Loss of heterozygosity (LOH) at locus 10q23.3 and mutation of the PTEN tumor suppressor gene occur frequently in both endometrial carcinoma and ovarian endometrioid carcinoma. To investigate the potential role of the PTEN gene in the carcinogenesis of ovarian endometrioid carcinoma and its related subtype, clear cell carcinoma, we examined 20 ovarian endometrioid carcinomas, 24 clear cell carcinomas, and 34 solitary endometrial cysts of the ovary for LOH at 10q23.3 and point mutations within the entire coding region of the PTEN gene. LOH was found in 8 of 19 ovarian endometrioid carcinomas (42.1%), 6 of 22 clear cell carcinomas (27.3%), and 13 of 23 solitary endometrial cysts (56.5%). In 5 endometrioid carcinomas synchronous with endometriosis, 3 cases displayed LOH events common to both the carcinoma and the endometriosis, 1 displayed an LOH event in only the carcinoma, and 1 displayed no LOH events in either lesion. In 7 clear cell carcinomas synchronous with endometriosis, 3 displayed LOH events common to both the carcinoma and the endometriosis, 1 displayed an LOH event in only the carcinoma, and 3 displayed no LOH events in either lesion. In no cases were there LOH events in the endometriosis only. Somatic mutations in the PTEN gene were identified in 4 of 20 ovarian endometrioid carcinomas (20.0%), 2 of 24 clear cell carcinomas (8.3%), and 7 of 34 solitary endometrial cysts (20.6%). These results indicate that inactivation of the PTEN tumor suppressor gene is an early event in the development of ovarian endometrioid carcinoma and clear cell carcinoma of the ovary.
The tumor suppressor gene PTEN/MMAC1, 3 located on chromosome arm 10q (10q23.3), was first reported in 1997 by Li et al. (1) . Frequent LOH at 10q23.3 and mutations of the gene have been found in various types of cancer (1, 2, 3, 4, 5) , and germ-line mutations of PTEN have also been associated with some familial neoplastic diseases such as Cowden disease, Lhermitte-Ducos disease, and Bannayan-Zonana syndrome (6, 7, 8) . Introduction of wild-type PTEN into the mutant PTEN glioma cell line results in growth suppression in vivo and in vitro (9) . PTEN encodes a phosphatase that dephosphorylates phosphatidylinositol-3,4,5-triphosphate. The function of the phosphatase is to interfere with the function of phosphatidylinositol-3,4,5-triphosphate, to inhibit cell death mediated by protein kinase B, and to encourage cell proliferation (10 , 11) . Furthermore, Di Cristofano and Pandolfi (12) reported that loss of function of just a single allele of PTEN is sufficient to confer a growth advantage. Their results indicated that PTEN mutation without LOH in the PTEN region or LOH in the PTEN region without mutation can reduce the function of PTEN. Recently, Perren et al. (13) examined the expression of the PTEN gene in breast cancer immunohistochemically and showed that hemizygous PTEN deletions were well correlated with lack of staining for PTEN protein.
PTEN gene abnormalities have been identified in various types of human carcinoma, including brain, endometrium, prostate, breast, thyroid, liver, lung (small cell carcinoma), and head and neck carcinomas and lymphomas (1 , 2 , 5 , 14, 15, 16, 17) . In particular, high frequencies of LOH at 10q23.3 and mutation of the gene have been reported in glioma, endometrial carcinoma of the uterus, and ovarian endometrioid carcinoma (1 , 2 , 18, 19, 20) . However, the roles of the PTEN gene in the carcinogenesis of glioma and endometrial carcinoma of the uterus seem to be different. Rasheed et al. (21) , Maier et al. (22) , and Davis et al. (23) reported that the PTEN gene mutations are restricted to high-grade rather than low-grade gliomas and may be associated with the transition from a low histological grade to anaplasia. In contrast, Risinger et al. (24) reported that, in the genesis of endometrial carcinoma, PTEN mutations are associated with early-stage, rather than late and metastatic, carcinomas. Furthermore, Maxwell et al. (25) have found frequent PTEN mutations in endometrial hyperplasia with and without atypia. These reports indicate that inactivation of the PTEN gene is an early event in the development of endometrial carcinoma of the uterus.
Endometriosis, the presence of ectopic endometrial tissue, is a common gynecological disease and is considered to be a benign tumor. Malignant transformation of endometriosis was first documented in 1925 by Sampsons (26) and has been thought to occur in 0.7–1.0% of all cases of endometriosis (27, 28, 29) . Genetically, ovarian endometrial cysts have a monoclonal origin (30) , and endometrioid and clear cell carcinoma of the ovary may arise through malignant transformation of ectopic endometrium (31) . These findings support the possibility that endometriosis is a precancerous form of certain types of ovarian cancer. The aim of the present study was to assess the role of LOH at the 10q23.3 locus and PTEN gene mutation in the multistep carcinogenesis of ovarian clear cell and endometrioid carcinoma. We examined ovarian endometrial cysts and ovarian endometrioid and clear cell carcinoma for LOH at the loci flanking the PTEN gene and mutation of the PTEN gene, using a laser-assisted microdissection method (32) . We found frequent LOHs at 10q23.3 and mutations of the PTEN gene in endometrial cysts and clear cell carcinoma of the ovary, as well as in endometrioid carcinoma of the ovary, suggesting that there is a sequential progression from ovarian endometrial cyst to endometrioid or clear cell carcinoma of the ovary.
MATERIALS AND METHODS
Cases and Microdissection.
We examined 20 endometrioid carcinomas, 24 clear cell carcinomas, and 34 solitary endometrial cysts of the ovary, which were resected at the University Hospital of Tsukuba (Ibaraki, Japan) between 1976 and 1998. We also examined 12 specimens of normal endometrium without endometriosis or leiomyoma. We used normal fallopian tubes or lymph nodes without metastases as normal controls for the endometrial cyst and carcinoma cases and myometrium as a normal control for normal endometrial tissue. Eight of the 20 cases of endometrioid carcinoma and 12 of the 24 cases of clear cell carcinoma contained apparently benign ectopic endometrium in the same ovary. All specimens were fixed with 10% formalin and embedded in paraffin. After histological examination, the ectopic endometrial cells, carcinoma cells, normal endometrium, and normal cells that were used as normal controls were microdissected with a Pixcell Laser Captured Microdissection System (Arcturus Engineering, Inc., Mountain View, CA; Fig. 1 ⇓ ). Finally, we microdissected about 20–40 cells from each specimen and extracted the genomic DNA.
Three microsatellite markers (D10S215, D10S541, and D10S608) were used to evaluate LOH on 10q23.3. All primers used in this study were obtained from Research Genetics (Huntsville, AL). Genomic DNA corresponding to DNA extracted from eight cells was subjected to PCR amplification in 10 μl of reaction mixture. The reaction mixture consisted of 3.6 units of Ex-Taq DNA polymerase (Takara, Tokyo, Japan); 200 μm dATP, dTTP, and dGTP; 20 μm dCTP; 5 μCi of[α -32P]dCTP (Amersham Life Science, Buckinghamshire, United Kingdom); 33.5 mm Tris-HCl (pH 8.8); 1.5 mm MgCl2; 16 mm (NH4)2SO4; 0.01% Tween 20; and 0.3 μl of each primer as supplied (20 μm each). PCR was carried out over 35 amplification cycles for 45 s at 94°C, 45 s at 55°C, and 60 s at 72°C in a Takara Thermal Cycler MP (Takara). The PCR products were resolved on a 6% denaturing polyacrylamide gel and visualized by autoradiography film (Kodak, Rochester, NY) exposure.
Nine exons of PTEN (except for the first primer set of exon 5) were amplified separately using the primer sets described by Risinger et al. (19) . We used ATCTTTTTACCACAGTTGCAC and GTCCCTTTCCAGCTTTACAG as the first primer set for exon 5. Genomic DNA corresponding to DNA extracted from eight cells was subjected to PCR amplification in 10 μl of the reaction mixture used in the LOH analysis. PCR was carried out as described previously (19) . The PCR products were resolved on a 0.5× Mutation Detection Enhancement gel (FMC Bioproducts, Rockland, ME) and visualized by autoradiography film (Kodak, Rochester, NY) exposure.
Shifted SSCP bands were excised from the Mutation Detection Enhancement gel. We extracted the DNA from the gel with distilled water and reamplified it using the original PCR primers. The reamplified PCR products were cloned into the pCRII TA vector (Invitrogen, San Diego, CA), according to the company’s instructions, and then sequenced with an ABI PRISM 310 Dye Terminator Cycle Sequencing Ready Reaction kit (Perkin-Elmer, Foster City, CA). For the cases with mutations, we repeated the microdissection, SSCP analysis and sequencing, and confirmed the results.
We examined the genotypes of 20 endometrioid carcinomas of the ovary at three highly polymorphic loci distributed at 10q23.3 (D10S215, D10S541, and D10S608). As Table 1 ⇓ shows, 8 of 19 informative cases of endometrioid carcinoma (42.1%) demonstrated LOH in the 10q23.3 region. To examine whether the LOH found frequently in this region in endometrioid carcinoma of the ovary was also present in clear cell carcinoma and endometrial cysts (which are thought to be associated histologically with endometrioid carcinoma of the ovary), we further studied 24 clear cell carcinomas and 34 endometrial cysts of the ovary (Table 1 ⇓ and Fig. 1 ⇓ ). Six of 22 informative cases of clear cell carcinoma (27.3%) and 13 of 23 informative cases of endometrial cyst (56.5%) demonstrated LOH at 10q23.3. None of the 12 specimens of normal endometrium showed LOH at D10S215, D10S541, or D10S608.
SSCP and Sequencing.
SSCP analysis of the PTEN gene was performed on 20 endometrioid carcinomas, 24 clear cell carcinomas, and 34 solitary endometrial cysts without carcinoma. We screened the entire PTEN coding region (9 exons; Fig. 2 ⇓ ). Four, 3, and 11 abnormally shifted bands that were detected from cases of ovarian endometrioid carcinoma, clear cell carcinoma, and endometrial cyst, respectively, were eluted from the gel, and the DNA extracted was sequenced (Table 2) ⇓ . There was one missense mutation (transversion), two nonsense mutations, and one deletion in four endometrioid carcinomas. Two missense mutations (transitions) and one deletion were detected in two clear cell carcinomas. Eight missense mutations (seven transitions and one transversion) and two deletions were detected in seven endometrial cysts. In the ovarian endometrioid carcinomas, the codons that showed the mutation were scattered between exons 1 and 9, but they appeared to cluster around the catalytic signature motif of the PTEN gene in the clear cell carcinomas and endometrial cysts (Fig. 3) ⇓ . Two of four cases of endometrioid carcinoma, one of two cases of clear cell carcinoma, and five of eight cases of solitary endometrial cyst were accompanied by LOH at 10q23.3.
Comparison of LOH at 10q23.3 in Ovarian Carcinoma and Synchronous Apparently Benign Endometrium.
We detected apparently benign ectopic endometrium synchronously in 8 of 20 endometrioid carcinomas and 12 of 24 clear cell carcinomas. These endometrial tissues were diagnosed as benign ectopic endometrium histologically and were also examined for LOH in the PTEN region (D10S215 and D10S541), the patterns of LOH being compared between them and the carcinomas. Five of eight endometrioid carcinoma cases synchronous with endometriosis and 7 of 12 clear cell carcinoma cases synchronous with endometriosis were informative. Of these five endometrioid carcinoma cases, three displayed LOH events common to both the carcinoma and endometriosis, one displayed an LOH event only in the carcinoma, and one displayed no LOH event in either lesion (Fig. 4) ⇓ . Of the seven clear cell carcinoma cases, three displayed LOH events common to both the carcinoma and endometriosis, 1 displayed an LOH event only in the carcinoma, and three displayed no LOH event in either lesion. In no case was LOH detected only in endometriosis.
In this study, we confirmed the high frequency of LOH at 10q23.3 in endometrioid carcinoma of the ovary (42.1%) and demonstrated high frequencies of LOH at 10q23.3 in solitary endometrial cysts (56.5%) and clear cell carcinoma of the ovary (27.3%). SSCP and DNA sequence analysis also displayed somatic mutations in the solitary endometrial cysts (20.6%), as well as in the endometrioid type (20.0%) and clear cell type (8.3%) of ovarian carcinoma.
Clinicopathologically, endometrial cysts of the ovary are thought to develop from ectopic endometrium in the ovary, and they have recently been confirmed to grow monoclonally (30) . Jiang et al. (33) found that <20% of endometrial cysts had LOH at the loci (chromosomes 9q, 11q, and 22q) of candidate tumor suppressor genes associated with ovarian cancers. We found more frequent LOH at 10q23.3 in endometrial cysts of the ovary than at these other loci (see Table 1 ⇓ ). Furthermore, we observed mutations of the PTEN gene in seven cases of solitary endometrial cyst (20.6%), and five of these seven were accompanied by LOH at 10q23.3 (Table 2) ⇓ . Maxwell et al. (25) reported somatic mutation of the PTEN gene in ∼20% of cases of endometrial hyperplasia, which is thought to be the precursor of endometrial carcinoma of the uterus, and the frequencies did not differ between hyperplasia with atypia and that without atypia. These data indicated the genetic sequence from endometrial hyperplasia to endometrial carcinoma of the uterus. The high frequency of PTEN gene mutation we observed in solitary endometrial cysts suggests a similar genetic association between solitary endometrial cyst of the ovary and ovarian endometrioid carcinoma. Our results indicate that LOH at 10q23.3 and mutations of the PTEN gene are very early events in the development of ovarian endometrioid carcinoma and also support the concept that solitary endometrial cysts of the ovary are a precancerous form of ovarian endometrioid carcinoma. In contrast, the frequency of PTEN gene alterations (including LOH on 10q23.3 and somatic mutations of the PTEN gene) in other types of ovarian carcinoma, such as serous or mucinous adenocarcinoma, has been reported to be very small compared with that in ovarian endometrioid carcinoma (20) . However, the frequency and significance of PTEN gene abnormalities in clear cell carcinoma of the ovary is still unclear. Obata et al. (20) reported LOH at 10q23.3 in one of seven informative cases of clear cell carcinoma but found no mutations of the PTEN gene in any of the cases they examined (0 of 8). We examined a large number of clear cell carcinomas (24 cases) and demonstrated LOH at 10q23.3 in 6 of 22 informative cases (27.3%); we found PTEN mutations in 8.3% (2 of 24). Although LOH at 10q23.3 and PTEN mutation occurred less frequently in clear cell carcinoma than in endometrioid carcinoma, there were no significant differences in their occurrences in the two types of carcinoma, and the frequency of LOH was still higher than in other histological types of ovarian carcinoma.
The mutation spectra of the PTEN gene in these three different types of ovarian tumors were of interest. Mutations in the endometrial cysts and clear cell carcinomas were concentrated at exons 5–6, which encode the phosphatase domain of the PTEN gene (Fig. 3) ⇓ . These results indicate the functional similarity of tumorigenesis between endometrial cyst and clear cell carcinoma of the ovary.
To confirm the sequences of carcinogenesis between endometrial cyst and ovarian endometrioid carcinoma or clear cell carcinoma of the ovary, we investigated LOH events at D10S215 and D10S541 (which flank the PTEN gene) in five endometrioid carcinomas and seven clear cell carcinomas that were thought to have occurred synchronously with endometrial cysts. We found cases that displayed LOH events in both the carcinoma and apparently benign cyst, or only in the carcinoma, but none of the cases showed LOH events only in the cyst (Fig. 4) ⇓ . Although there is a possibility that cystic lesions may be part of an extremely well differentiated adenocarcinoma, these results are thought to support the concept of sequential progression from endometrial cyst of the ovary to ovarian endometrioid or clear cell carcinoma.
In this study, we used a laser-assisted microdissection system (32) , which enabled us to collect not only tumor cells from carcinoma tissue but also the epithelial lining cells of endometriosis tissue, without contamination by nontumor cells such as lymphocytes, endothelial cells, and fibroblasts. This approach made it possible to analyze various genetic alterations in the endometrial cysts. We expect that this technique will be very useful for investigating genetic alterations in other tissues or precancerous lesions such as ulcerative colitis, lung fibrosis, or benign prostatic hyperplasia, because in these lesions, the dysplastic or atypical cells that are thought to be carcinoma precursors are widely mixed with numerous nontumor cells. Early genetic alterations in various precancerous cells detected by light microscopy can be readily identified by the tissue-microdissection method.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported in part by a Grant-in-Aid for Cancer Research from the Ministry of Health and Welfare of Japan.
↵2 To whom requests for reprints should be addressed, at Department of Pathology, Institute of Basic Medical Sciences, University of Tsukuba, Tennodai 1-1-1, Tsukuba-shi, Ibaraki 305-8575, Japan. Phone: 81-298-53-3350; Fax: 81-298-53-3150; E-mail:
↵3 The abbreviations used are: PTEN/MMAC1, phosphatase and tensin homologue deleted on chromosome 10/mutated in multiple advanced cancers 1; LOH, loss of heterozygosity; SSCP, single-strand conformational polymorphism.
- Received December 28, 1999.
- Accepted October 19, 2000.
- ©2000 American Association for Cancer Research.