507

Molecular Genetic Analysis of Clear Cell Adenocarcinomas of the Vagina and Cervix Associated and Unassociated with Diethylstilbestrol Exposure In Utero

Jeff Boyd, Ph.D.l”

Hiroyuki Takahashi, M.D., Ph.D.”’

Love11 A. Jones, Ph.D.’

Steven E. Waggoner, M.D!

Richard A. Hajek, Ph.D5

J. Taylor Wharton, M.D? Fu-shing Liu, M.O.‘ Takafumi Fujino, M.D., P m . ‘ J. Carl Barrett, Phn’

John A. McLachlan, Ph.D.3

’ Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania.

‘ Laboratory of Molecular Carcinogenesis, Na- tional Institute of Environmental Health Sci- ences, National Institutes of Health, Research Triangle Park, North Carolina.

‘ Laboratory of Reproductive and Develop- mental Toxicology, National Institute of Environ- mental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina.

’’ Section of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Chi- cago Medical Center, Chicago, Illinois.

” Department of Gynecologic Oncology, Univer- sity of Texas M. D. Anderson Cancer Center, Houston, Texas.

Presented at the 26th Annual Meeting of the Society of Gynecologic Oncologists, San Fran- cisco. California, February 1995.

Address for reprints: Jeff Boyd, Ph.D., Depart- ment of Obstetrics and Gynecology, University of Pennsylvania Medical Center, 778 Clinical Re- search Building, 41 5 Curie Boulevard, Philadel- phia, PA 19104.

Received July 12, 1995; revision received Octo- ber 4. 1995; accepted October 4, 1995.

BACKGROUND. Prenatal exposure to the synthetic estrogen diethylstilbestrol (DES) is associated with the subsequent development of clear cell adenocarcinoma of the lower reproductive tract in young women, and data concerning the molecular genetic alterations involved in the etiology of this tumor type have not previously been reported. Such knowledge would be of potential value by providing insight into the molecular mechanisms of hormonal carcinogenesis in general, as well as by suggesting molecular markers for risk assessment in the estrogen-exposed population. METHODS. A total of 24 samples of clear cell adenocarcinoma of the vagina or cervix, 16 associated with exposure in utero to DES and 8 with no history of DES exposure, were obtained as archival fixed and embedded tissue specimens. DNA was purified from these tissues and used to examine a number of biologically plausible molecular genetic endpoints for tumor specific alterations. RESULTS. No evidence was found for mutations in the K - m s or H-rus protoonco- genes, the Wilms’ tumor (WT1) tumor suppressor gene, or the estrogen receptor gene. Sporadic overexpression of the ,053 tumor suppressor gene was detected in some tumor cell nuclei by inimunohistochemistry, but in the absence of detectable ~753 gene mutation. Genetic instability as manifested by somatic mutation of niicro- satellite repeats was widespread in these tumors, with evidence of microsatellite instability in all DES-associated tumors examined, and in 50% of those tumors not associated with DES exposure. CONCLUSIONS. These data are consistent with the hypothesis that the induction of genoniic instability may be an important mechanism of DES-induced carcino- genesis. Cancer 1996; 77507- 13. 6 Arnericuri Cancer Societ),.

KEYWORDS vaginal neoplasm, cervical neoplasm, estrogen, diethylstilbestrol, DNA repair, ras, p53, WT1, estrogen receptor, mutation.

iethylstilbestrol (DES) is a synthetic, nonsteroidal estrogen that was D administered for pregnancy support to approximately two million women during the 1940s through the 1960s.’ The association of exposure in utero to DES and the development in young women of clear cell adeno- carcinoma of the vagina, an otherwise rare tumor, was recognized in 1971.‘ This tumor may also manifest in the cervix; the majority of cases occur between the ages of 17 and 21 years and affect approximately 0.01% of developmentally exposed women.3 The tumor is histologically indistin- guishable from the clear cell adenocarcinoma of the endometrium or ovary seen in older women,4 and it generally displays a higher metastatic

L 1996 American Cancer Society

508 CANCER February 1, 1996 I Volume 77 I Number 3

potential than the more typical squamous cell carcinoma of the vagina or cervix."'

The strong association of vaginal clear cell adenocar- cinoma in young women with DES exposure and the po- tential of DES to induce vaginal adenocarcinoma in devel- opmentally exposed rodents' implicate DES directly in the etiology of this tumor. The mechanism of DES carci- nogenicity is unknown; experimental studies suggest sev- eral possibilities, including somatic mutation of DNA by reactive metabolites,R.'' induction of aneuploidy through disruption of the microtubule spindle apparatus,'0"' and epigenetic effects on such processes as tumor promotion and the reprogramming of gene expression." The pres- ence of aneuploidy in some human cervicovaginal clear cell adenocar~inomas'~ and in the cervicovaginal epithe- lium of mice treated neonatally with estrogen14 suggests that the induction of chromosome instability may be a feature of estrogenic carcinogens. We have identified widespread gene amplification in mouse uterine adeno- carcinomas induced by DES,Is but to our knowledge, there are no published studies on the molecular genetic analysis of human tumors associated with DES exposure.

Such data would be potentially useful for several rea- sons, primarily in terms of furthering our understanding of the mechanisms of hormonal carcinogenesis. Estrogen is believed to contribute to the tumorigenic process for several common human cancers, including those of the endometrium and breast.I6 Although the epidemiologic data linking estrogen-related hormonal, reproductive, and menstrual events with risk for these cancers are well accepted, it is far less clear how estrogens exert their pathogenic effects. One approach to this problem is to address the hypothesis that estrogen-associated malig- nancies will share particular molecular genetic features. This strategy is hindered by the multifactorial, multigenic nature of human carcinogenesis," and thus the inherent difficulty of attributing any individual cancer case to es- trogen exposure. The clear cell adenocarcinomas from DES-exposed women provide a rare opportunity to par- tially circumvent this obstacle.

Additionally, this type of study could lead to a molec- ular marker for risk assessment in the DES-exposed popu- lation. Although the public health hazards associated with further exposure to DES have been largely eliminated, there are a number of compelling reasons for the contin- ued study of this patient population. For example, it is not apparent that the cancer incidence resulting from DES exposure has peaked. Although the majority of DES daughters have passed the age range for cervicovaginal clear cell adenocarcinoma development, few have reached the age range in which endometrial carcinoma typically occurs in the unexposed population, and endo- metrial carcinoma is seen at a much higher frequency than cervical or vaginal carcinoma in DES-treated female

mice.','* Similarly, breast cancer may still be a concern for this pop~1ation.l~ Finally, the identification of such markers would be of value in predicting risk for second- generation DES offspring (grandchildren), for whom little can be currently stated concerning potential health risk.20,2'

The purpose of this study was to examine biologically plausible molecular genetic endpoints for evidence of mutational alterations in human DES-associated cervico- vaginal malignancies. The ras and p53 genes are the most commonly mutated oncogene and tumor suppressor gene, respectively, in human cancers in general, and are also known genetic targets for chemical carcinogens in the human and rodent.",23 The Wilms tumor (WT1) tumor suppressor gene plays a critical role in the normal devel- opment of the reproductive tract, and germline mutations in this gene predispose to structural anomalies similar to those seen in DES daughters and sons,24 Germline varia- tion in the estrogen receptor gene could also confer sus- ceptibility to the tumorigenic action of DES; a number of human estrogen receptor variants have been described that may predispose to pathologic conditions, including ~ancer . '~ -~ ' Finally, genetic instability of microsatellite re- peat sequences has been observed in many human can- cer types, including estrogen-associated human endome- trial carcinoma^.'^^^"

MATERIALS AND METHODS All tumors utilized in this study were obtained as formalin fixed, paraffin embedded tissue specimens. A total of 24 tumors processed between 1965 and 1991 were retrieved; original pathology reports indicating clear cell adenocar- cinoma of the cervix or vagina were verified through the examination of hematoxylin and eosin-stained tissue sec- tions by at least two independent pathologists. Sixteen of these tumors were from patients with medical histories positive for DES exposure in utero, and the remaining 8 were from patients with medical histories nonspecific or negative for DES exposure. Eight of 16 DES-associated tumors were obtained from the surgical pathology ar- chives at the University of Texas M.D. Anderson Cancer Center in Houston. These tumors were used exclusively for the analysis of ras, p53, and WT1. The remaining eight DES-associated tumors and eight tumors unassociated with DES were obtained from the Clear Cell Adenocarci- noma Registry, University of Chicago, Chicago, Illinois. These tumors were used for the estrogen receptor and microsatellite instability analyses.

A stained tissue section from each block was used to determine the area of highest tumor cellularity for DNA preparation. One or more tissue plugs were then removed from each block with a punch biopsy tool, minced with a razor blade, and deparaffinized with a series of rinses in xylenes and ethanol. The tissue was digested com-

Molecular Genetics of DES-Induced CancerdBoyd et al. 509

pletely in TE buffer (pH 8.01, containing 400 pg/mL of proteinase K replenished daily and 0.5% sodium dodecyl sulfate, for 3-4 days at 54 "C. Following a standard series of phenol, phenollchloroform, and chloroform extrac- tions, the DNA was ethanol-precipitated, washed in 70% ethanol, resuspended in TE buffer (pH 8.0), and stored at -20 "C.

The mutational analyses of rus, p53, and WTl genes were performed on eight specimens of clear cell adeno- carcinoma of the vagina or cervix, all from patients with documented exposure to DES, obtained from the M.D. Anderson Cancer Center. Codons 12, 13, and 61 of the K-ras and H-rus genes were examined for all 38 known activating mutations using polymerase chain reaction (PCR) amplification of genomic DNA, slot-blotting, and radiolabeled allele specific oligonucleotide hybridization as previously described." DNA from human endometrial carcinoma cell lines with known ras gene mutations were used as positive controls.

Exons 4-9 of the p53 gene were examined for muta- tions by PCR amplification of genomic DNA using intron- based primers followed by single strand conformation polymorphism (SSCP) analysis as has been described.3z DNA from human gynecologic tumor cell lines with known mutations of the p53 gene were used as positive controls for each exon. Immunohistochemical analysis of p53 expression was performed on 6-pm thick tissue sections prepared from the same paraffin blocks used for the genetic analyses. Slides were deparaffinized, treated with 0.1 % trypsin for 5 minutes, and endogenous peroxi- dase was blocked with a 10-minute incubation in 0.3% HIO1. Subsequent steps of incubation with normal se- rum, incubation with primary and secondary antibodies, antibody detection, and counterstaining were carried out using a Vectastain Elite ABC kit according to instructions provided by the manufacturer. Incubation with the pri- mary antibody was for 16 hours at 4 "C in a humidified chamber, using a 1:lOO dilution of the mouse antihuman p53 monoclonal antibody DO-7 (Novacastra Laboratories Ltd., Newcastle upon Tyne, UK).33 Slides were rinsed be- tween all steps in Automation Buffer (Biomeda Corpora- tion, Foster City, CA).

Exons 4- 10 for the WT1 gene were examined for mu- tations by PCR amplification of genomic DNA using in- tron-based primers. Primer sequences were based on the published exonlintron sequence structure,34 and SSCP analysis was performed as described above. Variant bands were excised from gels and subjected to sequence analysis as previously described.35

The molecular genetic and immunohistochemical analyses of the ras, p53, and WT1 genes described above depleted the available tissue for most of the eight tumors used for these studies. Subsequent studies of the estrogen receptor gene and genetic instability were carried out

with a second set of 16 tumors obtained through the Uni- versity of Chicago Clear Cell Adenocarcinoma Registry. Eight of these tumors were from patients with exposure to DES in utero and eight were from patients with equivi- cal or undocumented exposure to DES.

The estrogen receptor gene was examined for muta- tion and polymorphic sequence variation by PCR ampli- fication of genomic DNA, using intron-based primers sur- rounding exons 1-8. Primer sequences were based on the published exonlintron sequence str~cture.~' SSCP analysis and variant sequence analysis were performed as de~cribed.'~

Genetic instability of microsatellite repeat sequences was assessed using a panel of 15 markers (5 dinucleotide, 5 trinucleotide, and 5 tetranucleotide repeats) as summa- rized in Table 1; all PCR primers were obtained from Research Genetics (Huntsville, AL). Microsatellite insta- bility was quantitated by PCR amplification of paired nor- mal and tumor DNA samples from the same individual, polyacrylamide gel electrophoresis, and comparison of product mobilities as previously described."

RESULTS Eight tumor specimens of clear cell adenocarcinoma of the vagina or cervix arising in women with exposure to DES in utero were analyzed for mutations in the rus, p53, and WT1 genes. Codons 12, 13, and 61 of the K-ras and H-rus protooncogenes were of the wild-type sequence only in all tumors (Fig. 1). Mutations in these codons were detected in all positive control DNA samples from tumors with known mutations that were analyzed con- comitantly. Furthermore, we have previously shown that the allele specific oligonucleotide technique used for rus gene analysis is extremely sensitive in detecting mutant alleles in DNA from fixed and embedded endometrial hyperplasia specimens that frequently have a low tumor to normal cell ratio.37

An SSCP analysis of exons 4-9 of the p53 tumor sup- pressor gene revealed products with wild-type mobility only in all 8 tumors (Fig. 2). All positive control tumor DNA samples with known mutations in each of these exons were detected as mobility shifts on SSCP gels. Using this same procedure, we have previously detected p53 mutations in the majority of human gynecologic sarco- mas tested.3z Immunohistochemical analysis of p53 pro- tein levels revealed a sporadic pattern of expression; five tumors exhibited strong nuclear overexpression in a small percentage (1-10%) of cells, with the majority of tumor cells showing no detectable p53 protein. Three tumors were completely negative for detectable p53 protein. Pos- itive control tissue sections from tumors with known p53 mutations, one from a chemically-induced mouse mam- mary adenocarcinoma and one from a human endome-

510 CANCER February 1,1996 I Volume 77 / Number 3

TABLE 1 Mierosatellite Instability in Clear Cell Adenocarcinomas

DESt Tumors DES- Tumors

1 3 4 5 9 12 14 15 2 6 7 8 10 11 13 16

Site' Age' Marker Dinucleotide

D?S119 D2S123 D5S117 D10S197 D l 6 1 6 4

Trinucleotide CTTR CTTlG

D2S422 D3S1752

W F D1S53? D4S243 D7S460 D16S310

DM-1

Tetranucleotide

v v v v v c v I: V C V V C V c c 19 22 18 18 23 18 25 16 11 83 30 26 28 77 24 24

O r 0 0 0 0 0 0 0 0 0 0 0 0 (3 0 c 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 I 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 (3 0

DES: di~thylsrilbestrol; if: vagina; C ten% closed circles: instabiliy: open circles: no instability * Sik of tumor origin. ' Age at diagnosis

trial carcinoma, both exhibited strong nuclear overex- pression in the majority of tumor cells.

SSCP and sequence analyses of exons 4-10 of the WTl tumor suppressor gene revealed no mutations in these 8 tumors. One tumor contained a sequence variant in exon 6 that did not alter the encoded amino acid; the presence of this variant in normal tissue DNA from the same individual is consistent with its classification as a rare polymorphism.

Sixteen additional cases of clear cell adenocarcinoma of the vagina or cervix, 8 of which were from patients with exposure to DES in utero, were examined for mutations or polymorphisms in the estrogen receptor gene and for evidence of genetic instability of microsatellite repeat se- quences. SSCP and sequence analyses of the entire estro- gen receptor coding region revealed known polymor- phisms in exons 1, 3, and 8, but no somatic or germline mutations. The polymorphisms were present at a fre- quency approximating that known to exist in the normal population. Notably, there were no tumor cases which contained the "B-region" polymorphic variant in exon 1 that has been associated with an increased risk of sponta- neous abortion.24

A survey of 15 microsatellite markers revealed evi- dence of genetic instability in 12 of 16 clear cell adenocar-

FIGURE 1. Mutation analysis of the ras protooncogenes. Autoradiogram on left indicates the presence of the wild-type allele (gly) for K-ras codon 12 in all tumor specimens. Autoradiogram on right indicates the absence of one possible activating mutation (asp) in all tumor specimens, with DNA from a positive control tumor cell line ( + C ) known to contain this particular mutation. The same experiment was repeated for all 38 activating mutations known to occur in codons 12, 13, and 61 of the K-ras and H- ras genes. -C: negative control derived from a polymerase chain reaction containing water instead of DNA.

cinomas examined (Table 1). Notably, all of the DES-asso- ciated cases were among those with microsatellite insta- bility. The extent of instability among tumors varied from 1 of 15 markers to 10 of 15 markers, whereas 4 of 8 tumors not associated with DES exposure displayed instability in

Molecular Genetics of DES-Induced CancerslBoyd et al. 51 1

FIGURE 2. Mutation screening of the p53 gene. Autoradiogram of single strand conformation polymorphism gel for exon 7 shows an invariant product mobility for six tumor specimens, with a positive control (+C) known to contain a point mutation in exon 7 showing altered gel mobility. -C: negative control derived from a polymerase chain reaction containing water instead of DNA.

from 1 of 15 to 6 of 15 markers examined. Instability was defined by altered mobility of a PCR product in tumor compared with normal DNA from the same individual, or by additional products of smaller or larger size present in the tumor compared with normal DNA (Fig. 3). No additional correlations between microsatellite instability and clinical or histopathologic characteristics of these tu- mors were observed.

DISCUSSION These data indicate that genetic instability of microsatel- lite repeat sequences is common in clear cell adenocarci- nomas of the vagina or cervix, especially those arising in women with a history of DES exposure in utero. This phenomenon has been described in a variety of different sporadic cancers and in virtually all of those associated with the hereditary nonpolyposis colon cancer syn- drome.38.3!l The somatic mutation of these polymorphic repeat sequences throughout the cancer cell genome is believed to reflect defective DNA mismatch repair. Inher- ited mutation in one of a family of four genes encoding DNA mismatch repair proteins is responsible for heredi- tary nonpolyposis colon and nuclear extracts from cancer cells harboring mutant alleles of these genes are defective in mismatch r e ~ a i r . ~ ’ - ~ ~

Although the number of tumor specimens examined in the present study was small, an unusually large per- centage (75%) of cervicovaginal clear cell adenocarcino- mas exhibited microsatellite instability, suggesting that altered DNA repair may represent a critical molecular feature of this tumor type. Consistent with this hypothesis are previous studies describing microsatellite instability in endometrial carcinoma^,'^^^^ especially those that may be categorized as type I, or estrogen-related tumors.’8 Further studies will be required to determine if this type of genetic instability is present in all DES-related cancers, as our data would suggest, and if “sporadic” and DES- induced clear cell cancers of the vagina and cervix share

FIGURE 3. Examples of microsatellite instability in diethylstilbestrol- associated clear cell adenocarcinomas. Genornic DNA samples from paired normal (N) and tumor (T) tissues were subjected to polymerase chain reaction amplification using the microsatellite repeat markers 028119 (A), 02S422 (B), and 0381752 (C). Arrows indicate alterations in the electro- phoretic mobility of polymerase chain reaction products from tumor corn- pared with normal tissue DNA.

512 CANCER February 1,1996 / Volume 77 / Number 3

a common molecular genetic profile. The common occur- rence of microsatellite instability in these cancers also provides a candidate genetic target(s) for the carcinogenic action of DES. A mutational analysis of the DNA mis- match repair genes would allow direct testing of this hy- pothesis. Alternatively, microsatellite instability in these tumors may reflect an indirect effect of estrogenic stimu- lation, with DNA replication errors accumulating as a re- sult of inappropriate cell division. This scenario would be consistent with the development of these tumors in women apparently unexposed to DES. Under this hy- pothesis, the clear cell cancers and accompanying micro- satellite instability would result from aberrant cell divi- sion caused by another factor, hormonal or otherwise.

Our data have also ruled out several plausible candi- date genes as direct or indirect targets of DES carcinoge- nicity. We could find no evidence for mutation of the rus or p53 genes, both of which are involved frequently in human and rodent carcinogenesis. The sporadic overex- pression of p53 protein that we observed in some tumor cells is likely a manifestation of the cellular response to unrepaired DNA damage.'' '' A more extensive survey of p53 mutation and expression status in clear cell adeno- carcinomas of the vagina and cervix supports this hypoth- e s i ~ . ~ ~ Additionally, these results indicate that the WT1 tumor suppressor gene, known to play a role in genitouri- nary tract development, is not mutationally altered in these tumors, and that polymorphisnis or mutations of the estrogen receptor gene do not underlie the suscepti- bility of some women to DES-induced carcinogenesis.

In conclusion, these results provide clues to the etiol- ogy of estrogen-induced human cancers. The direct or indirect induction of unrepaired DNA replication errors may represent a critical component of estrogen-induced carcinogenesis. The well documented potential of DES to damage DNA directly is consistent with the hypothesis that genes encoding DNA repair proteins may represent mutational targets of DES-induced mutagenesis. In addi- tion to the four hereditary nonpolyposis colon cancer- associated genes, tumor specific mutations have been de- scribed in the DNA polymerase p gene in human colo- rectal and prostate and in the DNA poly- merase 6 gene in colorectal carcinoma^.'^ To test this hypothesis, a mutation analysis of these genes in DES- associated clear cell adenocarcinomas is underway.

REFERENCES 1. Noller KL, Fish CR. Diethylstilbestrol usage: its interesting

past. important present, and questionable future. Med Cliri Norrlz A m 1974;58:793-810. Herbst AL, Ulfelder H, Poskanzer DC. Adenocarcinoma of the vagina: association of maternal stilbestrol therapy with tumor appearance in young women. N Erigl J Med

Melnick S, Cole P, Anderson D, Herbst A. Rates and risks of

2.

1971; 2841878-81, 3.

diethylstilbestrol-related clear cell adenocarcinoma of the vagina and cervix: an update. N Engl J Med 1987;316:514- 6. Sedlis A, Robboy SJ. Diseases of the vagina. In: Kurman RJ, editor. Blaustein's pathology of the female genital tract. 3rd edition. New York: Springer-Verlag, 1987:97- 140. Robboy SJ, Herbst AL, Scully RE. Clear cell adenocarcinoma of the genital tract in young females: analysis of 37 tumors that persisted or recurred after primary therapy. Cancer

Herbst AL, Norusis MJ, Rosenow PJ, Welch WR, Scully RE. An analysis of 346 cases of clear cell adenocarcinoma of the vagina and cervix with emphasis on recurrence and survival. Gyizecol 01zcol 1979; 7:111-22. Newbold RR, McLachlan JA. Vaginal adenosis and adenocar- cinoma in mice exposed prenatally or neonatally to diethyl- stilbestrol. Cancer Res 1982;42:2003- 11. Metzler M, McLachlan ]A. Peroxidase-mediated oxidation, a possible pathway for metabolic activation of diethylstilbes- trol. Biochem Biopliys Res Corizmun 1978; 85:874-84. Liehr JG, Randerath K, Randerath E. Target organ-specific covalent DNA damage preceding diethylstilbestrol-induced carcinogenesis. Curcinogenesis 1985;6:1067-9.

10. Tsutsui T, Maizumi H, McLachlan ]A, Barrett JC. Aneuploidy induction and cell transformation by diethylstilbestrol: a possible chromosome mechanism in carcinogenesis. Curzcer Res 1983; 43:3814-21.

11. Tucker RW, Barrett JC. Decreased numbers of spindle and cytoplasmic microtubules in hamster embryo cells treated with diethylstilbestrol. Cancer Res 1986; 46:2088-95.

12. Barrett JC. Molecular mechanisms of hormonal carcinogen- esis. In: Li JJ , Nandi S, Li SA, editors. Hormonal carcinogene- sis. New York Springer-Verlag, 1992:159-63.

13. Welch WR, Fu YS, Robboy SJ, Herbst AL. Nuclear DNA con- tent of clear cell adenocarcinoma of the vagina and cervix and its relationship to prognosis. Gyrzecol Orzcol

14. Hajek RA, Van NT, Johnston DA, Jones LA. In vivo induction of increased DNA ploidy of mouse cervicovaginal epithelium by neonatal estrogen treatment. Biol Reprod 1993;49:908-

4.

5.

1974;34:606- 10. 6.

7.

8.

9.

1983; 15:230-8.

.- I / .

15. Risinger 11, Terry LA, Boyd J. Use of representational differ- ence analysis for the identification of mdinZ oncogene am- plification in diethylstilbestrol-induced murine uterine ade- nocarcinonias. Mol Carcinog 1994; 11:13-8.

16. Henderson BE, Ross R, Bernstein L. Estrogens as a cause of human cancer: the Richard and Hinda Rosenthal Founda- tion award lecture. Cancer Res 1988; 48:246-53.

17. Boyd J, Barrett JC. Genetic and cellular basis of multistep carcinogenesis. Plzarnzncol Tlier 1990; 46:469-86.

18. Newbold RA, Bullock BC, McLachlan ]A. Uterine adenocarci- noma in mice following developmental treatment with es- trogens: a model for hormonal carcinogenesis. Cancer Res

19. Colton T, Greenberg ER, Noller K, Resseguie L, Van Ben- nekom C, Heeren T, et at. Breast cancer in mothers pre- scribed diethylstilbestrol in pregnancy. JAMA

20. Turusov VS, Trukhanova LS, Parfenov YD, Tomatis L. Occur- rence of tumours in the descendants of CBA male mice pre- natally treated with diethylstilbestrol. Iizt J Carzcer

1990; 50:7677-81.

1993;269:2096-100.

1992;50:131-5.

Molecular Genetics of DES-Induced CancerslBoyd et al. 51 3

21. Forsberg J-G, Halling A. Failure to detect a second-genera- tion effect in female mice after neonatal treatment with an estrogen (diethylstilbestrol). Acta Anat (Basel) 1992; 144:

22. Bos JL. ras oncogenes in human cancer: a review. Cancer Res 1989;49:4682-9.

23. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Muta- tions in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res

24. Hastie ND. Dominant-negative mutations in the Wilms tu- mour (WT1) gene cause Denys-Drash syndrome-proof that a tumour-suppressor gene plays a crucial role in normal genitourinary development. Hum Mol Genet 1992; 1293-5.

25. Lehrer S, Sanchez M, Song HK, Dalton J, Levine E, Savaretti P, et al. Oestrogen receptor B-region polymorphism and spontaneous abortion in women with breast cancer. Lancet

26. Karnik PS, Kulkarni S, Liu X-P, Budd GT, Bukowski RM. Es- trogen receptor mutations in tamoxifen-resistant breast cancer. Cancer Res 1994;54:349-53.

27. Smith EP, Boyd 7, Frank GR, Takahashi H, Cohen RM, Specker B, et al. Estrogen resistance caused by a mutation in the estrogen-receptor gene in a man. N Engl J Med 1994; 33 1: 1056-61.

28. Risinger 71, Berchuck A, Kohler MF, Watson P, Lynch HT, Boyd J. Genetic instability of microsatellites in endometrial carcinoma. Cancer Res 1993;53:5100-3.

29. Burks RT, Kessis TD, Cho KR, Hedrick L. Microsatellite insta- bility in endometrial carcinoma. Oncogene 1994; 9: 1163-6.

30. Duggan BD, Felix JC, Muderspach LI, Tourgeman D, Zheng J, Shibata D. Microsatellite instability in sporadic endometrial carcinoma. J Nut1 Cancer lnst 1994;86:1216-21.

31. Ignar-Trowbridge D, Risinger JI, Dent GA, Kohler MF, Ber- chuck A, McLachlan JA, et al. Mutations of the Ki-ras onco- gene in endometrial carcinoma. Am J Obstet Gynecol

32. Liu F-S, Kohler MF, Marks JR, Bast RC Jr, Boyd J, Berchuck A. Mutation and overexpression of the p53 tumor suppressor gene frequently occurs in uterine and ovarian sarcomas. Obstet Gynecol 1994;83:118-24.

33. Vojtesek B, Bartek J, Midgley CA, Lane DP. An immunohisto- chemical analysis of the human nuclear phosphoprotein p53: new monoclonal antibodies and epitope mapping using recombinant p53.J lmmunol Methods 1992; 151:237-44.

34. Gessler M, Konig A, Bruns GAP. The genomic organization and expression of the WT1 gene. Genomics 1992; 122307- 13.

35. Risinger 11, Berchuck A, Kohler MF, Boyd 1. Mutations of the E-cadherin gene in gynecologic cancers. Nut Genet

36. Ponglikitmongkol M, Green S, Chambon P. Genomic organi- zation of the human oestrogen receptor gene. EMBO J

37. Sasaki H, Nishii H, Takahashi H, Tada A, Furusato M, Tera- shima Y, et al. Mutation of the Ki-ras protooncogene in hu- man endometrial hyperplasia and carcinoma. Cancer Res

38. Loeb LA. Microsatellite instability: marker of a mutator phe- notype in cancer. Cancer Res 1994;54:5059-63.

103-6.

1994;54:4855-78.

1990; 335:622-4.

1992; 1671227-32.

1994;7:98-102.

1988;7:3385-8.

1993; 53:1906- 10.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

Modrich P. Mismatch repair, genetic stability, and cancer. Science 1994; 266: 1959-60. Leach FS, Nicolaides NC, Papadopoulos N, Liu B, Jen J , Par- sons R, et al. Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell 1993; 75:1215-25. Fishel R, Lescoe MK, Rao MRS, Copeland NG, Jenkins NA, Garber J, et al. The human mutator gene homolog MSHZ and its association with hereditary nonpolyposis colon can- cer. Cell 1993; 75:1027-38. Papadopoulos N, Nicolaides NC, Wei Y-F, Ruben SM, Carter KC, Rosen CA, et al. Mutation of a mutL homolog in heredi- tary colon cancer. Science 1994;263:1625-9. Bronner CE, Baker SM, Morrison PT, Warren G, Smith LG, Lescoe MK, et al. Mutation in the DNA mismatch repair gene homologue hMLHl is associated with hereditary non- polyposis colon cancer. Nature 1994;368:255-8. Nicolaides NC, Papadopoulos N, Liu B, Wei Y-F, Carter KC, Ruben SM, et al. Mutations of two PMS homologues in he- reditary nonpolyposis colon cancer. Nature 1994; 371:75-80. Liu B, Parsons RE, Hamilton SR, Petersen GM, Lynch HT, Watson P, et al. hMSH2 mutations in hereditary nonpolypo- sis colorectal cancer kindreds. Cancer Res 1994; 54:4590-4. Nystrom-Lahti M, Parsons R, Sistonen P, Pylkaanen L, Aalto- nen LA, Leach FS, et al. Mismatch repair genes on chromo- somes 2p and 3p account for a major share of hereditary nonpolyposis colorectal cancer families evaluable by link- age. A m 1 H u m Genet 1994;55659-65. Parsons R, Li G-M, Longley MJ, Fang W-H, Papadopoulous N, Jen J, et al. Hypermutability and mismatch repair defi- ciency in RER+ tumor cells. Cell 1993; 75:1227-36. Umar A, Boyer JC, Thomas DC, Nguyen DC, Risinger JI, Boyd J, et al. Defective mismatch repair in extracts of colorectal and endometrial cancer cell lines exhibiting microsatellite instability. 1 Biol Chem 1994; 269:14367-70. Koi M, Umar A, Chauhan DP, Cherian SP, Carethers JM, Kunkel TA, et al. Human chromosome 3 corrects mismatch repair deficiency and microsatellite instability and reduces N-methyl-M-nitro-N-nitrosoguanidine tolerance in colon tumor cells with homozygous hMLHl mutation. Cancer Res

Kastan MB, Onyekwere 0, Sidransky D, Vogelstein B, Craig RW. Participation of p53 protein in the cellular response to DNA damage. Cancer Res 1991;51:6304-11. Kessis TD, Slebos RJ, Nelson WG, Kastan MB, Plunkett BS, Han SM, et al. Human papillomavirus 16 E6 expression dis- rupts the p53-mediated cellular response to DNA damage. Proc Nut1 Acad Sci USA 1993;90:3988-92. Waggoner SE, Anderson SA, Luce MC, Takahashi H, Boyd J. P53 protein expression and gene analysis in clear cell adenocarci- noma of the vagina and cervix. Gynecol Oncol. In press, 1996. Wang L, Patel U, Ghosh L, Banerjee S. DNA polymerase p mutations in human colorectal cancer. Cancer Res

Dobashi Y, Shuin T, Tsuruga H, Uemura H, Torigoe S, Ku- bota Y. DNA polymerase p gene mutation in human prostate cancer. Cancer Res 1994;54:2827-9. da Costa LT, Liu B, el-Deiry WS, Hamilton SR, Kinzler KW, Vogelstein B. Polymerase 6 variants in RER colorectal tu- mours. Nut Genet 1995;9:10-1.

1994;54:4308-12.

1992; 52~4824-7.