The following is Part 2 of a three-part series of selected presentations from the 35th Annual Scientific Meeting of the American Pain Society (APS) held May 11-14, 2016, in Austin, Texas, US. Also see Part 1 and Part 3.
Despite tremendous advances in researchers’ understanding of the mechanisms underlying acute and chronic pain, it has been a struggle to translate that knowledge into new pain treatments for people. Many blame the difficulty on the gap between the animal models and behavioral tests used to study and assess pain, and the human experience of chronic pain.
Members of the APS Basic Science Shared Interest Group (SIG) gathered on Friday, May 13, in Austin for a dinner of pulled-pork sliders with coleslaw and a town hall-style discussion of the issue. The event was hosted by SIG co-chairs Michael Jankowski of Cincinnati Children’s Hospital and Steve Davidson of the University of Cincinnati, both in Ohio, US. A panel comprising Anne Murphy, Georgia State University, Atlanta, US; Todd Vanderah, University of Arizona, Tucson, US; and Michael Iadarola, National Institute of Dental and Craniofacial Research (NIDCR), Bethesda, US, made presentations while taking questions and comments from the audience of nearly 150 researchers.
Behavioral tests in pain research
First, Murphy and Vanderah co-presented an overview of behavioral tests currently used in animals in pain research. The presenters grouped the tests into three main categories: reflexive tests; measures of spontaneous pain; and behaviors that indicate an animal is “pain suppressed.” The vast majority of research uses reflexive tests such as withdrawal from a poke by a von Frey hair. More than 5,000 papers in the pain literature have reported using these tests, and another 3,000 reported using tests that could be considered both reflexive and spontaneous, such as withdrawal from a hot plate. Over 2,000 papers reported using purely spontaneous measures of pain, such as vocalization, while only 1,300 papers reported using pain-suppression tests such as wheel running or exploring.
Reflexive tests have their advantages: They can be used to test acute and inflammatory pain conditions; they are quick and inexpensive; and they are widely used. But they also have their disadvantages: The animals’ responses only reflect spinal cord and brainstem contributions to pain, so they “short-circuit” the contribution of cortical activity; they rely heavily on an observer, which could influence the results; and factors such as motor impairment or body temperature can produce false positives.
Audience members agreed: The overriding disadvantage to all behavioral tests in animals—and reflexive tests in particular—is that they don’t necessarily measure pain behaviors, such as spontaneous pain and widespread pain, that predominantly affect chronic pain patients. In using animal models, researchers also face the considerable challenge of observing behaviors natural to the species that somehow indicate whether the animal is affected by pain. Robert Coghill, Cincinnati Children’s Hospital, US, suggested that researchers really need an animal model of disability rather than readouts of pain behaviors.
There was also a broad consensus that measures of acute pain are relatively straightforward, but that the big gap between animal tests and humans exists in measurements of chronic pain in animals. In addition, Kathleen Sluka, University of Iowa, Iowa City, US, pointed out that measurement of pain-related behaviors is distinct from animal models of specific pain conditions, and each should be considered individually. Tests to measure pain behaviors should be appropriate to the model; for example, studies using a model of osteoarthritis should measure animals’ movements rather than withdrawal from a poke to the paw.
Vanderah then turned to the topic of standardization in behavioral animal testing. Mechanical probing with von Frey hairs is a widely used test of mechanical allodynia in pain research, yet how reproducible are the conditions from one lab to another? D.P. Mohapatra, Washington University, Saint Louis, US, noted that there are significant discrepancies among the four to five thousand papers using von Frey hairs. The force used, in particular, varies wildly from 0.1 grams of pressure to many times the weight of a 20-gram mouse, he said. And while four companies supply the von Frey hairs, the filaments all originate from the same manufacturer, with no information provided as to whether the hairs have been modified by the supplier.
Discussion continued over the all-or-none nature of a von Frey withdrawal response and whether measurement might ever be automated. Jessica Ross, a student in the Jankowski lab, pointed out that the test provides discreet, not continuous, data points, and so should be appropriately analyzed with nonparametric statistics, which few labs do. Michael Gold, University of Pittsburgh, US, said the key problem with von Frey hair testing is to find a standardized transducer of force that will produce the all-or-none withdrawal response. He mentioned that Geoffrey Bove, University of New England, Biddeford, US, had published a paper calling the monofilaments the field’s “tin standard” (Bove, 2006).
These inconsistencies led Vanderah to pose a question: Should funding agencies or universities require measures that would lead to more standardized testing? Laura Stone, McGill University, Montreal, Canada, objected to the idea of any “requirements” in research protocols, but recognized the value in guidelines. Ted Price, University of Texas at Dallas, US, noted that he was among a group of authors of such guidelines published in Pain in April (Andrews et al., 2016). Murphy went on to ask the group, Should studies routinely include at least two different behavioral tests? And should those include a spinal and a supraspinal measure? Attendees agreed that the testing depends on the questions being asked in the project and the specific model of disease being used; behavioral measures should be appropriate to the model.
Murphy also suggested that further testing be included in studies to rule out false positives, such as assessment of motor and autonomic function. Her next question was a familiar one that sparked lively debate and discussion: Should both male and female animals be used in pain studies? That question was posed at the 2015 APS Basic Science Dinner in the wake of new NIH requirements for male and female animals in preclinical research (see PRF related news and coverage of the discussion). This audience agreed that animals of both sexes should be routinely included in pain studies, but some voiced concerns about the costs and limitations involved.
Next, the group addressed the question of whether animal models of pain conditions can be predictive of how humans will be affected by a pain condition or by a therapeutic intervention. Becky Klein, Merck, Philadelphia, US, said that from her perspective in the pharmaceutical industry, researchers there do not expect preclinical results to necessarily predict outcomes in humans; rather, the idea is to understand the molecular mechanisms at play and identify potential drug targets from those insights.
Studying rodents is not the be-all and end-all
Iadarola then outlined the current procedure in development of new pain drugs: Identify a molecular target; develop the therapy in vitro and in vivo in animals; and conduct Phase 1 clinical trials in humans. He suggested that two new steps be introduced. The first, to be completed in conjunction with target identification, is to make sure that the molecule in question is expressed in human tissues at consequential levels, which could be achieved with in silico searches of existing gene databases.
Second, Iadarola suggested introducing large-animal testing before taking a drug to human clinical trials. Testing compounds on companion animals such as dogs has a number of advantages, Iadarola said. Notably, arthritis is a naturally occurring disease in dogs—not just a model. The complex behaviors and tissues of dogs are more similar to those of humans than those of rodents are, and dogs naturally interact with human observers. Success in a large transitional animal model, he added, “allows us to move forward with trials with a higher level of confidence.” The disadvantages: The dogs are a clinical population that naturally varies in disease presentation, and recruiting presents some hurdles.
Iadarola then presented work he has done in dogs with osteoarthritis, in which the animals were injected in the joint with resiniferatoxin, an agonist at transient receptor potential vanilloid type 1 (TRPV1), the capsaicin receptor. The toxin “sticks to the receptor and keeps it open, resulting in a toxic dose of calcium,” which effectively silences nociceptive afferent neurons for weeks or months, Iadarola said. The results were clear: According to owners’ assessments, dogs that received the treatment recovered mobility and displayed less pain than before treatment. Surprisingly, in some dogs the treatment was still effective 12 months after injection.
How can researchers strike up a collaboration that allows them to study large animals such as dogs? Iadarola suggested identifying a contact in a veterinary school and writing up an animal clinical protocol. The estimated cost of such testing is about $700 per animal, he said. George Wilcox, University of Minnesota, US, noted that, anecdotally, domesticated animals are considered by some better than veterinary pets for pain studies; Iadarola said he has successfully used goats in studies of osteoarthritis. Frank Rice, Integrated Tissue Dynamics, Rensselaer, New York, US, suggested pigs as an excellent model animal because, in addition to the advantages of other large animals, the skin structure of pigs is much more similar to human skin structure than to the furry skin of most large animals, and using pigs also allows for multiple skin-punch biopsies over time. Rice also suggested that in any large-animal study, researchers preserve some tissue—including skin, nerve, and dorsal root ganglia—for later histological experiments.
The roomful of researchers concurred: Better translational success will arise from preclinical research if scientists can address some of the challenges presented in the town hall, which will only come from such open and innovative discussions such as the one in Austin.
Stephani Sutherland, PhD, is a neuroscientist, yogi, and freelance writer in Southern California.