Incorporating the 3Rs (Refinement, Replacement and Reduction of animals in research) into the preclinical assessment of snake venom toxicity and antivenom efficacy

Abstract

Antivenom is the only effective treatment for snakebite and comprise immunoglobulins obtained from venom-immunised horses or sheep. Globally, more than 45 manufacturers make over 120 snake antivenoms; it is a regulatory requirement that the venom-neutralising efficacy of all antivenoms are assessed preclinically. The World Health Organisation (WHO) recommended preclinical tests of efficacy are the median lethal venom dose (LD50) and median effective antivenom dose (ED50) assays performed in mice. They result in substantial pain and suffering to the mice with death/survival as their metric. With NC3R-funding, we sought to apply the ‘Refine, Reduce and Replace’ principles of animal experimentation to these murine assays. Pain is a near-universal symptom of snake envenoming, and one of our objectives was to identify an effective analgesic that could be utilised without invalidating the assay results. The Mouse Grimace Scale and Activity scores were used to measure pain. We examined the effects of two opioid analgesics, buprenorphine and morphine, in a range of venom LD50 and ED50 assays. Both were effective at reducing pain scores, but death rates were higher in those which had received buprenorphine, hence morphine is preferable. We demonstrate that each venom exhibits a distinct set of lesions, the severity of which appears time and dose dependent, and that the observed murine pathological lesions show significant similarities to those reported in envenomed human victims. Applying the 3R principles, we have used pathological observations, in combination with ante mortem observations, to establish more humane end-points, consequently reducing the duration of LD50 and ED50 assays from 24 to 6 hours. In addition, we have implemented a ‘dose-staging’ element into experimental design in which one dose is given and the next dose(s) selected based on the results of the previous dose, reducing total mice required. To reduce the numbers of assays, and therefore mice, we have shown an excellent correlation between in vitro binding assays, cytotoxicity neutralisation assays and in vivo ED50 using antivenoms derived from the same pool of donor animals. Comparison of the results of in vitro binding assays between 35 different venom/antivenom combinations showed a poor correlation overall, but the correlation improved when each of five venoms were considered separately. The possibility of replacing the in vivo LD50 and ED50 tests, using a cell-based neutralisation assay was investigated using two cell lines from diverse tissues of origin, namely VERO epithelial-type cells and neural SH SY5Y. All venoms studied produced a cytopathic effect in both cell lines, with the VERO cells being more sensitive to viper venoms and SH SY5Y cells to the effects of elapid venoms, when both cell lines were grown in co-culture. However, variability of results made optimisation of a neutralisation assay inadequate for use as an alternative to in vivo tests.