Estrogen receptor alpha (ER-alpha) and beta (ER-beta) mRNAs in normal ovary, ovarian serous cystadenocarcinoma and ovarian cancer cell lines: down-regulation of ER-beta in neoplastic tissues.

Abstract

The prognosis in ovarian carcinoma, the most lethal of the gynecologic neoplasms, is poor and has changed little in the last three decades. Only a small number respond to antiestrogen therapy, although the classic estrogen receptor, ER-alpha, has been identified in ovarian surface epithelium, from which approximately 90% of ovarian cancers originate. We have previously shown that ER-beta mRNA is most abundant in human fetal ovaries, suggesting that it might play an important role in ovarian development. Therefore, we investigated the mRNA levels of both ERs in normal ovaries, ovarian serous cystadenocarcinomas, granulosa cells from patients undergoing in vitro fertilization (IVF), the ovarian surface epithelium cell line IOSE-Van, and the ovarian cancer cell lines SKOV3, HEY and OCC1. Northern blots of normal and neoplastic ovaries were hybridized with an ER-beta riboprobe that spans the A/B domain. We detected two major hybridizing bands at approximately 8 and 10 kb. An RNase protection assay using the same probe revealed a single band of the expected size. Hybridizing the same blot with an ER-alpha riboprobe showed a strong hybridizing band at approximately 6.5 kb. In ovarian cancer samples, ER-beta mRNA level was decreased when compared to normal ovaries. Using 25 cycles of RT-PCR followed by Southern blotting, we found equal amounts of ER-alpha and -beta mRNAs in normal ovaries in all age groups from 33 to 75 years; however, in ovarian cancer tissue, the level of ER-alpha mRNA was similar or slightly higher, comparable to 10(3) to 10(4) copies of plasmid DNA, but ER-beta mRNA levels were markedly decreased. Granulosa cells from IVF patients expressed high levels of ER-beta mRNA. The OSE cell line expressed a low level of ER-alpha, detectable after 40 cycles of RT-PCR and no ER-beta mRNA. SKOV3 showed a low level of ER-alpha and -beta mRNAs, whereas OCC1 showed a low level of ER-beta and a relatively high level of ER-alpha. HEY did not contain detectable amounts of either ER after 40 cycles of RT-PCR. We found no evidence of differential splicing or major deletions in almost the entire coding region of ER-beta in either normal ovaries or tumor samples.