A case-matched molecular comparison of extraovarian versus primary ovarian adenocarcinoma

1587

A Case-Matched Molecular Comparison of Extraovarian versus Primary Ovarian Adenocarcinoma Lynn D. Kowalski, M.D.1 Anisa I. Kanbour, M.D.2 Fredric V. Price, M.D.1 Sydney D. Finkelstein, M.D.2 Wayne A. Christopherson, M.D.1 Jan C. Seski, M.D.3 Gregory J. Naus, M.D.2 Judith A. Burnham, H.T.L.2 Amal Kanbour-Shakir, M.D., Ph.D.2 Robert P. Edwards, M.D.1

BACKGROUND. Extraovarian mu¨llerian adenocarcinoma (EOM) resembles primary ovarian carcinoma (POC) both histologically and clinically, yet little is known regarding the molecular genetic characteristics of this entity. The objective of this study was to compare the expression of three molecular markers of tumor behavior in EOMs and POCs.

METHODS. Forty-four patients meeting strict criteria for EOM were identified and

1

Department of Obstetrics, Gynecology, and Reproductive Sciences, University of Pittsburgh Health Sciences Center, Magee-Womens Hospital, Pittsburgh, Pennsylvania. 2

Department of Pathology, University of Pittsburgh Health Sciences Center, Magee-Womens Hospital, Pittsburgh, Pennsylvania. 3

Department of Obstetrics and Gynecology, Medical College of Pennsylvania, Hahnemann University, Allegheny Campus, Pittsburgh, Pennsylvania.

matched to POC controls for age, stage, tumor histology and grade, cytoreductive surgery, and survival. Immunohistochemistry was used to determine overexpression of p53 and HER-2/neu. DNA content was evaluated by flow cytometry. Direct DNA sequencing of exons 5–8 of the p53 gene was performed in nine EOM tumors. Statistical comparisons were made using chi-square, Kaplan-Meier, and Mantel-Cox log rank methods.

RESULTS. Overexpression of HER-2/neu was demonstrated in 59% (26 of 44) of the EOM group versus 36% overexpression (16 of 44) in the POC controls (P Å 0.05). Overexpression of p53 was noted in 48% of the EOM cases, similar to the 59% incidence observed in the control group (P Å 0.29). Missense mutations were found in 9 of 9 EOM tumors showing strong p53 nuclear immunostaining. No significant difference in the incidence of aneuploidy was observed when EOM cases were compared with POC controls (65% vs. 63%). High tumor grade was strongly associated with HER-2/neu overexpression in the EOM group (P Å 0.002). None of the parameters studied were predictive of prognosis within the EOM and POC groups. CONCLUSIONS. Although overexpression of p53 protein, p53 gene mutations, and abnormal DNA content were similar between EOMs and POCs, EOMs demonstrated almost twice the rate of HER-2/neu overexpression. This result suggests that distinct genetic events may be responsible for malignant transformation in EOMs versus POCs. Cancer 1997;79:1587–94. q 1997 American Cancer Society.

Abstract previously published in Gynecologic Oncology 1996;60:167. Presented in part at the Annual Meeting of the United States and Canadian Academy of Pathology, Washington, DC, March 23–29, 1996. The authors would like to thank Patricia A. Swalsky and Anke Bakker for their expert technical assistance with the p53 genotyping. Address for reprints: Lynn D. Kowalski, M.D., Department of Obstetrics and Gynecology, Washington University School of Medicine, 4911 Barnes Hospital Plaza, St. Louis, MO 63110. Received September 26, 1996; revision received December 17, 1996; accepted December 17, 1996.

KEYWORDS: ovarian neoplasms, peritoneal neoplasms, immunohistochemistry, p53 protein, c-erb B-2 protooncogene protein, flow cytometry, aneuploidy.

A

lthough histologically identical to primary epithelial tumors of the ovary, extraovarian mu ¨ llerian carcinomas (EOM) present with disseminated intraperitoneal carcinomatosis and normal-sized or absent ovaries. Considered rare in the past, recognition of EOM has increased in recent years, with 9 – 18% of laparotomies now performed for ovarian carcinoma revealing an EOM malignancy.1 – 4 Because all mu ¨ llerian structures derive from evagination of coelomic epithelium early in embryonic life,5 the epithelial layer of the ovary and the peritoneum share a common embryologic heritage. In addition, both EOM and primary ovarian carcinomas (POC) display similar aggressive clinical behavior and an associated poor prognosis.1 – 3 Therefore, a biologic distinction between EOM and POC remains obscure,

q 1997 American Cancer Society

/ 7b54$$0989

03-14-97 16:19:49

cana

W: Cancer

1588

CANCER April 15, 1997 / Volume 79 / Number 8

and, to date, the molecular genetic characteristics of EOM are poorly understood. Epithelial ovarian malignancies, like many other human carcinomas, are thought to arise from alterations in tumor suppressor genes and/or protooncogenes. The current literature supports the concept that transformation to the tumor phenotype is a stepwise accumulation of genetic events that lead to unregulated cell growth.6 These events have not yet been delineated for epithelial ovarian carcinomas, but advanced ovarian malignancies commonly demonstrate evidence of marked growth deregulation. Specifically, the most widely studied genetic alterations in POC are overexpression of the protooncogene HER-2/neu, overexpression and mutation of the tumor-suppressor gene p53, and tumor aneuploidy. The HER-2/neu protooncogene encodes a transmembrane tyrosine kinase growth factor receptor that is homologous to the epidermal growth factor receptor.7 In several human cancers and cancer cell lines, gene amplification rather than mutation has been shown to result in protein overexpression and thus malignant transformation.8 – 10 Approximately 25 – 30% of patients with breast and ovarian carcinoma overexpress HER-2/neu, and overexpression in both carcinomas has been associated with poor survival.8,11 In addition, immunohistochemistry in paraffin embedded tissue has been shown to be a reliable means of assessing protein overexpression and thus gene amplification.12 p53 exerts transcriptional control over several aspects of cell cycle regulation. Mutations in the p53 gene are the most common genetic aberration in all human cancers, found in at least 50% of tumors studied.13 Most p53 mutations have been found in exons 5 – 8, a highly conserved region of the gene, and lead to conformational changes that confer increased stability to the p53 gene product.14 As a result, protein overexpression, easily and reliably detected by immunohistochemistry, has been shown to correlate with missense mutations of the p53 gene.15 Approximately 50 – 60% of POCs test positive for p53 overexpression by immunohistochemistry.16,17 Tumor aneuploidy refers to abnormal multiples of cellular DNA content. It is well established that aneuploid tumors, whether determined by static or flow cytometric analysis, are associated with aggressive biologic behavior compared with tumors without demonstrable aneuploid cell populations.18 From 1989 to date, 386 primary ovarian tumors were evaluated for DNA content and entered into the study database at Magee-Womens Hospital. Of these, 183 (47.4%) had demonstrable aneuploid populations, and this proportion serves as the reference point for the analysis of ploidy in the current study.

/ 7b54$$0989

03-14-97 16:19:49

cana

To the authors’ knowledge, investigations of HER2/neu overexpression, p53 overexpression, and DNA ploidy in extraovarian malignancies are limited. This study, therefore, compares the relative frequencies of these tumor-associated genetic aberrations in EOMS and POCs.

MATERIALS AND METHODS Identification of Index Cases and Controls Sixty-two patients with a diagnosis of EOM were identified from the Magee-Womens Hospital Tumor Registry, the Department of Pathology database, and the database of a private gynecologic oncologist’s practice (W.A.C.). All cases were reviewed by two pathologists (A.K. and A.K.S.). Criteria for consideration included: 1) if one or both ovaries were present at the time of diagnosis they were normal in size, 2) gross assessment at the time of laparotomy revealed absent to no minimal ovarian surface involvement, with the bulk of disease involving peritoneal surfaces or omentum, and 3) on histologic evaluation, the tumor was mu ¨ llerian in origin and either did not involve the ovaries at all or only superficially (no evidence of cortical invasion). All tumors of mu ¨ llerian histology (i.e., papillary serous, clear cell, endometrioid, and mucinous subtypes) were included. Forty-six cases met criteria for inclusion, and tissue blocks were available for 44 cases. Reasons for exclusion included nine tumors of nonmu ¨ llerian histology, four patients who never underwent surgical evaluation, three patients who, on pathologic review, had primary ovarian tumors, and two patients with papillary serous lesions identified in the endometrium. All cases were matched to epithelial ovarian controls with regard to age, stage, histologic subtype and grade, residual tumor volume after cytoreductive surgery, and clinical outcome. Matching was performed in a rigorous fashion to achieve case and control groups that, in their clinical and pathologic characteristics, appeared as identical as possible. Clinicopathologic Data Clinical data were obtained from hospital and office records. Survival was defined as the number of months after histologic diagnosis to either date of death or date of last follow-up. Patients alive at the time of data analysis (June 30, 1995) were categorized by their disease status as either alive with no evidence of disease or alive with disease. Patients were staged according to International Federation of Gynecology and Obstetrics guidelines for ovarian carcinoma. Tumor grading was based on standard pathologic definitions regarding nuclear and histopathologic criteria. The

W: Cancer

Extraovarian Mullerian Adenocarcinoma/Kowalski et al.

high grade designation was assigned to poorly differentiated tumors and tumors of clear cell histology. Well differentiated and moderately differentiated lesions were designated low grade. Suboptimal debulking was defined as ú 2 cm of residual tumor volume remaining after initial cytoreductive surgery.

Immunohistochemistry p53 and HER-2/neu overexpression were evaluated utilizing a labeled streptavidin (LSAB II; Dako Co., Carpinteria, CA) technique. After pathologic review of all cases and controls, two paraffin embedded blocks from each case were selected for staining. Four-micron tissue sections, placed on positive ion-charged slides, were stained. The slides were then deparaffinized, treated with 3% hydrogen peroxide in methanol solution for 20 minutes to block endogenous peroxidases, and then washed in distilled water. A microwave antigen retrieval procedure using citrate buffers (BioGenix, San Ramon, CA) was performed on all slides. Slides were washed in Tris buffer for 10 minutes, then placed in a protein-blocking reagent for 15 minutes (Immunon; Shandon-Lipshaw, Pittsburgh, PA). All slides were incubated in primary antibody for 1 hour. For p53, a mouse antihuman p53 monoclonal antibody (p53, DO7; Dako Co.) was used that recognizes an epitope in the N-terminus of the human p53 protein (dilution 1:100). The mouse monoclonal antibody NCL-CB11 (Novocastra Laboratories, Vector Laboratories, Burlingame, CA, distributor), which binds specifically to the internal domain of the HER-2/neu oncoprotein, was selected for its proven effectiveness on formalin fixed, paraffin embedded tissue (dilution 1:40). The sections were covered with biotinylated antimouse immunoglobulins (Ig) for 15 minutes, followed by a Streptavidin-peroxidase complex step for 15 minutes (Dako Co.). Slides were then developed with the enzyme 3*3* diaminobenzidine (Sigma Chemical Co., St. Louis, MO). Slides were counterstained with Gill’s III hematoxylin, dehydrated, and mounted. All slides were reviewed by the primary author and a single pathologist (L.D.K. and A.K.). A hematoxylin and eosin stained section was examined for each block, and a negative control slide using nonspecific mouse IgG substituted for the primary antibody was performed on all blocks. Background staining was negligible. A positive control slide using infiltrating ductal breast carcinomas known to stain intensely for either HER-2/neu or p53 was used for each staining sequence. In addition, three cases of benign serous cystadenomas were incubated with each antibody with no staining noted. To reduce bias in assignment of status, each slide was examined on two separate occasions, and exami-

/ 7b54$$0989

03-14-97 16:19:49

cana

1589

nation was performed blinded to the primary tumor site (EOM vs. POC). For HER-2/neu, membrane staining was graded on a scale of 0 to 3: 0: no staining of tumor cells; 1: cytoplasmic staining; 2: light membrane staining; and 3: intense membrane staining. A grade of 0 or 1 on both sections was recorded as a negative response in the database. A grade of 2 or 3 in either or both sections was recorded as a positive response. p53 status was assessed based on the proportion of tumor cells with nuclear staining, rather than the intensity of staining.19 A few cases showed equivocal staining, and positive status was only assigned if ú50% of the tumor cells on a single slide were stained.

p53 Genotyping To determine the genetic basis for p53 overexpression as assessed by positive p53 immunohistochemical staining, a subset of nine tumors displaying strong and widespread staining were further evaluated by p53 mutational genotyping with polymerase chain reaction (PCR)/DNA sequencing as recently described.20 Tumor and normal tissue was scraped from 4-mm thick serial sections under stereomicroscopic control. Collected tissue was placed into microcentrifuge tubes, and exons 5 – 8 were directly PCR amplified in individual PCR reactions using primers as previously described. Amplified DNA was isolated by the GeneClean method (Bio101, La Jolla, CA) and then directly sequenced by dideoxy chain termination with S-35-dATP (United States Biochemical Corp., Cleveland, OH) using one of the amplifying oligonucleotides. Sequences were read from overnight-exposed autoradiograms of 6% polyacrylamide gels. Flow Cytometry Flow cytometric DNA content analysis was performed as previously described.21 Briefly, selected paraffin sections were deparaffinized in xylene, rehydrated through graded alcohols, and incubated in 0.5% pepsin (Sigma Chemical Co.) at pH 1.5 for 30 minutes at 37 7C. Suspensions of cells and nuclei were filtered through a 74-mm nylon mesh and incubated with RNAse for 30 minutes at 37 7C. The suspensions were then centrifuged and resuspended in propidium iodide (50 mg/mL) in 10 mM Tris buffer with 5 mM MgCl2 and 1 mg/mL sodium azide. DNA (propidium iodide) specific fluorescence of at least 30,000 nuclei was measured using an EPICS Profile flow cytometer (Coulter Corp., Hialeah, FL). Resultant histograms were analyzed for DNA content, with the most aberrant histogram being used for tumor characterization. Aneuploid cell populations were classified based on their DNA index (DI), and were subdivided into subtriploid (DI õ 1.5), hypertriploid (DI Å 1.5 – 1.85),

W: Cancer

1590

CANCER April 15, 1997 / Volume 79 / Number 8

TABLE 1 Clinicopathologic Characteristics of the Extraovarian Carcinoma and Primary Ovarian Carcinoma Study Groups

Median age (yrs) Tumor histology Serous Grade 1 Grade 2 Grade 3 Clear cell Endometrioid Mucinous Mixed Tumor stage Stage II Stage III Stage IV Median survival (mos)

EOM

POC

66 (range, 37–90)

65 (range, 30–80)

33 (75%) 2 (4.5%) 10 (23%) 21 (47.7%) 7 (15.9%) 2 (4.5%) 1 (2.3%) 1 (2.3%)

33 (75%) 3 (6.8%) 10 (23%) 20 (45%) 6 (13.6%) 2 (4.5%) 1 (2.3%) 2 (4.5%)

5 (11.4%) 34 (77.3%) 5 (11.4%) 27.8

5 (11.4%) 35 (79.5%) 4 (9.0%) 25

FIGURE 2. Immunohistochemical staining for the HER-2/neu oncoprotein in extraovarian mu¨llerian carcinoma demonstrating intense membrane specific staining.

EOM: extraovarian mu¨llerian carcinoma; POC: primary ovarian carcinoma.

FIGURE 1. Kaplan–Meier survival curve for matched extraovarian mu¨llerian carcinomas (EOM, solid line) and primary ovarian carcinomas (POC,

FIGURE 3. Immunohistochemical staining for the p53 oncoprotein in extraovarian mu¨llerian carcinoma demonstrating intense nuclear staining.

hatched line).

and hypertetraploid (DI ú 2.15). Remaining tumors were classified as diploid (no attempt was made to identify tetraploid cell populations and tetraploid tumors were classified in the diploid tumor category).

Statistical Analysis Statistical analysis was performed using Stat-View software, Version 4.5 (Abacus Concepts, Berkeley, CA). Chi-square and Fisher’s exact tests were used to test correlation of nominal variables. Survival curves were generated using the Kaplan – Meier product limit estimation and the Mantel – Cox test was used for compar-

/ 7b54$$0989

03-14-97 16:19:49

cana

ison of survival curves. Differences were considered statistically significant when P ° 0.05.

RESULTS Matching of Cases and Controls The clinicopathologic characteristics of the POC controls and the EOM cases are shown in Table 1. Statistical analysis revealed the two groups were well matched for survival as displayed in Figure 1. Approximately 89% of EOM cases (39 of 44) presented with advanced (Stage III or IV) disease; median survival was 27.8 months. Twenty-eight of the 44 EOM cases (64%) were high grade.

W: Cancer

Extraovarian Mullerian Adenocarcinoma/Kowalski et al.

Expression of HER-2/neu Protein Membrane specific staining with monoclonal antibody directed toward the HER-2/neu gene product was found in 59.1% (26 of 44) of EOM cases (Fig. 2). In contrast, HER-2/neu overexpression was only found in 36.4% (16 of 44) of matched POC controls (P Å 0.05). With respect to tumor histology, HER-2/neu overexpression was distributed among several cell types of EOM. Protein overexpression was detected in 19 of 33 papillary serous tumors (58%), 5 of 7 clear cell tumors (71%), 1 of 2 endometrioid tumors (50%), 0 of 1 mucinous tumors (0%), and 1 of 1 mixed cell type tumors (100%). HER-2/neu overexpression was also detected in different subtypes among the POC controls: 11 of 33 of papillary serous histology (33%), 3 of 6 of clear cell histology (50%), 1 of 1 of mucinous histology (100%), 0 of 2 of endometrioid histology (0%), and 1 of 2 of mixed type tumors (50%). Because EOM cases and POC controls were matched for grade, and HER-2/ neu overexpression was observed in several histologic subtypes within each group, differences in histology cannot explain the higher incidence of overexpression in the EOM cases. When the EOM cases were evaluated for survival, HER-2/neu overexpression was not correlated with prognosis (overall survival was poor, even for HER-2/ neu negative patients). However, high grade histology was significantly correlated (P Å 0.002) with HER-2/ neu overexpression in the EOM group (22 of 28). Other parameters that were analyzed with respect to HER2/neu status included positive family history for a firstdegree relative with breast or ovarian carcinoma, achievement of optimal surgical debulking, age õ 65 years, and increased preoperative CA 125. No significant associations were found by chi-square analysis. Expression of p53 Protein and Direct Sequencing of p53 Mutations Nuclear specific staining with a monoclonal antibody recognizing the p53 protein was measured in 47.7% of EOM cases and in 59.1% of POC controls (Fig. 3). The relative frequency of p53 overexpression was not significantly different (P Å 0.29) between the two groups and the observed frequency in POC controls was in accordance with previous studies utilizing immunohistochemistry.17,22 However, when analyzed for its relationship to overall survival, p53 overexpression was not correlated with prognosis for the patients with EOM (P Å 0.77). Separate analysis for the POC controls with regard to p53 overexpression also did not reveal a significant relationship (P Å 0.21). p53 overexpression was analyzed with respect to 2-, 3-, and 5-year survival and no significant prognostic correlation was detected for

/ 7b54$$0989

03-14-97 16:19:49

cana

1591

TABLE 2 p53 Mutations in Extraovarian Mu¨llerian Carcinomas Tumor

p53 mutation

Exon

10282 6563 6676 9326 13440 11347 11552 3318 8863

216 VM 273 RH 175 RH 284 TK 245 GC 213 RG 278 PR 273 RH 220 YC

6 8 5 8 7 6 8 8 6

either group. In addition, no significant associations were found when analyzing p53 status and tumor grade, family history, surgical debulking, age õ 65 years, or elevated preoperative CA 125. Direct sequencing of exons 5 – 9 was performed on nine cases that demonstrated strong nuclear staining with anti-p53 antibody. All showed missense mutations in p53 coding sequence (Table 2). Two tumors showed a mutation at codon 273, which is a previously established ‘‘hot spot’’ for p53 mutations in POCs.13 In fact, eight of the nine EOM cases sequenced contained mutations that were identical to published mutations in POC. A representative example of p53 mutational genotyping is shown in Figure 4. The authors did not sequence any of the matched POC controls because a large number of historic controls describing a wide range of mutations are available in the literature for comparison. The authors acknowledge that the true frequency of p53 alterations (sequence deletions, protein truncating mutations, and homozygous deletion) probably exceeds the current measurements due to the authors’ reliance on enzyme immunohistochemistry. Direct sequencing of EOM tumors not positive by immunohistochemistry may show genetic alterations that result in absence of p53 protein expression, as has been shown in the POC literature.15,23

Flow Cytophotometric Analysis No significant difference in the frequency of aneuploidy was identified in EOM cases (64.7%) compared with POC controls (62.8%). Analysis of the relative frequencies of aneuploidy subtypes also showed a striking similarity between EOM and POC tumors (Fig. 5). In addition, tumor DNA content did not correlate with histologic grade or any clinical parameter analyzed in this investigation.

DISCUSSION Defining the molecular makeup of EOM is important for several reasons: 1) elucidating the molecular events

W: Cancer

1592

CANCER April 15, 1997 / Volume 79 / Number 8

FIGURE 5. Distribution of ploidy categories among extraovarian mu¨llerian carcinomas (shaded boxes) and primary ovarian carcinomas (open boxes).

TABLE 3 Relative Frequencies (%) of Molecular Parameters in Extraovarian Mu¨llerian Carcinomas and Primary Ovarian Carcinomas

HER-2/neu overexpressiona p53 overexpressiona Aneuploidy

FIGURE 4. p53 mutational genotyping. Polymerase chain reaction-amplified DNA for exon 6 was directly sequenced using dideoxy chain termination. Note the presence of a misense p53 point mutation affecting codon 220 in Patient 1 (arrow indicates G to A base pair substitution leading to tyrosine replacement by cystine). Samples 2 and 3 were from normal controls.

of tumorigenesis in mu ¨ llerian malignancies; 2) classifying this disease as a subset of epithelial ovarian carcinomas versus assigning it a unique identity; and 3) determining independent predictors of survival and response to therapy. The results of the current study indicate that there are shared genetic events contributing to the malignant phenotype in EOM and POC tumors, but that there is also a unique distinction. Immunohistochemical analysis revealed no difference in overexpression of p53, but demonstrated a significant difference between cases versus controls in overexpression of HER-2/neu Table 3. Although the design of this study did not attempt to address the events leading to tumorigenesis in EOMs, this result strongly suggests that a distinction exists between EOM and POC tumors. The fact that the majority of EOM tumors overexpress HER-2/neu may have impli-

/ 7b54$$0989

03-14-97 16:19:49

cana

EOM

POC

P value

59.1 47.7 64.7b

36.4 59.1 62.8c

0.05 0.29 0.73

EOM: extraovarian mu¨llerian carcinoma; POC: primary ovarian carcinoma. a N Å 44 for both extraovarian mu¨llerian carcinomas and primary ovarian carcinomas. b N Å 34. c N Å 43.

cations in the understanding of tumor biology and in patient management. Several recent studies have reported success in utilizing anti-HER-2/neu gene therapy to suppress the growth of primary ovarian cell lines known to overexpress HER-2/neu,24,25 and the Gynecologic Oncology Group is currently conducting a Phase II trial testing anti-HER-2/neu antibody therapy for recurrent ovarian carcinomas that overexpress HER-2/neu.26 In addition, overexpression of HER-2/ neu may predict a poorer response to platinum-based chemotherapy,27,28 although this result remains to be confirmed in the clinical setting. A clear limitation of the current study was the lack of uniformity in chemotherapy dosing, scheduling of patient follow-up, and definition of recurrent disease between practitioners, which precluded any meaningful assessment of platinum response or disease free survival. Correlation of these clinical parameters with HER-2/neu status needs to be examined in future studies. The significant correlation of HER-2/neu positivity

W: Cancer

Extraovarian Mullerian Adenocarcinoma/Kowalski et al.

with high tumor grade supports many lines of evidence in both animal models and clinical correlation studies that overexpression enhances metastatic potential.8,29 – 31 The overwhelming majority of EOM tumors are either poorly or moderately differentiated.3,16,32 The high frequency of HER-2/neu overexpression in high grade EOM tumors suggests a significant relationship between HER-2/neu-mediated growth deregulation and the histologically malignant phenotype. In contrast to the HER-2/neu data, there was significant similarity in p53 status between EOM and POC. The number of shared identical p53 mutations (eight of nine) suggests similar interactions with endogenous or exogenous mutagens and/or the selection of tumor clones with growth advantages13 conferred by particular p53 mutations. This would not be surprising in light of the frequency of p53 mutations or alterations in many human tumors. The clinical implications of p53 overexpression warrants some discussion. In the EOM cases, p53 overexpression as detected by immunohistochemistry was not predictive of overall survival, or 2-, 3-, or 5-year survival. The POC tumors in the current study showed similar results. Whether a complete p53 mutational analysis of the EOM tumors would yield a more compelling association with survival remains speculative; however, in the literature regarding POC, which includes reports of p53 mutational analysis, the data do suggest a prognostic relationship. A report from the current study institution found that p53 mutation is a significant predictor of death at 3 and 5 years,20 but other groups using both immunohistochemical and sequencing methods have published both corroborating22,33 and conflicting17,34 data. The incidence of flow cytometrically determined aneuploidy was similar in case and control groups, even when analyzed for aneuploidy subtypes. Other studies have shown a correlation between p53 mutations and tumor aneuploidy,16,35,36 presumably through loss of the mitotic spindle checkpoint responsible for the faithful distribution of the normal DNA content to daughter cells. Therefore, it is not surprising that in the current study the frequencies of p53 overexpression and tumor aneuploidy were similar in both EOM cases and POC controls. The authors are not aware of any other direct molecular comparative studies between EOM and POC, but certainly other molecular parameters are worthy of study. However, there are numerous reports in the literature that examine the molecular composition of POC. Allelotyping studies of POC have reported frequent loss at 5q,37 11q,38 17q,39 13q,40 and frequent deletion of chromosome 9.41 In addition, alterations

/ 7b54$$0989

03-14-97 16:19:49

cana

1593

in genes other than p53 and HER-2/neu have been reported in POC; K-ras mutations are found in some mucinous and low malignant potential (LMP) tumors,42,43 and amplification of c-myc occurs in approximately 30% of ovarian carcinomas, with a greater frequency in the serous subtype.44 To date, there is no ‘‘ovarian carcinoma gene’’ and, in fact, the multitude of genetic alterations found in POC attests to the genetic instability and heterogeneity of these tumors. This makes a comprehensive genetic characterization of both EOM and POC nearly impossible. Further analysis of shared and distinct molecular genetic characteristics may help further define these two entities. In summary, EOM shares common p53 mutations and frequency of aneuploidy with POC but demonstrates twice the frequency of HER-2/neu overexpression. Although these two diseases share a similar natural history and poor survival, the need for specific therapies based on tumor biology warrants further investigation.

REFERENCES 1.

Killackey M, Davis AR. Papillary serous carcinoma of the peritoneal surface: matched-case comparison with papillary serous ovarian carcinoma. Gynecol Oncol 1993;51:171–4. 2. Fromm GL, Gershenson DM, Silva EG. Papillary serous carcinoma of the peritoneum. Obstet Gynecol 1990;74:89–95. 3. Bloss JD, Liao S-Y, Buller RE, Manetta A, Berman ML, McMeekin S, et al. Extraovarian peritoneal serous papillary carcinoma: a case-control retrospective comparison to papillary adenocarcinoma of the ovary. Gynecol Oncol 1993; 50:347–51. 4. Altaras MM, Aviram R, Cohen I, Cordoba M, Weiss E, Beyth Y. Primary peritoneal papillary serous adenocarcinoma: clinical and management aspects. Gynecol Oncol 1991; 40:230–6. 5. Stenchever M. Embryology. In: Droegemueller W et al., editors. Comprehensive gynecology. St. Louis: C.V. Mosby Company, 1988:11–8. 6. Fearon E, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61:759–67. 7. Gullick WJ. The role of the epidermal growth factor receptor and the c-erbB-2 protein in breast cancer. Int J Cancer 1990;(Supp 5):55–61. 8. Slamon DJ, Godolphin W, Jones LA, Holt JA, Wong SG, Keith DE, et al. Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science 1989;244:707–12. 9. Hudziak RM, Schlessinger J, Ullrich A. Increased expression of the putative growth factor receptor p185HER2 causes transformation and tumorigenesis of NIH 3T3 cells. Proc Natl Acad Sci USA 1987;84:7159–63. 10. DiFiore PP, Pierce JH, Fleming TP, Hazan R, Ullrich A, King CR, et al. Overexpression of the human EGF receptor confers an EGF-dependent transformed phenotype to NIH 3T3 cells. Cell 1987;51:1063–70. 11. Berchuk A, Kamel A, Whitaker R, Kerns B, Olt G, Kinney R, et al. Overexpression of HER-2/neu is associated with poor survival in advanced epithelial ovarian cancer. Cancer Res 1990;50:4087–91.

W: Cancer

1594

CANCER April 15, 1997 / Volume 79 / Number 8

12. Corbett IP, Henry JA, Angus B, Watchorn CJ, Wilkinson L, Hennessy C, et al. NCL-CB11, a new monoclonal antibody recognizing the internal domain of the c-erb B-2 oncogene protein effective for use on formalin-fixed, paraffin-embedded tissue. J Pathol 1990;161:15–25. 13. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 1994; 54:4855–78. 14. Finlay CA, Hinds PW, Tan T-H. Activating mutations for transformation by p53 produce a gene product that forms an hsc70-p53 complex with an altered half-life. Mol Cell Biol 1988;8:531–9. 15. Kupryjanczyk J, Thor AD, Beauchamp R, Merritt V, Edgerton SM, Bell DA, et al. p53 gene mutations and protein accumulation in human ovarian cancer. Proc Natl Acad Sci USA 1993;90:4961–5. 16. Kihana T, Tsuda H, Teshima S, Okada S, Matsuura S, Hirohashi S. High incidence of p53 gene mutation in human ovarian cancer and its association with nuclear accumulation of p53 protein and tumor DNA aneuploidy. Jpn J Cancer Res 1992;83:978–84. 17. Marks JR, Davidoff AM, Kerns BJ, Humphrey PA, Pence JC, Dodge RK, et al. Overexpression and mutation of p53 in epithelial ovarian cancer. Cancer Res 1991;51:2979–84. 18. Naus GJ. Cytophotometric DNA content analysis in malignancies of the female genital tract. In: Sciarra JJ, editor. Gynecology and obstetrics. Philadelphia: J.B. Lippincott, 1990:1–9. 19. Hall PA, Lane DP. p53 in tumour pathology: can we trust immunohistochemistry?-revisited! J Pathol 1994;172:1–4. 20. Finkelstein S, Przygodzki R, Pricolo V. p53/K-ras-2 topographic genotyping of colorectal adenocarcinoma. Mol Diag 1996;1:5–28. 21. Naus G, Simmerman R. Prognostic value of flow cytophotometric DNA content analysis in single treatment stage IB to IIA squamous cell carcinoma of the cervix. Gynecol Oncol 1991;43:149–53. 22. Henriksen R, Strang P, Wilander E, Backstrom T, Tribukait B, Oberg K. p53 expression in epithelial ovarian neoplasms: relationship to clinical and pathological parameters, Ki-67 expression, and flow cytometry. Gynecol Oncol 1994;53:301– 6. 23. Milner BJ, Allan LA, Eccles DM, Kitchener HC, Leonard RCF, Kelly KF, et al. p53 mutation is a common event in ovarian carcinoma. Cancer Res 1993;53:2128–32. 24. Deshane J, Siegal GP, Alvarez RD, Wang M, Cabrera G, Feng M, et al. A strategy to accomplish gene therapy for ovarian carcinoma based upon targeted tumor cell killing via an intracellular antibody directed against the erB-2 oncoprotein. Gynecol Oncol 1996;60(A):115–16. 25. Hung MC, Matin A, Zhang Y, Xing X, Sorgi F, Huang L. HER2/neu-targeting gene therapy-a review. Gene 1995;159:65– 71. 26. Bookman M, Karlan B, Birrer M. A Phase II evaluation of anti-HER2 antibody in recurrent or refractory ovarian cancer. Protocol GOG #160. Gynecology Oncology Group 1996. 27. Pietras RJ, Fendly BM, Chazin VR, Pegram MD, Howell SB, Slamon DJ. Antibody to HER-2/neu receptor blocks DNA repair after cisplatin in human breast and ovarian cancer cells. Oncogene 1994;9:1829–38. 28. Arteaga CL, Winnier AR, Poirier MC, Lopez-Larraza DM,

/ 7b54$$0989

03-14-97 16:19:49

cana

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

Shawver LK. p185c-erB-2 signaling enhances cisplatin-induced cytotoxicity in human breast carcinoma cells: association between an oncogenic receptor tyrosine kinase and druginduced DNA repair. Cancer Res 1994;54:3758–65. Yu D, Hung M-C. Expression of activated rat neu oncogene is sufficient to induce experimental metastasis in 3T3 cells. Oncogene 1991;6:1991–6. Yu D, Hamada J, Zhang H, Nicolson GL, Hung M-C. Mechanisms of c-erb B2/neu oncogene-induced metastasis and repression of metastatic properties by adenovirus 5 E1A gene products. Oncogene 1992;7:2263–70. Yu D, Wang S-S, Dulski KM, Tsai C-M, Nicolson GL, Hung M-C. C-erbB-2/neu overexpression enhances metastatic potential of human lung cancer cells by induction of metastasis-associated properties. Cancer Res 1994;54:3260–6. Bloss J. GOG Protocol #138: Phase II clinical trial of cisplatinum and cytoxan in extraovarian serous adenocarcinoma. Gynecologic Oncology Group Protocol 1995. Eltabbakh G, Belinson J, Kennedy A, Biscotti C, Casey G, Tubbs R. p53 expression in primary epithelial ovarian carcinoma: relationship to age and survival. Gynecol Oncol 1996; 60(A):146. Ikenberg H, Hollerbach E, Sauerbrei W, Pfisterer J, Kommoss F, Kiechle M, et al. p53 protein overexpression is not associated with poor prognosis in patients with ovarian carcinoma. Gynecol Oncol 1996;60(A):155. Cross SM, Sanchez CA, Morgan CA, Schimke MD, Ramel S, Edzerda RL, et al. A p53-dependent mouse spindle checkpoint. Science 1995;267:1353–6. Livingstone LR, White A, Sprouse J, Livanos E, Jacks T, Tisty T. Altered cell cycle arrest and gene amplification potential accompany loss of wild-type p53. Cell 1992;70:923–35. Tavassoli M, Steingrimsdottir H, Pierce E, Hiang X, Alagoz M, Farzaneh F, et al. Loss of heterozygosity on chromosome 5q in ovarian cancer is frequently accompanied by TP53 mutation and identifies a tumour suppressor gene locus at 5q13.1-21. Br J Cancer 1996;74:115–9. Gabra H, Watson J, Taylor K, MacKay J, Leonard R, Steel C, et al. Definition and refinement of a region of loss of heterozygosity at 11q23.3-q24.3 in epithelial ovarian cancer associated with poor prognosis. Cancer Res 1996;56:950–4. Chen K, Hsieh T, Schwartz P, Yang-Feng T. Allelic deletion mapping of putative tumor suppressor genes on 17q in sporadic ovarian cancer. J Obstet Gynecol 1995;21:619–24. Yang-Feng T, Han H, Chen K, Li S, Claus E, Carcangiu M, et al. Allelic loss in ovarian cancer. Int J Cancer 1993;54:546– 51. Devlin J, Elder P, Gabra H, Steel C, Knowles M. High frequency of chromosome 9 deletion in ovarian cancer: evidence for three tumour-suppressor loci. Br J Cancer 1996; 73:420–3. Mok S, Bell D, Knapp R, Fishbaugh P, Welch W, Muto M, et al. Mutation of K-ras protooncogene in human ovarian epithelial tumors of borderline malignancy. Cancer Res 1993;53:1489–92. Ichikawa Y, Nishida M, Suzuki H, Yoshida S, Tsunoda H, Kubo T, et al. Mutation of K-ras protooncogene is associated with histological subtypes in human mucinous ovarian tumors. Cancer Res 1994;54:33–5. Tashiro H, Miyazaki K, Okamura H, Iwai W, Fukomoto M. c-myc over-expression in human primary ovarian tumours: its relevance to tumour progression. Int J Cancer 1992; 50:828–33.

W: Cancer