Dear Arvind,

The following publication covers the answer to your question.

http://www.ncbi.nlm.nih.gov/books/NBK57273/

Translational Pain Research: From Mouse to Man.

Chapter 17Large Animal Models for Pain Therapeutic Development

Darrell A. Henze and Mark O. Urban.

The humane use of preclinical animal models plays a critical role both in understanding the basic biology of pain as well as in the development of therapeutic treatments to alleviate pain. Clinically relevant pain is the result of complex processes involving peripheral transduction and transmission as well as central modulation and processing that leads to the final conscious sensation of pain. Much has been learned about the mechanisms underlying the transduction and transmission of the pain signal within the nervous system through the use of cellular, biochemical, and molecular techniques (Millan 1999; Scholz and Woolf 2007; Zeilhofer 2005). However, understanding the actual experience of pain will always require an intact organism that can integrate the full range of external stimuli and internal cognitive and emotional states that drive and modulate pain.

Rodent models of pain have historically played a dominant role in the study of pain mechanisms (Negus et al. 2006;Walker et al. 1999). There are many good reasons for this, including the practicalities of low cost, simplified ethical concerns, and the scientific value of having a large database of prior research to provide context for new findings. A large historical database is particularly important in the field of drug discovery and development since the sensitivity and predictive validity of animal models can only be established through extensive testing in many contexts. For these reasons, the rat and mouse models will continue to be the workhorses driving basic research as well as drug discovery. Unfortunately, there are many ways that the biology of rodents may fail to accurately predict the biology and pharmacology of clinical pain conditions in humans (Blackburn -Munro 2004; Le Bars et al. 2001). Given the high cost of developing new therapeutics (Adams and Brantner 2006), there is a growing need to validate biological and pharmacological findings in non-rodent species that, while perhaps less tractable than rodents, address known or ill-defined differences between mice and men. It is hoped that through the humane study and evaluation of pain states in higher order preclinical species, we can better predict whether biological mechanisms and specific compounds have relevance for clinical pain. Ultimately, well-validated pain models in non-rodent species could enhance the speed, reduce the costs, and increase the probability of the successful development of new analgesic therapeutics offering enhanced efficacy and reduced adverse effects.

This chapter will review the gaps in current pain research using rodent models that may potentially be addressed using preclinical animal models of pain in higher species. Additionally, we will review the pain models and assessments that have been developed to date in “higher” and larger species and highlight areas where there is a need for the development of new models and methods of pain assessment.

Go to:

17.1. THE IMPORTANCE OF PREDICTIVE ANIMAL MODELS FOR DRUG DISCOVERY AND DEVELOPMENT

While it is always exciting to see a compound with a novel mechanism of action show great efficacy in a rodent model of inflammatory or neuropathic pain, the obvious ultimate goal is to identify novel compounds that are found to be safe and effective in humans. A major challenge faced in the development of new analgesics is establishing confidence in the predictive validity of preclinical models of pain. There are many examples where compounds that have demonstrated efficacy and tolerability in rodent models have failed to show sufficient efficacy or safety in the clinic (e.g., neurokinin-1 [NK-1] receptor antagonists; Hill 2000). Given the uncertainty about the predictive validity of the rodent models to determine efficacy and safety of compounds associated with novel mechanisms, it is logical to suggest that novel potential analgesic compounds should be tested in several clinical pain populations to evaluate safety and efficacy. Counterbalancing this logic, there is an ethical and fiscal responsibility to test new compounds in the clinic only under circumstances where there are preclinical data supporting safety and efficacy. Given that the success rate of bringing a new drug to market is less than 10% (Kola and Landis 2004), there clearly is a need to identify and validate preclinical model systems that will increase the probability of success. There has been a recent increased focus on assessing the predictive validity of preclinical model, and what must be done to understand and improve upon the current state (Negus et al. 2006; Whiteside et al. 2008). It is important to note that although existing animal models appear to do very well at predicting true positives, there is less confidence about their ability to predict falsepositive and false-negative outcomes (Rice et al. 2008).

Go to:

17.2. THE POTENTIAL VALUE OF LARGE ANIMAL MODELS

The preclinical animal models that are used to test novel compounds and mechanisms need to be able to predict accurately both the clinical efficacy and adverse effects of any new therapeutic. In most cases, safety studies are completed in both rodents as well as in some other larger species such as the dog or non-human primate. Since there are many rodent models of pain, it is relatively easy to calculate a therapeutic index in the rat or mouse. Unfortunately, there are very few validated large animal pain models in dogs or primates that would allow a similar therapeutic index calculation to be made in these species that may be more relevant to the human case.

One obvious reason why large animal models have a real potential to add value to the drug development process is that they are phylogenetically closer to humans than rodents. This is important for several reasons. The first is that at the molecular level many pharmaceutical targets have species specific variations in sequences or expression patterns that result in differences in affinity or potency for the target and functional importance of the target. In certain circumstances, it may not be possible to identify compounds that have identical or even similar affinity for the human versus rodent variants (e.g., calcitonin gene-related peptide (CGRP) receptor antagonists; Salvatore et al. 2008), in which case, higher order species such as the non-human primate will usually share a greater sequence homology with the human form. The second important reason is that at a systems level, the pharmacological response and network connectivity of the central nervous system varies. For example, pregabalin demonstrates robust efficacy in rodent models of neuropathic pain within 1 to 2 hours but it often takes days to weeks for robust efficacy in humans (Arezzo et al. 2008; Field et al. 1999). A third reason phylogenetic differences may matter is that organisms likely have experienced very different evolutionary pressures as they developed behavioral expressions of pain. These differences may confound reasonable comparisons of the amount of pain any given organism is experiencing. Rice et al. have pointed out that animals such as rats that often live in a hostile environment are usually viewed as potential prey and therefore may process and express pain in a very different way than predators, such as canines or primates, who live typically in a less hostile environment (Rice et al. 2008). Finally, it is well known that different species can metabolize compounds in different ways and different rates, with dogs and monkeys often being better of predictors of human metabolism (Lin 1999; Ward et al. 2005). Therefore, better estimates of efficacy and safety often can be obtained in large animals since sustained compound exposure can be an important contributor to the efficacy profile for a chronic pain drug.

A second general reason that large animal models benefit drug development is that they are, in fact, large and thus more suitable for certain types of pain measurements. For example, the increasing use of functional imaging technologies to reveal patterns of brain activation in pain states is much easier to accomplish in a large species such as a dog or primate. In addition, certain forms of chronic pain, such as osteoarthritis, appear to occur more naturally in large animals compared to lower species. This is particularly useful both for the ability to test potential therapeutics on naturally occurring diseases, as well as for the ability to test alternate delivery methods such as topical delivery on a joint of similar size and geometry to the human joint. However, note that the relatively large size of dogs and primates has a number of disadvantages, mostly due to the need for increased resources in housing and maintenance, as well as the increased amounts of test compound required to achieve desired exposures.

Go to:

17.3. MODELS OF ACUTE NOCICEPTION

While pathological pain, particularly chronic pain, is detrimental for an organism, acute pain is an important factor in protecting organisms from harm. Therefore it is important that there are animal models that can provide insight into the impact of analgesic compounds on acute nociception. In rodents, acute nociception is usually assessed using either a noxious heat stimulus (e.g., hot plate, tail withdrawal, Hargreaves test) or mechanical stimulus (Randall-Selitto test) applied to either a hindpaw or tail. While there are no reported quantitative assays to assess acute mechanical nociception in large animals, there are two assays that have been reported to assess acute thermal nociception: a canine thermal escape assay and a primate tail withdrawal assay. These assays may be used to evaluate whether novel compounds will result in blunting of the protective responses to noxious heat, with the primate model being of particular value due to the similarity of physiology compared to humans.

The canine thermal escape assay was recently described by Yaksh and colleagues (Wegner et al. 2008; Table 17.1). In this assay, a dog is trained to rest in a sling with its hindpaws resting on a clear plate. A high-powered lamp is then turned on to apply a heat stimulus to the anterior third of the metatarsal paw pad, and the latency from when the light is turned on is measured. Normal dogs withdraw their paws at about 9.3 seconds, which corresponds to a plate temperature of ~50°C. A variety of opioids were tested in this assay, and each showed a significant effect on the escape latency. A dose of 1 mg/kg (IV) morphine produced a latency very close to the cutoff-latency of 20 seconds. The sedating alpha2 adrenergic agonist dexmedetomidine (0.1 mg/kg) also produced a maximal possible effect. However, the phenothiazine sedative acepromazine (0.1 mg/kg) did not produce any increase in withdrawal latency, suggesting that sedative effects alone do not account for the increases in latency observed with opiate analgesics.

TABLE 17.1

Large Animal Models of Acute Nociception.

Similar in concept, the primate tail withdrawal assay has been used to assess sensitivity to acute application of noxious thermal stimuli (Dykstra and Woods 1986; Table 17.1). In this assay, a monkey is trained to sit in a primate chair. The distal tip of the tail is shaved and immersed in a water bath at 50°C or 55°C, and the time required for the subject to withdraw the tail is measured. A normal monkey will withdraw its tail from 55°C water in about 2–3 seconds, but a prior dose of 10 mg/kg morphine will increase the latency to longer than the cutoff time of 20 seconds. In contrast, monkeys that appeared sedated to the point of sleep with 30 mg/kg pentobarbital still withdrew their tails within 2–4 seconds, again suggesting that the increases in latency are not due to general sedative effects of the morphine (Dykstra and Woods 1986). One utility of this model for therapeutic development is to determine whether analgesic compounds of interest possess potentially undesirable effects on acute nociceptive thresholds and heat sensitivity. For example, the analgesic gabapentin has no effect on acute heat nociceptive thresholds in this assay (Figure 17.1A), although it is effective in inhibiting warm allodynia following sensitization of the tail skin with capsaicin (see below; Figure 17.1B). These findings are consistent with the known analgesic properties of gabapentin in terms of its selective effects on persistent pain associated with injury-induced sensitization (Urban et al. 2005).

FIGURE 17.1

(A) Effect of gabapentin (150 mg/kg, s.c.) in the rhesus monkey acute noxious heat tail withdrawal assay. Application of a noxious heat stimulus (52°C water) to the tail of rhesus monkeys results in a tail withdrawal response within 1–3(more...)

Go to:

17.4. MODELS OF ACUTE INFLAMMATORY PAIN

Clinically, acute inflammatory pain occurs in response to tissue injury due to surgery, burn , or the inflammatory response to an infection. Similar to acute nociception, acute inflammatory pain also provides a protective role, encouraging protective behaviors following a tissue injury. However, post-surgical pain and burn pain are two clinically important inflammatory pain states often targeted for therapeutic intervention. Some of the most common rodent models of acute inflammatory pain are the carrageenan model (Hargreaves et al. 1988), ultraviolet-B radiation burn (Bishop et al. 2007), formalin model (Dubuisson and Dennis 1977), and hindpaw incision model (Brennan et al. 1996). Both canine and primate models have been described that are at least in part analogous to these rodent models.

A commonly used canine model of inflammatory synovitis is achieved by intraarticular injections of uric acid crystals leading to inflammation and pain (Table 17.2). The unilateral injection of uric acid crystals into the hind limb stifle (knee) results in an acute synovitis and quantifiable lameness on the injected side (McCarty, Jr. et al. 1966). The lameness is typically quantified through a gait analysis using a force plate apparatus. The observed reduced weight loading of the affected limb is interpreted as reflecting a painful state of the joint since there is no structural change in the injected joint. The peak change in gait occurs approximately 4 hours post treatment and resolves within 24 hours (Rumph et al. 1993). A number of studies have demonstrated that non-steroidal anti-inflammatory drugs (NSAIDs; carprofen, etodolac, ketoprofen, and meloxicam), cyclo-oxygenase 2 (Cox2) inhibitors (deracoxib, firocoxib, ML-1,785,713), opioids (Butorphanol), and ketamine all show efficacy in this model (Borer et al. 2003; Cross et al. 1997;Drag et al. 2007; Hamilton et al. 2005; Hazewinkel et al. 2003; McCann et al. 2004; Millis et al. 2002). Additionally, in agreement with clinical studies demonstrating lack of efficacy of NK-1 antagonists (Boyce and Hill 2000), there is a report that an NK-1 antagonist failed to show efficacy in this model (Punke et al. 2007). Taken together, these data may support use of this canine model as predictive of the human clinical response for therapeutics targeting inflammatory pain.

TABLE 17.2

Large Animal Acute Inflammatory Pain Models.

In primates, a variation of the tail immersion assay described above has been used to evaluate inflammation-induced thermal allodynia (Table 17.2). The standard tail immersion procedure described above is modified by injecting the tail of the monkey with carrageenan, and the water temperature is reduced to a normally non-noxious temperature of 42°C. Under these conditions, a normal monkey will not withdraw its tail, but a monkey receiving a carrageenan injection will rapidly withdraw its tail due to the increased thermal sensitivity associated with the inflammatory response (Ko and Lee 2002). This model is sensitive to NSAIDs, opioids, and the Bradykinin receptor 1 (BK1) antagonist ELN441958 (Hawkinson et al. 2007; Ko and Lee 2002); and although the BK1 antagonist has not been clinically tested, the NSAID and opioid efficacy observed in this model suggests that it does have a positive predictive power for clinical inflammatory pain.

A recent report has begun to characterize a novel porcine ultraviolet-B (UV-B) irradiation model (Rukwied et al. 2008;Table 17.2). The UV-B irradiation model has already been developed in both rat (Bishop et al. 2007; Davies et al. 2005) and human (Harrison et al. 2004; Sycha et al. 2003). Because the pig has skin that is physiologically very similar to human skin, it may provide an excellent translational model system to study inflammatory pain associated with the skin. The initial description of the model examined physiological changes in erythema (flare or dermovasodilation) that are driven by activation of nociceptive C-fibers (Rukwied et al. 2008). This readout, while not evaluating a pain response per se, does allow for the pharmacodynamic assessment of compounds that are intended to modulate C-fiber nociceptor activation. Pharmacological validation of this idea was supported by showing that 1% lidocaine injected intracutaneously blocked the heat-induced flare response in UV-B irradiated skin (Rukwied et al. 2008).

Go to:

17.5. MODELS OF CHRONIC INFLAMMATORY PAIN

Clinically, chronic inflammatory pain, particularly pain due to osteoarthritis (OA), accounts for the largest single population of patients seeking analgesic therapies. Although NSAIDS and Cox2 inhibitors are included in the current standard of care, patients continue to seek more effective and safer treatments. There are a number of rodent models of chronic inflammatory pain that are commonly used to support the development of new therapeutics. Common rodent experimentally induced arthritis models include Complete Freund’s Adjuvant (CFA)-induced arthritis of the paw (Stein et al. 1988) and monoiodoacetate-induced arthritis of the knee (Fernihough et al. 2004). In these models, the pain responses are typically measured by assessing mechanical and/or thermal hypersensitivity of a hindpaw. While there are a number of large animal models of OA and rheumatoid arthritis, there is a surprising lack of data describing and validating the measurement of pain and efficacy of analgesics in these models.

The most commonly described large animal model of experimental induced osteoarthritis is the canine anterior (or cranial) cruciate ligament (ACL) transection of the hind limb stifle (knee) joint (Table 17.3). There are a number of different surgical methods used to rupture the ligament (Lopez et al. 2003; Marshall and Chan 1996; McDevitt et al. 1977; Pond and Nuki 1973) with various levels of secondary surgical trauma; but in all cases, once the ligament is transected, the joint immediately becomes unstable. Over weeks to months (Budsberg 2001; Dedrick et al. 1993;O’Connor et al. 1993), the joint shows signs of progressive OA including osteophyte growth, cartilage and meniscal damage, and ultimately changes in subchondral bone, thus sharing most features with human OA. One of the typical behavioral features of this model is that the dog shows a variable lameness of the affected limb. The degree of lameness is typically assessed using a graded or continuous rating scale used by an expert observer, or a more quantitative gait analysis is performed using a force-plate apparatus. It is often the case that lameness is interpreted and referred to as a behavioral correlate of pain in the joint, and the quantitative approach of the gait analysis would appear to make this a very good model to study OA pain. Despite these attractive features, there is very little activity in the literature that attempts to validate pain assessment methods to study analgesics in this model.

TABLE 17.3

Large Animal Chronic Inflammatory Pain Models.

A very recent paper using ACL transection reported testing the selective Cox2 inhibitor firocoxib in this model out to 18 weeks post lesion. Surprisingly, the study found no effects of treatment to enhance use and recovery of the affected limb assessed using a force-plate analysis. It was concluded that the deficits observed in the force-plate analysis were primarily due to the reduced structural stability of the lesioned joint but were not due to pain per se. Interestingly, animals in the study that were treated with firocoxib required statistically less rescue medication (butorphanol) in the days immediately after the ligament rupture. The need for rescue medication was determined by treatment group-blinded veterinary staff making subjective assessments of lameness on a five-point scale (Steiner et al. 2009). Taken together, these data suggest that the use of the force-plate analysis to read out pain is most likely confounded by structural effects in this model at time points out to 18 weeks post lesion. It is possible that force-plate analysis may be a more accurate measure of pain at longer time points post lesion (e.g., 1–2 years), since the relative contributions of structural effects versus direct pain effects may be different as the joint damage progresses over time (Budsberg 2001; Dedrick et al. 1993; O’Connor et al. 1993). The rescue medication data of Steiner et al. also suggest that there is a pain phenotype that can be observed and that the development and validation of specific rating scales may provide a way to assess pain in this model. Note that the lack of validity of the force-plate analysis in the early period following the ACL transection model contrasts with its apparent validity when used in the canine urate crystal induced synovitis model described above. This further underscores the hypothesis that models that include structural changes in the joint may confound pain assessment via force-plate analysis.

In addition to experimentally induced arthritis, both dogs and primates naturally develop osteoarthritis of the knee and hip joints as they age. Large-breed companion dogs are particularly susceptible to hip dysplasia and subsequent arthritic sequelae (Smith et al. 2001). The main challenge in attempting to take advantage of these populations of animals to study pain and the effects of specific analgesic agents is to have objective measures of the ongoing pain experienced by these animals.

Pain due to osteoarthritis in large-breed companion dogs is a common ailment seen in veterinary clinics. As a result, numerous randomized, placebo-controlled studies have been carried out to determine the potential benefit of various NSAIDs and Cox2 inhibitors to treat these animals (e.g., Budsberg et al. 1999; Peterson and Keefe 2004; Ryan et al. 2006; Vasseur et al. 1995). These veterinary studies use a variety of different subjective rating scales reported by the investigators as well as the dog owners. In some cases, a force-plate analysis is also done. In each case, the therapeutic was found to provide improvement over the placebo group using whichever measure was utilized, suggesting that this is a robust method to assess analgesic efficacy in natural model of OA in a large animal species. There are several challenges in leveraging these observations to enable studies in companion dogs for making critical go/no-go decisions about novel therapeutics ultimately intended for use in humans. The first challenge is that there does not appear to be any generally accepted and validated assessment instrument for the pain experienced in these animals. It is not clear how sensitive the various subjective scales and force-plate analyses are to therapeutic treatment, and obviously, the most sensitive assessment measures are desired. For example, natural osteoarthritis has an impact on joint structure. Since changes in force-plate analysis are probably a convolution of both structural deficits and painful sensation (see above), force-plate analysis may not be the most sensitive measure to assess pain alone. The second challenge is understanding whether this model and methods of pain assessment can generalize across different analgesic classes outside of the NSAID and Cox2 inhibitors. Ideally, work should be carried out to determine which assessment instruments are the most specific and sensitive to a variety of analgesic treatments. This type of information is critical to enable conclusive go/no-go decisions for a new therapeutic approach following testing in canines early in the development process.

Toward that end, at least two groups have begun developing and validating standardized pain assessment instruments for canines suffering from pain due to either osteoarthritis or bone cancer (Brown et al. 2007; Brown et al. 2008;Wiseman-Orr et al. 2004, 2006). These groups have followed sound and previously described methods for developing new tools for observers to assess the subjective states of others. Wiseman-Orr and colleagues have focused on the owners’ evaluation of the behavioral expression of affective states and developed an extensive questionnaire of 109 descriptors that are rated in a scale of 1 to 7 (the Glasgow University Veterinary School questionnaire, GUVQuest). In contrast, Brown and colleagues focused on transforming a known human clinical instrument, the brief pain inventory (BPI), into a form compatible for use in dogs, the canine brief pain inventory (CBPI). The final form is an 11-question instrument that is analogous to the BPI scale often used in human clinical trials. A strong advantage of both the GUVQuest and the CBPI is that they report the dog owner’s assessment of the dog’s pain over a period of time living with the animal. The simple fact that the pain assessment is not based on a single time period associated with a visit to the vet clinic increases the likelihood that this instrument is a more sensitive measure of functional outcome in real life. Currently, there are no published studies using the GUVQuest to assess its sensitivity to analgesic treatment. In contrast, the CBPI has been demonstrated to be able to detect the analgesic effect of carprofen after a 2-week course of treatment (Brown et al. 2008; Table 17.3). Hopefully further work will continue to describe the utility of these instruments, taking care to assess the sensitivity and predictive validity of the instruments to predict human clinical efficacy.

The existence of natural OA in primates has been documented in free-ranging rhesus monkeys after about 12 years of age (Chateauvert et al. 1989; Chateauvert et al. 1990; Kessler et al. 1986). Unfortunately, there have been no published results of any attempts to objectively quantify the pain that is expected to be associated with natural OA in primates. Therefore, although this animal model has very good face validity as a preclinical model of natural OA, any consideration of use for assessing the effectiveness of analgesics will require the development of validated pain assessment techniques. Perhaps some of the rating and measurement tools being used and developed in the clinic or in canines could be leveraged to jump start development of assessment tools for use in natural disease in primate species.

Go to:

17.6. MODELS OF CHRONIC NEUROPATHIC PAIN

Patients suffering from painful diabetic polyneuropathy (DPN) represent the largest clinical population of neuropathic pain patients in need of effective analgesics. The development of new therapeutics for DPN is primarily supported using rodent models of neuropathic pain including the spinal nerve ligation (SNL; Kim and Chung 1992) and chronic constriction injury models (CCI; Bennett and Xie 1988), although other rodent models of neuropathic pain are used as well (for review see Beggs and Salter 2006). The attractiveness of these models involving peripheral nerve ligation is that the surgical procedure is relatively straightforward and there is a good correlation between preclinical and clinical pharmacology for known analgesics. Although there is a rodent model of experimentally induced diabetes and associated diabetic neuropathy induced by streptozotocin treatment, there is a high degree of morbidity associated with this model, and there are questions regarding whether this is a true model of diabetic neuropathic pain (Bramwell et al. 2007; Fox et al. 1999). Additional rodent models of diabetic neuropathic pain include use of rats or mice that have a genetic predisposition for diabetes, including Zucker diabetic fatty rats and db/db leptin receptor-deficient mice (Obrosova et al. 2007; Sugimoto et al. 2008; Vareniuk et al. 2007). However, these models have not yet been fully validated and characterized in terms of their pharmacology and ability to predict analgesic effects.

There has been very limited exploration in developing neuropathic pain models in large animals. While there are no published reports describing canine experimental models of neuropathic pain, there have been a limited set of studies that have described a version of the rodent SNL model applied to non-human primates. Chung and colleagues ligated the L7 spinal nerve in rhesus monkeys, and ~14 days post lesion described the alteration in activity of spinothalamic tract (STT) neurons (Palecek et al. 1992) and behavior (Carlton et al. 1994). They found that following the L7 ligation, STT neurons showed increased activity in response to mechanical and thermal (both hot and cold) stimuli that were nominally non-noxious in the normal state (Palecek et al. 1992). In addition, the animals exhibited behavioral responses consistent with mechanical and thermal (both hot and cold) allodynia or hyperalgesia (Carlton et al. 1994). Interestingly, one additional monkey underwent a procedure analogous to the rodent CCI model and failed to develop symptoms of neuropathic pain. Upon post-mortem analysis, it was concluded that the larger size and thicker protective sheath of the primate spinal nerve prevented the nerve pathology typically observed in the rodent version of the CCI model (Palecek et al. 1992) Two additional studies using the L7 ligation model explored whether intraspinal administration of potentially analgesic compounds modulated the changes in STT neuron activity in the L7 ligated state. It was found that intraspinal administration of either the N-methyl-D-aspartate (NMDA)-antagonist dextrorphan (Carlton et al. 1997) or the kainate GluR5 receptor antagonist LY382884 (Palecek et al. 2004) reduced the activity of STT cells in both the normal and ligated state. Other investigators have used a primate L6 ligation model to induce a neuropathic state (Ali et al. 1999). In this study, a skin nerve preparation was prepared 14–21 days post ligation and used to study the excitability of the peripheral nociceptors. It was found that there was an increase in the spontaneous activity of C-fibers recorded from ligated animals. The C-fibers were also more sensitive to independent application of alpha-1 or alpha-2 adrenergic agonists. However, any behavioral effects associated with the neuropathic state were not reported in that study.

Unfortunately, the primate nerve ligation models described above, while interesting, are very labor intensive and raise concerns about the humane and ethical use of primate species in laboratory research. It is certain that it would require a significant effort to fully understand and validate these models as predictive preclinical models of neuropathic pain. An unevaluated or tested alternative to the surgical models described above would be to determine if a naturally expressed form of neuropathic pain is expressed in diabetic non-human primates. Spontaneously type-II diabetic rhesus monkeys show significant signs of distal peripheral neuropathy (Pare et al. 2007) that would be expected to be associated with neuropathic pain if the proper assessment methods were developed and employed. There is a significant opportunity to explore and develop useful pain assessment tools in a primate with naturally occurring diabetic neuropathy.

Go to:

17.7. ALGOGENIC/PHARMACODYNAMIC MODELS

The various pain models described above, in part, are attractive because they all have some degree of face validity with at least one clinically relevant pain patient population. However, this can be a two-edged sword in that the preclinical models also may share many of the complexities of the human clinical conditions that may confound understanding of the actual target-based activity of a specific analgesic compound in development. For the successful development of novel analgesics, it is important to understand if a compound is actually engaging the therapeutic target, regardless of demonstrated analgesic efficacy in a specific pain model. Fortunately, there are a number of large animal model systems that have been described that can be used to explore specific pain therapeutic targets and how they contribute to physiological and pathological pain.

One common approach is to build upon the primate tail immersion acute nociceptive or inflammatory pain model described above to create an algogen-sensitized tail immersion-withdrawal assay (Table 17.4). Similar to the inflammation-driven hypersensitivity induced by carrageenan injection into the tail, the injection or topical application of a variety of algogenic substances can lead to increased sensitivity to what is otherwise a non-painful stimulus. This approach has been shown to work for the selective transient receptor potential vanilloid 1 (TRPV1) agonist capsaicin (Butelman et al. 2003; Ko et al. 1998), the prostaglandin receptor agonist prostaglandin (Negus et al. 1993), and the non-selective bradykinin receptor agonist bradykinin (Butelman et al. 1995). In each case, when the agonist is either injected into or topically applied to the skin of the tail, the monkey will rapidly withdraw its tail from what would otherwise be a non-noxious thermal stimulus of 38°C-46°C. The hypersensitivity induced by these algogens can be inhibited by a variety of analgesic agents including various opioids and NMDA receptor antagonists (Butelman et al. 1995, 2003; Ko et al. 1998; Negus et al. 1993). It is interesting to note that the delta-opioid receptor agonist BW373U86 was effective in reversing bradykinin but not prostaglandin-E2 induced-hypersensitivity, suggesting that this approach may potentially be used to distinguish subtle differences in pain processing.

TABLE 17.4

Large Animal Algogenic/Pharmacodynamic Pain Models.

Capsaicin-induced allodynia and hyperalgesia is a commonly used model of experimental pain that has been used in rodents, human subjects, and non-human primates to evaluate pharmacodynamic antinociceptive effects of compounds of interest (Dirks et al. 2002; Joshi et al. 2006; Ko et al. 1998). Interestingly, a recent study by Joshi et al. (2006)suggested that this model may be used as a surrogate model of neuropathic pain, given the similar pharmacology observed in rodent models of capsaicin-induced and neuropathy-induced pain. Consistent with this notion, the anticonvulsant gabapentin is also efficacious in the rhesus monkey capsaicin pain model, suggesting this experimental pain model in non-human primates may also possibly be used as a surrogate model of neuropathic pain (Figure 17.1B). However, more studies will need to be performed to gain a better understanding of the pharmacology associated with this model and how closely the pharmacology tracks to mechanisms associated with neuropathic pain.

In addition to measuring stimulus-evoked reflex withdrawal responses, another way to assess the pharmacodynamic effects of compounds of interest is to take advantage of the axon reflex-induced flare or vasodilation response. It is commonly observed that direct activation of C-fibers with agents such as capsaicin leads to an increase in dermal vasodilation in the region surrounding the application of capsaicin (Willis 1999). Thus, compounds that target a mechanism that is involved in the detection and propagation of the capsaicin stimulus (e.g., TRPV1) or the expression of the vasodilation response (e.g., CGRP) may alter the capsaicin-induced vasodilation response (e.g., see Hershey et al. 2005; Table 17.4).

Go to:

17.8. SUMMARY

The development and use of rodent pain models has undoubtedly contributed to our understanding of the pathophysiology and pharmacology associated with acute and chronic pain states. However, despite significant research efforts using these models, few novel analgesics have been developed to date based on novel mechanisms identified in these models. Although many potential reasons for this can be identified, it seems clear that a disproportionate amount of research efforts have focused on rodent pain models as opposed to models in higher species, despite the advantages of using large animal models in terms of human translation. The relative paucity of large-animal pain model data to date illustrates the need to better characterize the existing models in terms of methods of pain assessment and pharmacology. In other cases, there are minimal or no large animal pain models that correspond to certain prevalent pain conditions such as in the case of painful diabetic polyneuropathy. The further characterization and development of large animal pain models that more closely represent the natural underlying disease in humans (e.g., osteoarthritis, diabetic neuropathy) will likely offer greater insight into novel mechanisms responsible for chronic pain and will increase the probability of success to develop novel therapeutics to treat pain.

Hoping this will be helpful,

Rafik

Hi Arvind,

With all due respect, your statement doesn’t meet any scientific or practical data. If one would carefully obey the most general ‘evidence-threshold’ approach (based on the assumption that the observer can detect signs of pain in the species of animal they work with) and follow every AEC-prescribed experimental manipulations (pointing at ‘pain equivalence’ maintenance), it does not matter rat or murine models were used for analgesic and anti-inflammatory activities.

All the above might be studied in detail from the review of John C. Schofield  and Virginia M. Williams  entitled “Analgesic Best Practice for the Use of Animals in Research and Teaching - An Interpretative International Literature Review”,  Food and Agriculture Organization of the United Nations, September 2002. Thus they directly mentioned that a problem with experimental investigations into preemptive medication appears to be the lack of a reliable animal model, and showed rats as the model for analgesic activitiy.  Among those - ovariohysterectomy in the rat was used model, which demonstrated a preemptive analgesic effect of morphine given 30 minutes before surgery.

Another study on using both rats and mice was of Jaiswal et al. It was published in Indian J Pharmacol. 2012; 44(4): 475–479, and entitled “Experimental evaluation of analgesic and anti-inflammatory activity of simvastatin and atorvastatin”. In that, analgesic activity of simvastatin and atorvastatin was assessed in tail flick model in rats, where it was compared with aspirin and tramadol and in acetic acid induced writhing in mice, where it was compared with aspirin. Anti-inflammatory activity of statins was evaluated using carrageenin induced paw edema and formalin induced arthritis in rats. The results obtained demonstrated that in the tail flick method, analgesic effect of tramadol was significantly more than the other drugs except at two observation times, when it was comparable to simvastatin and atorvastatin. Effect of simvastatin was found to be comparable to aspirin. In acetic acid induced writhing method, analgesic activity of simvastatin was comparable to that of aspirin while that of atorvastatin was significantly less. In carrageenin induced paw edema in rats, both simvastatin and atorvastatin showed anti-inflammatory activity which was comparable to aspirin.

Best wishes,

Ilya