β-Catenin/armadillo superfamily proteins are expressed in cells of all major tissue types including epithelial cells. They play important roles in cancer with some of its members as oncogenes while others act as tumor suppressors [1, 2]. δ-Catenin (gene designation as CTNND2), however, is a unique member of the family because it is primarily expressed in the central nervous system of normal individuals. However, δ-catenin is now well established as being overexpressed in prostate cancer [9, 26].
The frequently increased expression of the neuronal protein δ-catenin in peripheral prostate cancer tissue raised important questions such as what mechanisms are in place in cancer to result in a high level of δ-catenin expression. δ-Catenin gene amplification was observed in cervical cancer  and bladder cancer . Transcription factor Pax6 was found to play an important role for regulation of δ-catenin expression in developing eye and central nervous system . Recently, we showed that the ectopic overexpression of E2F1 and Pax6 positively upregulates δ-catenin expression in prostate cancer cells . Furthermore, increased translation efficiency by somatic mutations in the 5'-untranslated region of δ-catenin was observed in prostate cancer patients , further supporting the hypothesis that cancer cells implement multiple mechanisms to upregulate δ-catenin expression to advance tumor progression. In this study, we provided the first evidence that δ-catenin is capable of promoting the expansion of prostate cancer cells, altering gene profiles of prostate cancer cell cycle regulation and survival.
Regarding cell proliferation and colonization, earlier studies evaluating MDCK cells transfected with δ-catenin showed that there were no significant differences between MDCK cells expressing δ-catenin and control cells . Interestingly, ectopic expression of δ-catenin in NIH3T3 fibroblast cells inhibited cell division and induced cellular processes with branches . However, δ-catenin overexpression in pheochromocytoma (PC12) cells increased cell proliferation and promoted neurite outgrowth when treated with nerve growth factor . These studies, in addition to our current findings, suggest that δ-catenin effects on cell growth are context dependent. It is possible that preneoplastic cells may not tolerate high levels of stable δ-catenin expression, such as in MDCK cells , NIH3T3 cells , and mammary epithelial cells . We have also failed to develop stable cell lines using PZ-HPV-7 (non-cancer human prostate epithelial origin) and NL20 (non-cancer human lung epithelial origin) cells (Zeng and Lu, unpublished data). However, in tumor cells, such as PC12, CWR22Rv-1, PC-3 and NCI-H1299, stable cell lines with δ-catenin overexpression were not only successfully produced, but also showed increased cell viability, suggesting that δ-catenin does not transform normal cells but promotes cancer cell expansion.
The armadillo repeating units reveal the most significant homology among β-catenin superfamily proteins, while the sequences flanking the armadillo domain are quite variable [5, 7]. This feature is consistent with the hypothesis that sequences outside the armadillo domains play important regulatory roles to characterize each different member. δ-Catenin interacts with classical cadherins through the armadillo domains, and the NH2- or COOH-terminal sequences on their own do not localize to cell-cell junctions . In addition, the deletion of COOH-terminal 207 amino acids abolished the δ-catenin mediated process extension in NIH3T3 cells and compromised the roles of δ-catenin in promoting dendrite outgrowth in neurons [18, 33, 34]. Removing the NH2-terminal 280 amino acids also remarkably altered the effects of δ-catenin on 3T3 cell morphology . In our present study, we showed that while the full-length δ-catenin promoted prostate cancer cell colony formation in soft agar and tumor xenograft growth in nude mice, δ-catenin with the deletion of either NH2-terminal 280 amino acids or COOH-terminal 207 amino acids lost its ability to promote tumor development. These studies underscore the importance of these sequences outside the armadillo domain to the functions of δ-catenin in cancer.
Gene profiling can provide initial indications of what may be the potential molecular pathways that δ-catenin employs to contribute to tumor progression. We identified a number of genes that displayed changes in expression when δ-catenin is overexpressed in prostate cancer cells. Several of these genes, such as cyclin D1, cdc34, Bcl2L1, and HK2, are especially interesting. Both cyclin D1 and cdc34 are involved in G1 to S transition, and Bcl2L1 is a long isoform of Bcl2 which protects cells from undergoing apoptosis . HK2 phosphorylates glucose to produce glucose-6-phosphate, thus committing glucose to the glycolytic pathway. Expression of HK2 has been indicated in rapidly growing cancer cells [35, 36]. Nevertheless, we still do not know how an increased δ-catenin expression affects these pathways in prostate cancer development. Our future studies will dissect the signaling events to determine the mechanisms by which δ-catenin employs to promote prostate cancer cell growth and tumor progression.