Commentary: Can we prevent chronic osteoarthritis pain? A view from the bench

In this Special Issue on Prevention of Osteoarthritis (OA), Whittaker and colleagues thoughtfully discuss their views on strategies for preventing this common form of arthritis, which would likely have a high impact on reducing its associated individual and societal burden, given the paucity of available treatment options [1]. The authors highlight the fact that most of the focus in this arena has been on prevention of risk factors for the ‘disease’ of OA, i.e. structural features of joint pathology, while few clinical studies to date have considered the risk factors for the accompanying ‘illness’, i.e. symptomatic features of OA [1]. This should give us pause, since OA is worldwide one of the major sources of chronic pain, and current analgesics provide inadequate pain relief while their chronic use can cause significant side effects, especially in the older population [2]. Hence, if we could define windows of opportunity for early intervention and develop strategies for preventing the chronic pain associated with OA, this should be tremendously beneficial to patients and to society.

In this Commentary, we will briefly examine how basic and preclinical research may contribute to our understanding of early-onset pro-algesic mechanisms that lead to chronic pain, and may inform clinical research to help achieve the goal of preventing chronic pain. Many preclinical models of OA, in rodents as well as in large animals, model different risk factors of OA, predominantly post-traumatic OA, and less often, obesity and age-related OA [3]. For decades, the main outcomes in animal models were largely limited to structural joint pathology, but this is rapidly changing and preclinical studies increasingly report pain-related behavioral and functional outcomes in addition to structural outcomes [4]. Increasingly, it is being reported that measures indicative of sensitization – generally focusing on mechanical hypersensitivity at the knee or the hindpaw - become apparent early in the disease model [4]. For example, early onset mechanical hypersensitivity has been reported after destabilization of the medial meniscus (DMM), partial meniscectomy (PMX), or anterior cruciate ligament (ACL) rupture in the mouse knee [5]. Mechanical hypersensitivity (hyperalgesia, allodynia) is indicative of neuronal sensitization to mechanical stimuli, a process mediated by specific molecularly defined neuronal pathways, as briefly discussed below. Importantly, patients with OA also show signs of mechanical sensitization, as detected by quantitative sensory testing (QST), including reduced pain pressure thresholds and increased temporal summation, both at the OA joint as well as at anatomical sites away from the joint [6]. Very few studies to date have examined the temporal association between QST measures and pain in OA, but a recent clinical study by Carlesso et al. suggests that certain QST measures may be useful for predicting future patient outcomes [7]. In this community-based cohort consisting of 852 subjects all of whom were free of persistent knee pain at baseline, patients with low pain pressure thresholds and moderate temporal summation had elevated risk for developing persistent knee pain two years later. Interestingly, these features of pain sensitization had a large influence on phenotype formation, whereas psychological constructs (including depression and catastrophizing), sleep quality, or widespread pain did not. As the authors note, “Our findings indicate that while other studies have highlighted the importance of these latter factors, the presence of sensitization appears to have a far greater influence on the development of persistent pain” [7].

This key study suggests that QST measures may be used to predict who will develop the “illness” of OA. However, this research approach is in its infancy and a lot more information is needed on the temporal development of QST measures, as well as their relationship to pain and progressive joint damage in different types of OA. There are many unanswered key questions: (1) can nervous system sensitization, as detected by QST, be modified?; (2) does modifying nervous system sensitization have an impact on development of chronic pain?; and (3) are there optimal windows of opportunity in the disease course to do so?

Animal models provide a tool for dissecting the relationships between sensitization, pain, and joint damage.

Experimental animal models provide a powerful opportunity for studying the relationship between sensitization, pain, and joint damage, since they enable the longitudinal assessment of pain-related behaviors and underlying neurobiological mechanisms [4, 5].

The molecular mechanisms of sensitization of sensory neurons in the course of joint pathology are increasingly well understood. Mediators such as nerve growth factor (NGF) as well as specific chemokines and cytokines can all sensitize joint afferents, so that they are more likely to fire in response to normal mechanical stimuli associated with joint use [8]. This sensitization is accompanied and maintained by molecular changes in pain-sensing neurons in the dorsal root ganglia (DRG), from where the pain signals are transmitted to the central nervous system: first to the dorsal horn of the spinal cord, and then to higher regions of the neuraxis, eventually giving rise to the conscious perception of pain [8]. Additionally, neuro-immune interactions in the DRG can also promote chronification of pain [8]. An important step in trying to identify opportunities for modifying sensitization would be to target these peripheral pathways early on, and assess whether such interventions can prevent or delay development of persistent pain. We identified very few preclinical studies that used this approach: In one rat study, a small molecule inhibitor of the NGF receptor, TrkA, was administered for two weeks after meniscectomy (MNX) [9]. Upon withdrawal of treatment, hind paw mechanical allodynia rapidly developed, while onset of weight-bearing asymmetry was significantly delayed and, 2 weeks later, it was still less pronounced than in untreated rats [9]. Another study tested the effects of altering the timing of delivering a small molecule inhibitor of CCR2, the receptor for the pro-algesic chemokine, CCL2, after DMM surgery in mice. The authors reported that CCR2 blockade between weeks 1 and 4 after surgery was optimal to achieve inhibition of weight-bearing asymmetry, and this effect lasted for 8 weeks after treatment withdrawal [10]. Similarly, in murine collagenase-induced OA, administration of an antibody neutralizing GM-CSF from day 7 to day 14 prevented development of weight-bearing asymmetry for 28 days after treatment withdrawal [11]. Neither the CCR2 or the GM-CSF blockade study assessed mechanical hypersensitivity, but it is known that both CCL2 and GM-CSF produce mechanical sensitization when injected locally [12, 13].

These few studies highlight the complex interrelationships between sensitization and persistent pain, and suggest that specific sensitization pathways could be targeted early in order to delay the development of persistent pain. We recommend that such approaches should be explored in more depth, since they would provide valuable translational insight.

Obesity and fat content: modifiable risk factors for OA-associated sensitization and pain?

Most preclinical studies that assess mechanisms of sensitization use chemical or surgical models of OA [3]. However, as discussed by Whittaker and colleagues, obesity has emerged as a very strong and highly modifiable risk factor for OA pain. Indeed, 5 kg or 5% reduction in body weight is associated with significantly lower incidence rates of knee OA disease and illness after 6.5 years [1]. Diet-induced obesity increases the risk of symptomatic features of OA, including pain-related behaviors in mice [14]. While literature on early non-pharmacological interventions such as exercise or modulating fat content in animal models is scant, a few studies suggest that these may provide fruitful research avenues. For example, in the rat MIA model, exercise provided through treadmill running, starting on day 10 and continuing through day 49, prevented persistent tactile hypersensitivity and weight-bearing asymmetry [15]. Body weight was not reported in this study, but there was a beneficial effect of exercise on the subchondral bone. Similarly, in a 7-week murine DMM model, treadmill running attenuated hind paw mechanical allodynia [16]. Finally, a recent elegant study showed that lipodystrophy mice were protected from joint damage and knee hyperalgesia after DMM, even under a high fat diet [17]. Furthermore, adding back a mixture of subcutaneous and visceral adipose tissue from wildtype mice could overcome this protection, and resulted in these mice developing both joint damage and knee hyperalgesia.

These few studies in disparate models of OA suggest that strategies that target fat and obesity may indeed be useful for modifying sensitization and pain in OA, and when used early on may be able to prevent chronification of pain, as well as structural joint damage.

Conclusion and opportunities for future research

In conclusion, emerging studies in experimental models of OA suggest that neuronal sensitization may constitute a modifiable risk factor for developing debilitating chronic pain associated with OA. Optimal early targeting windows should be determined, since continuous treatment may not be necessary in order to gain these benefits. It should also be considered that these pathways and windows of opportunity may differ between the sexes. For example, two recent studies in surgical models reported that, compared to male mice, female mice developed weight-bearing asymmetry with milder joint damage [18, 19], suggesting that to prevent persistent pain one would have to intervene earlier in the structural disease progression process in females compared to males.

We should of course consider that pain is often fundamentally a protective mechanism and, as such, it may be important in preventing progressive joint damage. Future clinical success in treating joint pain will ultimately depend on whether we can produce pain relief in a fashion that does not produce unwanted side effects, and we can dissociate mechanisms of adaptive pain from those underlying non-adaptive chronic pain. We predict that it will be therapeutically fruitful to continue investigating the relationship between sensitization and development of persistent pain. In particular, we encourage the incorporation of measures of both sensitization and persistent pain along with assessment of structural joint damage in study designs. Since different models may employ disparate mechanisms with different rates of progression [20], it will be important to use different OA models that capture distinct phenotypes. In depth mechanistic studies will shed light on how different risk factors may lead to sensitization, and ultimately pain. Finally, incorporating therapeutic protocols that target different stages of disease will also be key to learning more about prevention in addition to treatment.