The PI discovered the first metastasis suppressor gene, nm23. Basic and translational research has investigated the role of Nm23 in the regulation of tumor metastasis. Eleven transfection studies have documented that overexpression of Nm23 in various tumor cell lines resulted in a 50-90% decrease in tumor metastatic potential in vivo. With regard to the biochemical mechanism of action of Nm23-H1, two potential contributors were previously identified: Nm23 proteins possess a hisitidine kinase activity, for which my group reported the Kinase suppresssor of ras (Ksr) to be a substrate. A second function of Nm23-H1 that is widely reported is that of a binding protein. Nm23 binds a number of proteins that contribute to an aggressive phenotype. A new project, described below, is returning to this question. We have now reported a third potential contributor to its suppression of metastasis: downstream changes in gene expression based on microarray analysis of control- and Nm23-H1 transfectants. We have identified several genes that were down-regulated in nm23 transfectants, but not in transfectants of mutant forms of nm23 which failed to suppress in vitro motility, including c-met, CTGF, EDG2, FZD1, L1CAM, NETO2, PTN and SMOH. Confirmatory experiments included the down-regulation of the gene set in: (a) wild type as opposed to nm23-M1 knockout mouse hepatocellular carcinoma tissues; (b)independent MDA-MB-435 breast carcinoma transfectants; (c) nm23-H1 transfectants of another breast carcinoma cell line, MDA-MB-231; (d) an cohort of breast carcinomas, stratified by nm23 expression level. Transfection of each of the genes into nm23-suppressed MDA-MB-435 cells revealed that only EDG2,and to a lesser extent c-met, overcame nm23-H1 inhibition of motility. The data suggest that Nm23 inhibits tumor motility by down regulating the receptors to widely available factors such as LPA (EDG2) and HGF (Met). In order to test the hypothesis that the downregulation of EDG2 is functionally involved in Nm23-H1 suppression of metastasis, EDG2 was re-expressed in Nm23-H1 overexpressing MDA-MB-435 carcinoma cells. In a spontaneous metastasis assay, EDG2 expression restored the incidence of pulmonary metastases to 90% in contrast to Nm23-H1 overexpressing lines (52%). The data establish loss of EDG2 as a downstream target of Nm23-H1. We are in negotiation with a company that has a preclinical inhibitor to EDG2 and plan to conduct preclinical efficacy experiments. Recent work has further explored changes in protein expression levels downstream of Nm23-H1 using ICAT analysis. In addition to most of the genes identified by microarray analysis, we were suprised to find that a set of mRNA processing proteins were also differentially expressed. Of these, we confirmed the differential expression of Ascinus 1, PolyA binding protein C1, hnRNPA2B1, Bop1 and Gemin5. Gemin5 was selected for further study based on its differential expression in an independent model system and the amount of its differential expression. As Gemin5 is involved in mRNA splicing, a splice array was performed on control and Nm23-H1 overexpressing MDA-MB-435 tumor cells. Differential splicing of multiple genes was identified. For three of the top spliced genes, transfection of Gemin5 into control transfectants restored the splicing patterns to those seen in the Nm23-H1 transfectants, suggesting that loss of Gemin5 expression. Transfection of Gemin5 into control transfectants coordinately reduced in vitro motility by 50%. The data suggest that the Nm23-H1 metastasis suppressor impacts mRNA splicing through its regulation of Gemin5 expression at the protein level. The data support the unique hypothesis that changes in the expression in the levels of alternative splicing proteins may give rise to some of the proteome instability inherent in metastatic tumor cells. A new project is exploring the contribution of Nm23 binding proteins to its regulation of tumor metastasis. This project will ask, in an unbiased manner, which proteins bind to a Flag tagged Nm23-H1 and M1 in vitro and in vivo. The 4T1 murine mammary metastasis model system was developed in the lab. Briefly, 4T1 cells are implanted in the mammary fat pad (mfp), a primary tumor is removed 10 days later, and metastases are evident in spleen, liver, lung and lymph nodes eight weeks post-injections. Transfection of 4T1 cells with Nm23-H1 and -M1 inhibited metastasis to the liver by 69% and 75%, respectively. In collaboration with Dr. Tim Veenstra we have immunoprecipitated Nm23 from 4T1 cells in vitro, primary in vivo tumors and metastases (in progress); mass spectrometry was performed to identify binding proteins. We identified 90 potential binding proteins which are being validated and will be subjected to further experimentation. Translational research on Nm23 proposed that elevation of Nm23 expression in micrometastatic or overtly metastatic breast or other carcinomas may limit colonization, motility and de-differentiation, with a clinical benefit. Analysis of the nm23-H1 promoter revealed a 400 bp region which controlled expression, and contained a cassette of transcription factors regulated by a glucocorticoid response element (GRE). Deletion studies showed that these sites were functional in regulating nm23-H1 transcription. Medroxyprogesterone acetate (MPA), an unusual agonist for GR, as well as the androgen receptor and progesterone receptor, elevated Nm23-H1 expression of breast carcinoma cell lines in vitro. MPA acted via a post-transcriptional mechanism using the GR, at pharmacologic doses. We have reported preclinical experiments to determine if MPA can halt metastatic colonization. Mice were injected iv with metastatic human MDA-MB-231 breast carcinoma cells, and permitted to develop micrometastases for one month. Mice were then randomized to vehicle or MPA, the latter given in a one month induction and subsequent bimonthly maintenance dose. Mice receiving MPA had significantly fewer gross metastases in the lung, a smaller proportion of mice with metastases and smaller metastases. Immunohistochemistry revealed that MPA treated mice had a greater proportion of pulmonary metastases with high Nm23 expression. Side effects included weight gain, but no effects on bone mineral density or mammary histology. The data indicate that agents elevating metastsis suppressor gene expression may be effective against metastatic colonization. A Phase II trial of MPA opened at Indiana University, (PI Kathy Miller) in 2007, funded by an Avon-NCI grant to Dr. Miller. Patients will be post-menopausal and have tumors that are hormone receptor negative, thus promoting utilization of MPA through the GR. Patients will be randomized to MPA or, in a second cohort, MPA plus metronomic IdoCM as an anti-angiogenenic agent. The primary objective is clinical benefit rate; secondary objectives include toxicity, PK, and surrogates of MPA effect in skin and tissue biopsies and the primary tumor. Our laboratory and that of Dr. Merino will participate in the tissue analysis.
Showing the most recent 10 out of 19 publications
