Editor’s note #1: This article, originally published 18 November 2022, has been updated to better articulate Dr. Ma’s perspective.
Qiufu Ma, PhD – a former professor at the Blavatnik Institute at Harvard Medical School, US, and now a professor at Westlake University, China – is likely most well known for his pioneering work characterizing how the traditional Eastern medicine practice of acupuncture can help manage inflammation. His laboratory, however, has also been hard at work mapping different pain pathways to identify those which may have clinical relevance – both in terms of understanding how chronic pain develops as well as for developing future treatments.
Yet as he and his colleagues have detailed the different cells and pathways involved with the experience of pain, Ma realized that many types of discomfort reported by real-world patients are not measureable by the common reflex-based assays used in preclinical studies. This, he maintains, might partially explain why so many promising treatment ideas fail once they reach clinical trials. In order to remedy the situation, he argues that the pain research community needs to revisit common methods of pain measurement – and subdivide the somatosensory system into two distinct, functional entities for exteroception and interoception. He penned a perspective article explaining his thinking in the January 2022 issue of Neuron.
Here, Ma speaks with freelance writer Kayt Sukel about how early human brain lesion studies and his own developmental biology guided his ideas, why it’s important to subdivide the somatosensory system by function, and what he believes pain researchers can do to more successfully translate pain research into effective therapies in the future.
Editor’s note #2: Dr. Ma recently participated in an IASP 2022 Global Year webinar titled, “A Functional Subdivision Within the Somatosensory System and Its Implications for Pain Research,” on 17 November 2022. A recording of this webinar is available at IASP’s Pain Education Resource Center.
What inspired you to write this perspective piece?
As I mentioned in the introduction of the Neuron article, we had a workshop in 2019 at the National Institutes of Health (NIH) [in the US] to discuss what was happening in clinical trials with pain treatments. So many had failed to produce effective treatments. On one hand, we had the opioid crisis. We needed new treatments to help people living with pain. But even though we had all of these great successes in animal studies – identifying promising pain genes and targets – we were rarely able to translate them into treatments for patients.
At the NIH workshop, we talked about how to build a better animal model – what animal model reflects clinically relevant pain? How can we better measure joint pain, muscle pain, or deep tissue pain, since the vast majority of studies measure pain on the skin because it’s more convenient to do it that way. The other questions discussed were what behaviors denote the sensory and emotional experience of pain, and how to measure those. As I was thinking about these questions and participating in the meeting, it all went back to early insight we gained from developmental biology studies. We needed to think about somatosensory processing in a different way.
How so?
Over the past three decades, many investigators have been studying how the sensory system can sense all kinds of different stimuli – the majority being external stimuli. Whether it’s cold selective or mechanical selective, a heat or pressure pain, these studies look at how the somatosensory system responds to external stimuli. But about 10 to 15 years ago, studies from my lab made it clear that there’s a hierarchical organization of those different cell types.
I had been at Harvard Medical School since 1999. From 1999 to 2015, my lab had been working on sensory neural development, particularly for those neurons marked by the developmental expression of the nerve growth factor receptor TrkA. Human patients with TrkA mutations lose their sense of pain and temperature. A key finding is that the TrkA lineage neurons are developmentally and anatomically segregated into two large groups, based on the expression of transcription factor RUNX1. Neurons with persistent RUNX1 expression exclusively innervate the epidermis and hair follicles, whereas neurons with transient RUNX1 expression innervate throughout the body.
This kind of anatomical organization could reflect a functional subdivision. You have one line in the somatosensory system that is there to sense external danger, as RUNX1-persistent neurons do. That’s exteroception. You have cells with unlimited expansions of sensory channel receptors that say, “Here is danger,” and, “You should withdraw to avoid an injury.” Then you also have a second line of neurons, interoceptive neurons, such as RUNX1-transient neurons that fire in response to receiving an actual injury.
Evolution is so beautifully designed. We have cells that give us this first-line defense. Even in the 1950s, Henry Beecher understood there is deeper, lasting pain due to injury that is more emotional and more sensitive to morphine. That’s not what these first-line defense neurons are doing under normal conditions, even though some of these neurons do produce short-lasting sharp pain. That is what first gave me the idea that we should be subdividing the somatosensory system, not based on how different cell types respond to different stimuli, but rather on the purpose, function, and meaning of that response.
Consider if you touch a hot plate. If you remove your hand quickly enough, there won’t be too much of an injury or even any emotional consequence of touching it. That’s the first line of defense that helps you prevent, or at least limit, an injury. But if you are in a situation that results in a deeper burn injury with tonic pain – the kind of pain that causes emotional distress in humans – that’s a different experience. But researchers, the vast majority of the time, just measure that first-line reflex using common reflexive assays. That means we are often missing very important information about what is happening during these different dimensions of pain.
Does the biology support this functional subdivision?
I first want to highlight some human brain lesion studies that will help us to understand the significance and purpose of different animal behaviors based on how they will be impacted by analogous lesions seen in humans. Back in the early 1900s, Henry Head discovered that lateral thalamus lesions lead to a loss of sensory discrimination and a loss of unpleasantness when people were given a moderate noxious stimulus that didn’t really cause damage (Head and Holmes, 1911). But other studies show when there is a lesion in the medial thalamus, people retain sensory discrimination as well as the unpleasantness and withdrawal responses to those same moderate stimuli. However, when things get more intense, the patient with medial thalamic lesions has a pain indifference. There’s no emotional distress. Thus, these human studies suggest that sensory and emotional experience requires cortical structures with: 1) the lateral thalamic pathway critical for exteroceptive sensory discrimination, via connection to the somatosensory cortex; and 2) the medial thalamic pathways are more important for interoceptive affective suffering and emotional distress to bodily injury, via connection to the anterior cingulate cortex (ACC).
We now know that a lot of that defensive first-line withdrawal behavior can be sufficiently mediated by subcortical circuits. When you remove the cerebral cortex in animal models, the animals retain the expected withdrawal behavior. That’s a big problem – because it means even if the animals have cortex-mediated conscious motivation to withdraw from the external danger, a loss of such motivation could easily be masked by the presence of an independent subcortical circuit that can subconsciously drive defensive responses. In other words, such defensive behavior does not necessarily measure the sensory and emotional experience of pain that requires cortical structures.
Several investigators then reported one form of interoceptive behavior that requires the ACC, such as persistent licking in response to bodily injury or irritation caused by inescapable intradermal injection of formalin or bee venom. Our lab showed that persistent licking responses can also be evoked by skin burn injury or by inescapable skin pinch, both of which produce tonic pain and suffering in humans. In collaboration with Martyn Goulding at the Salk Institute (California, US) and other investigators, we then discovered that there are distinct circuits from the skin to the spinal cord that drive reflexive-defensive reactions to external threats versus interoceptive self-caring licking responses to internal body injury, thereby supporting the anatomical and functional subdivision of the somatosensory system.
What do you hope other researchers will take away from your suggestion to subdivide the somatosensory system based on exteroception and interoception to improve research studies?
Reflexive withdrawal is not the same as tonic pain or affective pain. These two different responses and/or experiences rely on different parts of the somatosensory system. One part is driving that defensive reaction to a stimulus and involves subcortical circuits. Emotional distress to ongoing bodily injury and its associated tonic pain require the cortex, particularly the ACC. It’s true that there is crosstalk between these two systems, but they are responsible for distinct functions.
That said, after reviewing the literature, I’d say that researchers should rely on an assay where the behavior they are measuring requires some kind of cortical structure to generate it. For example, when you give a really painful injury that results in sustained guarding, persistent licking, or real-time operant escape, those are behaviors that require the ACC and/or the somatosensory cortex, which could measure the sensory and/or emotional experience of pain.
Will you then throw away the reflexive-defensive assays?
The answer is no. The functional segregation between exteroceptive and interoceptive neural circuits, observed in naïve conditions, can be disrupted under pathological conditions. Primary sensory afferents associated with exteroceptive discrimination and/or reflexive responses, such as low threshold mechanoreceptors and Mrgprd+ polymodal nociceptors, also send excitatory inputs to spinal neurons associated with interoceptive affective pain, but these inputs are normally masked via feedforward inhibition. Following central sensitization and/or disinhibition induced by nerve lesions, these afferents can now gain the ability to drive affective pain and/or pain comorbidities. Thus, we may need a cocktail of drugs that blocks both exteroceptive and interoceptive circuits for effective chronic pain treatment.
I believe the pain field, as a whole, needs to work on developing better models as well as new assays that more effectively reflect the sensory and emotional experience of pain. Most behaviors we’ve discussed are associated with evoked pain, not spontaneous pain, which bothers patients the most. We need more assays to measure spontaneous pain. We also need more studies to measure pain from deep tissues, such as in muscles, joints, bones, and visceral organs, rather than predominantly focusing on pain from the skin. With these new assays, I think we will see that our studies will result in more successful translations.
Kayt Sukel is a freelance writer based outside Houston, Texas.