BACKGROUND: The metastasis-suppressing role of the NM23 gene in the metastatic spread of solid tumors is still debated. We examined the role of NM23 in tumor development and metastatic dissemination by using transgenic mice that lack mouse NM23 (NM23-M1) in two mouse models of hepatocellular carcinoma (HCC) that recapitulate all steps of tumor progression. METHODS: We induced HCC in mice that contained (NM23-M1(+/+)) or lacked (NM23-M1(-/-)) NM23-M1 by diethylnitrosamine injection or by a crossing scheme that transferred a transgene that leads to liver expression of simian virus 40 large T antigen (ASV mice). We used microscopic examination and immunohistochemistry to analyze tumor progression. Expression of Nm23 protein isoforms (Nm23-M1 and Nm23-M2) and several tumor markers was analyzed in the primary tumor and in metastases by Western blotting. The statistical significance of differences in the incidence of Nm23-M2 overexpression in null mice relative to that in wild-type mice was tested by a one-sided Fisher's exact test. The statistical significance of differences in the incidence of metastases was examined using one-sided chi-square tests. All other statistical tests were two-sided. RESULTS: In both models, Nm23-M1 and/or Nm23-M2 were overexpressed in the primary liver tumors compared with nontumor liver tissue; however, the lack of the NM23-M1 gene had no effect on primary tumor formation in either model. ASV mice developed pulmonary metastases that were positive for the Hep-Par 1 antibody, which recognizes a specific hepatocyte antigen, whereas the few pulmonary nodules that developed in diethylnitrosamine-injected mice were negative for this antigen. Statistically significantly more ASV/NM23-M1(-/-) mice than ASV/NM23-M1(+/+) mice developed lung metastases (69.2% versus 37.5%; difference = 31.7%, 95% confidence interval = 13.1% to 50.3%; P<.001). In ASV/NM23-M1(+/+) mice, immunohistochemical staining for Nm23-M1 was highly heterogeneous among the primary liver tumors, but weak or negative among lung metastases. CONCLUSIONS: The lack of NM23-M1 expression promotes metastasis in the SV40 animal model of liver carcinogenesis.
BACKGROUND: The metastasis-suppressing role of the NM23 gene in the metastatic spread of solid tumors is still debated. We examined the role of NM23 in tumor development and metastatic dissemination by using transgenic mice that lack mouseNM23 (NM23-M1) in two mouse models of hepatocellular carcinoma (HCC) that recapitulate all steps of tumor progression. METHODS: We induced HCC in mice that contained (NM23-M1(+/+)) or lacked (NM23-M1(-/-)) NM23-M1 by diethylnitrosamine injection or by a crossing scheme that transferred a transgene that leads to liver expression of simian virus 40 large T antigen (ASV mice). We used microscopic examination and immunohistochemistry to analyze tumor progression. Expression of Nm23 protein isoforms (Nm23-M1 and Nm23-M2) and several tumor markers was analyzed in the primary tumor and in metastases by Western blotting. The statistical significance of differences in the incidence of Nm23-M2 overexpression in null mice relative to that in wild-type mice was tested by a one-sided Fisher's exact test. The statistical significance of differences in the incidence of metastases was examined using one-sided chi-square tests. All other statistical tests were two-sided. RESULTS: In both models, Nm23-M1 and/or Nm23-M2 were overexpressed in the primary liver tumors compared with nontumor liver tissue; however, the lack of the NM23-M1 gene had no effect on primary tumor formation in either model. ASV mice developed pulmonary metastases that were positive for the Hep-Par 1 antibody, which recognizes a specific hepatocyte antigen, whereas the few pulmonary nodules that developed in diethylnitrosamine-injected mice were negative for this antigen. Statistically significantly more ASV/NM23-M1(-/-) mice than ASV/NM23-M1(+/+) mice developed lung metastases (69.2% versus 37.5%; difference = 31.7%, 95% confidence interval = 13.1% to 50.3%; P<.001). In ASV/NM23-M1(+/+) mice, immunohistochemical staining for Nm23-M1 was highly heterogeneous among the primary liver tumors, but weak or negative among lung metastases. CONCLUSIONS: The lack of NM23-M1 expression promotes metastasis in the SV40 animal model of liver carcinogenesis.