Differential transforming abilities of non-secreted and secreted forms of human fibroblast growth factor-1.

Fibroblast growth factor (FGF)-1(1-154), the precursor for acidic FGF-1(21-154), is a potent angiogenic polypeptide, the structure of which lacks a signal peptide sequence for secretion. To investigate the biological significance of this structural feature, we have attempted forced secretion of FGF-1 through fusion of the entire FGF-1 coding frame with the signal peptide (sp) from the hst/KS3 gene, a secretory member of the heparin-binding growth factor family. We also studied the transforming ability of the signal-less forms of FGF-1 comprising FGF(1-154) and FGF-1(21-154). The presence of a soluble and biologically active form of FGF-1 was readily detected in the conditioned medium of NIH 3T3 cells transfected with sp-hst/KS3:FGF-1(1-154) as demonstrated by Western blot analysis and DNA synthesis assays, whereas sp-hst/KS3:FGF-1(21-154) was not detectable in conditioned medium even though the protein was detected in cellular extracts. The secreted form of sp-hst/KS3:FGF-1(1-154) stimulated the proliferation of human umbilical vein endothelial cells in vitro and was able to induce receptor-mediated tyrosine phosphorylation. Furthermore, the forced secretion of biologically active FGF-1 resulted in NIH 3T3 cell transformation as demonstrated by altered morphology in vitro, the formation of discrete colonies in soft agarose, growth under serum-free conditions, and ability to rapidly form highly vascular tumors in vivo. Interestingly, sp-hst/KS3:FGF-1(21-154) also mediated the transition to a transformed phenotype despite the inability to detect extracellular FGF-1 in the media conditioned by these NIH 3T3 cell transfectants. Although the transfection of FGF-1(21-154) yielded similar NIH 3T3 cell morphologic changes, these transfectants did not grow under serum-free conditions or yield colonies in soft agarose, and formed tumors in vivo with delayed kinetics. Furthermore, the FGF-1(1-154) NIH 3T3 cell transfectants did not exhibit morphologic changes, and this may be due to the inability of mRNA to express protein. These data suggest that although non-sp forms of FGF-1 may alter the monolayer phenotype of NIH 3T3 cells in vitro, the ability of FGF-1 to transform NIH 3T3 cells requires the function of a sp-directed secretory pathway and suggests that this pathway increases tumorigenicity in vivo.