Possible involvement of loss of heterozygosity in malignant transformation of ovarian endometriosis

Abstract

Objective

Although histological and epidemiological studies have suggested that endometriosis has malignant potential, the molecular mechanism underlying the malignant transformation of endometriosis is poorly understood. Ovarian cancer arising from endometriosis (OCEM) may provide an ideal model for genetic studies. To investigate the genetic alterations during transformation of ovarian endometriosis into cancer, we analysed loss of heterozygosity (LOH) and mutations of tumour-related genes in OCEM cases (n = 12) that fulfilled the histological criteria and in solitary ovarian endometriosis (n = 12).

Methods

Each paraffin-embedded section was microdissected to isolate the endometriotic epithelium, transitional epithelium, and cancer cells. Extracted DNA was amplified by nested PCR. LOH was identified with fluorescence-labelled microsatellite markers. Mutations of tumour-related genes PTEN, TP53, and K-ras were examined by bidirectional DNA sequencing.

Results

With 13 microsatellite markers on six chromosomes, we detected 31 and eight LOH events in OCEMs and in solitary ovarian endometriosis, respectively. In the OCEM group, 18 LOH events were found in cancer cells alone, while 13 LOH events were found in all cancer, transitional, and endometriotic tissues. High frequencies of LOH were detected on chromosomes 9p, 10q, and 13q. However, only two point mutations were detected in cancer cells in one case.

Conclusion

Our study suggested that accumulated LOH events on some chromosome loci may be involved in malignant transformation of endometriosis, while mutations in certain tumour-related genes are not.

Introduction

Endometriosis is defined as the presence of ectopic endometrial tissue outside of the uterine cavity, most commonly in the ovaries and peritoneum. It is a common gynecological disorder, affecting 10−15% of all women of reproductive age [1], [2]. Although endometriosis generally follows a benign course, its malignant potential and its specific association with ovarian cancer have long been suspected. Several large follow-up studies have indicated that women with endometriosis have an increased risk of ovarian cancer, and the risk is most pronounced for women with early diagnoses and long-standing endometriosis [3], [4], [5], [6]. A series of pathologic studies reported that 5−10% of women with ovarian endometriosis have ovarian cancer, with clear cell and endometrioid cancers being the predominant endometriosis-associated ovarian cancers (EAOCs) [4], [7], [8], [9], [10]. In surgical specimens, a frequent coexistence of endometrioid/clear cell ovarian cancers with endometriosis has been observed. Vercellini et al. [11] supported this observation by reporting that endometriosis is complicated in 26.3% of women by ovarian cancer of the endometrioid subtype. In approximately 60% of EAOCs, cancer is adjacent to or arises directly from endometriosis tissue, suggesting that malignant transformation of endometriosis occurs [9], [12].

Although the histological and epidemiological evidence supports the possibility that endometriosis is the precursor to certain types of ovarian cancer, the current data are insufficient to demonstrate a causative relationship between endometriosis and ovarian cancer. Because endometriosis and ovarian cancer share some similar risk factors, protective factors, and/or antecedent mechanisms, it is difficult to adequately control the confounding factors when investigating the relationship between endometriosis and ovarian malignancy. On the other hand, the great difficulties involved in sample collection and the cellular heterogeneity of specimens hamper the molecular analysis of specific causative loci common to endometriosis and cancer [13].

Based on histological observations and clinically documented cases, the frequency of the malignant transformation of ovarian endometriosis is estimated at about 0.7–1.6% [4], [7], [14]. Sampson [15] described the first case of the malignant transformation of endometriosis with the following criteria: [1] the coexistence of carcinoma and endometriosis at the same site, which [2] bear a similar histological relationship to each other, and [3] the exclusion of a second malignant tumour elsewhere. Scott [16] later added another criterion: [4] the morphological demonstration of continuity between benign and malignant endometrial epithelium. The specimens of ovarian cancer arising from endometriosis (OCEMs) would be the ideal subjects for molecular analysis of the malignant transformation of endometriosis because such cases provide direct histological evidence of the transition, in direct continuity, from benign to malignancy. However, it is extremely difficult to identify a case fulfilling the above criteria. For instance, cancer might have obliterated the tissue of origin and eliminated the histological evidence of endometriosis in specimens. The demonstration of a histological continuity between endometriosis and cancer requires the extensive sectioning of surgical specimens. Therefore, there is little direct evidence from OCEMs to date.

The present study investigated whether there is a common lineage between ovarian endometriosis and ovarian cancer and sought molecular genetic evidence for the malignant transformation from endometriosis to cancer. Twelve cases of OCEMs were analysed in this study. The cases fulfilled the histological criteria and had clinical histories of endometriosis. For each case, microdissection was performed to separate different types of cells from paraffin-embedded sections. We also analysed 12 cases of solitary ovarian endometriosis. This study investigated the genetic alterations common to epithelial ovarian cancers, including loss of heterozygosity (LOH) and mutations of the tumour-related genes PTEN, p53, and K-ras.