Pain Research in Animals: Why Do It, and What Can It Tell Us?

ThumbnailJeffrey Mogil, PhD, is the E.P. Taylor Professor of Pain Studies and the Canada Research Chair in the Genetics of Pain in the Department of Psychology at McGill University in Montreal, Canada. Mogil has long studied pain genetics using animal models, and has contributed significantly to the evolution of animal experimentation in pain research. Freelance journalist Stephani Sutherland recently spoke with Mogil about the rewards and challenges of studying pain in animals. Below is an edited transcript of their conversation.

Why do researchers use animal models to study pain?

Basically, we use animals because we don’t have any other choice if we want to understand the causes of pain. There are two main types of experiments one can do in science. One can measure things that are associated with some condition or disease—like chronic pain—and compare them to the measurements of those same things in people who don’t have pain, for example. That type of experiment has a major limitation, which is that you’re only seeing correlations, and there’s really no way to prove that anything you’ve measured is the cause of the pain.

In order to find out the cause of things, you have to do another type of experiment, which involves manipulating factors one by one. You hold everything else constant, and you make a change in one factor, and then you see what impact that has on pain, for example. You simply can’t do that type of experiment in humans for two reasons. One, it’s impossible to hold everything constant with human beings. And two, for most of the factors we would like to manipulate, we can’t do it practically or ethically in humans. We might want to get rid of a particular brain area, or give a particular drug with unknown actions, or activate a particular set of neurons, and see what happens. You can’t do that in people. To perform the important experiments that show you the cause of pain, there’s simply no way around animals in research.

What is the importance of animal, or “preclinical,” research in developing new pain therapies?

All advances in biomedicine come from animal research—they all do. We preclinical researchers feel a little bit underappreciated by the public. The public doesn’t hear much about our work; they only hear about the final experiment, which is the clinical trial of a drug. There’s this idea that pain research is the clinical trial at the end, but that’s just the final study in a series of studies—sometimes hundreds of studies going back for decades. Clinical trials don’t happen unless there’s an entire literature of basic science—preclinical research in animals—that supports the idea of the clinical trial. Without that work, there will be no new clinical trials of anything. I think that’s important to remind everybody.

What animals are most commonly used in pain research, and why?

The animals that we’ve settled on to do experiments throughout biomedical research are basically small mammals—mice and rats. Why? The fact that they’re small is good, because you can keep a lot of them in a room, and it’s affordable to house and feed them. They’re mammals so they’re pretty close to humans—not as close as primates but closer than fish or birds. How much like a human? I don’t really know. Certainly the default assumption of all of biomedical research is that mice and rats are close enough to humans to make it worth our while to study them and generalize the results to humans. Does that work? It works most of the time. Does it work all of the time? No, of course not.

Talk about the specifics of working with animals in pain research. How can you tell if animals are in pain, considering they can’t talk?

That is a really interesting question that I’ve devoted a lot of my career to. One thing that’s a little bit surprising is that everyone would think that it would be pretty easy to tell if an animal was in pain. It turns out to be pretty complicated, but it’s not because they can’t talk. I’ve never quite understood why people think there’s a fundamental difference between humans and animals in this respect. Yes, humans talk. You ask them how much pain they’re in, and they say seven on a scale of one to ten. Well, what does that mean? How do I know that their seven is my seven, and how do I know that they’re not trying to exaggerate their pain so I’ll take them more seriously, or minimize their pain because they don’t want a fuss made over them? I would point out that a human saying the word “seven” is no different, in theory, than an animal licking its hind paw. They’re both pain-related behaviors that have to be interpreted.

So that’s what we do: we try to interpret animal behaviors as indicating pain. The problem then becomes, which behaviors? And what turns out to be interesting and frustrating is this: if we give a stimulus that we have every reason to believe is going to produce pain but very short-lasting pain—up to a few hours, say—then animals make very vigorous behaviors, and it all works out very well. They lick and shake their paws (assuming the noxious stimulus is put into the paw). They make obvious behaviors that you can simply count up, and the more behaviors they make the more pain they’re in, and this turns out to be fairly straightforward. We know that animals are doing these behaviors for about the same amount of time that humans would.

The problem arises when we use stimuli or injuries that we have reason to believe would cause pain for a much longer period of time: for days, weeks, or even months. Of course, these are the injuries that we really want to study, because they are clinically relevant. It’s hard to convince yourself you’re really studying something like arthritis or diabetic neuropathy if you inject a substance that causes pain in an animal for 30 minutes. That’s not what arthritis is like, and that’s not what painful diabetic neuropathy is like. So we try to produce injuries that, in theory, would put these animals in pain for longer periods of time.

What occurs then?

An interesting and strange thing happens: when you use such injuries, the animals don’t make any behaviors at all. They simply don’t do anything, and you can’t tell which of the mice are in pain after these chronic injuries. What researchers have resorted to is to try to get evidence by provoking pain in these animals—they’ll poke at them with fibers or heat beams, and then you can demonstrate that, indeed, mice and rats respond more vigorously and more quickly and to lower levels of force or lower temperatures. But the problem with that, of course, is that these are reflex withdrawals. Although clinical pain patients sometimes show these changes in reflex responses, a lot of times they don’t. Even if they do, it’s not really their major problem, which is spontaneous pain: pain that is just there, even without being provoked by a painful stimulus.

The big question in animal pain research right now continues to be, how do you measure spontaneous pain in animals? What are the behaviors to look at? It’s here that just in the past few years there have been some really important advances. Our little contribution to this was to point out a few years ago that mice and rats—and other species too—grimace in pain just like humans do, and that this can be used to measure spontaneous pain, at least for a certain amount of time. Other methods have been proposed too. People are starting to change what they’re measuring in animal research in ways that more closely parallel what human clinical pain is really like.

Interestingly, in one way, rats and mice are not the best choice we could have made to do pain research, because they are both prey species. Prey species have every reason to hide their pain from potential predators like us. So in some sense, one might imagine that part of the problem we’re having is that we’ve actually settled on the wrong small mammals, and maybe we actually should be using big mammals that aren’t prey. But of course, that’s simply not going to happen for practical reasons.

What are the ethical guidelines for using animals in pain research, and how is this regulated at academic research institutions?

In the US and Canada, the national government sets regulations that institutions have to adhere to, and the institutions have what are known as Institutional Animal Care and Use Committees. To do any experiments on any vertebrate animal, one needs to write a protocol to tell the committee exactly what you’re going to do, exactly how many animals you’re going to use, and what your plan is to minimize the number of animals you use and to keep the amount of pain you inflict on the animals at a minimum and still be consistent with your experimental goals. The committee makes recommendations, and if the work is approved, then the committee continues to monitor the research and inspect the animal facilities. So the regulatory process is fairly laborious, but necessary. We take our responsibility to our animals very seriously, and we pay a lot of attention to make sure that everything is done in the most efficient and modern way possible.

I have two motivations for making sure the animals are as happy as they can be. One is because it’s an ethical imperative, but the other is that I would be shooting myself in the foot if we did anything to make our animals upset or stressed beyond normal levels that would be expected in such experiments. We have every motivation scientifically to take as good care of them as we possibly can.

There is a common misconception: people have a greatly exaggerated idea of the amount of pain that’s being used in pain research experiments. We wouldn’t want to use very high levels of pain, because what we do know about animal behavior is that when animals are in severe pain they don’t do anything; they simply freeze. When they freeze, there’s nothing to measure at all, and so you couldn’t actually do an experiment if the pain levels were too high. The pain levels need to be manageable in order to be relevant to the human situation, but also so that the experiments work.

Are there alternatives to using animals in research?

Some people say that you don’t need animals—that you can just model a disease on computers or study it in cells. The problem with this idea of modeling chronic pain is that we have no idea what to model. To make a computerized system that would be predictive of anything, you would have to know a lot more about the system than we do. So that’s just a fantasy. In terms of cells in a dish, you can do very important experiments on cells in a dish, and people do all of the time. It’s just that real brains are not cells in a dish. They’re far, far more complicated than that. What people think about experiments on cells in a dish is that they’re a good place to start, but you would absolutely not know if anything you were learning was relevant until you took it to a whole organism and did an experiment in a live, breathing, behaving organism, and then we’re back to mice and rats.

Why haven’t animal studies yielded more new pain treatments?

The big hurdle is that pain is a phenomenon that’s mediated by the brain—the most complicated object in the universe. People greatly underestimate the complexity of the problem. We’ve made great advances in biomedicine, but most of those great advances have been made on organ systems that are much, much simpler, such as the heart or the liver. When you consider pain, you really have to remember that the size of the problem is much, much bigger. And we do have pain-relieving drugs that work. Frankly, finding them has depended more on luck and less on brilliance, but surely there have been advances in pain drug development. But it hasn’t gone as fast as everyone would have hoped; this is true.

Why is that? Some people blame the animal models; part of the problem in pain research is that we spent way too much time on reflex symptoms of pain and not nearly enough time trying to measure spontaneous pain. Other people think that the problem in translating animal work to humans is actually with the clinical trials–that something has changed in the way clinical trials are run that make it harder to demonstrate that drugs work. There’s reason to believe that this is true. To be honest, it’s probably a combination of all of these things. This is a very lively topic of discussion currently in the field.