Direct and indirect effects of recombinant human granulocyte-colony stimulating factor on in vitro colony formation of human bladder cancer cells.

Although the present experimental use of recombinant human granulocyte-colony-stimulating factor (rG-CSF) has been proven to alleviate the myelosuppression induced by antitumor chemotherapy, it is also believed to stimulate growth of some nonhematopoietic tumor cells. We investigated both the direct and indirect effects of rG-CSF on in vitro colony formation of human bladder cancer cell lines using a modified human tumor clonogenic assay. Peripheral blood mononuclear cells (PBMC) were used as feeder cells (a mixture of 5 x 10(4) monocytes/dish and 5 x 10(5) lymphocytes/dish obtained from healthy donors). Human bladder cancer cell lines KK-47, TCCSUP and T24, all derived from human transitional-cell carcinomas, were incubated continuously with various concentrations of rG-CSF ranging from 0.01 ng/ml to 10 ng/ml both with and without PBMC for 7-21 days. The concentrations of rG-CSF used were chosen as being in the range of achievable serum concentrations in patients treated with rG-CSF. At the end of incubation, colonies were counted under an inverted phase-contrast microscope, and an increase in the number of colonies in comparison with the control was used to evaluate the effects of rG-CSF. Results were expressed as a percentage of controls. rG-CSF in the upper layer at concentrations ranging from 0.1 ng/ml to 10 ng/ml stimulated the colony formation of all the cancer cell lines tested in the absence of PBMC in the feeder layer, whereas cells with PBMC in the feeder layer were significantly stimulated more than those without PBMC in the feeder layer (P < 0.05) up to a certain concentration, which varied from cell line to cell line. At higher concentrations of rG-CSF, no further stimulation but, on the contrary, a decrease in colony formation was observed in cells with PBMC in the feeder layer in all the cell lines tested. Colony formation in KK-47 and T24 cell lines was significantly inhibited at 5 ng/ml and/or 10 ng/ml rG-CSF compared with cells without PBMC in the feeder layer. Our results suggest that rG-CSF may have both direct and indirect stimulatory effects on the growth of human bladder cancer cell lines in vitro. The results obtained also raise the possibility of adverse effects of rG-CSF in bladder cancer patients whose malignant cells may be directly and indirectly stimulated by this factor while it is being used clinically to alleviate the myelosuppression induced by antitumor chemotherapy.