We have now cloned the mRNA encoding a N-terminal truncated splice isoform of the prohormone processing enzyme, carboxypeptidase E with Mol. Wt. 40kD, (CPE-deltaN) from hepatocellular carcinoma (HCC) cell line. Northern blot analysis shows this mRNA (1.83 kb in size) is present in metastatic hepatocellular carcinoma, ovarian and pancreatic cell lines. 40kD CPE-deltaN lacks the N-terminus and hence the secretory pathway sorting signal found in wild-type CPE. Wild-type CPE is localized in secretory granules; however, CPE-deltaN is translocated from the cytoplasm to the nucleus of metastatic cancer cells where it could bind to a nuclear protein complex to promote the expression of metastatic genes in a histone deacetylase (HDAC1/2)-dependent manner. In Panc-1 pancreatic cancer cells, knockdown of all forms of endogenous CPE inhibited cell growth and invasion; in contrast, ectopic expression of 40kD N-terminal truncated CPE isoform significantly promoted proliferation, colony formation, invasion and migration, whereas overexpression of WT-CPE had only a small effect. These results suggest that the 40kD CPE- deltaN isoform plays a role in cell growth and metastasis. In collaboration with Dr. Arbersfeld (Tel Aviv University), we have shown that 40kD CPE-deltaN can activate the canonical Wnt pathway in HEK293 cells, resulting in increased beta-catenin expression. Beta-catenin functions with T-cell factor/lymphoid enhancer factor in the nucleus to activate expression of Wnt target genes, some of which are known to promote cancer cell proliferation. It is well known that such a mechanism involving the wnt pathway could lead to colorectal cancer progression. We have also examined the extracellular role and mechanism of action of CPE in tumor cell growth and survival from several cancer types using purified recombinant CPE. We showed that rat pheochromocytoma cells, a neuroendocrine tumor cell line (PC12) secretes CPE and addition of an anti-CPE neutralizing antibody in the cell medium resulted in increased cytotoxic effects and poor survival of the cells under metabolic stress (nutrient starvation and hypoxia). This loss of function experiment demonstrates that CPE is involved in maintaining the durability and resilience of neuroendocrine tumors under this type of stress. In gain of function experiments, we found that HCC cells, that do not synthesize much CPE, and showed significantly less cytotoxicity under these metabolic stress conditions when purified recombinant CPE protein was added to the culture medium. This effect was also observed when CPE was treated with 5microM GEMSA, a specific and potent inhibitor of CPE, indicating that the extracellular role of CPE in imparting resistance to the cells during metabolic stress is independent of its enzymatic activity. We found that treatment of HCC cells under metabolic stress, with CPE, resulted in increased phosphorylation of ERK1/2 and an increase in the expression of the survival gene BCL-2, at the mRNA and protein levels. Thus CPE is a tumor pro-survival factor during metabolic stress, acting through ERK-signaling. In addition, CPE treatment caused an increase of phospho-GSK3beta; (Ser9) and active-beta-catenin, suggesting the involvement of the canonical Wnt signaling pathway. Several other genes (TNF, NF-kappa,beta, I-kappa,beta,alpha, and IL-8), which could support tumor cell survival were also up-regulated in the CPE-treated HCC cells under metabolic stress. We demonstrated the ability of extracellular CPE to inhibit migration and invasion of a very aggressive fibrosarcoma cell line, HT1080, suggesting that CPE has anti-metastatic effects in these cells. However, this anti-metastatic effect was not observed in HCC cells. The mechanism underlying the inhibition of migration or invasion by CPE is less clear. Since the Wnt pathway components can mediate cancer cell invasion, one can speculate that the negative regulation of the Wnt pathway by CPE that we reported previously could be responsible for the inhibition of migration and invasion observed with CPE treatment. Our studies also indicated that CPE can drive tumor cell survival through ERK-BCL-2 signaling, as well as activate the wnt pathway during metabolic stress. Thus CPE can have different effects in different cancer cell types. Hence, the level of expression of CPE, the tumor environment and contributions from other pathways, all dictate the final phenotype of the tumor (Murthy et al, Cancer Lett. 2013). Recently, we have developed an assay that can quantitatively measure CPE/CPE-deltaN mRNA copy numbers in exosomes prepared from cell culture media and serum from normal and cancer patients. Our results on HCC, pancreatic, ovarian and cervical cancer cell lines show significantly elevated levels of CPE/CPEdeltaN mRNA in high versus low metastatic cells of these different cancer types. Moreover, the size and exosome numbers were no different for high and low metastatic cells. Work is now in progress to evaluate the use of this assay for serum exosomes to distinguish between patients with cancer versus normal controls.
