We have begun characterizing the signal transduction pathways used by the c-met receptor in cells in which ligand (HGF-SF) stimulates motogenesis in the absence of mitogenesis. Primary targets (within 10-15 minutes) were identified as PI-3 kinase, GAP, PLC gamma, src, and MAP kinase, substrates which are also activated upon growth factor activation of mitogenic receptor systems. Following HGF-SF treatment, the 85 kD subunit of PI-3 kinase is phosphorylated on tyrosine and PI-3 kinase activity rapidly associates with the c-met receptor. A number of these substrates are implicated in cytoskeletal rearrangements and may be important in the motogenic response to the factor. We have also identified a number of colon carcinoma lines which express unamplified levels of constitutively tyrosine phosphorylated c-met protein. In these and other (gastric) cell lines which express amplified levels of activated receptor protein, we have determined that receptor activation is not due to the autocrine production of ligand.
We have previously shown that, in mouse NIH/3T3 cells, it is necessary to coexpress the gene for human hepatocyte growth factor/scatter factor (HGF/SFhu) with its receptor, the human met protooncogene (methu), to activate the transforming activity of the receptor (S. Rong, M. Bodescot, D. Blair, T. Nakamura, K. Mizuno, M. Park, A. Chan, S. Aaronson, and G. F. Vande Woude, Mol. Cell. Biol., 12: 5152-5158, 1992). In this study, we report that exceptionally high levels of the ligand and its receptor are expressed in tumor cell explants after several tumor passages through nude mice. Confluent tumor cells explanted after the second passage in nude mice can express 1700 units/ml/10(6) cells/72 h of scatter activity as determined in Madin-Darby canine kidney cell scatter assays. The motogenic factor produced by these cells is easily purified by heparin-Sepharose chromatography, and the purified factor efficiently induces tyrosine phosphorylation of Methu in YaOvBix2NMA human ovarian carcinoma cells. To account for the unusually high level of HGF/SFhu and Methu expression, we propose that normal levels of Methu receptor are inefficient at transducing the signal(s) required for transformation of mouse cells. Therefore, high levels of Methu receptor are required for tumorigenesis, and corresponding high levels of the ligand are required to induce the signal. Consistent with this model, endogenous mouse scatter factor is not detected in conditioned medium from cells transformed by overexpression of the Metmu receptor.
Hepatocyte growth factor (HGF) and its high-affinity receptor, c-met, have been found to increase in the whole kidney of the rat following several types of renal injury or renal hypertrophy. In an attempt to determine whether the upregulation of this growth factor and its receptor is selective for the regions of greatest anatomic change, and therefore likely to be important in regulating renal tubular hyperplasia and/or hypertrophy, we examined their expression in liver, whole kidney, and subsections of the kidney following either sham operation, transient ischemia of one kidney, or unilateral nephrectomy. The message for HGF was increased in both liver and kidney by all surgical procedures tested, including sham operation, and was seen predominantly in the outer cortex, the site of least morphological change. However, c-met was not upregulated by sham operation or in the liver, but rather was selectively upregulated only in the kidney in both the hypertrophy and hyperplasia (ischemia/reflow) models. Renal subsections revealed that this increase was confined to the renal medulla, with the greatest change in the outer medulla. Thus induction of the message for HGF can occur nonselectively and at sites distant to the injurious stimulus, whereas the target for HGF, c-met, is upregulated selectively at the site of greatest tubular injury or hypertrophy. These results support a role for HGF/c-met in regulation of these renal tubular events.
