Basic researchers rarely set foot in the clinic, and clinicians are often unaware of the research possibilities for the patients they see. How to bridge the divide? Anne Louise Oaklander, MD, PhD, has tried to answer that question. Oaklander is an associate professor of neurology at Harvard Medical School in Boston, US, and an attending neurologist at Massachusetts General Hospital, where she directs the nerve injury unit and neurodiagnostic skin biopsy service, and cares for patients with some of the most difficult forms of chronic pain. In addition, she runs a research laboratory investigating chronic pain and itch disorders. She is passionate about carefully characterizing patients’ symptoms, bringing attention to under-recognized pain syndromes, and grounding mechanistic research in clinical reality. Oaklander welcomed Megan Talkington into her office, where they discussed what researchers can learn from patients, the satisfaction of discovery, and the challenges of communicating new findings. The following is an edited transcript of their conversation.
Could you tell our readers about your research interests?
Our lab has a number of projects ongoing, with the common theme being sensorineural dysfunction. I am particularly interested in diagnosis, as this provides the basis for treatment, and I enjoy discovering new neuropathic syndromes, new diseases. It may seem unlikely that in the year 2012, there are still diseases to be discovered, but in fact, there are. That is the central focus of my career, and something that brings me great satisfaction.
The first disease that we helped to put on the map is neuropathic itch, and specifically post-herpetic itch. I remember as a postdoctoral fellow seeing patients who had had shingles and whose major problem was not pain but unremitting itch. I remember standing in the hallway while a patient showed me that she had scratched off one of her eyebrows because of her itch. I thought, wow, I need to look into this, because I don’t remember seeing any papers or review articles on post-herpetic itch.
Then, years later at Massachusetts General Hospital, I met an extraordinary patient whose shingles-induced itching was so unrelenting that she had scratched entirely through her skull and into her brain. She had given herself an epidural abscess that was pressing on her brain, and part of her brain that had herniated out through her skull was necrotic and had to be removed. Skin was grafted into her dura to cover the hole, but even after this major neurosurgery, she still could not stop scratching, and she scratched through several of these dural grafts during her hospitalization. She ultimately was left with severe brain damage, including personality changes, hemiparesis of half of her body, sensory changes, and seizures.
We published a detailed psychophysical and neuropathological study of what in her nervous system could possibly permit this seemingly insane behavior [
Oaklander et al., 2002]. I should mention that she had been scratching for 10 months before I met her; indeed, she was presumed to have OCD [obsessive-compulsive disorder] and was under the treatment of a psychiatrist—which was the wrong diagnosis and not helpful to her.
We established that she was scratching skin that was nearly devoid of nociceptive innervation. Skin biopsy and quantitative sensory testing showed that she was insensate to almost all stimuli on the part of her scalp that had been affected by shingles, and that she only perceived itch. We concluded that the self-injury was driven by the conjunction of unremitting neuropathic itch and loss of protective pain sensation, which normally limits scratching. We had great difficulty getting this paper published; it was rejected from the leading medical and neurology journals with itch viewed as “an unsuitable topic.” However, the patient and I were later interviewed by both Science magazine and The New Yorker, so this paper put neuropathic itch on the map.
We later studied itch in several populations of former shingles patients [
Oaklander et al., 2003], and used information from patients to create a rat model of post-herpetic itch [
Brewer et al., 2008]. Actually, the model had already been created by Robert Yezierski and his colleagues at the University of Florida. I was merely the person who realized that their model of spinal cord injury that they had interpreted as modeling pain was more likely modeling neuropathic itch, because those rodents scratched through their skin into their internal organs, just like my patient.
Are there other syndromes out there that you are trying to characterize?
Yes! Current efforts include rediscovery of a neuropathic pain syndrome that is right in front of us yet largely unrecognized. I refer to Tarlov cysts, which are small cysts that form only on the sensory nerve roots. They form in the arachnoid space that is pulled distally when the cell bodies of primary afferent neurons migrate out of the spinal cord to form the dorsal root ganglia during embryogenesis. This space allows cerebrospinal fluid to track along the sensory nerve roots through tenuous connections. In some cases, fluid is forced into the neural tissues when intracranial pressure increases (during a cough or bowel movement, for instance), and this gradually causes cysts to form. These cysts in some cases damage the sensory axons and cell bodies in the dorsal root ganglia, and the most common symptom they produce is neuropathic pain.
These perineurial cysts were first described in 1937 by neurosurgeon Isadore Tarlov, who discovered them at autopsy. He had no information about the patients’ symptoms, so he opined in his first paper that they did not cause clinical symptoms. He later treated patients who had these cysts during life and recognized that they are a cause of radiculopathy, much like a herniated disk. But his later papers did not receive adequate attention, because everyone is taught to cite the first paper. So a medical myth was born that Tarlov cysts are irrelevant lesions, and it remains widespread today. In fact, radiologists often see Tarlov cysts on MRIs but don’t report them, because they have been taught these cysts are an incidental finding of no medical significance—just as a dermatologist might not report freckles.
I learned about Tarlov cysts from patients. My first patient with Tarlov cysts had vulvodynia—pain in the vulva present for decades and often attributed to psychiatric causes. Many treatments had been tried, including vulvar vestibulectomy, or amputation of the outer parts of her vulva, a standard treatment; however, she had never seen a neurologist before me. When I ordered spinal cord imaging, it revealed large Tarlov cysts. At first I had no idea if these were related. Since there was no mention of Tarlov cysts in the textbooks or recent literature, I had to obtain and read the historical papers, which opened my eyes and enabled me to reformulate her illness as neurologic rather than psychiatric.
This is important because Tarlov cysts can be treated and in some cases cured completely using surgical or percutaneous procedures that definitively collapse or remove the cysts, but only if the correct diagnosis is recognized and the clinician knows the treatment options. So Tarlov cysts are a curable cause of painful radiculopathy—and they are not rare, it turns out. But because so few physicians know about them, most patients never get adequate diagnosis and never get treated.
I proposed symposia on Tarlov cysts to be presented at the American Pain Society meeting, and for two years in a row the scientific program committee turned down my request because, they said, no one had ever heard of this, so how important could it be? In fact, pain practitioners should welcome this information as they could learn to do the percutaneous CT-guided cyst aspiration, which has been published as an effective and definitive treatment.
What research are you doing on Tarlov cysts?
Because I am a woman who specializes in neuropathic pain, I was sought out by so many Tarlov cyst patients that it became inescapable for me to make the correct diagnosis and begin to investigate the pathophysiology. In 2010, several colleagues and I published the first paper on Tarlov cysts in the pain literature [
Hiers et al, 2010]. A subsequent paper in press in the New England Journal of Medicine describes a detailed study of a nurse who had 21 Tarlov cysts, and despite many medical visits for her chronic neuropathic pain, the cause remained undiagnosed for most of her adult life. When she died unexpectedly of leukemia, I arranged for a very detailed autopsy and neuropathologic study. I hope that once the NEJM paper appears, the pain community will take another look. And I hope that the neurology, neurosurgery, radiology, and gynecology communities will become aware of this disease entity as well so that more patients can get correct diagnosis and treatment.
I obtained the world’s first-ever research grant on Tarlov cysts, and we conducted the largest study of radiologically identified cysts, and the first report of cyst symptoms, from a cohort of 500 symptomatic patients. A big part of the problem is that the average Tarlov-cyst patient is a woman with chronic pelvic pain, but the average spine clinician is a man, which has impeded forthright communication needed to establish the link between symptoms and cause. Some of the patients don’t even mention their pelvic pain and dysfunction, and it’s not a topic that many male physicians feel comfortable asking their female patients about. This has created a “don’t ask, don’t tell” situation that neither side—the clinicians nor the patients—has been able to bridge, to bring this condition to medical and public awareness.
Are you investigating any promising new treatments for neuropathic pain?
We do have a major treatment effort underway on noninvasive brain stimulation for neuropathic pain. I’ve chosen to get involved in devices for several reasons, including the fact that the brain works predominantly by electrical transmission. And while there are a number of medications effective for neuropathic pain, they are limited by systemic side effects. The advantage of this therapy is that it usually affects only a very small area under the electrode, and there are few, if any, systemic side effects. So these devices are better tolerated by most people.
My postdoctoral fellowship in pain was in the neurosurgery department at Johns Hopkins, where I was exposed to the implanted neural stimulators for pain. Noninvasive brain stimulation translates this to the outside of the body. The technology that is furthest advanced is called transcranial magnetic stimulation, or TMS, which involves using high-strength electromagnets to generate electrical pulses that travel through the skull, dura, and cerebrospinal fluid to trigger action potentials in the cortex. The technique is being used to investigate and treat several neurologic dysfunctions including stroke, movement disorders, and visual problems. TMS has been approved by the FDA [US Food and Drug Administration] for treating major depression. For treating neuropathic pain, it is the motor cortex that is often targeted. It’s not clear if it’s the firing of the motor axons themselves that produce the pain relief or whether it is the second-order synapses affecting the thalamus, for instance. But several clinical trials find TMS is effective for a number of neuropathic pain syndromes and other syndromes that may or may not be neuropathic, including fibromyalgia.
The way TMS works is that the part of the cortex I intend to target is mapped to a location on the outside of your head. Then I hold a paddle containing electromagnetic coils against the skull, and when I’m in the right location, I press a button to trigger electromagnetic pulses that stimulate action potentials in underlying cortex.
How long does the effect last?
TMS has to be repeated to have any long-lasting or therapeutic effect. Typically patients return five days weekly for several weeks. If they find that it’s effective in decreasing their pain, some come back periodically. It changes the electrical connectivity in your brain, so, just like other forms of stimulation, it needs to be redone periodically.
Setting this up has been a huge task, and it took a year to get approval from our Institutional Review Board. But the data are encouraging for neuropathic pain, and the need is so great, that I have pushed forward.
What else is going on in your research?
We also work on animal models, but we tend to do the animal models after we have characterized the human condition. The greatest need is not for more animal models; it’s for better characterization of the actual human diseases. One of the problems of animal models is that, because the phenotyping is so rudimentary for many neuropathic pain conditions, it’s not clear how relevant the animal models are.
For instance, the chronic constriction injury (CCI) model, the most widely used neuropathic pain model, is a wonderful model, but it doesn’t correspond to any particular human disease. It has both nerve injury and inflammatory components, and it changes over time as the sutures that are used are absorbed. So all of the wonderful studies that are done in CCI… Not to dismiss them, because we have learned a tremendous amount about how the nervous system works from them, but the model corresponds loosely at best to an actual human disease or illness.
My lab developed a model of complex regional pain syndrome type I (CRPS-I), formerly known as reflex sympathetic dystrophy, after discovering that patients with CRPS-I have evidence of peripheral nerve injury. It’s a myth that patients with CRPS-I don’t have nerve injury—most of them have undiagnosed nerve injury. Hardly any among the clinicians who treat CRPS-I patients are nerve experts. Working in a neurosurgery department was the key for me. Neurosurgeons are good at localizing nerve injuries because before cutting into a patient, they need to be sure where the lesion is. You can’t just open a patient up and say, woops, I was wrong; let’s try another place.
Another reason why the nerve injuries in CRPS-I remain unrecognized is that patients often have injuries to small sensory nerve branches that are not routinely tested by nerve conduction study, or partial injuries that involve only a few of the axons. So the injured limb may still work reasonably well, and there will still be sensation in the affected area, giving a misleading impression of no nerve injury. Furthermore, EMG [electromyography] and nerve conduction study, the usual tests for diagnosing nerve injury, are completely insensitive to small-fiber function. Many CRPS-I patients have normal EMG and nerve conduction studies, even in the face of injury and dysfunction of their nociceptive fibers.
Once we published in 2006 in Pain that skin biopsies detected small-fiber injuries in CRPS-I patients [
Oaklander et al., 2006], we then created a rodent model using a clinically relevant cause of human CRPS-I, needlestick nerve injury (NNI) [
Siegel et al., 2007]. Like CCI, it’s very easy to do: It involves sticking a regular needle, of the kind that sometimes causes CRPS-I in patients, through the tibial nerve of a rat. We based this on the spared nerve injury (SNI) model developed by Isabelle Decosterd and Clifford Woolf here at Mass General, which has the advantage over CCI of being more distal, so that individual nerves are lesioned and the effect on nearby “spared” zones can be documented. We use that feature, but instead of cutting and ligating the entire tibial nerve, which doesn’t often happen clinically, we pass a needle through it once.
The other unique feature of the NNI model is that, unlike most neuropathic pain models, only a subset of lesioned rats develop the pain phenotype. About 50 percent of NNI rats are left without mechanical allodynia, swelling, or other signs of CRPS that are seen in the other half. This also mimics the human condition, as only a minority of patients with a particular injury develop CRPS. This model permits us to look for differences in the rats that have the pain phenotype and the rats that don’t have the pain phenotype [see
Klein et al., 2011].
This has been a big problem in basic science pain research—most models cause hyperalgesia in virtually all lesioned rodents. The few that don’t develop hyperalgesia are typically discarded from analysis, so it becomes very difficult to know how many of the biological changes you detect are actually related to pain, or whether they might be nonspecific signs of nerve injury.
You are in a unique position because you are active in both the clinic and the laboratory. Is there anything that you wish other pain researchers knew about the clinical situation—or, conversely, that you wish your clinical colleagues understood about research?
There are many people who do both research and clinical care—institutions such as Harvard are full of them. But most are not in the pain field. So the problem is how to get more overlap between pain research and pain clinicians. There is a robust basic science research community in pain, but not enough people doing high-quality clinically relevant pain research. I hope that I am among them, and I try to influence basic science pain researchers to consider more clinically relevant questions, and to interest clinicians in the research opportunities in chronic pain. I mentor residents here, and I’m a core preceptor in the neurology residents’ clinic—and I try to encourage some of these young neurologists to consider a career focusing on pain and itch.
Over your own career, how has your thinking about pain evolved?
When I was a neurology resident, pain was not something that was discussed or taught. I shared the common suspicion of patients who complained of chronic pain, perceiving them as a clinical nuisance at best, drug addicts at worst—and as people to be moved out of my office as quickly as possible.
Even as a peripheral nerve scientist, I had only vague awareness of C fibers and Aδ fibers. My research as a graduate student focused on myelinated axons, and I wasn’t alone in that. Most neuromuscular neurologists still focus almost exclusively on myelinated axons. There are very few academic neurologists interested in unmyelinated fibers.
What got you interested in pain?
It was squarely the patients. I owe my whole career to my patients.
What changed your mind about pain patients?
I met these people! I got to know them, and I saw that, in fact, most of them were not nuts, or at least no more nuts than I was. And in fact, most were not drug addicts. Sure, you run into the occasional drug addict, but you do in every area of medicine.
I saw that most were hard-working, honest people just like anyone else striving to live their lives in the face of severe disability from chronic pain. We in medicine knew so little about this that it was embarrassing. We didn’t know the cause of their symptoms, we didn’t have effective treatments, and many patients were marginalized by the medical environment and had to struggle to get help. The more I learned, the more I saw that the status quo was wrong, and that these patients deserved to have smart researchers and smart clinicians provide them with the same expertise that we expect in other areas of medicine.
You speak in the past tense—that clinicians didn’t know the causes of symptoms, and that they had nothing to offer patients. Do you still find that you don’t have good treatments to help your patients?
In most cases we can help them. The issues are not so much lack of medications, although we do need more and better medications. The problems are at the delivery level: Most physicians don’t know how and when to use the tools we already have. That has pulled me out of the lab, and I now am willing to talk to journalists to help educate the public and medical colleagues. I’ve realized that it’s not enough that Iknow how to treat these patients, but I need to help spread the word so that more of my medical colleagues know as well.
If you could make a to-do list for basic researchers studying pain, what would be at the top?
I would tell them to hook up with their clinical colleagues. Go spend time in the clinic, and meet some patients, meet the physicians. Find out what the issues are, and what the phenomena are. Because then they would see the disconnect.
If you’re going to devote your career to basic science pain research on animal models, you owe it to yourself, to your research sponsors, and to the patients to take some time to investigate what the clinical problems are. There are wonderful opportunities for basic scientists willing to look into diseases that are not as heavily studied. Instead of just using tried-and-true animal models, think about creating new models for which there is pressing clinical need.
Thank you so much for sharing your experiences with the PRF.
Thank you.
_____________________________________________________
PRF Related Content:
View Anne Louise Oaklander’s
profile on the Pain Research Forum (requires member log in)
Additional Reading:
