Molecular pathogenesis of endometrial cancer1 George L. Mutter, MD Associate Professor of Pathology, Harvard Medical School Associate Pathologist, Division of Women’s and Perinatal Pathology Department of Pathology, Brigham and Women's Hospital 75 Francis Street, Boston, MA 02115, USA e-mail: [email protected] Website: www.endometrium.org Introduction. Genetic changes which cause endometrial cancer may be discovered by simple comparison of tumor with normal tissues, and correlating clinicopathologic with genotypic features. More difficult is the specific elaboration of those interacting events which may transpire in a particular sequence during a protracted interval of carcinogenesis. A small burden of premalignant cells makes them elusive targets for study, and their histopathologic plasticity complicates achieving diagnostic consensus and reproducibility between laboratories. Despite these formidable obstacles, the last decade has seen an explosion of new data regarding endometrial carcinogenesis. Cancer subtypes have been clearly divided along genetic and clinical lines, and a flood of information about those genetic changes which cause endometrial cancer has been forthcoming from many sources, including non-gynecologic tumor systems. The diagnosis of premalignant endometrial lesions, long a confusing and contentious issue among pathologists, has achieved objectivity from biomarker studies and histomorphometric analysis. These data have yielded a scientific basis for standardization and revision of endometrial precancer diagnosis in a routine diagnostic setting. Lastly, experimental access to preclinical stages of premalignant endometrial disease, a phase of tumorigenesis which is the probable mediator of endocrine risk factors, affords an opportunity for cancer chemoprevention. Genetic Subgroups of Endometrial cancer. Sporadic endometrial adenocarcinoma may be classified into dichotomous genotypic classes defined by polarized frequencies of inactivation of specific genes (Table I). These genetic pathways of endometrial carcinogenesis are generally paralleled by endometrioid (Type 1) and nonendometrioid (Type 2) subtypes that comprise distinct clinicopathologic entities7;8. Approximately 70-80% of newly diagnosed cases of endometrial cancer in the U.S. are of endometrioid histology. Type 1 cancers have been associated with unopposed estrogen exposure, and are often preceded by premalignant disease. Progression to carcinoma is highly inefficient, involving complex interactions between multiple genetic events (PTEN, KRAS, microsatellite instability) and ambient hormonal selection factors. In contrast, Type 2 endometrial cancers have papillary serous or clear cell histology with a very aggressive clinical course. Hormonal risk factors have not been identified, nor is there a readily observed premalignant phase. Type 2 tumors characteristically present as fully developed malignancies and have P53 tumor suppressor gene defects. For purposes of discussion Type 1 and Type 2 sporadic endometrial adenocarcinoma might just as well be considered separate diseases. 1

These notes are reproduced from Mutter GL, Ince TA. Molecular Pathogenesis of Endometrial Cancer. In: Fuller A, Seiden MV, Young R, editors. Uterine Cancer: American Cancer Society Atlas of Clinical Oncology. Hamilton,Ontario,Canada: B.C.Decker, 2003.

Table I: Altered Gene Function In Sporadic Endometrioid and Non-Endometrioid Endometrial Adenocarcinoma Gene p53 PTEN K-ras β-Catenin MLH1 P27 Cyclin D1 P16 Rb Bcl-2 Bax Estrogen & Progesterone receptors

Alteration mutation loss of function mutational activation mutation (immunoreactive) Microsatellite Instability (epigenetic silencing) loss of function increased expression loss of function loss of function loss of function loss of function positive immunoreactivity

Endometrioi d 5-10% 55% 13-26

NonEndometrioid 80-90% 11% 0-10%

25-38%

rare

17%

5%

68-81% 41 - 56% 20-34% 3 - 4% 65% 48%

76% 19% 10% 10% 67% 43%

70-73%

19-24%

Refs 2 3;5;6 8;9 12-14

4;19-21

23 23 15;23 23;27 15;28 15;32 15

A third group are those extremely rare endometrial adenocarcinomas which present as a manifestation of multi-cancer heritable syndromes (Table II) 10;11. The low frequency with which these are encountered in clinical practice belies their high level of scientific interest, as familial presentation affords unique opportunities to dissect those causal genetic events which may also be effective in a sporadic setting. The PTEN inactivation seen in Cowden’s syndrome3;12, and DNA mismatch repair defects of Hereditary Nonpolyposis Colon Cancer1;15, are frequent accompaniments to sporadic endometrial carcinogenesis16. Table II: Heritable Conditions associated with Endometrial Cancer1 Condition Cowden Syndrome

Gene(s) PTEN

Hereditary Nonpolyposis Colon Cancer

Mismatch Repair MLH1, MSH2,MSH6

Phenotype endometrial, endocrine, breast cancer colon, endometrial cancer

Refs 3

4

An overview of the most prevalent genetic changes in endometrial carcinoma can be obtained by detailed review of a relatively small number of individual genetic changes. P53.

p53 is a tumor suppressor gene that is inactivated in half of all human tumors17. As a multifunctional transcription factor p53 plays a central role in cell cycle regulation and apoptosis pathways. Despite substantial effort it has not always been possible to link p53 mutations with a

particular histologic pattern or clinical outcome in many tumor types. Endometrial cancer is one of the exceptions, in which p53 mutations are almost exclusively associated with a specific histological subtype, and not distributed stochastically among half of the cases. Type 2 tumors, which are generally of serous papillary or clear cell histology frequently (>90%) harbor p53 mutations 18. A much lower p53 mutational rate in Type 1 tumors (