This is the seventh in a series of Forum interviews with PRF science advisors.
Stephen McMahon, PhD, casts a wide net to catch the molecular mediators of pain. He is the Sherrington Professor of Physiology at King’s College London in the UK, and his laboratory at the Wolfson Centre for Age-Related Diseases uses techniques from molecular biology and electrophysiology in animals to genome profiling in patients. He has a longstanding interest in identifying molecules that cause pain after tissue damage or in disease. Recently, he has been developing strategies to use human pain models and tissues in the earliest stages of research to boost the chances that preclinical findings will translate into effective therapies. McMahon is an architect of structures that facilitate collaboration, directing the London Pain Consortium and also serving as the academic leader of Europain, a research partnership between universities and pharmaceutical companies across Europe. He is co-editor of Wall and Melzack’s Textbook of Pain, 5th Edition, and the 6th Edition, to be published in August 2012. McMahon spoke with Megan Talkington about his far-flung lines of inquiry and the changing ways in which researchers can work together to advance understanding and treatment of pain. The following is an edited transcript of their conversation.
Can you tell our readers about your general approach to pain research?
We have pursued a bunch of different topics, but one strong theme that has run through my research has been trying to identify novel pain mediators. By that, I mean molecules that are released in the vicinity of damaged or diseased tissues and act directly or indirectly to drive the pain signaling system.
There is considerable evidence that a great deal of chronic pain is sustained by the presence of these mediators over months and years, or even longer. In a variety of conditions, manipulations that temporarily interfere with this peripheral process of mediators driving nociceptors are effective in transiently relieving chronic pain. You can put local anesthetic into a diseased joint, and people feel better. You can take the joint out and replace it with a metal joint, and not everyone, but the majority of people get better—suggesting very strongly that there are mediators there.
Throughout much of the 1990s, we were focused on the idea that nerve growth factor might be one of these mediators. We pursued that avidly, along with several other people around the world, and I think the work, together, provided a very strong preclinical base for the idea that NGF might act as a pain mediator, particularly in inflammatory conditions. Part of NGF’s action is to directly excite nociceptors, and part of its action is to be retrogradely transported and drive alterations in gene expression that probably sustain the chronic pain state.
After that, there was a gap in the research, and then about five years ago people started testing NGF-blocking antibodies in the clinic. This is one of the few examples we have where a very comprehensive preclinical analysis preceded clinical testing and was found to be successful. There were Phase 2 and 3 trials underway that were halted because of adverse events, but in March the FDA [US Food and Drug Administration] gave approval to start them again. The published efficacy in osteoarthritis pain and smaller trials in other conditions is, I think, extremely encouraging that this way of doing pain research will work.
Are there still things to learn about NGF and pain?
Yes. One of the reasons the trials were halted was that it looks as if the use of anti-NGF for any reason may be associated with an increased risk of joint replacement in some patients. The numbers are small, and, as I say, the FDA has given its approval for trials to continue. But the biology behind those problems is not understood, and that almost certainly will spur a lot more research.
I suspect that if these things are sold as drugs, there will be a renewed effort to examine other features of NGF biology that haven’t been explored yet. What we don’t know, given the amount of work that’s been done, is surprising, actually. The relative role of different NGF receptors is not definitively established. The dependence on transcriptional change for driving chronic pain states is not strictly established. There really are many things that haven’t been definitively answered. And I’m sure new things will be thrown up by the clinical trials, because when thousands of people are exposed to drugs, sometimes interesting things happen. I was just talking to a colleague and saying I really think now is the time to go back and address the questions that slipped through the net in that first decade when we worked on NGF.
What are some of the other pain mediators you are looking at?
I think there’s a very good body of evidence suggesting that ATP may be a mediator released peripherally in some pain states, and that it acts on P2X3 receptors, which are pretty selectively expressed on nociceptors. Again, this is not just my belief—there’s quite a large body of experimental work supporting the idea. That’s been slow in coming to fruition, partly because the chemical design of highly selective receptor antagonists proved very difficult. There is a whole family of P2X receptors expressed ubiquitously in the brain, and only one, P2X3, is selectively expressed on peripheral nociceptors. The trick has been to get molecules that selectively antagonize the putative pain-producing effects of ATP. I’m not trying to promote the whims of any one company, but we got some of Roche’s compounds, and they are extremely efficient, highly selective P2X3 receptor antagonists, and they certainly have efficacy in some preclinical models that we have examined.
Ironically, the success in the chemistry department at Roche came at exactly the time the company decided to abandon some of its neuroscience efforts. The compounds are now being taken forward by a spinout company, Afferent Pharmaceuticals, and they are starting clinical trials. So the role of P2X3 in pain will get explored in the clinic, and I think there is still more we might do to understand the biology of that molecule.
In the last couple of years, I have shifted gears again, but still under the heading of exploring potential pain mediators. There is a subclass of cytokines known as chemokines that have been well described in immunological processes but poorly described from a sensory neurobiological perspective. I think they are very good candidates to be peripheral mediators that drive abnormal sensory behavior. We published a paper last year [Dawes et al., 2011] showing that one of these chemokines, CXCL5, which had no published record of being important in pain, was a pain mediator in some situations.
So the pursuit of mediators that are dysregulated in chronic injury states and drive pain-related behavior has preoccupied me for quite a long time. The ways in which we do things have changed a bit; I’m more focused on doing experiments on humans now rather than just starting with animal tissues. I think if you make some effort to get hold of pathological tissues, then you can ask many questions of human samples that you might more traditionally ask using experimental animals. But the theme is the same: I think there will be a variety of mediators that play roles in different pain states.
Of course, none of this is to dismiss the central changes—in the spinal cord, brain stem, and brain—in chronic pain states. They actively contribute to the chronic pain condition. It’s just that I believe that, for many patients, there are mediators that are driving those central changes, and if you take away the peripheral mediators, the central changes will have less impact on their own. As a therapeutic lever, the mediators are important.
What are you doing outside of pursuing peripheral mediators?
I have a few new things that have caught our attention. I don’t know whether this is like a blackbird being attracted to things that glitter. But we’ve spent a few years working in the arena of genetics, particularly from the perspective of human genetics, and recently we’ve started to explore the role of epigenetics in chronic pain states. I don’t know how this will turn out. For the peripheral mediators, I think there is no doubt—there are multiple lines of evidence suggesting that there are mediators that drive a lot of chronic pain behavior. Epigenetics is a much newer field—there’s lots of talk, but not much real data yet.
My interest in genetics started about five years ago after a chance encounter with Tim Spector, who runs a twins unit at St. Thomas’s Hospital in London and has a registry of thousands of monozygotic and dizygotic twins. That led to us to do a classical hereditability study. The idea is that all of the twins have common environments for much of their early life. But the identical twins share 100 percent of their genes, and the non-identical twins share, on average, 50 percent. You take a trait (anything you like—hair color, blood pressure, susceptibility to pain), and you ask whether or not that is more similar amongst identical than non-identical twins. The degree to which it is similar allows you to get estimates of heritability—how much of the trait is conditioned by the genes these people carry.
We looked at experimental pain traits—the response to experimental stimuli in healthy twins. We did that because otherwise there is a tendency to end up studying genes that predispose people to a particular disease—if you try to find out whether low back pain is heritable, you may well find genes that are driving structural changes that lead to low back pain. Studying experimental pain is a way of getting a purer signal, although, of course, there is the issue of how applicable that will be to chronic pain settings.
As a child of the ’60s, I was never very keen on the nature bit of the nature versus nurture argument. But I think it is now irrefutable that genes are driving many aspects of our behavior and physiology. In our heritability studies, it turned out that, for many of the pain traits we looked at, 50 percent or so of the variance appeared to be explained by genetic factors. There may be a few caveats there, but that is the straightforward interpretation.
Are you also studying the twins to investigate epigenetic factors?
Yes. One thing we are looking at is identical twins that are discordant for a particular trait. The idea is that the twins share genetic identity. But they may have different environmental inputs affecting their physiology and their behavior. Some of those things may play out through epigenetic mechanisms. We are collecting identical twins who have very different sensibilities to experimental pain. We think that will be a very useful system for asking whether these people have changes in methylation or acetylation of their DNA that may be consistent with their altered pain sensitivity.
In animals, we are looking at the impact of a number of chronic pain models on genomewide acetylation, and then we’ll want to ask whether the changes in acetylation are functionally important to the emergence of chronic pain. It would be reasonable to ask whether some intractable forms of chronic pain in patients are associated with epigenetic marks that are driving abnormal gene expression and abnormal behavior of neuronal circuits. That is something we are very keen to explore.
When you first start experiments, what you have is lots of optimism and no data! This is the optimistic phase—hope springs eternal; this could be the answer, and if it’s not, well, then we’ll all drown our sorrows in some other way. But for a couple of years, I think it will be fun to explore the possibilities.
Besides your own projects, you lead two prominent networks of pain researchers—can you tell us about that?
I have been lucky in the last 10 years to be involved in a slightly new wave of doing science. Ten or 20 years ago, the Wellcome Trust did some very interesting demographic research on publications in biological research. They found that the typical research paper had one and a half authors, from one institute, whereas, of course, now a typical research paper has a whole football team on it, from three or four institutes. It really has been a sea change. Science has become collaborative, which is obviously a good thing. When I was a student, most laboratories were one man and his dog—and it usually was a man (and it wasn’t quite a dog; it was usually a medical student).
Well, clearly that’s changed, and funding agencies have driven that change a bit. They’ve offered a number of opportunities to interact. The European Union has tried a few things—they’ve tried big networks and small networks. I don’t know that there is a consensus about what works best, but I do think there is a feeling that networks of a modest number—five or 10 laboratories, 15 maybe—can be highly productive and synergistic. We’ve been involved in several of those approaches. Our own London Pain Consortium was a response to a grant call to set up exactly that—a consortium of people interested in one topic.
More recently, we received a large grant from the European Union to run a bigger consortium between pharmaceutical companies and academics called Europain, which has about 20-odd members, to study issues in the biology of pain that are limiting development in the field.
One of the things these kinds of networks have made possible is genetic studies. To identify genes that have very small influences on pain susceptibility, you need to coordinate quite large cohorts of patients—typically thousands of subjects. One of the reasons that genetics has become possible is the lucky coincidence of technology that allows screening across the whole genome in an affordable way, and the emergence of these consortia, which tend to have budgets that enable them to tackle these bigger problems.
In a way, I’ve got no idea whether large-scale genetic studies will be productive or not. There hasn’t been a formal GWAS [genomewide association study] in chronic pain published yet. We have a number of studies of candidate genes or approaches that have combined GWAS concepts with other preclinical datasets, and they haven’t revealed factors with very big effect sizes.
It’s not yet clear how many genes there will be and what effect sizes they will exert. But one of the reasons I find it appealing to work in this area is that, in theory, you’re getting away from your own prejudices. Most of the research that preclinical scientists have undertaken in the past has been to pursue what, in a cop show, would be called a “hunch”: You intellectualize a problem, and you have a smart solution that is worthy of testing. Obviously, that has been very successful, but it has also been very limited. The genetics offers a way of shining a light on factors that we wouldn’t otherwise come across.
What other possibilities do you see in Europain?
There are several clinical members in Europain, and I think that offers an opportunity for the most radical changes. Many preclinical scientists could do interesting experiments and provide useful insights about clinical material or clinical data, if they were able to work with it. But for logistical reasons, it’s very difficult to do experiments on humans. By having members in Europain who work with and collect patient groups, there is a great opportunity for preclinical science to synergize with some of those patients. I believe we can ask more powerful questions by applying our preclinical approaches in a clinical setting.
One example is a surgeon, Henrik Kehlet in Copenhagen, who has a very interesting approach to looking for risk factors for the development of chronic pain. Some patients get chronic pain as a consequence of surgery, and some surgical interventions lead quite frequently to a chronic pain state. Because these people’s pain arises only after the surgery, you can do a prospective study: You can study people in their normal, pre-surgical state, then you can see who gets the chronic pain condition.
In our own laboratory, we have a longstanding interest in the use of several models in human volunteers, particularly UV radiation as a model of inflammatory challenge. One of the things we are planning to do with Henrik as part of the Europain project is to ask whether or not the immunological response to UV challenge given in individuals before they have surgery is a predictor of their susceptibility to developing chronic pain after surgery.
There is quite a broad sweep of activity in the consortium. One question that has affected the pain field for a long time concerns the robustness and appropriateness of our animal models. So one of Europain’s work packages (as we say in Eurospeak) is focused on improving existing animal models, and validating them from a pharmacological perspective.
Other than the increase in collaborative projects, have you noticed other important shifts in the way pain researchers do science?
The first paper I ever sent off, one of the reviewers wrote, this paper would be enhanced if they did some immunostaining. And I wrote to the editor and said, we don’t do that. And he said, fine, we’ll publish it. The world has certainly changed. Now, the way science is done, you are forced to become a jack-of-all-trades. Actually, I am not sure that it’s for the best. But it does surprise me that you’re only really limited by your imagination now—and money. The technical advances are so prolific.
Your ability to measure almost anything you want means that it’s now much easier to do proper hypothesis-driven research. It used to be much more a matter of following your nose and doing what was possible. That puts the onus on scientists to really decide what are the important questions.
Is there a project you would love to do but that has not been possible yet?
One of the things that has changed over the last couple of decades is the study of the brain generally. In neuroscience, and in pain research, it has become less and less obligatory to ever talk about the activity of nerve cells. I’m sure it will soon be possible to go to an entire neuroscience meeting where no one talks about nerve activity—they just talk about, say, gene expression—whereas the functional unit of organization of the brain is the neuron, and its activity is encoded in discharges. If there were a Holy Grail, it would be to be able to monitor that activity. We can’t do that very easily in conscious humans—we can do it indirectly through imaging technologies, but that doesn’t have the resolution we need to really do proper systems biology. So if there is a fantasy, I think, it would be to harness enough power to define the circuitry underlying pain states.
I understand that the next edition of Wall and Melzack’s Textbook of Pain, which you edited with Martin Koltzenburg, is due out in August. Are there major changes to watch for?
There are significant revisions. One thing that didn’t exist in the last edition is a chapter dedicated to the immune system and pain. We had something on peripheral mediators that involved some immunology, but no central immune reactions. The last edition was six years ago, so the field obviously existed then, but it has grown dramatically in the interim.
The molecular biology of pain has also dramatically progressed, so several chapters have changed considerably. And we have had a fresh look at the topic of animal models.
Obviously, there are many new developments that are represented in the chapters, but it’s more of an evolution rather than a radical rethink. In the last edition, we had reorganized the whole thing, and this time we’ve gone more for a progression than a fundamental shift in the organization of the content.
Besides your textbook, are there particular books or papers that you think everyone in pain research ought to read?
I think they should go back to very simple things, actually. One of the problems you see very much with your own students is that, on the things that relate to their own experiments, they know everything, inside and out. But pain is such a broad church that it is difficult to put the area where an individual works into an appropriately broad context.
I would recommend some really simple books. I thought The Challenge of Pain by Pat Wall wasn’t a book that trivializes science, but it tried to highlight what the questions were. I think there is a lot of value in understanding the breadth of the problem when you are doing your day-to-day experiments—seeing how your experiments fit into the bigger picture.
I think people are under more time pressure now than they used to be. And it is a guilty secret that one of the things that’s easy not to do, when you are under time pressure, is read. In a sense, it doesn’t matter what you are reading, but you ought to be reading something. We have a lot of pain research where I’m based in King’s College in London, and then the Pain Consortium itself encompasses a lot of pain research. So we put on quite a number of tutorials for our students, where we try to highlight a range of developments that are taking place in the field by getting them to critically evaluate that work. As I say, if you keep reading sensible things, it will be useful. The real danger is not to read enough. It’s difficult to keep up, and students need to work hard to become at least au fait with the broad thinking in a variety of disciplines that are applied to pain.
Is there anything you wish I had asked but didn’t?
I thought you were going to say, “What do you think are the hottest areas?” And there are a few things I keep my eyes on. I certainly think it looks as if mechanotransduction is being unraveled now. That will have obvious important functional consequences and will be of great academic interest.
It’s been a classic race to find mechanotransducer proteins. People are screening for them and desperately trying to clone them, and the first ones are out—Piezo1 and 2. Clearly, progress is being made.
Also, I applaud the systems neuroscientists, the imagers, the people trying to apply EEG and evoked potential studies to the whole of brain function; I think they have the greatest opportunity for advancing knowledge in a way that hasn’t been possible so far.
The more integrative things… This is really the power of science. However much we like to feel our own contributions are fascinating and seminal, it’s the cumulative nature of science that really makes it go places.
Thank you for sharing your thoughts with our readers.
Good to talk with you.
PRF Related Content:
News: FDA Gives the Green Light to Restart NGF Antibody Trials (13 Mar 2012)
News: Novel Ion Channel Senses Painful Touch (23 Feb 2012) (on Piezo proteins)
News: EFIC VII: Taking Aim at Cancer and Bone Pain (12 Oct 2011) (includes coverage of McMahon’s research on P2X3)
News: Sunburn Pain? Check Your Chemokines (12 Jul 2011)
View Stephen McMahon’s profile on Pain Research Forum (requires member log in)
Denk F, McMahon SB. Chronic pain: emerging evidence for the involvement of epigenetics. Neuron. 2012 Feb 9; 73(3):435-44.
Norbury TA, MacGregor AJ, Urwin J, Spector TD, McMahon SB. Heritability of responses to painful stimuli in women: a classical twin study. Brain. 2007 Nov; 130(Pt 11):3041-9. Epub 2007 Oct 11.
McMahon SB, Bennett DL, Priestley JV, Shelton DL. The biological effects of endogenous nerve growth factor on adult sensory neurons revealed by a trkA-IgG fusion molecule. Nat Med. 1995 Aug; 1(8):774-80.
Ronald Melzack and Patrick Wall. The Challenge of Pain. Penguin Press,1982 (First Edition).
Other Forum Interviews:
Pain and Its Control: A Conversation with Allan Basbaum (6 June 2012)