The switch/sucrose non-fermentable (SWI/SNF) subunit ARID1A (AT-rich interactive domain 1A gene) has been recently postulated as a novel tumor suppressor of gynecologic cancer and one of the driver genes in endometrial carcinogenesis. However, specific relationships with established molecular alterations in endometrioid endometrial cancer (EEC) are currently unknown. We analyzed the expression of ARID1A in 146 endometrial cancers (130 EECs and 16 non-EECs) in relation to alterations in the PI3K-Akt pathway (PTEN expression/KRAS/PIK3CA mutations), TP53 status (TP53 immunohistochemistry) and microsatellite instability. To discriminate between microsatellite instability due to somatic MLH1 promoter hypermethylation or germline mutations in one of the mismatch repair genes (Lynch syndrome), we included a 'Lynch syndrome set'. This set included 21 cases with confirmed germline mutations and 15 cases that were suspected to have a germline mutation. Loss of ARID1A expression was exclusively found in EECs in 31% (40/130) of the EEC cases. No loss of expression of the other subunits of the SWI/SNF complex, SMARCD3 and SMARCB1, was detected. Alterations in the PI3K-Akt pathway were more frequent when ARID1A expression was lost. Loss of ARID1A and mutant-like TP53 expression was nearly mutually exclusive (P=0.0004). In contrast to Lynch-associated tumors, a strong association between ARID1A loss and sporadic microsatellite instability was found. Only five cases (14%) of the 'Lynch syndrome set' as compared with 24 cases (75%, P<0.0001) of the sporadic microsatellite-unstable tumors showed loss of ARID1A. These observations suggest that ARID1A is a causative gene, instead of a target gene, of microsatellite instability by having a role in epigenetic silencing of the MLH1 gene in endometrial cancer.