The development of new anticancer treatments relies heavily on animal models, which are essential for assessing the efficacy and toxicity of drug candidates before clinical translation. Nevertheless, the success rate of molecules in development remains low, in part because preclinical data struggle to predict actual toxicity and efficacy in humans. Compounding this scientific challenge is an ethical requirement: the so-called 3Rs guidelines (replace, reduce, refine) mandate the preservation of animal welfare by limiting any form of suffering and reducing handling. In this context, automated digital ventilated cages (DVC®) offer the possibility of continuously monitoring animals' locomotor activity directly within their home cage, without human intervention. The value of this technology, however, had remained little explored in oncology.
The authors assessed the usefulness of continuous locomotor activity monitoring in mice using the DVC® platform, employing two reference compounds in oncology, cisplatin and cyclophosphamide. SCID mice, which exhibit severe combined immunodeficiency affecting B and T lymphocytes and are commonly used as hosts for the xenotransplantation of human tumors, were housed in groups of four, across twelve cages in total. They were monitored by standard clinical examination and by digital measurement, before and after treatment, over a period of seventeen days.
The platform detected statistically significant effects of cisplatin on the activity of the mice, both in the short and long term, as well as trends for cyclophosphamide. The differences between treated and control groups were particularly pronounced at night, the period of peak animal activity during which staff are generally unavailable for checks. Analysis of the twelve-hour nighttime phase proved the most relevant for discriminating treatment effects. Cisplatin exerted a more chronic effect on activity, detectable from day 7 to day 13, whereas cyclophosphamide showed a more transient effect at the doses tested. Notably, standard daytime clinical assessment—whose most contributive variable was coat deterioration rather than impaired mobility—proved poorly sensitive for detecting these disturbances in overall activity; conversely, body weight variations provided complementary information.
The authors acknowledge certain limitations, in particular the small number of cages per group, the inability to measure individual rather than average activity, and the potential stress induced by the dual clinical and digital monitoring. They nevertheless emphasize that this continuous, non-invasive monitoring, which requires little human intervention, can be integrated into a multimodal surveillance approach in order to better capture drug effects on antitumor efficacy, toxicity, and safety, and thereby improve translation to clinical studies. Further work, drawing on other models and other compounds, remains necessary to confirm these observations.