Epidemiological studies show a strong link between postmenopausal hormone replacement therapy and decreased incidence of colorectal cancer. The colon cancer cell line, COLO 205, develops sensitivity to 17beta-oestradiol (E(2)) in apoptosis assays with increasing passage number (>40), and we hypothesised that genes selectively regulated in multiply passaged cells were likely to be important in E(2)-related apoptosis. Gene array analysis was used to compare the patterns of genes up- or down-regulated in E(2)-sensitive and -insensitive cells. For some genes, changes in mRNA expression were confirmed by protein expression analyses. Changes found in response to E(2) in multiply passaged cells, but not minimally passaged cells, included induction of growth arrest and DNA damage-inducible protein 153 (GADD153), and repression of Kirsten-Ras 2B (K-Ras-2B), metastasis inhibition factor NM23 and vascular endothelial growth factor. A second group of genes was regulated with E(2) exposure in both cell types, and is unlikely to be critically involved in E(2)-associated apoptosis. These included up-regulation of butyrate response factor 1 (BRF1) and down-regulation of c-jun and the breast cancer associated ring domain gene known as BARD1. By comparing control arrays from the two cell populations, cAMP-response element-binding protein (CBP), which is associated with steroid receptor-dependent target gene transcription and the oncoprotein, tyrosine kinase-T3 (TRK-T3), were up-regulated whereas retinoic acid receptor alpha (RARalpha) was down-regulated in multiply passaged cells. This study provides evidence for selective regulation of genes in colon cancer cells by E(2), indicates which of those regulated are likely to be involved in induced apoptosis, and suggests genes likely to be responsible for facilitation.

Epidemiological studies of postmenopausal hormone replacement therapy show a reduction in the risk of developing colon cancer, and animal studies using 17beta-oestradiol (E(2)) demonstrate a decreased incidence of chemically-induced colon cancer. Using the colon cancer cell line, COLO205, we found that E(2) induced a dose-dependent increase in DNA fragmentation and nuclear condensation, significant effects being seen at 10(-12 )mol/l. BSA-conjugated E(2), which cannot enter cells, was ineffective at inducing apoptosis in COLO205 cells, indicating that E(2) was not acting through a cell-membrane receptor. E(2) did not induce the morphological changes characteristic of differentiation. Using RT-PCR we found that the oestrogen receptor alpha (ERalpha) isoform was absent in the COLO205 cell line in contrast to CACO-2, LoVo and SW620 cells, but mRNAs for ERbeta1, -beta2, -beta5 and -beta6 isoforms were detected. Western immunoblotting results showed full-length ERbeta protein but no detectable ERalpha in COLO205 cells. In normal human colon tissue samples immunoreactive ERbeta was found but ERalpha was barely detectable. Expression of ERbeta was lost in some colon cancer specimens and reduced in others. We conclude that E(2), through ERbeta, at concentrations found during replacement therapy, may inhibit the development of colon cancer by inducing apoptosis.