Abnormal gait, due to inflammation but not nerve injury, reflects enhanced nociception in preclinical pain models

Abstract

Validation of gait analysis has the potential to bridge the gap between data from animal pain models and clinical observations. The goal of these studies was to compare alterations in gait due to inflammation or nerve injury to traditional pain measurements in animals. Pharmacological experiments determined whether gait alterations were related to enhanced nociception, edema, or motor nerve dysfunction. Gait was analyzed using an automated system (DigiGait) after injection of an inflammatory agent (carrageenan; CARR or FCA; Freund's complete adjuvant) or nerve injury (axotomy; AXO, partial sciatic nerve ligation; PSNL, spinal nerve ligation; SNL or chronic constriction injury; CCI). All models caused significant alterations in gait and thermal (inflammatory) or mechanical (nerve injury) hyperalgesia. Both indomethacin and morphine were able to block or reverse thermal hyperalgesia and normalize gait in the CARR model. Indomethacin partially blocked and did not reverse paw edema, suggesting that gait alterations must be primarily driven by enhanced nociception. In nerve injury models, AXO, PSNL, CCI, and SNL caused changes to the largest number of gait indices with the rank order being AXO > PSNL = CCI > > SNL. Gabapentin and duloxetine reversed mechanical hyperalgesia but did not normalize gait in any nerve injury model. Collectively, these data suggest that pain is the primary driver of abnormal gait in models of inflammatory but not nerve injury-related pain and suggests that, in the latter, disruption in gait is due to perturbation to the motor system. Gait may therefore constitute an alternative and potentially clinically relevant measure of pain due to inflammation.

Introduction

Assessment of pain in rodent models of inflammatory and neuropathic pain has focused on evoked stimulus–response assays to determine tactile, mechanical, or thermal thresholds. Recently, there has been a trend to evaluate alternative measurements that assess nonevoked or spontaneous pain; examples include weight bearing (Bove et al., 2003), conditioned responding (Johansen et al., 2001), and operant tasks (Vierck et al., 1995). Gait analysis also has been partially characterized as a pain measurement in rats and mice (Bozkurt et al., 2008, Clarke et al., 1997, Coulthard et al., 2002, Gabriel et al., 2007, Möller et al., 2008, Simjee et al., 2007, Vincelette et al., 2007, Vrinten and Hamers, 2003, Yu et al., 2001). These measurements remove experimenter bias and have the potential to be more clinically relevant (Matson et al., 2007, Negus et al., 2006, Vrinten and Hamers, 2003). In addition, validation of gait analysis has the potential to bridge the wide gap between data from animal pain models and clinical observations.

Gait has been studied in animals and humans. It has provided an understanding of locomotor activity and is now a common clinical assessment that is sensitive to relatively minor changes associated with disease, injury, or rehabilitation (Clarke et al., 1997, Coulthard et al., 2002, Coulthard et al., 2003, Hawkins, 2002, Rocha et al., 1999, Whittle, 1996, Yu et al., 2001). Gait is composed of three distinct phases that make up stride (Fig. 1C): the first is brake, which is defined as the duration between initial and maximum paw-surface contact. The second phase is propulsion, which is defined as the duration between maximum and final paw-surface contact. The third phase is swing, which is defined as the duration of time that the paw is not in contact with the surface. The most common and simple rodent gait analysis is footprint analysis; this requires application of ink, food dye, powdered charcoal, or photo-developer to the paws, and static gait parameters are then measured from the resulting footprints. This approach is simple and reasonably sensitive, but there are practical limitations (Klapdor et al., 1997) and the outcome may not reflect coordinated movement (Hampton et al., 2004). Recently, a number of automated technologies have been proposed to, first, overcome these limitations and, second, to automate measurement of dynamic gait in rodents.

Alterations in gait have been shown to be associated with disease (Hampton et al., 2004), aging (van der Leeden et al., 2006), and substances that impair motor coordination (Kale et al., 2004). Importantly, gait alterations have also been reported in animal models that cause pain, such as osteoarthritis (Clarke et al., 1997), chronic inflammation (Coulthard et al., 2002), and nerve injury (Vrinten and Hamers, 2003). While these studies suggest that gait alterations are related to pain, other factors, such as the presence of edema, joint instability, or motor fiber damage, were not ruled out. In the present series of experiments, we sought to extend the above findings by conducting a comprehensive assessment of multiple rat models of inflammatory (i.e., carrageenan and FCA) and neuropathic pain (i.e., axotomy, partial sciatic nerve ligation, spinal nerve ligation, and chronic constriction injury) using the commercially available, automated gait analysis system DigiGait (Mouse Specifics, Boston, MA). In addition, via pharmacological manipulation, we wanted to determine whether these gait changes were the result of pain as opposed to the presence of other factors such as edema or motor nerve dysfunction (Table 1).