Editor’s note: In May 2020, the Kavli Foundation awarded the Kavli Prize in Neuroscience jointly to neuroscientists David Julius, University of California, San Francisco, US, and Ardem Patapoutian, Scripps Research Institute, La Jolla, California, US. In 1997, Julius and colleagues identified and cloned the capsaicin receptor, now known as the transient receptor potential vanilloid type 1 (TRPV1) receptor. Subsequently, he identified and characterized multiple other TRP channels with roles in pain and somatosensation. Patapoutian’s team was responsible for cloning PIEZO1 and PIEZO2, mechanically sensitive ion channels that transduce sensations of touch and, sometimes, pain. The researchers sat down with PRF contributor Stephani Sutherland to talk about the discovery of TRPV1 and the PIEZO channels, implications of their work for analgesic drug development, and more. Below is an edited transcript of the conversation.
David, your research group cloned the TRPV1 channel in 1997, and although people had been studying capsaicin for many years, that discovery made it possible to study this receptor at the molecular level. Can you give a historical perspective on what that was like?
David Julius: At that time, molecular pain research was still a pretty small area. Most people involved were doing physiology, behavior, or both. There was a group at the Sandoz Institute, associated with University College London, who were among the first to really start thinking about looking at genes and proteins that define somatosensory neurons – colleagues such as Humphrey Rang, Stuart Bevan, John Wood, and Janet Winter.
But the foundation for studying capsaicin in pain emerged from the work of Nicholas Jancsó and colleagues in Hungary in the 1940s. His group established the idea that capsaicin is a selective activator of a subset of somatosensory neurons; it’s a defining functional hallmark for nociceptors. That’s really what lit the match in that area, and why people used capsaicin for so many years as a kind of pharmacological test of the nociceptor. Of course, this was especially interesting in the context of the widely appreciated sensorial effects of chili peppers and other capsicum peppers!
And then, for a long time, there was a lot of mystique around this putative receptor. I remember, even just a few months before we cloned TRPV1, that there was an article suggesting that capsaicin didn’t really have a specific receptor, that somehow it interacted with the lipid membrane, which, of course, didn’t really go far toward explaining selectivity of action. Perhaps except for Peter Blumberg’s characterization of resiniferatoxin and some structure-activity analysis of capsaicin and related vanilloids, there wasn’t a lot of pharmacology around it; there wasn’t an incontrovertible functional basis to really suggest that there was a discrete receptor constituting a specific site of action.
But we weren’t the only ones to realize that identifying this receptor, if we could do it, would be one of the first big molecular hallmarks of the nociceptor. At the time, the capsaicin receptor was to the nociceptor what the T cell receptor had been to the lymphocyte – it had the promise of being a defining functional hallmark. And of course, then the question was, Once you identify it, what toehold will that give you in understanding its contribution to the function of the nociceptor and to pain sensation?
So that was the backdrop. Many people realized that understanding the molecular basis for capsaicin sensitivity would be an important contribution, and a number of people had tried to do this. It was a time when there were a lot of new channels being cloned, and every new channel that came out, people would put those into some kind of cell and ask whether it could be activated by capsaicin. There were a lot of efforts with the candidate-gene approach, and none of those things panned out. In the end, it really took a direct function-based cloning approach, which was spearheaded by Mike Caterina when he was a fellow in my group.
Was it immediately clear to you after identifying TRPV1 that it was activated not only by capsaicin but by temperature?
David Julius: No – in retrospect, maybe it should’ve been obvious, but it wasn’t. There had been some speculation years ago that the site that was activated by capsaicin might also be activated by endogenous agents, such as bradykinin. And then, everybody was thinking that there should be some endogenous capsaicin-like molecule, a specific chemical activator that capsaicin was mimicking. And of course that turns out in some ways to be true.
The idea of heat really just came up through experimentation, where we just kept asking, “What causes pain? What kind of things can we throw at these oocytes that are expressing the receptor?” Eventually we tried hot and cold, and that’s when we figured it out. It makes sense in retrospect, but it wasn’t something where we said, “Oh, this is probably a heat receptor.”
Just before that, Peter McNaughton was one of the first to actually record heat-evoked currents from sensory neurons. Amazingly, even though people had looked at heat-evoked currents in sensory nerve fiber recordings for years, nobody before him had really put in much effort in terms of patch clamping cells and asking what heat-evoked currents looked like. We looked at those data once we started seeing heat responses with TRPV1, and you could see that there were some similarities. But there wasn’t much background on the biophysical or pharmacological aspects of those heat-evoked responses.
Ardem, let’s switch to your work now. Speaking of mystique, I think there was a sense during those early days of molecular research that studying mechanically sensitive currents or channels was incredibly daunting, partly because of potential movement artifacts. What was it like to work in that area and to identify the PIEZO channels?
Ardem Patapoutian: Yes, mechanosensation has its own similar history. Jim Hudspeth, who studies hearing, was the first to really show that these mechanically gated channels must exist and are the ones responsible for hair cell mechanotransduction for hearing. The logic there was that the time between the stimulus and electrical response is so fast that a second-messenger system – like that in taste and olfaction, and even vision, where a GPCR affects channel modulation – is too slow to account for such signaling. So that set the tone, in a way, saying that a channel directly activated by mechanical force must exist.
Another person who contributed early on was Fred Sachs, who first recorded from stretch-activated channels that we now think were PIEZO1 channels. There was a lot of controversy over whether what he was recording was an artifact of messing around with the membrane. The field oscillated between thinking it was an artifact to thinking everything is mechanosensitive. This has parallels with temperature as well, because when we talk about temperature and pressure, of course these affect biological functions.
So in that sense, when you find a molecule like TRPV1 or the PIEZOs, it settles most of these questions, because you can do much more straightforward measurements of what activates it, find the structure of the protein, etc. That’s why we think finding a receptor is so important, especially if you can heterologously express it and look at it carefully to answer the question of what really activates it and what the properties of the response are. But to me, the biggest thing is, once you have this molecule, it enables you to find novel biology or answer questions that you really want to answer in vivo, which is the role of TRPV1 in acute and chronic pain, or in our case, the role of PIEZOs in mechanosensation.
In terms of finding the PEIZOs, the most important decision Bertrand Coste, a postdoctoral fellow in the lab, and I made was, instead of looking for mechanically activated channels in DRGs [dorsal root ganglia], we took a more reductionist approach. This is a very important point, because the key for any biological question is to ask, How reductionist should I be? For us, it was finding a cell line that had this mechanically sensitive activity and doing a loss-of-function RNAi [RNA interference] screen to find it. That was the important decision. Because finding it in difficult-to-manipulate and heterogeneous DRG neurons was much more challenging.
Looking at the bigger picture of sensory physiology, could you each discuss the “labeled-line” theory, which postulates that certain sensory neurons convey particular sensations. Where do we stand now with our understanding of it? And how does this relate to the role of PIEZO channels in the phenomenon of allodynia?
Ardem Patapoutian: The answer is that it’s still very complicated. To a certain extent, of course it’s correct: There’s a labeled line for many aspects of somatosensation, and it was known for a long time that there’s a nociceptive labeled line. And there’s clear evidence that so-called low-threshold touch neurons can, in conditions of inflammation or injury, send a nociceptive signal, resulting in allodynia.
Our recent study, which was in collaboration with Alex Chesler, who’s a fantastic investigator who trained with David Julius, and Carsten Bönnemann, at NIH, included both human and mouse analyses, because they identified humans who don’t have PIEZO2. Bringing it full circle, they used capsaicin to induce inflammation, and the people without PIEZO2 didn’t have tactile allodynia after application of capsaicin. This shows that, even though acute noxious sensation is not very strongly affected by the loss of PIEZO2 in both humans and mice, this tactile allodynia appears to be through PIEZO2. [See PRF related news.]
David Julius: We’ve all wrestled with this issue in the pain field. When I came into this field, totally from the outside, the real question was specificity versus patterning, which I’ve always considered to be a bit of an academic food fight, to some degree, but there is some importance to it. Let’s boil it down to two practical questions: Does one think that the primary afferent is a place that’s specifically targetable, pharmacologically, for intervening in pain? Or do you have to go to the CNS to really do that?
The molecular studies over the years by many of us really show that there is specificity at the level of the primary afferent neuron – that all sensory stimuli are not necessarily de-convoluted just at the level of the CNS, but that there is already some specific recognition of signals and routing of signals at the level of the periphery. Not all primary afferents are created equal in terms of their function and their genetics. Maybe that was already known from fiber recordings, but the molecular biology really shows that most clearly.
So the molecular studies really do give some credence to the idea that there’s some aspect of labeled-line specificity. But this is not yet clear for all nociceptive modalities. For example, there are still a lot of unknowns with regard to how we discriminate between itch and pain. Are those really separate entities? Do they use the same pathways? Do they share some channels like TRPA1?
So there’s clearly molecular discrimination and specificity at the level of the primary afferent, but it’s still a work in progress to know how somatosensory stimuli are encoded to produce distinct psychophysical sensations. In contrast to certain other sensory systems – such as taste – that show beautiful labeled-line organization at the periphery, the pain pathway has to deal with the phenomena of hypersensitivity, allodynia, and hyperalgesia, which are critically important to the warning and guarding protective aspects of the system. That makes it a little bit harder to separate things out because the signaling gets complex, and the modalities are harder to separate in the setting of pathophysiology. Understanding mechanisms behind these phenomena are, of course, key to developing therapies.
Where are we now in terms of using knowledge about TRPV1 and PIEZO channels to find treatments for pain?
David Julius: One of the challenges is that pain is a hard space for pharma companies to work in. They came back to pain and then abandoned it in favor of things like neurodegenerative diseases and cancer. My guess is that there’s some appetite for coming back again because of the opioid epidemic, and a realization that you can’t use morphine as the gold standard against development of all other analgesics; it’s just not realistic. Developing pain drugs is tough, because you have a substantial placebo effect, and you probably need different analgesics for different types of pain syndromes.
There’s still a lot of promise for TRP channels. TRPV1 antagonists have had some on-target side effects, which are in some ways predictable: hyperthermia, which is mild and transient, as well as a diminished ability to acutely sense noxious heat. But recent trials suggest that TRPV1 antagonists may score as well as NSAIDs in osteoarthritic knee pain. I also believe that TRPA1 remains a good and rational target for inflammatory pain, and there are still numerous companies that are exploring that space.
Ardem Patapoutian: Agreed. For PIEZO, there are difficulties in targeting it with compounds, but the wide-ranging roles that it plays in mechanosensation also mean that blocking it would have massive on-target side effects.
I want to come back to the issue of appetite for pain research in pharma that David raised. Whether it was correct or not, I think much of the industry reacted to this idea that mouse models don’t translate very well to human clinical pain. There was a review article published over 15 years ago saying that, and lots of pharma companies started moving away from pain for that reason. I think that was a bit unfair because, compared to other indications in neuroscience, this is probably less of a problem.
But that’s why we’re very excited about connecting our research to human genetics. This whole idea that tactile allodynia is due to PIEZO2 gives this target an immediate plausibility that it has a good chance of working in humans. I’m not trying to “sell” this as a great target, and I highlighted its major issues, which are not insignificant. But as the field goes forward, the more connection there is to human genetics, the better it will be.
David Julius: Recently, there seems to be a little bit of a sentiment that we’ve had these molecular players for 10 or 20 years and there have been no new drugs – that it’s time to face the failure, stop looking at specific molecules, and start looking at ways we can screen for molecules that may affect a panoply of different receptors at once and have multitargeted effects. I disagree with that. If there’s a motivation to do research in the way that we’re doing it, it’s really to bring a mechanism-based platform to rational drug design: You understand what a molecule does in a process, and you pursue that, if it makes sense, as a way to develop a drug. If you say we’re not going to do rational, mechanism-based drug discovery, then basically we’re back to random screening, in a sense.
We’re as impatient as anybody; wouldn’t it be great if a drug came out against one of the targets that the field had worked on – a new analgesic? That would be fantastic, and I’d love to see that happen. But 10 or 20 years is not a lot of time to identify targets and overcome issues associated with drug development. Nonetheless, I believe that we still have to pursue the idea that you’re going to use basic, mechanism-based research to understand how things work, as a way of finding a logical approach to developing medicines. Perhaps the development of anti-CGRP and CGRP receptor antibodies, or anti-NGF antibodies for treating migraine or arthritic pain, are apt examples.
What are the big unanswered questions that motivate you and keep you excited about working in this field?
Ardem Patapoutian: After cloning the PIEZOs and characterizing some of the phenotypes, we have multiple options. You can keep looking for other sensors for acute pain sensation, and that’s a question we’re very interested in. You can also use these PIEZOs to understand more about upstream spinal cord circuitry.
But what excites me most is this idea that we do not know how mechanosensation is used in different cells for basic physiology and pathophysiology. I call PIEZOs professional mechanosensors in the sense that they seem to only sense membrane tension; wherever we see expression, there is a mechanosensory role. We want to see what cell types, what neurons, use PIEZOs to sense their mechanical environment and respond to it. And indeed, PIEZOs seem to involve many physiological processes including touch, pain, breathing, and regulating blood pressure.
David Julius: We remain fascinated with the structural work; that has really been a wonderful project that we have pursued with my UCSF colleague, Yifan Cheng, and there are still a lot of big questions to ask. For molecules like TRPV1 or TRPA1, because these channels are really signal integrators, there are a lot of interesting questions in terms of the structural biophysics, how they appreciate these different stimuli, and how they integrate them to regulate channel function – those are really at the heart of what they do. From a practical point of view, I think this work has been very helpful to people who think about drug design.
That’s on the more reductionist end. The other end of things is to realize that different types of pain syndromes have different cellular and molecular underpinnings. The field is really in a position now to ask, How do different types of pain syndromes differ? What types of cells and molecules do they use predominantly in transmitting certain types of nociceptive signals? So we’re interested in exploring what the specific differences are between different types of pain syndromes and how many of the players are redundant or different, and what that means not only for acute pain, but most importantly, for pain hypersensitivity.
Congratulations to both of you for winning the Kavli Prize in Neuroscience. What has that been like?
Ardem Patapoutian: It’s wonderful on so many levels. It’s a great honor, and a particular honor to get it with David. But of course, it’s great for everyone who’s worked in my lab, all the trainees. I think they’re getting a big kick out of it, as are our collaborators.
David Julius: I’d echo everything Ardem said. It’s really great for the field of somatosensory and pain research to have the Kavli Institute recognize this as an area they think is prizeworthy. It’s also an opportunity to appreciate the many wonderful contributions that our trainees and collaborators brought to this field, and continue to do so in their own laboratories.
Stephani Sutherland, PhD, is a neuroscientist and freelance journalist in Southern California. Follow her on Twitter @SutherlandPhD
Featured image credit: The Kavli Foundation. Photograph credit: Courtesy of UCSF (Julius); Courtesy of Ardem Patapoutian.