The beneficial influence of E2 in the maintenance of healthy bone is well recognized. However, the way in which the actions of this hormone are mediated is less clearly understood. Western blot analysis of ERα in osteoblasts clearly demonstrated that the well characterized 66-kDa ERα was only one of the ERα isoforms present. Here we describe a 46-kDa isoform of ERα, expressed at a level similar to the 66-kDa isoform, that is also present in human primary osteoblasts. This shorter isoform is generated by alternative splicing of an ERα gene product, which results in exon 1 being skipped with a start codon in exon 2 used to initiate translation of the protein. Consequently, the transactivation domain AF-1 of this ERα isoform is absent. Functional analysis revealed that human (h)ERα46 is able to heterodimerize with the full-length ERα and also with ERβ. Further, a DNA-binding complex that corresponds to hERα46 is detectable in human osteoblasts. We have shown that hERα46 is a strong inhibitor of hERα66 when they are coexpressed in the human osteosarcoma cell line SaOs. As a functional consequence, proliferation of the transfected cells is inhibited when increasing amounts of hERα46 are cotransfected with hERα66. In addition to human bone, the expression of the alternatively spliced ERα mRNA variant is also detectable in bone of ERα knockout mice.

These data suggest that, in osteoblasts, E2 can act in part through an ERα isoform that is markedly different from the 66-kDa receptor. The expression of two ERα protein isoforms may account, in part, for the differential action that estrogens and estrogen analogs have in different tissues. In particular, the current models of the action of estrogens should be reevaluated to take account of the presence of at least two ERα protein isoforms in bone and perhaps in other tissues.