DPUK collaboration identifies possible new target for drug repurposing in dementia
Jodi Watt, a Postdoctoral Researcher at the University of Glasgow, is part of a team involving researchers at the University of Glasgow, Cardiff University, University of Cambridge, and St George’s, University of London, who have identified a possible new drug target for repurposing in dementia. Here, they explain how the team has used research cohort data, national clinical data, and genetic data to identify this target, and why we should be excited but cautious about the findings.
Repurposing is a term that is used to describe taking medications used for one intervention and applying them to an alternative purpose. The idea of repurposing medications is appealing in dementia, as it can save both time and money – highly beneficial in the face of a global health and socio-economic issue. Whilst the traditional drug development pipeline can enable specific targeted drug design, we actually already know a fair bit about the underpinnings of dementia, and specifically to my work, the bi-directional relationship with cardiovascular health. Additionally, there are lots of drugs on the market which are designed to address cardiovascular ill-health. Which in turn begs the question – can we repurpose cardiovascular drugs for dementia?
The idea of repurposing drugs isn’t new – some of the most prescribed medications in modern healthcare were originally developed for other conditions. For example, sildenafil (tradename Viagra), probably most known as a medication for erectile dysfunction, was actually being developed for angina.
In our study, we took three different approaches to trying to identify possible drug targets. The commonality between each of these approaches is that they all employed very sizable datasets. First, we used UK Biobank – a research cohort of ~500k participants – to see if we could identify any drugs that showed a negative association with all-cause dementia, i.e had a possible role in decreasing the risk of dementia. Our first pass, using the full cohort, only identified drugs with positive associations with dementia, the opposite of what we were looking for! But this did allow us to sense-check our methodology, as we would expect medications used in neurodegenerative disease to be associated with dementia. We then took our analyses into a sub-cohort of UK Biobank, where everyone had some level of cardiovascular risk. In this instance, we were more successful and identified one drug with a negative association with all-cause dementia – a diuretic (water pill) used for high blood pressure: bendroflumethiazide.
This finding was an exciting one, as to the best of our knowledge, it was a novel one. However, of course, validation is the cardinal rule of cutting-edge science. As such, we took this finding into the Secure Anonymised Information Linkage databank (SAIL), which provided several advantages – it used NHS data, negating certain biases of the UK Biobank cohort, it was significantly bigger than UK Biobank, and it also had prescription data which allowed us to carry out dose-response analysis. Within this cohort, not only were our results successfully validated, but higher participant numbers allowed us to be more confident in these findings.
This was all well and good, but there are some questions about how thiazides – the drug family that bendroflumethiazide is part of – work. As such, we used an analysis technique called Mendelian randomization to assess associations of a genetic proxy for thiazide action, with all-cause dementia and clinical and imaging markers of vascular impairment. Here, we found an association between this thiazide proxy and a reduced risk of cerebral small vessel disease, a key risk factor for dementia.
Our results are really exciting, as they demonstrate the power of large-scale datasets and such a triangulated approach in identifying possible medications for repurposing. We can’t fully dismiss the issues of reverse causation and indication bias, and we don’t know how well our findings will generalise to other populations. Nevertheless, the consistency of these findings suggests that this would be an interesting class of drugs for future research, and for a repurposing trial.