Differential effects of transforming growth factor-beta on proliferation of normal and malignant rat liver epithelial cells in culture.

Transforming growth factors (TGF-betas) have been shown to cause both stimulatory and inhibitory effects on cellular growth in a variety of normal and neoplastic cells. The nature of the inhibitory effects of TGF-beta on proliferation of different cell types is at present unclear. We have used freshly isolated rat hepatocytes, a normal diploid rat liver epithelial cell line (NRLM), and a subline (AFB) derived from it which was transformed in vitro by aflatoxin B1 to study the nature of TGF-beta-induced growth inhibition and its alteration following chemically induced neoplastic transformation. TGF-beta had a vastly different effect on proliferation of normal rat liver epithelial cells (both freshly isolated and NRLM cells) compared to aflatoxin B1-transformed cells. TGF-beta at 20 pg/ml caused 83% inhibition of colony formation of NRLM, whereas the growth of AFB cells was unaffected by TGF-beta at concentrations as high as 10 ng/ml. A parallel dose-dependent inhibition of DNA synthesis by TGF-beta was observed in both primary hepatocytes and NRLM cells at concentrations between 10 pg and 10 ng/ml. No inhibition of DNA synthesis was observed in AFB cells. Furthermore, TGF-beta did neither induce anchorage-independent growth of NRLM cells nor affect the growth of AFB cells in soft agar. TGF-beta-induced inhibition of the NRLM cells was irreversible in nature, since treated cells were unable to proliferate and form colonies upon removal of TGF-beta from the medium. Also, NRLM cells showed, after 4 days in the presence of 20 pg of TGF-beta per ml morphological changes characterized by cytoplasmic hypertrophy and the formation of abundant liposomal derivatives, some of which resemble lipofuscin. The finding that TGF-beta caused a high degree of irreversible inhibition of NRLM cells emphasizes the need for caution in interpreting data from inhibition studies, since most assays presently used are designed for assessing growth stimulation in vitro and do not adequately distinguish between the possible cytotoxic and/or cytostatic action of growth inhibitors.