Although ovarian cancer only comprises approximately 3% of all female malignancies, it accounts for a disproportionately high mortality number, making it the leading cause of death from gynecologic cancers. Serous ovarian carcinoma (SOC) comprises approximately 70% of all ovarian cancers. Its aggressive characteristics are caused by its propensity for serosal organ involvement, discovery at an advanced stage, and peritoneal spread.
There has been an ongoing debate about the pathogenesis of SOC. Many studies have focused on the ovarian surface epithelium, mainly supported by epidemiologic findings: a reduction in the number of ovulations as a result of multiple pregnancies, breast feeding, and the use of oral contraceptives lead to a decreased risk for developing ovarian cancer.1,2 Constant ovulation-induced damage and repair of the ovarian surface was suggested as the source of malignant transformation of ovarian epithelium.3,4 However, experimental or histopathological evidence is lacking, and a precursor lesion was never identified in the ovary itself. Recently, evidence has accumulated supporting the model of origin for a subset of high-grade ovarian serous carcinoma in the distal fallopian tube.5,6 Several studies identified an early form of serous carcinoma in the fimbriated end of the fallopian tube of female BRCA-mutation carriers, and this precursor lesion was termed serous tubal intraepithelial carcinoma (STIC).7–9
We recently described an alternative hypothesis regarding the site of origin for SOC in which endometrial intraepithelial carcinoma (EIC) was proposed to be the precursor lesion.10 Endometrial intraepithelial carcinoma was originally designated as the precursor lesion of uterine papillary serous carcinoma (UPSC) and was found near or adjacent to UPSC in 50% to 90% of the cases.11,12 Endometrial intraepithelial carcinoma is noninvasive, often multifocal in origin, and can also be found on the surface of the ovary and in the fallopian tube.11,13 It was presumed that coexistence of multifocal serous carcinoma of the ovaries and endometrium originated from a single rather than from multiple sites.14 Therefore, we postulate that transtubal migration of the loosely cohesive cells from EIC foci in the endometrium may be the basis for the development of any type of intraperitoneal serous carcinoma. This is strengthened by the observations that hysterectomy without oophorectomy and tubal ligation both have been associated with reductions in the risk for ovarian cancer; odds ratios have ranged from 0.03 to 0.8 for hysterectomy and from 0.2 to 0.9 for tubal ligation.15–20
Mutation and altered expression of the tumor suppressor protein p53 may be involved in the development of EIC and SOC. TP53 mutation analysis revealed identical mutations in both EIC and the concordant serous pelvic or ovarian carcinoma in a small cohort of patients, supporting a clonal origin.21 In a different study, identical TP53 mutations were shown in tumor foci from patients with peritoneal dissemination and ovarian involvement in association with a noninvasive serous endometrial lesion.14
The present study aims to further clarify the possible relationship between SOC and the presence of concurrent EIC in the endometrium by performing protein expression analysis using immunohistochemistry, TP53 mutation analysis, and in situ DNA ploidy analysis.
MATERIALS AND METHODS
Patients with unilateral or bilateral SOC, diagnosed and surgically treated at the Radboud University Nijmegen Medical Center during the period 1993 and 2009 were selected (N = 192). Without revision of the primary histology at this stage, cases were selected with (1) a noninvasive or superficially invasive serous lesion, a hyperplastic lesion with/without atypia, or EIC in the endometrium described in the original pathology report, and (2) tissue from both ovaries, the fallopian tubes in toto, and the uterus were available for revision. Thirty-eight cases fulfilled these criteria, and hematoxylin-and-eosin–stained sections of all tissues were extensively reviewed for the presence of EIC by an expert gynecopathologist (J.B.), using previously published criteria.11,22 A total of 9 cases with SOC and concurrent EIC were identified and included in this study. For each case, patients’ demographics and clinicopathological findings were determined from our institution’s medical and operative records, including age, International Federation of Gynecology and Obstetrics (FIGO) stage, and the extent of the ovarian carcinoma, lymphovascular space invasion, and the number of endometrial slides reviewed.
For assessment of protein expression by immunohistochemistry, 4-μm tissue sections were prepared from archival, formalin-fixed, paraffin-embedded tissue of both SOC and EIC and processed using standard techniques. A p53 monoclonal antibody (clone DO-7, Neomarkers, Fremont, CA) was used at a dilution of 1:250. A monoclonal antibody for Ki-67 (clone MIB-1, Dako, Glostrup, Denmark) was used at a dilution of 1:200; and for estrogen receptor (ER; clone SP-1, Thermo Scientific/Lab Vision, Fremont, CA), a monoclonal antibody was used at a dilution of 1:100. A monoclonal antibody for progesterone receptor (PR; clone PgR636, Dako) was used at a dilution of 1:500. After the application of appropriate secondary antibodies, a standard streptavidin-biotin technique (Vector Laboratories, Burlingame, CA) with diaminobenzidine as chromogen was used. All slides were counterstained with hematoxylin. Protein expression was scored as both the percentage of positively stained cells and the intensity of staining (−, negative; +, slight; ++, moderate; and +++, strong). For Ki-67, the percentage of immunostaining was assessed by average of 5 high-power fields. All slides were analyzed by 2 independent observers (T.R. and L.vK.).
Analysis of p53 Mutation Status
Both the SOC and its concordant EIC were analyzed for TP53 mutations. In all cases, normal endometrium/myometrium and normal ovarian stroma served as reference tissue to allow the identification of possible somatic mutations. Tissue sections (7 µm) were mounted on glass-foiled polyethylene naphthalate membrane–covered glass slides (Leica, Rijswijk, The Netherlands) and stained with cresyl violet (Applied Biosystems/Ambion, Austin, TX) according to suppliers’ instructions. After drying, sections were covered with PinPoint solution (ZymoResearch, Orange, CA). Laser-capture microdissection (LMD6000, Leica Microsystems, Wetzlar, Germany) was performed to isolate clusters of cells for genetic analysis. DNA isolation was performed as previously described.23 Genomic DNA was amplified by polymerase chain reaction using M13-tailed primers designed to amplify exons 5 to 8 of TP53.23 Polymerase chain reaction products were sequenced from both strands using M13 primers. Data were analyzed using Chromas software (Version 1.45, Griffith University, Queensland, Australia). Candidate mutations found by the software were compared with a reference database for cancer-associated TP53 mutations (International Agency for Research on Cancer TP53 Database, http://www-p53.iarc.fr/). Samples were scored as TP53 mutation–positive only if an identical mutation was identified in both the forward and reverse sequences.
DNA Ploidy Assessment in Tissue Sections
Measurement of the DNA ploidy status was performed as described previously.24,25 In brief, 7-μm thick paraffin-embedded tissue sections were incubated with primary antibody directed against Ki-67 (clone MIB-1, Dako). Subsequently, nuclear DNA was stoichiometrically stained using DRAQ-5 (Biostatus Limited, Leicestershire, UK). For each case, representative areas were selected by an experienced gynecopathologist (J.B.) and consisted of the following: (1) areas classified histopathologically as EIC with (2) areas containing normal endometrium/myometrium as diploid reference tissue, and (3) areas containing SOC with (4) areas containing normal ovarian stroma as diploid reference tissue. Images of DRAQ-5 and Ki-67 were acquired simultaneously using fluorescence microscopy. Automated recognition of nuclei was performed, and if required, results of automatic segmentation were interactively corrected. For nuclei in EIC and SOC, DNA index (DI) values were calculated using the respective diploid reference tissues (DI = 1.00 means diploidy). To be able to differentiate between proliferating euploid cells and aneuploid cells, only Ki-67–negative nuclei were analyzed. Those nuclei in EIC and SOC with DI greater than 1.25 were considered aneuploid. From these, the percentage of aneuploid cells (extent of aneuploidy) and the average DI value (degree of aneuploidy) were calculated.
All data analyses were performed using SPSS software (version 16.0, SPSS Inc, Chicago, IL). The Fisher exact test was used to calculate P values for association of p53 expression between EIC and SOC. To compare the extent and degree of aneuploidy between EIC and SOC, the Sign test was used.
Nine cases of previously diagnosed SOC fulfilling our criteria were identified. In each case, noninvasive or microinvasive EIC was confirmed in the uterine specimen, whereas none of the cases exhibited a STIC in their fallopian tubes after examination. In 2 of the 9 cases, EICs were partially microinvasive. However, based on the total extensiveness of the EIC in the various histopathological slides in these 2 cases, and after review by our expert gyneco-pathologist, the lesions in the endometrium did not meet the criteria for minimally invasive UPSC. Therefore, we denoted these lesions as microinvasive EIC. Table 1 summarizes the results of histopathological examination of the endometrium and fallopian tubes of the included patients. A median of 5 endometrial sections (range, 2–10) of each uterus were evaluated. The mean age of the patients at presentation was 66.4 years (range, 53–82 years). Furthermore, SOC with coinciding EIC only present in an endometrial polyp was identified in 1 case (case 6), whereas all patients with both EIC and concordant SOC had metastases to omentum, peritoneum, and/or serosa of the uterus and fallopian tubes.
Immunohistochemical findings in both intrauterine EIC and its concordant SOC are summarized in Table 2. Immunostaining for the 4 protein markers (p53, Ki-67, ER, and PR) revealed almost identical expression patterns and similar intensities in each pair of EIC and coincident SOC. In 6 of the 9 cases, both EIC and its concordant SOC showed strong nuclear staining for p53. Interestingly, in the other 3 cases, p53 immunostaining was absent in both EIC and in SOC. The probability of these findings by chance is P = 0.012. Percentages of Ki-67 nuclear staining ranged from 5% to 90%, although the percentage of positive staining was almost identical in each pair of EIC and SOC. The same holds true for the steroid hormone receptors (ER and PR); although most cases showed no expression of PR in EIC and SOC, 2 of the 9 cases showed nuclear staining for both; whereas in 6 cases, nuclear expression of ER was found in both coincident lesions. Two representative cases with their immunostaining pattern for both EIC and SOC are presented in Figure 1.
TP53 Mutations in EIC and SOC
The similarities in p53 protein expression between EIC and coinciding SOC led us to investigate whether this was a result of TP53 mutations in exons 5 to 8 and whether these mutations were similar in EIC and coinciding SOC. In all 9 cases, we were able to successfully analyze exons 5 to 8 of the TP53 gene in both the tumor and its concordant EIC. Identical TP53 mutations were found in both EIC and SOC in 3 of the 9 cases (33%; Table 3: cases 3, 7, and 8), and coincided with high p53 protein expression in these lesions (Table 2). The adjacent normal tissue in these cases did not contain mutated TP53, indicating that the mutations were EIC and tumor specific. Although the detected mutations were identical between EIC and the corresponding SOC, the type of mutations between cases was different. In 2 cases (cases 1 and 6), we found a TP53 mutation in EIC, whereas the corresponding SOC showed a wild-type TP53 sequence. In the 4 other cases, the sequencing data revealed no mutations in exons 5 to 8 in either the EIC or the ovarian carcinoma.
In situ DNA Ploidy Analysis
The DNA ploidy status of both EIC and concordant SOC were analyzed for each case. For each patient, 3 microscopic fields were measured for each lesion (EIC and SOC), with a mean of 57 nuclei measured per field (range, 19–115). Both the percentage of aneuploid cells and the degree of aneuploidy were calculated. The degree of aneuploidy was expressed as the mean amount of DNA per cell (ie, DI) compared to normal diploid endometrial and ovarian control tissue in nonproliferating Ki-67–negative cells. DNA ploidy analyses of individual nuclei demonstrated an increase in the number of aneuploid nuclei in 8 of the 9 SOCs compared to their corresponding EIC (P = 0.039; Fig. 2). In addition, the DI per nucleus in SOC was higher (ie, more aneuploid) compared to EIC (in 8 of 9 cases; P = 0.039). The combined probability of these findings by chance is P = 0.0015. When EIC contained aneuploid cells, the DNA content per cell in the associated tumor was at least comparable or higher. In one case (case 9; Fig. 2), the EIC contained more aneuploid cells with a higher DNA index compared to its coinciding ovarian tumor.
The pathogenesis of SOC in women has been subject to extensive research and controversy. The traditional view holds that ovarian cancer arises from Müllerian epithelium on the ovarian surface or from intracortical inclusion cysts; however, evidence at the clinical, histopathological, or DNA level is lacking to prove this concept. Here, we provide support for our previously described hypothesis10 that endometrial EIC is a likely precursor lesion for SOC based on immunohistochemical staining patterns, TP53 mutation, and DNA aneuploidy analyses.
Our immunohistochemistry data provide first evidence for a possible relation between EIC and coinciding SOC. We found that immunostainings for 4 proteins (p53, MIB-1, ER, and PR) revealed almost identical staining patterns and similar intensities for each pair of EIC and corresponding SOC. These findings are concordant with recently published data in which similar expression profiles were shown between EIC and extrauterine deposits in the ovaries, fallopian tube, omentum, or peritoneum.22,26,27 Our histopathologic examination and immunohistochemical data alone cannot unequivocally distinguish between a monoclonal or multicentric origin. Endometrial intraepithelial carcinoma in the uterine cavity, although without myometrial invasion, can be associated with extensive extrauterine carcinomatosis.26,28–30 Several theories have been proposed to explain the relationship between intrauterine disease and extrauterine disease, including early lymphatic spread and synchronous primary tumors (multicentricity). Endometrial intraepithelial carcinoma may present with focal lymphovascular invasion and lymphovascular metastasis may explain their extrauterine spread in some cases. However, only 3 of the 9 cases of EIC in this study with concordant SOC showed lymphovascular invasion. Importantly, EIC can present with extensive peritoneal metastases. Shedding of the (pre)malignant tumor cells into the uterine cavity after which the cells are transported through the fallopian tube lumen onto the ovaries and other pelvic peritoneal surfaces is the most likely mechanism. This is also substantiated by infrequent findings of so-called in-transit deposits of serous carcinoma cells in the fallopian tube. Alterations in cell surface adhesion molecule expression, including E-cadherin and β-catenin, is associated with the loosely cohesive nature of these (pre)malignant cells resulting in implantation at distant sites.31–33
The most widely accepted model of evolution of primary cancers to metastases is the clonal evolution model in which tumors develop by a process of linear clonal evolution driven by accumulation of somatic genetic alterations. Mutation of the TP53 tumor suppressor gene has been detected in a diverse array of tumor types and is the most commonly altered gene in human malignancies known to date.34 Although the TP53 gene consists of a total of 11 exons, approximately 90% of mutations occur in exons 5 to 8.34 Studies have already shown a high prevalence of TP53 mutations in UPSC and its precursor EIC: up to 90% and 80%, respectively.21,29 Importantly, molecular studies on a small number of cases have recently found identical TP53 mutations in uterine EIC lesions and their corresponding extrauterine tumor deposits, suggesting a clonal relationship.14,21,26 In the present study, we identified identical TP53 mutations in each pair of EIC and concordant SOC in 3 of the 9 cases (33%), highly suggestive for a clonal origin. In 2 cases, TP53 mutations were found in EIC and not in their corresponding SOC. However, both mutations have a relatively rare incidence in human tumors and more specifically in ovarian carcinoma (Table 3), and one of the mutations is a silent mutation (case 6) not causing changes in the amino acid sequence of the p53 protein. We speculate that these 2 mutations do not provide a major advantage for the malignant progression of EIC toward SOC. In 4 other cases, mutation analysis revealed no TP53 mutations in exons 5 to 8 in EIC or SOC. In addition, we expected to find TP53 mutations in those cases with immunohistochemically negative staining, most likely based on frameshift mutations. However, although approximately 90% of human TP53 mutations occur in exons 5 to 8, we cannot rule out the possibility of p53 function altering mutations within other exons of the TP53 gene, which were not investigated in this study.
It can be argued that EIC found in the endometrium represents intraepithelial spread from serous peritoneal or ovarian carcinomas. This cannot entirely be refuted on purely morphologic and immunohistochemical characteristics. DNA aneuploidy, an aberrant chromosome number, has been suggested as a useful marker for neoplastic progression of premalignant lesions at different localizations, including esophagus, skin, head and neck, and colon.35,36 It has been demonstrated previously that most of ovarian carcinomas are aneuploid.37,38 Our finding of a greater-than-diploid DNA content in nondividing epithelial cells in both EIC and SOC indicates aneuploidy in both lesions. More importantly, the severity of DNA aneuploidy in SOC was higher compared to the associated EIC in 8 of the 9 cases, suggesting an accumulation of DNA aberration.
In conclusion, it should be emphasized that the data of this study are presented as a first indication for EIC as a possible precursor lesion for SOC, based on a retrospective analysis of the endometrium with a limited number of samples and markers. In contrast, the view of ovarian carcinogenesis has been challenged lately through evidence that SOC arises in the fimbriae of the fallopian tube.39,40 This most widely accepted concept at present originated with the discovery of STIC in the fimbriae of BRCA mutation carriers. Subsequently, STICs were reported in women with ovarian or peritoneal serous carcinomas, regardless of family history.41 Based on our findings, we propose that EIC is another credible source of SOC along with fallopian tube precursors. Total analysis of the endometrium has not been part of the routine workup in patients with ovarian cancer. In most cases, limited tissue blocks of the endometrium were available or only a minimal fragment was archived. Therefore, we expect that the incidence of coexisting EIC and SOC that we report in this study is a significant underestimation. Only a prospective in-depth analysis of the total endometrium will reveal its true incidence. Despite this pitfall of our study, our data support a new concept of the origin of SOC. With the endometrium as a possible origin of SOC, it is not difficult to understand the preventive effect of tubal ligation and hysterectomy. In addition, all mechanisms that lead to “incessant” ovulation (pregnancy, breast feeding, and oral contraceptives) also have a clear impact on the endometrium and may thus interfere with the development of intrauterine premalignancies. Although the significance of the endometrium as a source of pelvic and ovarian serous carcinoma has to be further validated in a preferably larger longitudinal prospective study, these data could have clinical implications for ovarian cancer management in the future, as its precursor may reside in the endometrium.
1. Gwinn ML, Lee NC, Rhodes PH, et al.. Pregnancy, breast feeding, and oral contraceptives and the risk of epithelial ovarian cancer. J Clin Epidemiol. 1990; 43: 559–568.
2. Moorman PG, Calingaert B, Palmieri RT, et al.. Hormonal risk factors for ovarian cancer in premenopausal and postmenopausal women. Am J Epidemiol. 2008; 167: 1059–1069.
3. Cramer DW, Welch WR. Determinants of ovarian cancer risk. II. Inferences regarding pathogenesis. J Natl Cancer Inst. 1983; 71: 717–721.
4. Fathalla MF. Incessant ovulation—a factor in ovarian neoplasia? Lancet. 1971; 2: 163.
5. Crum CP, Drapkin R, Miron A, et al.. The distal fallopian tube: a new model for pelvic serous carcinogenesis. Curr Opin Obstet Gynecol. 2007; 19: 3–9.
6. Piek JM, Kenemans P, Verheijen RH. Intraperitoneal serous adenocarcinoma: a critical appraisal of three hypotheses on its cause. Am J Obstet Gynecol. 2004; 191: 718–732.
7. Callahan MJ, Crum CP, Medeiros F, et al.. Primary fallopian tube malignancies in BRCA-positive women undergoing surgery for ovarian cancer risk reduction. J Clin Oncol. 2007; 25: 3985–3990.
8. Crum CP, Drapkin R, Kindelberger D, et al.. Lessons from BRCA: the tubal fimbria emerges as an origin for pelvic serous cancer. Clin Med Res. 2007; 5: 35–44.
9. Finch A, Shaw P, Rosen B, et al.. Clinical and pathologic findings of prophylactic salpingo-oophorectomies in 159 BRCA1 and BRCA2 carriers. Gynecol Oncol. 2006; 100: 58–64.
10. Massuger L, Roelofsen T, Ham MV, et al.. The origin of serous ovarian cancer may be found in the uterus: a novel hypothesis. Med Hypotheses. 2010; 74: 859–861.
11. Ambros RA, Sherman ME, Zahn CM, et al.. Endometrial intraepithelial carcinoma: a distinctive lesion specifically associated with tumors displaying serous differentiation. Hum Pathol. 1995; 26: 1260–1267.
12. Hendrickson M, Ross J, Eifel P, et al.. Uterine papillary serous carcinoma: a highly malignant form of endometrial adenocarcinoma. Am J Surg Pathol. 1982; 6: 93–108.
13. Zheng W, Schwartz PE. Serous EIC as an early form of uterine papillary serous carcinoma: recent progress in understanding its pathogenesis and current opinions regarding pathologic and clinical management. Gynecol Oncol. 2005; 96: 579–582.
14. Kupryjanczyk J, Thor AD, Beauchamp R, et al.. Ovarian, peritoneal, and endometrial serous carcinoma: clonal origin of multifocal disease. Mod Pathol. 1996; 9: 166–173.
15. Chiaffarino F, Parazzini F, Decarli A, et al.. Hysterectomy with or without unilateral oophorectomy and risk of ovarian cancer. Gynecol Oncol. 2005; 97: 318–322.
16. Cramer DW, Xu H. Epidemiologic evidence for uterine growth factors in the pathogenesis of ovarian cancer. Ann Epidemiol. 1995; 5: 310–314.
17. Hankinson SE, Hunter DJ, Colditz GA, et al.. Tubal ligation, hysterectomy, and risk of ovarian cancer. A prospective study. JAMA. 1993; 270: 2813–2818.
18. Irwin KL, Weiss NS, Lee NC, et al.. Tubal sterilization, hysterectomy, and the subsequent occurrence of epithelial ovarian cancer. Am J Epidemiol. 1991; 134: 362–369.
19. Loft A, Lidegaard O, Tabor A. Incidence of ovarian cancer after hysterectomy: a nationwide controlled follow up. Br J Obstet Gynaecol. 1997; 104: 1296–1301.
20. Rosenblatt KA, Thomas DB. Reduced risk of ovarian cancer in women with a tubal ligation or hysterectomy. The World Health Organization Collaborative Study of Neoplasia and Steroid Contraceptives. Cancer Epidemiol Biomarkers Prev. 1996; 5: 933–935.
21. Baergen RN, Warren CD, Isacson C, et al.. Early uterine serous carcinoma: clonal origin of extrauterine disease. Int J Gynecol Pathol. 2001; 20: 214–219.
22. Jarboe EA, Miron A, Carlson JW, et al.. Coexisting intraepithelial serous carcinomas of the endometrium and fallopian tube: frequency and potential significance. Int J Gynecol Pathol. 2009; 28: 308–315.
23. Blokx WA, Ruiter DJ, Verdijk MA, et al.. INK4-ARF and p53 mutations in metastatic cutaneous squamous cell carcinoma: case report and archival study on the use of Ink4a-ARF and p53 mutation analysis in identification of the corresponding primary tumor. Am J Surg Pathol. 2005; 29: 125–130.
24. Fleskens SJ, Takes RP, Otte-Holler I, et al.. Simultaneous assessment of DNA ploidy and biomarker expression in paraffin-embedded tissue sections. Histopathology. 2010; 57: 14–26.
25. van der Avoort IA, van de Nieuwenhof HP, Otte-Holler I, et al.. High levels of p53 expression correlate with DNA aneuploidy in (pre)malignancies of the vulva. Hum Pathol. 2010; 41: 1475–1485.
0. For the complete list of references, please contact T.Roelofsen@obgyn.umcn.nl.
Keywords:Copyright © 2012 by IGCS and ESGO
Endometrial intraepithelial carcinoma; Ovarian carcinoma; Lesion of origin; Precursor