Normal haematopoietic cell regulation involves interaction between marrow stromal cells and haematopoietic progenitor cells which may be facilitated by specific recognition and adhesion. Some leukaemogenic events might produce a selective growth advantage by altering this regulatory network, possibly by diminishing the capacities of cells to adhere to stromal elements. Using an in vitro culture system which allows investigation of adhesion to stromal layers and subsequent colony formation by blast colony-forming cells (B1-CFC) in normal marrow and Ph+ chronic myeloid leukaemic (CML) blood, we compared the adhesive properties of normal and malignant progenitor cells. We present evidence that altered adhesive interactions between primitive progenitor cells and marrow stromal cells occur in CML.

Summary A prospective randomised study was carried out to compare the effect of mesna (2-mercaptoethane sulphonate sodium) with that of forced diuresis in preventing cyclophosphamide induced haemorrhagic cystitis in marrow transplant recipients. Sixty-one consecutive BMT recipients were randomised for treatment with forced diuresis or mesna. The incidence of macroscopic haematuria was significantly lower in the mesna treated group (X2 = 4.03, P <0.05). No specific side effects of mesna were detected. The lymphopenia induced by cyclophosphamide in the aplastic recipients was similar in the mesna and forced diuresis groups suggesting that mesna has no effect on the lymphocytotoxic activity of cyclophosphamide, although 6 out of 7 episodes of graft failure documented in the study occurred in mesna treated patients. As a result of this study our present policy is to use mesna in all BMT recipients but to continue careful documentation of the incidence of graft failure.

Inhibition of apoptosis (genetically programmed active cell death) by p210 BCR-ABL expression is a mechanism that might contribute to clonal expansion in chronic myeloid leukaemia (CML). Since cell death following exposure to ionizing radiation and many chemotherapeutic agents can occur by the apoptotic pathway, inhibition of apoptosis would be expected to confer a relative resistance to these treatments. Similarly, cells deprived of growth factors in vitro die by apoptosis, and inhibition of apoptosis would therefore be expected to allow cells to survive better in growth factor-deprived conditions. We found that the survival of normal and CML myeloid progenitors was the same after in vitro incubation in deprived conditions and after treatment with X-irradiation or glucocorticoids. We also found that mature cells in colonies produced by CML progenitors (CFU-GM) did not survive better than those produced by normal progenitor cells. Flow cytometric analysis of propidium iodide-stained cells provided a direct indication that the degree of apoptosis may correspond to the degree of deprivation. These results suggest that inhibition of apoptosis may not be the primary mechanism whereby BCR-ABL influences the expansion of the malignant clone in CML.

Several clinical oncology units are studying the roles of in vitro hematotoxicology in phase I evaluations. At the same time, the European Center for the Validation of Alternative Methods (ECVAM) is supporting a validation study of the CFU-GM assay. It is important that these activities be coordinated so that high performance, optimized technical protocols are used for prospective and retrospective clinical evaluations. The EROTC, the NCI and ECVAM could provide support for these coordinated efforts. There is an opportunity for medical oncologists involved in early clinical trials to participate in the evaluation of in vitro tests and their clinical application . Fundamental to acceptance of these assays by oncologists and regulatory scientists, they must predict clinical outcome for myelosuppressive agents and then improve phase I design and performance. These achievements would justify more aggressive dose escalation schemes using guidance from in vitro studies without compromising patient safety. Success in predicting neutropenia might also stimulate the research required to understand how to predict other hematologic toxicities, such as a thrombocytopenia. The complexity of a validation study in hematotoxicology is that it seeks to predict the level of exposure that causes neutropenia, in contrast to other validation studies that have sought to classify a xenobiotic as toxic or not. It may be that the clinical relevance of a new assay is not just a yes-no answer. This important distinction came from the realization that the xenobiotic tolerance in other organ systems of the body must be the same or greater than marrow in order for myelosuppression to be a clinical consequence of exposure. Pharmacological principles of system exposure and toxicity that are integrated into the prediction model provide the links to clinical oncology. It is also important to anticipate future applications of in vitro hematotoxicology. If the maximum tolerated level of drug exposure for human hematopoietic cells can be predicted, then in vitro hematotoxicology could play an important role in new drug discovery. One concept involves screening for compounds that show efficacy at the IC level that predicts maximum tolerated exposure levels in the human. 'Therapeutic index based' drug discovery has been applied to the tallimustine family with some success.