How the Immune System Cranks Up Pain
For decades, physicians trying to relieve pain have focused on opioid drugs like morphine. Their rationale made sense: Electrical pain signals travel along sensory nerve cells from the periphery—that’s our fingers, toes, arms, legs—to a little structure in the spinal cord called the dorsal horn and from there up to the brain, where we actually feel the hurt. Opioid drugs fit into receptors on nerve cells all along the way, dampening pain signals, albeit not completely and not without serious side effects.
But in recent years, scientists have discovered that a very different kind of cell with no obvious connection to the traditional view of pain also plays a major role.1 These are the glial cells, which live in the central nervous system, where they outnumber nerve cells nine to one. Glia comes from the Greek for “glue,” so named because early neuroanatomists thought that glial cells were like glue or “bubble wrap” that just support the supposedly more important nerve cells. There are other, less lofty, translations of glia, too, including “slime,” or even less respectfully, “snot.” Glial cells are now thought to play key roles in a number of psychiatric illnesses such as schizophrenia and depression, perhaps Alzheimer’s, as well as addiction, multiple sclerosis, brain cancer, aging, and sleep.2, 3, 4, 5, 6, 7, 8 And now, pain. Yet they were completely overlooked by neuroscientists for decades.9, 10
“Like medieval astronomers who were shocked to learn that the earth is not the center of the universe, neuroscientists today are facing a similar revelation about neurons,” wrote neuroscientist R. Douglas Fields in a 2011 article for Scientific American Mind.11 For more than a century, scientists had been clinging to the neuron doctrine, the idea that all information in the nervous system is transmitted by electrical impulses over networks of neurons linked through synaptic connections.
But this bedrock theorem is deeply flawed, as Fields noted. New research proves that some information bypasses the neurons completely, flowing without electricity through networks of cells called glia. And that is completely upending the understanding of every aspect of brain function in health and disease.
It is now clear that neurons are not the only players that drive the establishment and maintenance of common clinical pain. This recognition is crucial because it offers a completely new treatment approach, as pain researcher Clifford Woolf of Children’s Hospital in Boston put it, an approach that is sorely needed because current drugs don’t work all that well against pain.12
In 2009, during a freak October snowstorm in Boulder, Colorado, I visited the glial cell research laboratory of Linda Watkins at the University of Colorado. Watkins is a vigorous woman in her late 50s, who, with her husband and colleague, Steve Maier, thinks nothing of biking 100 miles in a day. She runs 4 miles several times a week and lives in jeans, sneakers and fleece jackets. The four-footed joy of her life is her yellow Lab, Ike, whose toys—a green stuffed moose, bones, and water bowl—lie on the floor outside her office. Watkins seems both mom and mentor to her crew of “labbies,” the young scientists who carry out the experiments she dreams up to test her theories. When the freak snowstorm dumped 17 inches of snow on Boulder, Watkins looked on with amusement as her labbies poured out into the hallway, shrugged into jackets and gloves, and had a massive snowball fight in the courtyard.
Their high spirits were contagious—one young woman rushed back in, giggling. “I forgot my mittens,” she said as she dug into a box of surgical latex gloves, wiggled her fingers into them, and dashed back to the fray. Within a half hour, they all returned, rosy-cheeked, to their labs. Perched on a cluttered hallway table with her computer on her lap as her students worked nearby, Linda Watkins’s face lit up and her eyes shone as she talked about the things she was discovering about glial cells, which come in three basic types—astrocytes, oligodentocytes, and microglia.
Astrocytes, which look a bit like stars, carry nutrients and waste to and from blood vessels and mediate communication among neurons.13 Although they come from nerve cell progenitors, they act like immune cells, pumping out pro-inflammatory substances called cytokines. One of the first clues about the role of glial cells in pain came in 1994 when researchers at the University of Iowa administered a poison to animals in whom pain had been induced. The poison was designed to selectively kill astrocytes. With their astrocytes knocked out, the animals exhibited far less chronic pain.14 Oligodendrocytes also come from nerve cell progenitors; their job is to coat axons with myelin, a protective, fatty insulation that dramatically increases the speed at which pain signals travel along neurons. (In the peripheral nervous system, the task of coating nerves with myelin falls to a slightly different kind of glial cell called a Schwann cell.) The third type of glial cells is the microglia, which come from—and are—genuine immune cells. They fight infection, help repair damaged cells, and are crucial to brain functioning.15
Like many things in biology, glial cells have a Jekyll-and-Hyde personality—a good side and a dark side. A very dark side, as it turns out. Normally, glial cells are benign and quiescent. But if they become activated by pain signals from nerves, they send out chemical signals that increase pain. Sometimes, they even add insult to injury: They steal some of the very morphine a patient takes for pain relief and use it for evil, further revving up pain signals.16 When the body senses that it is under attack—from things as diverse as physical trauma, chemotherapy, diabetes, direct nerve damage, inflammation, pain transmitters like substance P, and even bits of blood that leak out of blood vessels—chemical distress signals that Watkins calls “ alarmins” land on an important receptor on the glial cell surface called TLR-4. (TLR-4 stands for toll-like receptor 4.)17
Once that landing occurs, glial cells swing into action, pumping out a swarm of chemicals, most importantly, the cytokines IL-1 (interleukin 1), IL-6 (interleukin 6), and TNF (tumor necrosis factor). These cytokines are also both good guys and bad guys. IL-6, for instance, is a great immune stimulant. When a person gets an infection, IL-6 signals the brain to produce fever, which raises body temperature, which in turn helps kill bacteria. IL-1 is a good actor, too: It helps white blood cells flock to the site of an infection to fight bacteria. But these pro-inflammatory cytokines, particularly Il-1, are devils as well as angels. In addition to their helpful immune-boosting role, they act as neuroexcitatory molecules, that is, molecules that excite nerve cells and thus boost pain.
When triggered by an activated TLR-4 receptor, pro-inflammatory cytokines float around in the spinal cord until they land on nerves carrying pain signals up to the brain, explained Watkins. This TLR-4 stimulation amplifies the original pain signal, keeping the pain pathway in a constant state of activity. As a result, nerve cells wind up firing faster and faster, generating ever more pain signals headed for the brain. In part through this mechanism, what started out as short-term, acute pain turns into long-term, chronic pain.18 The good news, noted Watkins, is that more than 200 animal studies show that preventing glial activation can reduce pain.19 In people with fibromyalgia, too, researchers have shown that stopping microglial cells from pumping out cytokines (with low-dose naltrexone, a drug normally used to block opioids) can reduce pain.20, 21
One of the first questions Watkins asked as she began delving into TLR-4 receptors was this: If activation of TLR-4 cranks up production of IL-1 and ultimately, more pain, could drugs that block IL-1 also block pain? The answer was yes, but there was a downside. If a drug blocks only IL-1, other cytokines take over.22 Which made Watkins wonder if a different way of blocking Il-1 might work better. Perhaps, she thought, she could use a cytokine called IL-10 to block its cousin Il-1. Indeed, it turns out that Il-10 can block not just IL-1, but TNF and IL-6 as well. By reducing all three, IL-10 can slow down the revved-up chronic pain cascade.
In an ideal world, doctors could get more IL-10 into the spinal cord by simply giving a drug containing IL-10, thereby reducing pain. But IL-10 drugs do not cross the blood-brain barrier, the network of blood vessels around the brain and spinal cord that has evolved to keep potentially dangerous substances out. But what if there were another way to get IL-10 where it is needed, around the spinal cord? Watkins decided to try another idea: gene therapy—injecting a gene that makes high quantities of IL-10 into the intrathecal space around the spinal cord.
To be sure, the gene therapy approach is controversial, and has had a bad name ever since 18-year-old Jesse Gelsinger died in an early gene therapy trial at the University of Pennsylvania in 1999. (In that trial, scientists used a virus to carry the needed gene into Gelsinger’s body; Watkins’s approach does not use a virus, a decided advantage.) And pharmaceutical companies may look askance at gene therapy.
“If you can’t put it in a pill and take it twice a day, they aren’t interested,” says a glial scientist, sighing.23 Nonetheless, the approach does have promise. In a series of experiments, Watkins showed that, administered this way, IL-10 can indeed shut down IL-1 in animals and make neuropathic pain disappear—for at least three months at a time.24, 25, 26 (In 2011, researchers from the University of Michigan found that a different kind of gene therapy reduced pain in a small, human study.27)
Watkins is convinced there may be other ways to boost IL-10 as well, including with a drug that closely mimics a chemical, adenosine, that we all make naturally in our bodies. To explore this approach, Watkins started with rats in which nerve pain had been triggered by a poke on the paw with a tiny mechanical hair. It typically takes 8 to 10 grams of force to elicit the normal pain reaction—the rat lifts its paw away from the stimulus.
Once it’s clear that a rat has a normal response to pain, the rat is put under anesthesia and given surgery on one leg to constrict the sciatic nerve. This causes a partial nerve injury similar to the sciatica that leads to chronic pain in people. After the operation is over, chronic neuropathic pain develops in the rats over the next 10 days, then stays stable for about three months.
What happens next in the rats is much like what happens in people with chronic, neuropathic pain. The animal—or person—becomes far more sensitive than before, even to very mild pain stimuli. Now, when the rat’s paw is stimulated by the mechanical hair, it lifts its leg sooner and sooner, and after much less forceful poking—clear signs of hyper-reactivity to pain stimuli. In technical language, this is called allodynia. Mere touch has become exquisite pain. Stanford University pediatric anesthesiologist Elliott Krane demonstrates this vividly in a TED talk, showing how, in people, the slightest touch on the skin with a feather feels like the burning agony from a blow torch.28 Once Watkins is sure that her rats have allodynia, she injects the space around their spinal cords with either an adenosine-like drug called ATL313 or a placebo.
It’s almost like magic. The rats that get the adenosine-like drug quickly go back to normal in terms of responsiveness to pain stimuli. In other words, their hypersensitivity to pain—their allodynia—is gone. The other rats—the ones that got placebo—remain hypersensitive. The beneficial effect seems to last for at least four weeks.29
And there’s still more to the unfolding TLR-4 story. TLR-4 receptors also seem to play a role in the body’s processing of the opioid, morphine. Morphine has an attraction to opioid receptors on nerve cells, as we’ve seen. But it also has an affinity for TLR-4 receptors on glial cells. And when it lands on theses receptors, it increases pain rather than decreasing it. This unfortunate phenomenon may help explain opioid-induced hyperalgesia, an increase in pain caused by the very opioids patients hope will make the pain go away.
But there’s a silver lining in all this complexity. The very fact that morphine goes to two different receptors—the ones on nerve cells and the ones on glial cells—suggests that, in theory, scientists could tinker with things to keep morphine’s good, pain-reducing effects on nerve cells and block its bad effects on glial cells. Which is precisely what Watkins is now trying to do.30 In experiments with rats, she has found that blocking glial cells means that rats do not become tolerant to morphine. (Tolerance occurs when a drug becomes less and less effective over time.) Blocking TLR-4 receptors may also be a way to reduce drug dependence and the withdrawal symptoms that occur when opioids are discontinued abruptly.
“With our little rat addicts,” said Watkins, “if we make them go ‘cold turkey,’ they get withdrawal symptoms.” But if she gives them glial inhibition—by blocking their TLR-4 receptors—they get very little withdrawal.31 So far, there’s only one anti-glial drug in human trials that appears to reduce these morphine complications: Ibudilast (also known as AV-411 and MN-166), which is on the market in Japan for asthma.32, 33, 34 Studies in rats show that the drug, now being developed by MediciNova, crosses the blood-brain barrier, can be given orally, reduces glial activation, and helps combat neuropathic pain.35 It appears to reduce the risk of developing tolerance to opioids as well. It’s currently being tested as a treatment for addiction and medication overuse in headache pain.36
Across the country from Linda Watkins, another glial cell researcher, Joyce De Leo, a pharmacologist at Dartmouth Medical School, is also pursuing glial cell biology in search of better pain treatments. In 2005, Joyce De Leo and her team were actually the first to discover the role of TLR-4 in pain.37, 38 Her team looked at rats given experimental spinal cord injuries. Untreated rats showed the classic revving up of pain signals. But “knockout” rats (those that had been genetically altered so that they lacked TLR-4 receptors) showed much less hypersensitivity. So did mice with spinal cord injuries that were given a form of DNA that blocked the gene that makes TLR-4. But De Leo was more interested in a potential pain blocker called propentofylline, a caffeine-like substance that protects nerve cells from the damage in strokes.39 She has shown that propentofylline can block both microglial and astrocyte activation—and reduce pain.40, 41 She has shown that an antibiotic called minocycline, which prevents microglial activation, can reduce chronic pain in rats.42, 43, 44 (A different drug, fluorocitrate, can do the same thing.45, 46) And she has shown that propentofylline can help with morphine tolerance in rats.47
A few years ago, things were looking great. Hopes for propentofylline were soaring. Armed by the encouraging animal studies and backed by Cambridge, Massachusetts-based Solace Pharmaceuticals, De Leo set out on a big proof-of-concept trial of propentofylline—in human patients, not rats. She selected people suffering from a stubborn pain syndrome, postherpetic neuralgia, which often follows shingles, a painful condition caused by the chickenpox virus. The drug, under its Solace name, SLC022, was hot. And unlike Watkins’s intrathecal injections, this was a pill, just what Big Pharma wanted. In 2009, Solace trumpeted the $12 million trial on its website, noting that the study design was based on a large body of preclinical and clinical data.48 The study was meticulously designed. It was double-blind, placebo-controlled, with 185 people per arm. This meant that some patients got the drug, some a dummy pill, and neither patients nor doctors knew until the code was broken who got what. The trial took place at multiple medical centers, another way of reducing researcher bias. Patients got the pills three times a day. Tests showed that, as hoped, propentofylline did reach the central nervous system.
Then things fell apart. The drug just didn’t work, said De Leo with frustration and disappointment in her voice. This happens in science. Perhaps postherpetic neuralgia was the wrong pain condition to study. Perhaps scientists simply need better ways to tell whether glial cells are being activated. And, of course, drugs that work beautifully in rats sometimes don’t work at all in humans, and nobody knows why.49
Whatever the explanation, the study results were tough to swallow, raising the question of whether glial cells will live up to their promise as good targets for human pain therapy.50 Pharmaceutical companies, which are already cutting back on neuroscience drugs because the brain and central nervous system are so complex and expensive to study, might get further discouraged.51
Still, the bad news “doesn’t mean I will lose my enthusiasm totally,” De Leo told me.52 And other pain researchers remain excited about targeting glial cells. In addition to the TLR-4 receptors, scientists are studying other receptors, including one called P2X4. When nerve cells are damaged, microglia go ballistic, ramping up spinal cord pain.53, 54 But when the P2X4 receptor is blocked, pain is blocked, too, Japanese researchers led by Kazuhide Inoue of Kyushu University have shown.55 Inoue, a neuropharmacologist, said he remains very optimistic about targeting glial cells for pain.56 Microglial cells also have receptors for chemicals called fractalkines, which are also pumped out by damaged nerve cells. When rats in chronic pain are given a drug that blocks fractalkine receptors, the animals are freed from their chronic pain.
All this work makes for an astonishing transition in medical science, according to neuroscientist R. Douglas Fields of the National Institutes of Health. “It’s as if a door has been cracked open into a room filled with an entirely new stock of drugs to cure chronic pain.”57
It’s not just glial cells that have captured pain researchers’ imaginations. T cells, the so-called managers of the immune system, are now believed to play a key role, too, explains Michael Costigan, a neurologist and pain researcher at Children’s Hospital in Boston. Microglial cells are part of the innate immune system, an evolutionarily older part of the immune system that finds and attacks invaders like viruses and bacteria by engulfing and eating them.
“Essentially, they are the first responders, executioners and trash collectors, all rolled into one,” Costigan said.58 By contrast, T cells, which orchestrate the entire immune response, and B cells, which make antibodies, are part of the adaptive immune system, a more sophisticated and recent evolutionary development. In fact, the adaptive immune system is downright clever. Cells in the adaptive system shuffle their genes to produce many different cells, each of which has the unique ability to identify and lock onto a particular protein on the outer shell of a particular invader. The body is constantly producing millions of these cells.
Astonishingly, it’s a random process—at first. By sheer chance, one immune cell will glom onto the “right” protein on a particular invader. Then things become nonrandom. That special immune cell then cranks up its cell division, making lots and lots of cells that also fit perfectly onto that particular invader. Meanwhile, T cells signal other cells to gobble up what’s left of the invading germs. And, importantly, the winning T-cell line that has emerged doesn’t “forget” its achievement. The highly specific T cells now stick around, “remembering” their successful battle, ready to attack if that invader ever enters the body again. All of which, scientists now think, plays into the problem of pain as well.
T cells are not normally found in a healthy central nervous system, as microglia are. In fact, the central nervous system is supposed to be an immunologically privileged site—no T cells allowed. But in some autoimmune diseases such as multiple sclerosis, T cells do find their way into the central nervous system. Once there, they make a big mistake. They “see” the myelin sheath surrounding nerve cells as foreign (“nonself”)—that is, they see it as an invader, and proceed to attack it; without the myelin sheath, the nerves then can’t function normally. Costigan and others now think that something similar may happen with chronic pain, specifically, that T cells somehow do gain entry to the spinal cord and other parts of the nervous system and work with microglia to ramp up pain signals.
Costigan’s team demonstrated this dramatically in 2009. Like Linda Watkins, Costigan created a peripheral nerve injury in rats, in his case, in both adult and newborn rats. After the injury, the adult rats showed ongoing neuropathic pain, but the babies did not. (In humans, too, while babies feel acute pain, they do not develop the long-term, neuropathic pain that is a disease in itself.) Why? Costigan suspected that T cells might be the culprits, ramping up pain in adults, whose adaptive immune systems had had enough time for T cells to mature, but not in newborns. In newborns, he thought, T cells might simply not be present, and thus could not have entered the nervous system. Sure enough, when Costigan looked at spinal cord tissue, he found that T cells had infiltrated the dorsal horn in the spinal cords of adult rats—and not in the newborns.59 Boosting this theory is the fact that a particular strain of mice that do not have T cells also do not develop neuropathic pain.60
That strongly suggests, Costigan says, that T cells—and the genes that control them—are involved in neuropathic pain. Once again, this suggests that neuropathic pain may be as much an immune problem as a nervous system problem. And that suggests that damping down T cells, just like damping down microglia, may be yet another way to combat neuropathic pain.
And there’s yet another important way in which the immune system may be involved in some types of chronic pain. Researchers at the Mayo Clinic in Rochester, Minnesota, have recently found that in some people with chronic pain, antibodies to a potassium channel may contribute to—and may possibly even be one cause of—the chronic pain.61, 62 In other words, in some people with apparently normal nervous systems, targeting these antibodies could open yet another new avenue for pain treatment.63
Glial cells are becoming so important in neuroscience research today that, to hammer home this point, I can’t resist telling this story, originally reported in the 2000 book Driving Mr. Albert.64
In 1955, when the brilliant physicist Albert Einstein died, a pathologist named Thomas Harvey performed the autopsy of his brain, holding in his hands the three pounds of convoluted tissue that changed our view of physics forever. After finishing his task, Harvey irreverently took Einstein’s brain home, where he kept it floating in a plastic container for the next 40 years, as neuroscientist R. Douglas Fields wonderfully retold the story in a 2004 article in Scientific American and again in a 2009 book, The Other Brain.65, 66
Every now and then, Harvey would dole out small brain slices to scientists and pseudoscientists around the world who probed the tissue for clues to Einstein’s genius. But when Harvey reached his 80s, he placed what was left of the brain in the trunk of his Buick Skylark and embarked on a road trip across the country to return it to Einstein’s granddaughter. Four pieces of the famous brain wound up in the hands of a distinguished neuroanatomist at the University of California at Berkeley, Marian C. Diamond, who found nothing unusual about the number or size of its neurons.
What she did find, in the association cortex, which is responsible for high-level cognition, was a surprisingly large number of glial cells. In fact, this “excess” of glial cells seemed to be the only difference between an average brain and Einstein’s. Perhaps, Fields speculated, glial cells might be the cellular basis of genius. And perhaps they could be the next key to unraveling the mysteries of pain.