Trials of disease-modifying drug therapies for dementia, stroke, multiple sclerosis and other neurological conditions keep failing. We need new models for these diseases.
Scientists have been trying to understand the human brain for centuries – perhaps millennia. Yet compared to other areas of medicine, there’s been a massive lack of progress in treating and understanding all sorts of neurological conditions.
In fact, neurological disease is the big black hole in medical research at the moment – and there’s a lot of reasons why. Some of it comes down to bad science, though now there’s a huge push (by journals, institutions and funding bodies) to improve rigour in research. But the one thing that we haven’t been able to figure out – or perhaps we’ve been politely ignoring – is the animal models of disease that we use to test potential therapies. One of the critical solutions to improving drug discovery is to overhaul our animal models of neurological disease, and therefore the way we test these drugs.
We have a big problem in medical research: even though we’ve been working on neurological conditions such as multiple sclerosis, stroke, psychiatric disease and traumatic brain injury for decades, we have continually failed to translate research at the lab bench into effective human drugs.
The failure rates for clinical trials are disproportionately high in this area of research. So much so that drug companies don’t want to invest in neurological conditions anymore, because so many of their incredibly costly clinical trials are failing. This is an absolute travesty, and makes it even harder to solve the issue.
Let’s consider stroke research, for example. There are well over 1,000 treatments developed for stroke that showed promise in pre-clinical trials, yet only one was successful in clinical trials. My former boss, leading stroke researcher Emeritus Professor David Howells, did what any great researcher would do – he asked why. The big red flags that emerged were flaws in the quality of the preclinical research – and these include the suitability of the animal models themselves.
We know that some of the failure of these clinical trials is because there’s variability between humans – for any individual with a neurological disease or condition, there can be huge variability in the regions of the brain (and therefore the types of brain cells) that may be affected. We also know humans often have multiple conditions at the same time, which might affect how they respond to different therapies. And we all have very different life experiences, which can affect patterns of brain wiring, and physiological variables such as the stress response. The brain is incredibly complex – there are many nuances between individuals, and we can’t deny that this is part of the problem.
But we also need to take a step back and consider the science that happens before we get to the clinical trial. And the science that is informing these clinical trials is often imperfect. We might screen a drug to see if it works in an iPSC-derived human neuron in a petri dish. Perhaps it works – excellent. Then we need to consider how it works in an intact brain full of neural networks, complete with different cell types and brain regions serving different functions. And then we need to make sure that that drug which helps your brain doesn’t injure other parts of your body, such as your kidneys.
We can’t just test a drug in a petri dish full of brain cells, see that it works, and then trial it in a human. You must first establish how that drug is going to behave in an intact organism: one with a working metabolism, with a functioning detoxifying liver, and replete with blood proteins that might affect the pharmacokinetics of the drug and alter its efficacy. We call this final stage of drug testing that informs clinical trials “preclinical research”.
Read also: Testing robot help for stroke victims
The confronting part is that preclinical research is done in animals – typically fish, mice or larger mammals. And unfortunately, these animal models of disease don’t always sufficiently represent the human condition. Our models are limited by how much we understand the disease and the organ to begin with – and the brain remains a bit of a “final frontier” compared to the rest of our bodies. For example, to test a potential stroke therapy, you must give the animal an artificial stroke via surgery. To test a potential multiple sclerosis therapy, multiple sclerosis-like symptoms might be triggered by introducing an emulsifier, virus or protein that triggers demyelination in the brain. To test a drug for dementia, mice (who normally do not get dementia) must have a piece of human genetic code (that is linked to approximately 1% of dementia cases) incorporated into their DNA. What these animal models contribute to our understanding of disease and for testing putative drugs is critically important. However, it must be acknowledged that each of these models of disease are somewhat different to the disease process that happens in a human. We account for those differences as best we can in our interpretation of the results, but it remains imperfect.
The essential ingredient to improve drug discovery for any neurological condition (and for many non-neurological conditions) is through improving the animal models of those diseases.
Believe me, no one wants science to do away with animal testing more than any medical researcher. But the reality is, with the technology and knowledge that we have right now, there’s no way around it. But we can do better – we must do better – and that’s going to be the game-changer.
One of the things I’ve been working on is trying to develop a new model of stroke that is more reminiscent of the spontaneous stroke that happens in humans. It is estimated that even modest improvements in animal modelling might achieve savings of ~$1.9bn in the estimated $14.6bn research cost of developing a new stroke therapy.
When we develop drugs, researchers tend to use the same animal model of disease that other scientists have established and used before them, perhaps without questioning whether it is sufficiently representative of the disease that we are trying to treat in humans.
However, just because our forebears have done something a certain way doesn’t mean that we can’t question it. Challenging the status quo is difficult at the best of times – especially if you are challenging the work of people who have been in the field for many decades, and you are relatively new. But we need to have the courage to challenge the status quo. We can do things better. We need to be questioning the validity of our animal models of disease, and where possible, make models of neurological conditions better represent what happens in the human brain. That’s the next big thing.
Dr Lila Landowski is a neuroscientist and lecturer at the University of Tasmania, a Director of the Australian Society for Medical Research, a Director of Epilepsy Tasmania, one of Science and Technology Australia’s “Superstars of STEM,” and regular guest expert with ABC radio. Her current field of research is stroke and fatigue.