Gatekeeper for endometrium: the PTEN tumor suppressor gene.

The complexity of genetic alterations in invasive cancers and the heterogeneity of tumor cell populations hamper attempts to translate molecular understanding at the genetic level into accurate diagnostic and therapeutic approaches. Delineation of molecular changes in the initial stages of tumor development is, therefore, highly desirable because such information can be translated into more effective cancer prevention and treatment strategies. Significant advances have been made in understanding the molecular events that occur as normal tissues evolve— often into precancerous lesions, some of which eventually progress to cancer. In some cancers, such as colorectal carcinomas and gliomas, details of distinct molecular pathways have been identified. In most cancers, however, no consistent pattern of genetic alteration is convincingly linked to the initiation or progression of the disease. In the case of endometrial cancer, molecular details are beginning to emerge that may eventually help advance our understanding of the complex histopathology of this disease. Two major pathogenetic variants of endometrial carcinoma, endometrioid and serous, seem to evolve via divergent pathways and contain distinct genetic abnormalities. Serous endometrial carcinoma, the less common variant, is highly aggressive and estrogen independent. More than 90% of serous tumors contain p53 mutations (1). The estrogen-related endometrioid adenocarcinomas account for up to 80% of the endometrial cancer cases and appear to arise via a progression pathway. Frequently, the advanced carcinomas are temporally and spatially associated with endometrial hyperplasia, which constitutes a range of heterogeneous precancerous lesions. Complex atypical hyperplasia (CAH), the most advanced component of this histopathologic spectrum, is believed to be the direct precursor lesion of endometrioid carcinoma. Subsets of endometrioid carcinoma were found to contain K-ras (also known as KRAS) mutations (15%–20%) (2) and the microsatellite instability (MI) phenotype (20%–30%) (3), which is indicative of DNA mismatch-repair defects. Both of these abnormalities were also detected in CAH. In 1997, several groups (4–6) reported mutational inactivation of the tumor suppressor gene PTEN in 33%–55% of lowgrade as well as in high-grade endometrial cancers of endometrioid histology. More importantly, two independent studies (7,8) described mutations in the PTEN gene in about 20%–30% of CAH, the putative precursor lesion of endometrioid carcinoma. Their observations strongly suggest the involvement of the gene in the initial stages of tumor development. A comparative analysis of the CAH lesions with and without synchronous carcinomas provided evidence that mutation of the PTEN gene is an early event and may precede the development of the MI phenotype (8). In this issue of the Journal, Mutter et al. (9) report a much higher frequency of mutations in both adenocarcinomas (83%) and endometrial hyperplasias (55%). As Mutter et al. suggested, this high frequency may be due to a combination of preselection bias for the biopsy specimens of endometrioid histology by computerized morphometric imaging analysis and the denaturing gradient gel electrophoresis technique used for mutation detection. Homozygous inactivation of the gene occurred in 33% of the tumors, whereas 50% of the tumors contained monoallelic mutations. Furthermore, by use of immunohistochemistry (albeit on a different set of biopsy specimens), the authors detected loss of expression or decreased expression of PTEN in most, if not all, cancers and precancers and in a group of poorly classified lesions that are intermediate between normal epithelium and precancer. In the future, quantitative analysis of PTEN expression and its association with the genotype in the same tumor biopsy specimens would help clarify whether hemizygous inactivation of the gene in endometrioid carcinomas and precancerous lesions relates to the partial or total loss of its expression. One PTEN transgenic study (10) revealed that PTEN haploinsufficiency in mice generates hyperplastic foci in the endometrium that closely resemble the histopathologic spectrum of human endometrial precancers. Mutter et al. (9) also explored the possible association between the widely accepted endocrine risk for endometrial carcinoma and PTEN expression. They reported loss of expression of PTEN protein in the estrogen-exposed endometrial epithelium. In two of the seven tissue samples examined, PTEN-negative glands were interspersed among PTEN-positive areas; in contrast, contiguous groups of PTEN-negative glands were observed in both cancerous and precancerous lesions. Exposure to estrogen unopposed by the differentiating effect of progestin is considered a risk factor for the development of endometrial hyperplasia and carcinoma. The molecular basis of this observation is not known, but it likely involves interaction between the hormone and specific gene(s). The scattered loss of PTEN expression in unopposed estrogen-exposed endometrium, if borne out by future molecular studies on an expanded set of samples, would elucidate one such interaction at the earliest stage of transition of the normal endometrial epithelium to a precancerous state. Also, it would provide an exciting challenge to decipher the molecular details of interactions between genetic and hormonal events in an epithelium that is subjected to orderly cyclic proliferation and regression under the influence of hormones. In this context, it is noteworthy that serous tumors that arise via a different genetic pathway are hormone independent and do not harbor PTEN mutations. It is possible that the two histologic variants arise from two different endometrial stem cells. Information on the transcriptional regulation of the PTEN gene would provide insight about its role (or the lack of one) in the evolution of endometrial cancers arising from two distinct genetic pathways.