Nitrosoureas are the drugs most effective in the treatment of patients with intracerebral malignant glioma. Their limiting toxicity is delayed myelosuppression. A prospective, randomised crossover study of recombinant human granulocyte-macrophage colony-stimulating factor (rhGM-CSF) was performed in patients receiving BCNU for relapsed glioblastoma, to investigate whether the resulting haematological toxicity profile could be modified by rhGM-CSF. Adequate data for analysis were obtained in 13 patients. Following BCNU, the nadir neutrophil count was higher in 12 out of 13 patients during the rhGM-CSF-protected cycles compared with the unprotected cycles. The median nadir was also significantly higher (1.79, CI 0.76-3.52, P < 0.005). Five episodes of neutropenia (< 2 x 10(9) l-1) occurred during the unprotected cycles compared with none in the rhGM-CSF-protected cycles (P = 0.076). There was no evidence of any effect on platelets. This result shows that the haematological toxicity profile following therapeutic doses of BCNU can be modified. It suggests that rhGM-CSF and other growth factors should be investigated for clinical efficacy in chemotherapy using nitrosoureas.

The functional importance of mast cells in modifying tumour cell behaviour has long been recognised and recent evidence suggests that mast cell: tumour interactions contribute to the growth and development of specific cancers. We have used in vitro studies to examine such interactions in relation to tumour cell proliferation. Soluble mast cell products (MCP) were prepared from Furth mastocytoma cells at 106 cells/ml. The addition of MCP (10% v/v) to cultures of human melanoma cells (SK 23 and DX3.LT5.1), breast carcinoma cells (8701-BC) and normal skin fibroblasts resulted in increased rates of cell proliferation. Although histamine H2 receptors are expressed by the melanoma cells, the addition of exogenous histamine (5 and 25 μM) had no effect on the proliferative rate of the human melanoma cell lines, or on carcinoma cells derived from breast and prostate. The data indicate that exogenous histamine was not mitogenic for three tumour cell cultures and that other factors derived from mast cells appear to be essential for the increased proliferative response.

Mast cells were shown to accumulate around the periphery of the invasive and metastatic rat mammary adenocarcinoma (MTLn3), and histological evidence of mast cell degranulation was observed during the later stages of this model. To assess the physiological role of mast cells in vivo we have used the mast cell-stabilising compound FPL 55618 applied i.p. daily at 1 mg kg-1 for 23 days. Using groups of 12 rats we have found that this compound inhibited tumour growth at the primary site by as much as 70% in most of the treated animals compared with the control group which received equivalent volumes of saline. When the drug treatment was stopped after 23 days, tumour growth of the test group accelerated over the next 7 days and reached a similar tumour size to that of control animals. Histological studies of the tumour and contiguous host tissue at day 24 of the experiment revealed numerous extra-tumoural mast cells often showing signs of degranulation at several sites around the tumour periphery in the control animals. Such observations were not seen in those animals receiving FPL 55618 where, in contrast to controls, numerous intact mast cells were often seen within the tumour mass. Following cessation of the MC-stabilising treatment progressive mast cell activation was evident within 2-4 days, primarily at the tumour periphery. In vitro studies have shown that drug concentrations equivalent to five times the in vivo dose had no effect on the proliferative rate or viability of the MTLn3 cells. Moreover, the proliferative rate of these cells in culture was significantly increased when exposed to soluble mast cell products. Thus our data indicate that a mast cell-stabilising compound has significant benefits in reducing tumour growth in vivo, an observation which supports the concept that mast cell:tumour cell interactions are important for the growth and invasive properties demonstrated by this model of breast carcinoma.