8). The mobility of the fixed material was slightly slowerthan that of DNA from fresh tissue, as also reported by Goelzet a!. (15). The reasons for this decrease in mobility are notclear, but this was found with all of the DNA samplesextractedfrom paraffin blocks. The banding patterns of fresh and fixedtissues were otherwise very similar, and comparison of theresults with the 2 enzymes showed that DNA methylationpatterns were not altered by formalin fixation and could beanalyzed by methylation-sensitive endonucleases.
2967
Fig. 4. Electrophoresis of DNA extracted for different lengths of time inSolution A. Paraffin blocks of poorly fixed specimens from 3 different breasttumors (a to c) were cut with the microtome, rehydrated, and incubated for 24 hin Solution A. The 3 samples were then centrifuged, and the pellets wereresuspended in fresh Solution A. The supernatants and resuspended pellets wereput back at 3TC for an additional 60 h. DNA was extracted from each of theresulting 6 samples but was not spooled after the phenol extraction step. Eachsample was electrophoresed through 1% agarose, and the gel was stained withethidium bromide. For each tumor, Lane 1 is the supernatant recovered after thefirst incubation, and Lane 2 is the resuspended pellet. kb, kilobase(s).
itates were grossly indistinguishable from those of intact DNA,but they did not stick to glass rods. This unspoolable materialwas precipitated overnight in 2 volumes of ethanol at —20°C.DNAs were quantitated by UV absorbanceafter removing thephenol by dialysis. The proportion of total DNA from eachsample that could be spooled decreasedas the length of fixationwas increased (Table 1). Although 57% of the total DNAextractable after 12 h of fixation could be spooled, only 8% ofthe material recovered after 110 h was spoolable.
The rate of solubilization of the DNA during incubation inSolution A and the percentage of total DNA recovered afterincubation in this solution were not altered when the length offixation was varied from 12 h to 5 days (results not shown).
Importance of DNA Spooling. Since one of the effects ofincreasing the length of fixation was to increase the amount ofunspoolable DNA in the preparation, we wanted to determineif this unspoolable material was still suitable for analysis. Equalamounts of spoolable and unspoolable DNA from each of thesamples listed in Table 1 were analyzed by dot blots with the c
Research. on November 19, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from
DNA FROM FORMALIN-FIXED TISSUE
aba b c dh
kb
23.3-
Fig. 7. Electrophoresis of restriction DNA fragments from fixed and unfixedtissues. DNAs from a fresh (Lane a) or fixed (Lane b) bladder tumor specimenwere digested with MspI, electrophoresed, and stained with ethidium bromide.kb, kilobase(s).
nization that allowed the recovery of over 95% of total DNApresent in embedded tissues. We have also characterized theeffect of stromal density on the rate of recovery of the DNAfrom fixed tissues.
Despite these improvements we were not successful in obtaming DNA suitable for analysis in every case. This was mostoften due to tissue autolysis or necrosis. Although our initialsuccess rate in obtaining usable high-molecular-weight DNAfrom old paraffin blocks was less than 50%, this was improvedto almost 100% when care was taken to select blocks of wellfixed viable tissues. The amount of usable material will undoubtedly depend on the quality of fixation of surgical pathology specimens available at any given center. Unfortunately,intact DNA was not obtained from tissues that had been exposed to picric acid containing (Bouin) or mercuric chloridecontaining (Zenker, B5) fixatives.
The number of surgical pathology specimens kept as formaIm-fixed, paraffin-embedded tissue is considrable, and thismaterial is available throughout the world. The ability to extractintact high-molecular-weight DNA from such specimens bringsa new dimension to the use of molecular biology for the studyof tumor biology and as a diagnostic tool. Probes for immunoglobulin genes(21—23)and more recently T-cell receptor genes(24—27)have been used for the typing of B- and T-cell lymphomas, and this technique can now be expanded for use onformalin-fixed samples. Because DNA methylation patterns of
k
23.
9.
6.
4.
2.
1.
1.
1.
0.
Fig. 6. Effect of DNA spooling on electrophoretic mobility of restrictionfragments.DNA samples used for Dots a, b, c, d, and h in Fig. 5 were digestedwith Mspl, electrophoresed, and stained with ethidium bromide. kb, kilobase(s).
DISCUSSION
The results of our studies clearly show that DNA suitable fordot blotting and Southern blotting analysis can be obtainedfrom formalin-fixed, paraffin-embedded material. The resultsalso show that DNA methylation patterns are preserved intactduring histological processing. The procedure we have developed contains the following modifications over a simpler onerecently reported by Goelz et a!. (15). (a) Degraded low-molecular-weight DNA was removed by sequential extractions withSolution A; this was also an effective way to remove tissuepigments (e.g., hemoglobin). (b) Chemically modified DNAunsuitable for restriction endonucleasedigestion or hybridization studies was removed by spooling out the intact DNA. (c)We have brought attention to the fact that the amount of DNAusable for hybridization studies decreasedmarkedly when thetime of fixation in formalin was increased to as little as 5 days.(d) We used a simple and effective method for tissue homoge
2968
Research. on November 19, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from
DNA FROM FORMALIN-FIXED TISSUE
2. Brodeur, G. M., Seeger, R. C., Schwab, M., Varmus, H. E., and Bishop, J.M. Amplification of N-myc in untreated human neuroblastomas correlateswith advanced disease stage. Science (Wash. DC), 224: 1121—1124,1984.
3. Cavanee, W. K., Hansen, M. F., Nordenskjold, M., Kock, E., Maumenee, I.,Squire, J. A., Phillips, R. A., and Gallie, B. L. Genetic origin of mutationspredisposing to retinoblastoma. Science (Wash. DC), 228: 501—503,1985.
4. Santos, E., Martin-Zanca, D., Reddy, E. P., Pierotti, M. A., Della Ports, G.,and Barbacid, M. Malignant activation ofa k-ras oncogene in lung carcinomabut not in normal tissue of the same patient. Science (Wash. DC), 223: 661—664, 1984.
5. Feinberg, A. P., and Vogelstein, B. A technique for radiolabelling DNArestriction endonucleasefragments to high specificactivity.Anal. Biochem.132:6—13,1983.
6. Riggs, A. D., and Jones, P. A. 5-Methylcytosine, gene regulation, and cancer.Adv. Cancer Res., 40: 1—30,1983.
7. Little, C. D., Nau, M. M., Carney, D. N., Gazdar, A. F., and Minna, J. D.Amplificationand expressionof the c-myconcogenein human lung cancercell lines. Nature (Lond.),306: 194—196,1983.
8. Krontiris, T. G., DiMartino, N. A., CoIb, M., and Parkinson, D. R. Uniqueallelic restriction fragments of the human Ha-rat locus in leukocyte andtumor DNAs ofcancer patients. Nature (Lond.), 313: 369—374,1985.
9. PaIbo, S. Molecular cloning of ancient Egyptian mummy DNA. Nature(Lond.),314: 644—645,1985.
10. Fraenkel-Conrat, H. Reaction of nucleic acid with formaldehyde. Biochim.Biophys. Ada. 15: 307—309,1954.
11. Hoard, D. E. The applicability of formol titration to the problem of endgroup determinations in polynucleotides. A preliminary investigation.Biochim. Biophys. Acta, 40: 62—70,1960.
12. Haselkorn, R., and Doty, P. The reaction of formaldehyde with polynucleotides. J. Biol. Chem. 236: 2738—2745,1960.
13. Jackson, V., and Chalkey, R. Separation ofnewly synthesized nucleohistoneby equilibrium centrifugation in cesium chloride. Biochemistry, 13: 3952—3956, 1974.
14. Jackson, V. Studies on histone organization in the nucleosome using formaldehydeas a reversiblecross-linkingagent. Cell, 15:945—954,1978.
15. Goelz, S. E., Hamilton, S. R., and Vogelstein, B. Purification of DNA fromformaldehyde-fixedand paraffin-embeddedhuman tissue.Biochem.Biophys.Res. Commun., 130: 118—126,1985.
16. Jones, P. A., and Taylor, S. M. Cellular differentiation, cytidine analogs, andDNA methylation. Cell, 20: 85—93,1980.
17. Southern, E. M. Detection of specific sequences among DNA fragmentsseparated by gel electrophoresis. J. Mol. Biol., 98: 503—517,1975.
18. Feinberg, A. P., and Vogelstein, B. Hypomethylation distinguishes genes ofsome human cancers from their normal counterparts. Nature (Lond.), 301:89—92,1983.
19. Denhardt, D. T. A membrane-filter technique for the detection of complementary DNA. Biochem.Biophys.Res. Commun.,23: 641-646, 1966.
20. Meinkoth, J., and WahI, G. Hybridization of nucleic acids immobilized onsolid supports. Anal. Biochem., 138: 267—284,1984.
21. Korsmeyer, S. J., Hieter, P. A., Ravetch, J. V., Poplack, D. G., Waldman,T. A., and Leder, P. Developmental hierarchy of immunoglobulin generearrangementsin human leukemicpre-B-cells.Proc. Nail. Acad.Sci. USA,78: 7096—7100,1981.
22. Arnold, A., Cossman, J., Bakhshi, A., Jaffe, E. S., Waldman, T. A., andKorsmeyer, S. J. Immunoglobulin-gene rearrangements as unique clonalmarkers in human lymphoid neoplasms. N. EngI. J. Med., 309: 1593—1599,1983.
23. Cleary, M. L., Warnke, R., and Sklar, J. Monoclonality of lymphoproliferative lesions in cardiac-transpisnt recipients: clonal analysis based on immunoglobulin-gene rearrangements. N. Engl. J. Med., 310: 477—482,1984.
24. Aisenberg, A. C., Krontiris, T. G., Mak, T. W., and Wilkes, B. M. Rearrangement of the gene for the beta chain of the T-cell receptor in T-cell chroniclymphocyticleukemia and related disorders. N. Engl. J. Med., 313: 529—533,1985.
25. Bertness, V., Kirsch, I., Hollis, G., Johnson, B., and Bunn, P. A. T-cellreceptor gene rearrangements as clinical markers of human T-cell lymphomas. N. EngI. J. Med., 313: 534—538,1985.
26. Weiss, L M., Hu, E., Wood, G. S., Moulds, C., Cleary, M., Warnke, R., andSkIar, J. Clonal rearrangements ofT-cell receptor genes in mycosis fungoidesand dermatopathiclymphadenopathy. N. EngI. J. Med., 313: 539—544,1985.
27. Waldman, T. A., Davis, M. M., Bengiovanni, F., and Korsmeyer, S. J.Rearrangementsof genes for the antigen receptor on T cells as markers oflineage and clonality in human lymphoid neoplasm& N. Engl. J. Med., 313:776—783,1985.
28. Vogelstein, B., Fearon, E. R., Hamilton, S. R., and Feinberg, A. P. Use ofrestrictionfragment length polymorphismsto determinethe clonal origin ofhuman tumors. Science (Wash. DC), 227: 642—645,1985.
29. Burton, K. A study of the conditions and mechanism of the diphenylaminereaction for the colorimetric estimation of deoxyribonucleic acid. Biochem.J., 62: 315—323,1956.
2969
abcd
@ r@@-
.
1.35—
: @-
Fig. 8. Southern blotting of fresh and fixed DNA after digestion with restriction enzymes. DNA extracted from fresh (Lanes a and c) and fixed (Lanes b and‘0samplesofauterineleiomyomawasdigestedwitheitherMspI(Lanesaandb)or HpaII (Lanes c and d). The restricted DNAs were electrophoresed, transferredto nitrocellulose, hybridized to 32P-labeled c-Ha-rat, and autoradiographed. kb,kilobase(s).
certain genesare tissue specific (seeRef. 6), methylation studiesof metastatic tumors may allow us to determine their tissue oforigin. Viral infections, which often produce nonspecific morphological changes only, can be readily detected by using theappropriate DNA probes. In certain cases, the method can beused to determine if 2 tumors arose from the sameor differentclones (28) and thus assist the pathologist in determining if theappearance of a new tumor in a cancer patient represents a newprimary or a metastasis. Perhaps even much more significant
is the fact that the importance of genetic alterations such as
geneamplification, deletion, DNA methylation, and restrictionfragment length polymorphism on tumor behavior can now bestudied retrospectively and that this is likely to provide accurateand objective methods to obtain prognostic information forpatients.
ACKNOWLEDGMENTS
We thank Dr. Para Chandrasoma from the Los Angeles CountyHospital for providing us with paraffin blocks offixed tissue specimensand Dr. Emil Bogenmann from the Children's Hospital of Los Angelesfor supplying us with the fresh leiomyoma specimen. We also thankLigaya Pen for her technical assistance.
REFERENCES
1. Klein, G. Specific chromosomal translocations and the genesis of B-cellderived tumors in mice and men. Cell, 32: 311—315,1983.
kb
2.2-1.8—
Research. on November 19, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from
1986;46:2964-2969. Cancer Res Louis Dubeau, Lois A. Chandler, Julie R. Gralow, et al. Pathology SpecimensSouthern Blot Analysis of DNA Extracted from Formalin-fixed
Updated version
http://cancerres.aacrjournals.org/content/46/6/2964
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://cancerres.aacrjournals.org/content/46/6/2964To request permission to re-use all or part of this article, use this link
Research. on November 19, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from
