Assessing the welfare of mice used in cancer research

Mice are widely used in cancer research. This may cause pain and distress to the animals but few studies have assessed how much pain this causes and whether this is a major welfare concern. To apply appropriate humane endpoints (and effective cost-benefit analyses of projects) we urgently need objective assessments and not the relatively arbitrary guidelines that are used currently. We have already obtained preliminary evidence that bladder cancer causes pain in mice, but variation in tumour growth meant we were unable to establish firm guidelines on an appropriate (humane) end-point. In this study we will use several different approaches to address these issues. We will assess long-term behaviour changes (using automated analyses developed in our previous project), peripheral sensitivity changes and new techniques designed to address the problem of assessing the affective component of pain (how it makes animals feel). Parallel imaging studies will monitor cancer growth, so that we can determine how these different measures of pain and nociception change in animals with tumours. The measures of subjective/affective state are especially important, since it remains possible that the behavioural and other signs we attribute to pain actually have limited welfare consequences. We will use two methods, the conditioned place preference paradigm (CPP) and conditional drug discrimination. In the CPP paradigm, whether mice are feeling pain as tumours develop can be inferred by whether their preference for an environment where previously they were exposed to an analgesic (morphine) progressively increases. If mice are ‘feeling' pain, in conditioned drug discrimination studies they should be able to be trained to use changes in this (due to analgesic treatment) as a cue to perform a task for a food reward. If the degree of pain is of little consequence, or if they do not have an internal state of "what it is like to be in pain", they would not show enhanced morphine seeking in the CPP paradigm, and would fail to distinguish between the analgesic and saline in drug discrimination trials. Correlating theses measures of subjective state with changes in behaviour, peripheral sensitisation and measures of tumour growth will give us a clearer picture of the pain caused by different types of tumour. This will allow us to refine endpoints in cancer studies. Of potentially even greater importance is that we will have established new methods for assessing the welfare consequences of other potentially painful procedures applied in laboratory rodents.

• Further Funding: NC3Rs Project grant, Neoplasia and pain in laboratory animals, January 2005, £212,204