Translational Pain Assessment: Could Natural Animal Models Be the Missing Link? Mary P
Total Page：16
File Type：pdf, Size：1020Kb
Comprehensive Review Translational pain assessment: could natural animal models be the missing link? Mary P. Klincka, Jeffrey S. Mogilb, Maxim Moreaua,c, B. Duncan X. Lascellesd,e, Paul A. Flecknellf, Thierry Poitteg, Eric Troncya,c,* Abstract Failure of analgesic drugs in clinical development is common. Along with the current “reproducibility crisis” in pain research, this has led some to question the use of animal models. Experimental models tend to comprise genetically homogeneous groups of young, male rodents in restricted and unvarying environments, and pain-producing assays that may not closely mimic the natural condition of interest. In addition, typical experimental outcome measures using thresholds or latencies for withdrawal may not adequately reflect clinical pain phenomena pertinent to human patients. It has been suggested that naturally occurring disease in veterinary patients may provide more valid models for the study of painful disease. Many painful conditions in animals resemble those in people. Like humans, veterinary patients are genetically diverse, often live to old age, and enjoy a complex environment, often the same as their owners. There is increasing interest in the development and validation of outcome measures for detecting pain in veterinary patients; these include objective (eg, locomotor activity monitoring, kinetic evaluation, quantitative sensory testing, and bioimaging) and subjective (eg, pain scales and quality of life scales) measures. Veterinary subject diversity, pathophysiological similarities to humans, and diverse outcome measures could yield better generalizability of findings and improved translation potential, potentially benefiting both humans and animals. The Comparative Oncology Trial Consortium in dogs has pawed the way for translational research, surmounting the challenges inherent in veterinary clinical trials. This review describes numerous conditions similarly applicable to pain research, with potential mutual benefits for human and veterinary clinicians, and their respective patients. Keywords: Preclinical animal models, Natural, Experimental, Validity, Applicability, Translation 1. Introduction clinical development today than they were in the 1970s.15,134 The Major goals of pain research are improving our understanding of credibility of efficacy data obtained from animal disease models has lately been called into question, in biomedical research in pain pathology and identifying novel molecular therapeutic 88,134,148 14,93–95,121,142 targets for better clinical pain management. Despite huge general and in the pain field specifically. scientific and technological advances, and tremendously increased Scientific research also faces a “reproducibility crisis,” which may research and development costs, drugs are more likely to fail in be a contributing factor in the translational crisis. Failure rates in the clinical phase are around 90% to 95%,8,59,122 due in part to the challenges of interpreting animal model data. The predictability of Sponsorships or competing interests that may be relevant to content are disclosed basic research varies depending on the understanding and at the end of this article. complexity of disease biology, whereas for therapeutics targeting a Animal Pharmacology Research Group of Quebec (GREPAQ), Department of infectious diseases, success rates are high, for diseases involving Veterinary Biomedical Sciences, Facult ´ede m ´edecinev ´et´erinaire,Universit ´ede complex mechanisms, such as neurological diseases and cancer, b Montr ´eal,Saint-Hyacinthe, QC, Canada, Department of Psychology, Alan they can be as low as 2.3%.59 For pain studies, the likelihood of Edwards Centre for Research on Pain, McGill University, Montreal, QC, Canada, c Osteoarthritis Research Unit, Universit ´ede Montr ´ealHospital Research Center eventual FDA approval of a drug entering phase I studies has been 59 (CRCHUM), Montreal, QC, Canada, d Comparative Pain Research and Education reported at 10.7%. Some have blamed animal models for these Centre and Comparative Medicine Institute, College of Veterinary Medicine, North translational difficulties. Indeed, a lack of tangible benefit over a long e Carolina State University, Raleigh, NC, USA, Center for Translational Pain enough period of time could lead one to question both the Research, Department of Anesthesiology, Duke University, Durham, NC, USA, f Comparative Biology Center, Medical School, Newcastle University, Newcastle commitment of substantial funding to, and the ethics of, animal 121,148 Upon Tyne, United Kingdom, g CAPdouleur Change Animal Pain, Sainte Marie de use for this research. However, the lack of success is almost R´e, France certainly also associated with other factors, and both the initial *Corresponding author. Address: Animal Pharmacology Research Group of Quebec compound development and in vitro screening programs should be (GREPAQ), Department of Veterinary Biomedical Sciences, Facultedem´ ´ edecine examined, as well as the animal models used. v´et ´erinaire,Universit ´ede Montr ´eal,1500 des v ´et´erinairesSt, P.O. Box 5000, Saint- Hyacinthe, QC, Canada J2S 7C6. Tel.: (450) 773-8521; ext.: 8399; fax: (450) 778- The main challenges of animal models are not only recreating 8103. E-mail address: [email protected] (E. Troncy). disease conditions but also defining measurable and clinically Supplemental digital content is available for this article. Direct URL citations appear translatable efficacy parameters. For a pain model to be valid, it in the printed text and are provided in the HTML and PDF versions of this article on should encompass key elements of the human pain experience the journal’s Web site (www.painjournalonline.com). and measure the pertinent aspects of that experience. Given that PAIN 158 (2017) 1633–1646 an animal model consists of a subject, a method of pain induction © 2017 International Association for the Study of Pain (ie, an assay), and an outcome measure(s),93 each of these http://dx.doi.org/10.1097/j.pain.0000000000000978 components should reflect, as closely as possible, the clinical September 2017· Volume 158· Number 9 www.painjournalonline.com 1633 1634 M.P. Klinck et al.·158 (2017) 1633–1646 PAIN® condition to which the results are to be applied. Some proposed (increased complexity in the design and interpretation of studies strategies for improving model face and predictive validity include using these subjects). Effectiveness of new analgesic treatments the use of so-called “natural” animal models of painful disease (ie, in canine or feline painful conditions would be anticipated to veterinary patients),21,93,121,128 and the use of nonevoked predict similar benefits in human patients. Using these sponta- outcome measures representative of important clinical pain neous models as part of the drug development process (eg, as phenomena.94,155 The purpose of this review is to examine how has recently been done in canine osteoarthritis and cancer143) the modeling of human pain using animals might be improved, by could yield higher “translatability” (see examples in section 3.2 for considering how aspects of the model may influence represen- osteoarthritis; also recent FDA granting of a Fast Track tation of the pain experience and its interpretation. To what extent Designation for recombinant HER2-expressing Listeria mono- are experimental models representative of, and natural models cytogenes immunotherapy for pediatric osteosarcoma, based on consistent with, the pain condition of interest? How do methods encouraging results, including prolonged survival, in dogs with of pain evaluation compare between these types of animal the disease [https://globenewswire.com/news-release/2016/04/ models? Finally, we will discuss limitations and practical aspects 28/834081/0/en/FDA-Grants-Advaxis-Fast-Track-Designation- of studying natural animal models. for-ADXS-HER2-for-Patients-with-Newly-Diagnosed-Non-Met- astatic-Surgically-Resectable-Osteosarcoma.html]). However, it also raises major practical issues relating to the recruitment of 2. Modeling the human pain experience sufficient subjects and this is discussed in more detail below. 2.1. Experimental animal models Papers discussing specific natural animal models exist for feline interstitial cystitis as a model of visceral pain,23 for feline Similarities in the neuroanatomy and physiology of pain across diabetes mellitus as a model of neuropathic pain,92 for mammalian species, as well as evolutionary evidence, argue for osteoarthritis in many species,87,99 and for various tumors in 139 parallel pain experiences in humans and animals. However, dogs20,81,115,149 and cats.116,149 However, this barely scratches many authors have noted limitations of experimental mod- the surface of the huge overlap of veterinary and human painful 14,80,93,95,104 els that may inhibit extrapolation of findings to humans. diseases. See Figure 1 for some examples of painful veterinary For instance, patients with clinical pain are often middle aged or conditions. older, and most are women; the persistent use of young, male, The most accessible veterinary patients for study are the rodent subjects, although less expensive and more convenient for commonly kept species, ie, domesticated animals (pet dogs and the experimenter, neglects the demonstrated modulatory effects of cats, and livestock including horse, swine, sheep, and cattle). Sex 95 organismic factors on pain and analgesic responses. It is usually and age distributions either of particular patient