Animal models for evaluating the myelosuppressive effects of cancer chemotherapeutic agents.

An important objective of new anticancer drug discovery programs is identification of agents that are less myelosuppressive than those currently available. We have developed several animal models to evaluate these drugs for myelosuppression. Our screening model measures changes in neutrophil counts in mice as an indicator of myelosuppression. This model correctly predicted the myelosuppressive effects of 13 (76%) of 17 known agents. Cisplatin, carboplatin, spiroplatin, and marcellomycin caused no reduction in the neutrophil counts, representing four (24%) of 17 false negatives. Our secondary evaluation system is the more labor-intensive murine CFU-C assay on femoral bone marrow cells from drug-treated mice. Known myelosuppressive drugs such as mitomycin C, doxorubicin, and BCNU, as well as the false negatives from the mouse neutropenia model (cisplatin, carboplatin, spiroplatin, and marcellomycin) caused marked inhibition of colony formation 24 h after dosing; bleomycin was inactive. Advanced evaluations are performed using ferrets in which neutrophil counts can be monitored in the same animal for 28 days after treatment. Mitomycin C, doxorubicin, and BCNU caused significant reductions in the neutrophil counts whereas bleomycin had no effect. Importantly, cisplatin and marcellomycin also caused significant reductions in the neutrophil counts. Although the mouse neutropenia model is a rapid assay, there is potential for false-negative predictions. It is important that other test systems be used for more advanced evaluation of drugs identified by this model as being less myelosuppressive than reference drugs. The mouse CFU-C and ferret hematology models are suitable for this purpose in that they can identify the false-negative predictions as well as identify less myelosuppressive drugs such as bleomycin.