Regulation of Hedgehog signalling during development and carcinogenesis

Abstract

Hedgehog (Hh) signalling is one of the most important signalling networks in the body and, when miss-regulated, has been implicated in many human tumours, particularly that of skin (Basal Cell Carcinoma, BCC) and brain (medulloblastoma). BCC most commonly arises due to mutations in the gene sequence of the Hh receptor protein PATCHED1 (PTCH1). The PTCH1 gene is reported to have four first exons and we showed in paper I that PTCH1 is transcriptionally regulated by three independent promoters. Additionally, we provided evidence of a single functional Gli-consensus site in the PTCH1 promoter region and revealed that transcriptional activation, mediated via the Hh signalling pathway, is fully dependent on this single Gli-binding site.

Earlier studies have revealed that one of the PTCH1 isoforms shows induced levels in BCC, but so far no connection between these isoforms and breast tumourigenesis exists. However, several lines of data implicate a role for Hh in breast carcinogenesis, and in support of this we found that an altered expression of the human Hh effector protein, GLI1, in the mammary epithelial cells of mice could induce the formation of heterogeneous mammary tumours (paper III).

GLI1-induced mammary tumour cells showed high expression of the progenitor cell marker, Bmi-1 and cytokeratin 6 as well as proliferative, cell survival and metastatic markers. GLI1-induced mammary tumours did not fully regress after removal of the transgenic expression. However, the mammary tumour appearance shifted towards a more stromal epithelial content. Prior to tumour formation GLI1-induced the formation of hyperplastic lesions and defective terminal end buds in both male and female mammary glands. In addition, transgenic GLI1 mice were unable to feed their pups (Paper II). This failure could molecularly be explained by a reduced alveolar differentiation and milk gene expression during pregnancy. In addition, mammary glands with induced expression of GLI1 showed an increased expression of the GLI1 response gene Snail, coupled with reduced expression of E-cadherin and Stat5. Removal of the transgenic expression did not restore the lactation ability of these mice.

We further demonstrated that the involution process was impaired in the GLI1 transgenic mammary glands, with cellular residuals residing in the mammary ducts. Interestingly, embryonic mammary and salivary gland development, according to branching morphogenesis share common mechanisms and when GLI1 expression was induced in salivary epithelial cells, hyperplastic lesions with basaloid content appeared in 100% of the transgenic salivary glands (Paper IV). These lesions showed strong homology with the hyperplastic lesions which appeared in the GLI1-induced mammary glands, suggesting that GLI1 targets analogous cells in different tissues.

Additionally, we demonstrated that salivary acinar differentiation was blocked, cystic lesions formed and the ductal structures appeared more prominent in the GLI1 transgenic salivary glands when compared to wild type. The lack of salivary acinar differentiation in combination with the extensive morphological alterations, indicated a diminished secretion of salivary fluid in the GLI1 transgenic mice. This is analogous to the mammary gland where induced expression of GLI1 resulted in the appearance of immature and undifferentiated alveolar secreting cells at parturition, which resulted in an inability to secrete milk. The GLI1-induced mammary changes could not be restored after termination of the oncogenic expression, while GLI1-induced salivary gland lesions could.