An animal model of chronic inflammatory pain: Pharmacological and temporal differentiation from acute models

1. Introduction 

Rheumatoid and osteoarthritis comprise the two most common forms of arthritis, with over seven million sufferers in the UK alone (Arthritis Research Campaign, 2004). Although arthritis is defined as inflammation of the joint, the primary feature with which patients present in the clinic is chronic pain. Currently available therapies used in osteoarthritis (OA) and rheumatoid arthritis (RA) fail to adequately alleviate pain in many patients, and side effects of the treatments often limit their use. There remains an unmet need for effective treatment of chronic inflammatory pain. Current pharmacological treatment of chronic pain relies on drugs whose mechanisms were originally established more than half a century ago. The substantial investment in pain research by pharmaceutical companies has failed to deliver novel therapies based on new mechanisms, despite positive animal model data and a better understanding of the pain generation and processing pathways. However, pre-clinical and clinical failures are seldom fully reported in detail, making a direct correlation between animal model success and clinical failure difficult to interpret. There is therefore a clear need for pre-clinical models of pain which are better able to predict a positive clinical outcome.

The original models of inflammatory pain attempted to recapitulate the multiple site arthritis found clinically by using a systemic injection of inflammatory mediator (adjuvant) into the base of the tail in rats (see Rubens-Duval et al., 1996, Rainsford, 1982, Millan, 1986; and Colpaert, 1987 for reviews). This resulted in a polyarthritis comprising profound inflammation and hyperalgesia, initially at the site of injection. However, this model also produced a delayed T-cell mediated hypersensitivity reaction (Pearson and Wood, 1959, Wall, 1984) with multiple joint involvement and subsequent development of lesions of the eyes, ears, nose and genitals, as well as lymph node enlargement and impairment of liver metabolism. These global effects are not reflective of the pathology observed in clinical RA or OA and therefore the model was subsequently modified to restrict the exposure of the inflammatory mediator to a small area.

These restricted exposure models commonly comprise the subcutaneous injection of an inflammatory mediator, into the plantar surface of the rat hind paw, with the resulting inflammation and pain being measured using a variety of readouts, including von Frey filaments (Chaplan et al., 1994), paw withdrawal, (Randall and Selitto, 1957) and weight bearing (Clayton et al., 1997). Further adaptations of this employed intra-articular injection of inflammatory agents with some success (Bileviciute et al., 1993, Mapp et al., 1993, LaBuda and Fuchs, 2001), although this was not subsequently universally employed for the drug discovery process. Almost universally, the inflammation and pain are recorded within several hours of the insult and many clinically active anti-inflammatory and analgesic drugs have been shown to be efficacious in these models, reversing inflammation, hypersensitivity, or both.

Given the apparent similarity of pharmacological effects in these models previously, often the more acute model is employed to determine efficacy of novel chemical entities. Despite the apparent correlation between efficacy in these models of acute inflammatory pain and clinical benefit, research strategies based on use of the models have not led to the discovery of new therapies in recent years. Thus, in spite of the success of the COX-2 inhibitors, which show close correlation between effectiveness in animal models and in the clinic, many of the newer agents reported to be active in these pre-clinical models appear to fail when taken forward into the clinic. Animal models that are more predictive for human inflammatory pain states are therefore desirable for drug discovery. Clinically, inflammatory pain is most often associated with chronic conditions such as OA, chronic lower back pain and RA in which any inflammation or plastic neuronal changes in the peripheral and central nervous system would have been occurring for some time. Typically, animal models used to assess novel therapies employ an acute readout. The inflammation and pain induced by intra-plantar injection of adjuvant, while significant within 12h, continues to rise over subsequent days and may peak anywhere between 1 and 2 weeks following a single injection (Stein et al., 1988, Nagakura et al., 2003). This suggests that testing novel agents for analgesia or anti-inflammatory activity within hours of such insults may be inappropriate because the underlying mechanisms responsible for these effects may develop and change over a longer period of time.

The pathophysiology of joint pain resulting from adjuvant injection includes a peripheral sensitisation, which can occur within hours of the insult (Schaible and Schmidt, 1988). Further central sensitisation also occurs following continued nociceptive primary afferent discharge as a result of the initial insult, and sensitisation to adrenaline and sympathetic discharge may also occur with time (Sanjue and Jun, 1989). These data suggest that a chronic animal paradigm in which the inflammatory insult has had time to induce centrally mediated changes, may result in a more predictive model of the pathology present in man. Additionally, many of the clinical inflammatory pain conditions result from afferent barrage and central sensitisation following joint damage/distortion or deterioration which is not well modelled by injection of adjuvant to the plantar surface of a rat hind paw. Demonstration of drug effects in a chronic model in which the insult is confined to the joint capsule may therefore be more suitable for evaluating new therapies aimed at pharmacological treatment of chronic inflammatory rheumatoid pain.

The aims of this study were to develop a model of chronic inflammatory pain, which was restricted to a joint, showed a greater chronicity compared with typical acute or semi-acute models and which responded to clinically relevant therapies. We tested the utility of this model using two novel therapeutic approaches, namely antagonism of the NR2B receptor, and inhibition of iNOS, found quite different effects when comparing results from traditional models with those from this model of chronic inflammatory joint pain.

2. Methods 

3. Joint pain studies 

4. Intra-plantar CFA studies 

4.4. NR2B antagonist/iNOS inhibitor studies 

Twenty three hours after CFA injection, Ro 25–6981 (3–30mg/kg, s.c.), GW274150 (3–30mg/kg, p.o.), or vehicle (saline) were administered (n=7/group), and weight bearing and paw volume measured 1h later.

4.4.1. Drugs 

The compounds used in this study, the supplier, together with their mechanism of action are detailed in the table below.

	Agent	Mechanism	Supplier
	CFA	–	Sigma
	Ibuprofen	NSAID	Sigma
	Morphine	Opioid	Sigma
	Dexamethasone	Steroid	Sigma
	Ro 25-6981	NR2B antagonist	Sigma
	Rofecoxib	COX-2 inhibitor	GSK
	Etoricoxib	COX-2 inhibitor	GSK
	GW274150F	iNOS inhibitor	GSK

4.4.2. Data analysis 

Data from these studies were analysed using the Statistica v 5.1 software package. Study data was analysed by ANOVA followed by post-hoc Duncan’s test. Statistically significant differences are expressed as P values less than 0.05.

5. Results 

5.1. Joint pain model 

Following CFA injection animals appeared to be of normal general health and demonstrated no marked changes in weight gain or grooming. Although animals displayed inflammation of the knee injected with CFA, the inflammatory response was restricted to the stifle joint and effects similar to those observed in tail-base CFA administration were not observed. All data is expressed as a percentage of the contralateral limb reading to negate differences due to absolute weight of the animal contributing to noise in the data.

In the initial study, injection of 100μg CFA into the rat knee induced a time dependent shift in weight bearing such that the animals carried the majority of their body weight on the contralateral leg (Fig. 1). This effect was apparent by 6h post-injection (p<0.05), although after 7–14 days the body weight distribution appeared to be returning toward saline levels, although the weight bearing effect was still significantly different from saline at all time points (p<0.05).

View Large Image Download to PowerPoint

Fig. 1. Effect of intra-articular injection into the rat knee joint with 100μl CFA (closed squares) or saline (open squares) on weight bearing. Data are mean±s.e.m. All CFA values are significantly different from saline (p<0.05) except 3h.

In the subsequent time course study (Fig. 2A/B) a robust and consistent difference in weight bearing and joint diameter were observed following injection of the larger 150μg CFA. Sham saline injections did not affect weight bearing or joint diameter. This time-course study demonstrated an initial peak in hyperalgesia following intra-articular CFA injection, resulting in approximately 85% shift in body weight to the contralateral hind limb, which resolved slightly over the following two weeks to approximately 55% shift in body weight distribution and then remained at this level for the rest of the study. All values from CFA-treated animals were significantly different from saline (p<0.05), except on day 0 (before injection). Similarly, the left knee joint diameter showed an initial peak at day 5 post-CFA (150% of contralateral). This slowly resolved to 130% of contralateral by day 30 and then sharply declined to 110% by day 35, and was then maintained at this level for the rest of the time-course. All CFA values are significantly different from saline (p<0.05), except on day 0 (before injection). Tissue samples described earlier were taken at various times in the progression of the model.

View Large Image Download to PowerPoint

Fig. 2. (A/B) Effect of intra-articular injection into the rat knee joint with 150μl CFA (closed squares) or saline (open squares) on (A) weight bearing and (B) joint diameter. Data are mean±s.e.m. All CFA values are significantly different from saline (p<0.05) except Day 0.

The right and left stifle joints were assessed for histopathological changes (Table 1). Changes were graded (Table 1) from 1 (very slight) to 5 (very marked). Sham treated joints (not tabulated) showed either no notable changes or minimal synovial changes, with the predominant cell type present being adipocytes. All CFA treated joints showed pathological changes. Twenty four hours following CFA injection the synovial histology appeared largely unchanged (Fig. 3) with the major cell type present being adipocytes, although inflammatory cells comprising mainly neutrophil polymorphs were present. On day 5, synovitis was characterised by a moderate to marked inflammatory cell infiltration, still comprised mainly of neutrophil polymorphs. This was modified at the later time points (days 16 and 35) to include increased numbers of mononuclear cells (lymphocytes, macrophages, plasmacytes) together with fibroblasts, fibrocytes and fewer neutrophils. Fibroplasia was present to a modest degree at these time points, while numbers of adipocytes was greatly reduced. By days 62–90 the cellularity was decreased and the fibroplastic component was more pronounced and mature, indicating a more chronic stage of inflammation. Very slight erosion of articular cartilage was noted in most animals. Bone changes were minimal and noted in joints from only a few rats. In all CFA-injected joints, initial periarticular inflammatory changes, involving the tendons, muscles and subcutis were evident, particularly at the early time points. Sham treated rats showed minimal to mild inflammatory changes peaking at day 5 and then declining to “background” levels.

Table 1. Numerical evaluation of the pathological findings following CFA or saline injection over the second time course study

Day 1

Animal number

31R

31L

32R

32L

33R

33L

34R

34L

35R

35L

Synovial inflammation

–

3

–

3

–

3

–

3

–

3

Periarticular inflammation

–

3

–

2

–

4

–

3

–

2

Cartilage: erosion/destruction

–

1

–

1

–

1

–

1

–

1

Bone: resorption/destruction

–

–

–

–

–

–

–

–

–

–

Day 5

Animal number

36R

36L

37R

37L

38R

38L

39R

39L

40R

40L

Synovial inflammation

–

3

–

3

–

4

–

4

–

4

Periarticular inflammation

–

3

–

4

–

3

–

4

–

3

Cartilage: erosion/destruction

–

2

–

2

–

2

–

2

–

2

Bone: resorption/destruction

–

–

–

–

–

–

–

–

–

–

Day 16

Animal number

41R

41L

42R

42L

43R

43L

44R

44L

45R

45L

Synovial inflammation

–

4

–

3

–

3

–

4

–

4

Periarticular inflammation

–

2

–

2

–

3

–

3

–

3

Cartilage: erosion/destruction

–

2

–

2

–

2

–

2

–

3

Bone: resorption/destruction

–

1

–

–

–

2

–

2

–

3

Day 35

Animal number

46R

46L

47R

47L

48R

48L

49R

49L

50R

50L

Synovial inflammation

–

2

–

2

–

3

–

2

–

3

Periarticular inflammation

–

3

–

2

–

3

–

3

–

2

Cartilage: erosion/destruction

–

2

–

2

–

2

–

2

–

2

Bone: resorption/destruction

–

–

–

–

–

–

–

–

–

–

Day 62

Animal number

51R

51L

52R

52L

53R

53L

54R

54L

55R

55L

Synovial inflammation

–

2

–

2

–

2

–

2

–

2

Periarticular inflammation

–

2

–

2

–

2

–

2

–

3

Cartilage: erosion/destruction

–

2

–

1

–

1

–

1

–

1

Bone: resorption/destruction

–

–

–

–

–

–

–

–

–

–

Day 90

Animal Number

56R

56L

57R

57L

58R

58L

59R

59L

60R

60L

Synovial inflammation

–

2

–

3

–

2

–

3

1

3

Periarticular inflammation

–

2

–

3

–

2

–

2

–

2

Cartilage: erosion/destruction

–

2

–

2

–

2

–

1

–

1

Bone: resorption/destruction

–

1

–

2

–

–

–

–

–

–

Changes were graded (Fig. 3) from 1 (very slight) to 5 (very marked). First row numbers/letters in each block refer to the animal number and left (L) or right (R) knee joint. Sham treated joints (not tabulated) showed either no notable changes or minimal synovial changes.

View Large Image Download to PowerPoint

Fig. 3. Histology plates, showing changes in synovium over the time-course of the CFA-induced response. Days 1, 5, 16, 35, 62 90 and sham injected.

5.2. Drug studies 

Based on the temporal profile of both hypersensitivity and inflammation in the initial time-course studies, two time-points were chosen at which to test compounds. These were days 13–17, when hypersensitivity was observed in the presence of marked inflammation, and also days 55–59, when hypersensitivity was present with minimal joint diameter differences and little inflammation assessed histologically.

Two validation studies (Fig. 4) were performed to test gold standards having differing mechanisms of action. Activity was assessed at two time points (days 14–16) and (days 55–59), as detailed above.

View Large Image Download to PowerPoint

Fig. 4. (A–D) Effect of vehicle (filled square), morphine (t.i.d., 3mg/kg, filled circle), ibuprofen (t.i.d., 30mg/kg, open circle), etoricoxib (b.i.d., 5mg/kg, filled triangle) or dexamethasone (b.i.d., 1mg/kg, open triangle) on CFA-induced weight bearing over days 14–16 (A) and 55–59 (B) and joint diameter over days 14–16 (C) and 55–59 (D) compared to sham (open squares). All compounds resulted in a significant (p<0.05) reversal of both hypersensitivity and inflammation.

At the early time point (days 14–16) all drugs reversed both the CFA effects on hypersensitivity and inflammation. The effects on inflammation were apparent within 1h of the first dose (p<0.05) and on subsequent dosing days, (Fig. 4C), however, the hypersensitivity effects were small at this time, but significantly increased (p<0.05) on the second day of dosing (Fig. 4A). The apparent inflammation at the later time point assessment (days 55–59, Fig. 4D) was vastly reduced compared with the earlier time points (days 14–16), but a marked and consistent hyperalgesia was still evident. Following dosing over days 55–59 there was still a clear anti-hypersensitivity effect of all compounds following 5 days administration, with all compounds producing a significant reversal of hypersensitivity (p<0.05) over days 55–59 (Fig. 4B).

Administration of the COX-2 inhibitor rofecoxib produced a significant and dose dependent reversal of the CFA-induced hypersensitivity (Fig. 5A) over the early phase of the model (p<0.05, ANOVA). Post-hoc analysis demonstrated a dose of 5mg/kg to be maximally effective to reverse hypersensitivity, with no further improvement at 10mg/kg. Effects of rofecoxib on hypersensitivity at the late phase of the model were identical to those observed during the early phase (data not shown). While a significant effect was observed following the first dose at both phases, the effects were maximal after 4 days of twice daily dosing. Doses of 1–10mg/kg of rofecoxib all produced very similar effects against joint diameter (p<0.05) over days 13–17 during the early phase of the model (Fig. 5B). There was little inflammation present during the late phase for rofecoxib to have any further effects (data not shown). Administration of both 5 and 10mg/kg of rofecoxib produced indistinguishable effects on both hypersensitivity and inflammation. For these reasons 5mg/kg of rofecoxib was chosen as an appropriate dose to use as a model standard comparator in future drug studies.

View Large Image Download to PowerPoint

Fig. 5. COX-2 inhibitor, rofecoxib (0.3–10mg/kg, p.o., b.i.d., ×5 days; days 13–17) effects on (A) hypersensitivity, (B) joint diameter in the chronic inflammatory joint pain model. Vehicle (open square), 0.3mg/kg (closed square), 1mg/kg (open circle), 3mg/kg (closed circle), 5mg/kg (open triangle), 10mg/kg (closed triangle).

Administration of Ro 25-6981 produced a significant reversal of the CFA-induced hypersensitivity in the joint pain model only at the higher dose of 30mg/kg, over both early and late phase of the model (Fig. 6A, p<0.05). While a significant effect was observed following the first dose at both phases, the effects were not maximal until 4 days of twice daily dosing (Fig. 6A). The lower dose of Ro 25-6981 was without significant effect against hypersensitivity. Neither dose of antagonist had a significant effect against knee-joint diameter in this model (Fig. 6B, p>0.05). The COX-2 inhibitor rofecoxib, used as a positive control during the early phase in this study resulted in a significant reversal of both hypersensitivity and inflammation (p<0.05).

View Large Image Download to PowerPoint

Fig. 6. Effect of the NR2B antagonist Ro 25-6981 (3–30mg/kg s.c., b.i.d., ×5 days; days 13–17 or days 55–59) on (A) hypersensitivity, (B) joint diameter in the chronic inflammatory joint pain model. Ro25-6981 vehicle (open squares), Ro25-6981 vehicle /d55-59 Ro25-6981 (closed squares), rofecoxib vehicle (open circle), Ro25-6981 10mg/kg (closed circle), Ro25-6981 30mg/kg (open triangle), rofecoxib 5mg/kg (closed triangle).

GW274150F failed to reverse the hypersensitivity induced by intra-articular injection of CFA (Fig. 8A). There was however a significant reversal of joint diameter (Fig. 8B, p<0.05) in keeping with previous inflammation and acute hyperalgesia studies with this molecule.

View Large Image Download to PowerPoint

Fig. 8. Effect of the iNOS inhibitor GW274150F (3–30mg/kg p.o., b.i.d., ×5 days; [days 13–17]) on (A) hypersensitivity, (B) joint diameter in the chronic inflammatory joint pain model. GW274150F vehicle (open squares), rofecoxib vehicle (closed squares), GW274150F 3mg/kg (open circles), GW274150F 10mg/kg (closed circles), GW274150F 30mg/kg (open triangles), rofecoxib 5mg/kg (closed traingles).

5.3. Intra-plantar model 

Ro 25-6981 produced a modest reversal of the intra-plantar CFA-induced effect on weight bearing, with a maximal effect of ∼50% reversal at 30mg/kg, although this effect was not statistically significant as a result of the large variability in this data set (Fig. 7). There was no effect on paw oedema.

View Large Image Download to PowerPoint

Fig. 7. Effect of the NR2B antagonist Ro 25-6981 (3–30mg/kg s.c.,) on intra-plantar CFA induced hypersensitivity (A) and paw oedema (B).

GW274150F produced a significant (p<0.05) and dose related reversal of hyperalgesia in the intra-plantar CFA model, with a maximum effect of 53% reversal at 30mg/kg, although all dose levels were significantly different from vehicle (Fig. 9A). A small non-significant effect of 30mg/kg GW274150F was observed against paw oedema (Fig. 9B).

View Large Image Download to PowerPoint

Fig. 9. Effect of the iNOS inhibitor GW274150F (3–30mg/kg p.o.) on intra-plantar CFA induced hypersensitivity (A) and paw inflammation (B).

6. Discussion 

Here we have described a model for chronic inflammatory joint pain. Standard drugs for the treatment of inflammatory pain were effective at reducing both hypersensitivity and inflammation during the early phase of the model (days 13–17), and to further reduce hypersensitivity during the later phase (days 55–59) in which significantly less inflammation was present. Importantly, this model is capable of differentiating between molecules that result in apparent hypersensitivity in the more acute inflammatory pain models. Thus, the NR2B antagonist, Ro25-6981 which had modest affects to reverse hypersensitivity in the intra-plantar CFA model caused a marked reversal of hypersensitivity in the more chronic model of joint pain. Furthermore, the iNOS inhibitor, GW274150F significantly reversed hypersensitivity in intra-plantar CFA but was ineffective in the chronic joint pain model, despite having anti-inflammatory effects. Our results indicate that a chronic model of inflammatory joint pain, such as described here, may provide a better predictive tool over previously described acute models, for investigating novel therapies for chronic inflammatory pain indications. Progress of such novel mechanisms to the clinic will be the ultimate test of the model.

Previous monoarthritis models have used a variety of doses of CFA to induce inflammation and pain (Donaldson et al., 1993). In the present study an initial dose of 100μg of CFA into the knee joint resulted in hypersensitivity, similar to that seen following intra-plantar CFA using the same weight bearing model (Clayton et al., 1997). However the effects reduced with time, particularly after day 4, such that the hypersensitivity was less than 50% of that present at 6h post-CFA. Previous data from other labs (Donaldson et al., 1993) have demonstrated that the effects of intra-articular CFA are dose dependent. In our studies we have shown a larger concentration of CFA (150μg) resulted initially in a higher degree of hypersensitivity, followed by a stable and robust hypersensitivity for at least 90 days. Additionally, joint diameter demonstrated an initial peak on day 5, which showed modest resolution by day 30, and thereafter marked resolution back to pre-injection levels. Contralateral knee diameters did not change over the time-course of the study.

These data therefore suggests that changes occurring as a result of the CFA injection into the joint either from the peripheral afferents, and/or dorsal horn, spinal cord, or other higher centres, mediate plastic changes which evoke and sustain hypersensitivity in the absence of marked inflammation, indicative of a chronic pain-like response following an inflammatory insult. Reports of ongoing changes occurring at primary afferents and second order neurones are widespread in the literature. Schaible and Schmidt (1988) have found an increase in sensitivity of mechanosensitive nerves occurring within hours of an inflammatory insult, which relates to the initial response and beginning of sensitisation of these neurones. Stein et al. (1988) have shown that the inflammatory response to CFA does not peak until day 16 after injection and peak hypersensitivity to paw pressure occurred at day 12. This and other studies demonstrate that CFA-induced inflammation will eventually resolve. We have also shown that hypersensitivity persists well beyond the initial peak of inflammation, during which there was a resolving inflammatory appearance using histological techniques. We have also shown ongoing ‘pain’, assessed by weight bearing, of a similar magnitude, both in the presence of marked and resolving inflammation, suggesting that hypersensitivity can still occur and be maintained well beyond the initial inflammatory insult.

Histopathological findings indicate an appreciable development of pathology associated with CFA-induced inflammation, following a typical course from acute to chronic synovitis. Sham-injected rats showed minimal changes in the early stages only, which predominantly relate to injection artefacts. These data agree with the behavioural findings, suggesting the presence of ongoing hypersensitivity despite the reduction in inflammation at the later time points. The results suggest that injection of 150μg of adjuvant into the knee of the rat is a model of chronic inflammatory-induced pain. Furthermore, the histological data suggest that the 24h time-point, typically used for intra-plantar CFA evaluation of potential analgesics effective in inflammatory pain, may be inappropriate because the inflammatory insult driving the extent and type of inflammatory cells present are markedly smaller and different, in comparison with the later time-points. This suggests that the nociceptive barrage at 24h may be far less than at later time-points, resulting in considerably less central sensitisation. Furthermore, although marked hyperalgesia is apparent in the more acute models, this may reflect an acute inflammatory response rather than the pathological pain state present in clinical chronic pain.

The second aim of this study was to demonstrate validation of the chronic inflammatory joint pain model, using standard drugs, to reduce inflammation and/or hypersensitivity. Standard analgesics were administered at two different time points based on the time-course studies. These included days 14–16 when there was a marked and consistent joint diameter difference, and days 55–59, when the joint diameter was largely resolved. The longer dosing period for the later time point was to allow for the chronic nature of the model at this time and therefore allowing more possibility of observing a drug effect. However, it is unlikely that dosing for this period would be sufficient to observe any potential bone/cartilage morphology changes. The drugs tested (morphine, ibuprofen, etoricoxib and dexamethasone, as well as rofecoxib were able to reduce both hypersensitivity and inflammation in this chronic model. The ability of morphine to reduce inflammation was surprising although there is evidence in the literature consistent with opioids having anti-inflammatory effects (Stein et al., 2003, Machelska et al., 2002, Brack et al., 2004, Pourpak et al., 2004). Our data validates this chronic model since a variety of clinically active molecules having different mechanisms of action caused a reversal of CFA-induced hypersensitivity. Together with the chronic hypersensitivity phenotype, this paradigm can thus be used as an alternative chronic inflammatory pain model, which may better predict clinically active molecules from novel target classes.

The third aim of this study was to compare and contrast effects of two potential novel pain targets, NR2B antagonists and iNOS inhibitors, in the intra-plantar CFA model of inflammatory pain and the more chronic inflammatory joint pain model.

NMDA receptor antagonists have long been considered a potential target for the treatment of pain, particularly neuropathic pain. Their therapeutic potential has however been limited due to their unacceptable side effects (e.g., motor deficits, sedation and psychotomimetic). More recently the identification of distinct NMDA receptors comprising different subunits with differing neuronal distributions has allowed a more refined pharmacological approach targeting only those NMDA receptors containing the NR2B subunit. This approach has demonstrated positive effects in models of neuropathic and inflammatory pain which are devoid of the typical pan-NMDA receptor subunit antagonist side effects. In the present study we found only variable and modest effects of the NR2B antagonist Ro 25-6981 in the intra-plantar CFA model. This is in contrast with a number of other studies which have demonstrated analgesic effects with NR2B receptor antagonists in acute models of pain such as intra-plantar carrageenan (Taniguchi et al., 1997, Boyce et al., 1999, Claiborne et al., 2003). In those studies punctate stimuli such as paw withdrawal pressure or thermal stimulation were used, which may be surmountable by antagonists acting via peripheral receptors present on primary afferent nociceptive fibres. It is well known that NMDA receptor up-regulation occurs in pathological pain states in a time dependent manner (Bolay and Moskowitz, 2002, Miki et al., 2002, Heidinger et al., 2002, Loftis and Janowsky, 2003). We suggest that the reason for the modest effects in our intra-plantar CFA model may be that unstimulated, ongoing hypersensitivity in the weight bearing readout may be less influenced by acute NR2B antagonism. Furthermore, at 24h post-injection of CFA, up-regulation of NMDA NR2B receptors may be ongoing. We proposed that NR2B receptor antagonists may be more effective in a chronic incident pain model in which receptor changes are established and in which puntacte nociceptive stimuli are not administered, as in the chronic inflammatory joint pain model described herein. Indeed following administration of the NR2B antagonist Ro 25-6981 in our chronic joint pain model, the higher dose of 30mg/kg resulted in complete time-dependent reversal of hypersensitivity. The lack of effect against inflammation suggest a block of pain signalling through inhibition of NMDA driven excitatory signalling of primary nociceptive afferents and second order dorsal horn spinal cord neurones known to be up-regulated in ‘pain’ states, rather than through an anti-inflammatory mechanism (Chazot, 2000).

Despite the evidence supporting a role of nitric oxide and iNOS (Wu et al., 1997, Wu et al., 1998) and the positive findings of the iNOS inhibitor GW264150F in acute inflammatory models the present study failed to demonstrate an effect of the iNOS inhibitor GW274150F in the chronic joint pain model. Evidence suggests that intra-plantar injection of inflammatory mediators result in a local iNOS elevation (unpublished data), but this was markedly reduced by three weeks post-CFA insult. Given this temporal iNOS expression, and the positive effects in our acute, but not chronic model of inflammatory pain, it can be concluded that iNOS plays a major role in the initiation of inflammation and inflammatory pain, but that other factors mediate the pain during chronic phases. Interestingly, the inflammation was reduced in the joint pain model following GW274150F administration, suggesting that while iNOS may be integrally associated with inflammation and subsequent pain, alleviation of the inflammation alone is not sufficient to reduce hypersensitivity over the time-course of this study.

In summary, the present study has demonstrated that injection of CFA into the knee capsule resulted in a marked inflammatory response with concomitant behavioural hypersensitivity. Additionally, hypersensitivity was maintained over a period of 90 days, despite a significant reduction in inflammation (assessed behaviourally and histopathologically), suggesting central neuronal plasticity as a major component of the ongoing hypersensitivity. The present data also demonstrate the positive action of clinically relevant molecules to reduce both inflammation and hypersensitivity. This chronic model may therefore better reflect the chronic pain seen in many clinical arthritis conditions and thus be more predictive for evaluating novel analgesics. Differential effects of an NR2B antagonist and an iNOS inhibitor were observed between the traditional intra-plantar CFA model and the chronic joint pain model. The NR2B selective NMDA receptor antagonist inhibited chronic inflammatory pain despite disappointing effects in the acute intra-plantar CFA model, adding further weight to the possibility of targeting NR2B receptor antagonists as a novel therapy for chronic inflammatory pain, without the complication of pan-NMDA receptor antagonist side effects. Furthermore, inhibition of iNOS was ineffective against hypersensitivity in the chronic model, despite there being an anti-inflammatory effect. This suggests a role of iNOS in the initiation of inflammatory pain, but not in maintenance following establishment of a chronic condition.

In conclusion, the pharmacological data from our chronic model demonstrate differences from that observed in the acute inflammatory models, suggesting that the chronic paradigm may reflect an alternative to better model the chronic pain state observed clinically, and demonstrate enhanced predictive validity for clinically efficacious new chemical entities.