Reversal of Monoarthritis-induced Affective Disorders by Diclofenac in Rats

PATIENTS with arthritis primarily seek medical assistance due to persistent pain caused by the inflammation and destruction of joints.¹  This pain results from a complex interaction between the peripheral and central nervous systems. In general consensus, chronic pain provokes neuroplastic alterations that contribute to the maintenance and amplification of painful sensation and to the onset of related disorders.²  One of the most common consequences of chronic pain is the development of anxio-depressive disorders, manifested as helplessness, pessimism, rumination, and catastrophism.³  These affective alterations originate due to the disruption of central mechanisms and should be carefully monitored, because their presence is directly related with pain severity.⁴  Indeed, such modifications establish a vicious circle whereby chronic pain triggers profound emotional changes, which in turn enhance pain perception.⁴  It is therefore critical to determine whether these plastic changes are reversible, because this could strongly condition the treatment strategies used. One potentially effective approach is to explore the effect of topically administering nonsteroidal anti-inflammatory drug in animal models which are among the first-line treatments to alleviate joint pain⁵  and are frequently administered topically to avoid the side effects associated with their oral administration.⁶ 

We focused on the pathway formed by the spinal cord (SC), paragigantocellularis nucleus (PGi), locus coeruleus (LC), and prefrontal cortex (PFC). The LC is involved in ascending and descending pain pathways and constitutes the main source of noradrenaline in the central nervous system, a neurotransmitter implicated in pain, emotion, stress, depression, anxiety, and other disorders.⁷  The LC receives ascending nociceptive inputs from the SC through the PGi⁸  and it projects to forebrain structures similar to the anterior cingulate cortex in the PFC, which is implicated in cognition and pain-related emotion.^9–13  Although it is clear that the anatomical and modulatory link between these structures may mediate the integration of pain processing at the emotional level, the underlying molecular and cellular mechanisms remain poorly understood. The extracellular signal–regulated kinases 1 and 2 (ERK_1/2), members of the mitogen-activated protein kinase superfamily, have been used as markers of activity in the SC after noxious stimulation and tissue injury.¹⁴  Beneficial results were demonstrated in a clinical trial using a mitogen-activated protein kinase inhibitor to treat neuropathic pain,¹⁵  suggesting this kinase as a target in the treatment of pathological pain.¹⁶  Interestingly, ERK_1/2 activation is also strongly implicated in the regulation of pain-associated disorders. Indeed, recent studies indicate that ERK_1/2 is activated in the anterior cingulate cortex of the PFC after tissue or nerve injury,^9,10,13  suggesting that ERK_1/2 is involved in both the sensorial and emotional aspects of pain.

We hypothesized that an effective treatment of the inflammatory condition is able to reverse the emotional and molecular responses produced by this condition. Thus, we used an experimental rat model of rheumatoid arthritis (monoarthritis) and we assessed the subsequent effects of diclofenac treatment on pain-induced anxiety, depression, and ERK_1/2 activation in the SC–PGi–LC–PFC pathway.

Adult male Harlan Sprague–Dawley rats (n = 116) weighing 250 to 300 g were provided by the Experimental Unit of the University of Cádiz, Cádiz, Spain (registration number: ES110120000210). The animals were housed 2 to 4 per cage, with an ad libitum access to food and water, and they were maintained on a 12-h light–dark cycle at 22°C and with 45 to 60% humidity. All experimental procedures were carried out in accordance with the European Communities Council Directive of September 22, 2010 (2010/63/EC), Spanish Law (RD 1201/2005), and ethical guidelines for the study of experimental pain in animals.¹⁷  The experimental protocols were reviewed and approved by the Institutional Ethical Committee for animal care and use (Cádiz, Andalucía, Spain).

Monoarthritis was induced as described previously.¹⁸  In brief, anesthesia was induced and maintained with isoflurane (4 and 2%, respectively; Abbott, Madrid, Spain), and the rats were injected in the left tibiotarsal joint with 50 μl of complete Freund’s adjuvant (CFA) solution containing 30 mg of desiccated Mycobacterium butyricum (Difco Laboratories, Detroit, MI) diluted in the vehicle solution (3 ml paraffin oil, 2 ml saline, and 500 μl Tween-80). Control rats were injected with the vehicle solution alone and any animal exhibiting sign of polyarthritis was excluded from the study.

After CFA or vehicle injection, all the animals were monitored weekly for 4 weeks, measuring their body weight, rectal temperature (Chy 580BR Thermometer; Bioseb, Vitrolles, France), paw volume (using a plethysmometer apparatus; Ugo Basile, Comerio, Italy), and inflammation score. The inflammation score was a subjective scoring based on the signs of inflammation at the injection site and locomotion, whereby: 0 corresponds to the absence of inflammation and 4 to severe inflammation with persistent flexion of the affected limb and motor activity effects.¹⁹ 

Each week, nociceptive behavior was assessed in both the ipsi- and contralateral paws. Physiological evaluation of movement-induced nociception (allodynia) was performed using the ankle-bend test,²⁰  which involves the assessment of squeak and/or struggle reactions in response to five alternate flexions and extensions of the ankle joint. Higher scores (score 2) indicate squeak responses to moderate manipulations of the inflamed joint, whereas lower scores (score 0) indicate the absence of a response to manipulation. The reactions recorded in response to each extension or flexion are summed to obtain the ankle-bend score, giving a maximum value of 20, and a high ankle-bend score is indicative of allodynia. In addition, the presence of mechanical secondary hyperalgesia was determined using the paw-pinch test.²¹  In brief, increasing pressure was gradually applied to the dorsal side of the paw using a graded motor-driven device (Ugo Basile) and beginning at 30 g of pressure. Three measurements were taken for each paw at 5-min intervals and the average value determined, with a 250 g cutoff applied to prevent damage to the paw. Secondary hyperalgesia is indicated by a reduction in the pressure-provoking withdrawal.

Pain-induced emotional and affective changes were assessed in a separate set of animals at 4 (monoarthritis [MA]4D), 14 (MA14D), and 28 days (MA28D) after inducing monoarthritis. Anxiety-like behavior was evaluated by using the marble-burying and elevated zero maze (EZM) tests.

In the marble-burying test,²²  after a 30-min acclimatization, the rats were placed individually in a transparent plastic cage (51 × 22 × 15 cm) illuminated by a 100-W light and containing bedding (5 cm deep) in which 20 marbles were arranged in four columns and five rows. After 30 min, the rats were removed and the number of buried marbles was counted, considering the marbles to be buried if they were at least two thirds covered with bedding. A larger number of buried marbles are taken as an indicator of increased anxiety-like behavior.

The EZM consisted of a black circular platform 105 cm in diameter that was elevated 65 cm above the floor. This maze was divided into four equal length quadrants: the two opposing open quadrants had 1-cm-high clear lips to prevent falls, whereas the two opposing closed quadrants were enclosed by 27-cm-high black walls. Each 5-min trial was carried out under the same lighting conditions and commenced with the animal being placed in the center of a closed quadrant. Spontaneous Motor Activity Recording and Tracking software (Panlab S.L.U., Barcelona, Spain) was used to analyze the percentage time spent in the open arms and the total distance travelled by each rat. An increase in the amount of time spent in the closed areas is correlated with anxiety-like behavior.

Depression-like behavior was evaluated by using a modified version of the forced swimming test (FST), as described previously.²³  When confined to a limited space, rodents engage in vigorous escape behavior. When it becomes clear that escape is impossible, these animals adopt an immobile posture, performing only the necessary movements required to keep their head above the water. Accordingly, in the FST, rats were placed for 15 min (pretest) in a clear cylindrical plastic container (46 cm high and 20 cm in diameter) filled with 30 cm of water (25° ± 1°C). After 24 h, the rats were once again exposed to the same conditions for 5 min (test) and they were videotaped to score their immobility (floating without struggling, using small movements to maintain the head above water), swimming (actively moving limbs more than is required to maintain the head above water), and climbing (active forepaw movements in and out of the water, often directed at the wall of the tank).²³  Customized software (Red Mice, Cádiz, Spain) was used to analyze the videos and to determine the predominant behavior at 5-s intervals. The total counts for each behavior during a 5-min test were averaged and although longer periods of immobility are taken as an indication of depression-like behavior, changes in the time spent climbing or swimming have been correlated with alterations in the availability of noradrenaline and serotonin, respectively.²³  As a positive control of antidepressant activity in the FST, the antidepressant desipramine (20 mg/kg; Sigma-Aldrich, St Louis, MO) was administered intraperitoneally to naive rats at 23.5, 5, and 1 h before testing. The behavioral tests were separated by a 3-day interval.

Fresh tissue samples from the ipsilateral SC segments L3–L6, PGi, LC, and PFC regions were collected at 4, 14, and 28 days after injection of the CFA or vehicle. All the samples were processed for Western blotting and after the tissue was lysed, an aliquot (50 μg) was separated on a 10% polyacrylamide gel and then transferred to a polyvinylidene difluoride membrane (BioRad, Hercules, CA). After washing in Tris-buffered saline containing 0.1% Tween-20 (TBST), the blots were blocked with 5% bovine serum albumin (Sigma) in TBST and probed overnight at 4°C with rabbit anti-phospho-ERK_1/2 (1:5,000; Acris Antibodies, Herford, Germany), mouse-anti ERK_1/2 (1:2,000; Cell Signaling Technology, Danvers, MA), and mouse anti-tubulin (1:250,000; Sigma) antibodies diluted in 5% bovine serum albumin–TBST. After thorough washing, these primary antibodies were detected by incubating for 1 h at room temperature with IRDye 800CW goat anti-rabbit (green) or IRDye 680LT goat anti-mouse (red) secondary antibodies (1:10,000; LI-COR^®, Lincoln, NE). After three final washes with TBST, the antibody binding was detected by using a LI-COR Odyssey^® two-channel quantitative fluorescence imaging system (LI-COR^®). Digital images of Western blots were analyzed by densitometry using the ImageJ free access software (National Institutes of Health, Bethesda, MD). The data were expressed as pERK_1/2 expression levels relative to those of total ERK_1/2, as no significant differences in the loading control (tubulin) were observed. As no differences in pERK_1/2 expression were observed between the ipsilateral and contralateral sides for the PGi, LC, or PFC, these values were combined and averaged.

To induce analgesia, 10 mg of sodium diclofenac (equivalent to 1 g of commercial Voltaren Gel^®; Novartis, Basel, Switzerland) was applied topically twice daily for 3 to 5 days to the ipsilateral paw of control and monoarthritic rats until analgesia was achieved, beginning 21 days after CFA injection. Pure Vaseline (Acofarmaderm, Acofarma S.A., Barcelona, Spain) was applied to the control and monoarthritic rats and served as a control of diclofenac application. A plastic Elizabeth collar was fixed around the neck of each animal to prevent ingestion of the cream/vaseline.²⁴  The effect of diclofenac on the signs of inflammation and nociceptive behavior was analyzed after the last topical application (see Inflammation Assessment and Nociceptive Behavioral Assessment sections). As indicators of paw inflammation, photographs and footprints of the hind paw were taken to evaluate the shape and area of paw. At the end of the experiment, Western blot procedures were performed to evaluate the effect of diclofenac in the pattern of ERK_1/2 activation in the SC, PGi, LC, and PFC. Next, topical application of diclofenac/vaseline was repeated in another set of animals to evaluate the effect of nonsteroidal anti-inflammatory drug analgesia on pain-induced affective changes following the protocols described in the Anxiety- and Depression-like Behavior section.

To study the possible site of action of diclofenac, an additional group of animals was organized to receive local topical treatment of diclofenac in the contralateral paw (right paw) following the same protocol as described in the first paragraph of this section. Thus, the following extra experimental groups consisted of control and monoarthritic animals receiving vaseline (Cont + VAS and MA + VAS) and a group of monoarthritic animals receiving diclofenac treatment (MA + DIC). Before and after vaseline/diclofenac administration, baseline measures were taken for the assessment of the paw volume, ankle-bend score, and paw-pinch test values. Afterwards, Western blot procedures were performed to evaluate the effect of contralateral paw administration in the pattern of ERK_1/2 activation in the SC, PGi, LC, and PFC.

All data are presented as the means ± SEM and all the results were analyzed using STATISTICA 10.0 (StatSoft, Tulsa, OK) or GraphPad Prism 5 software (GraphPad Software, Inc., La Jolla, CA) using either a Student t test (unpaired or paired, two-tailed) or a one-way or two-way ANOVA with or without repeated measures, followed by the appropriate post hoc tests (Student–Newman–Keuls or Dunnett tests). The independent variables were monoarthritis (between-groups) and treatment (between-groups). The level of significance was set at a P value of less than 0.05. Detailed information regarding statistical results is shown in tables 1-4.

The injection of the vehicle alone (control group) did not provoke any inflammatory reaction in rats and these animals exhibited a normal behavior. By contrast, stable monoarthritis was induced by CFA injection, with marked inflammatory signs within several hours of induction that persisted for 4 weeks. Monoarthritic rats displayed: (1) normal body weight gain; (2) normal body temperature, except in the second week when a significant increase was observed (P < 0.001; monoarthritis vs. control for week 2 by ANOVA followed by Student–Newman–Keuls post hoc test; fig. 1A); (3) a significant increase in the ipsilateral paw volume evident each week (P < 0.001 for each time point; monoarthritis vs. control by ANOVA followed by Student–Newman–Keuls post hoc test; fig. 1B); and (4) a higher inflammation score in the ipsilateral paw that persisted until the fourth week (fig. 1B). Significantly, none of these features were observed in the contralateral paw of these monoarthritic rats. When the movement-induced nociceptive behavior was evaluated, CFA injection provoked higher ankle-bend scores (allodynia) throughout the experimental period (fig. 1C) and monoarthritic rats exhibited significantly stronger mechanical hyperalgesia in the ipsilateral paw (P < 0.001 for weeks 1 and 4, P < 0.01 for weeks 2 and 3; monoarthritis vs. control by ANOVA followed by Student–Newman–Keuls post hoc test; fig. 1C). No painful reactions were observed in the control rats or in the contralateral paw of monoarthritic rats. Descriptive statistics is shown in table 1.

To determine whether chronic pain associated with joint inflammation induced anxiogenic-like behavior, the rats were subjected to the marble-burying and EZM tests at several time points during the development of monoarthritis (fig. 2A). Unlike MA4D rats, more marbles were buried by MA14D (P < 0.05 by one-way ANOVA followed by Dunnett post hoc test) and MA28D (P < 0.01 by one-way ANOVA followed by Dunnett post hoc test) rats compared with the marbles buried by their corresponding controls, indicative of anxiety-like behavior.

In the EZM maze test, there was no difference in the percentage of time spent in the open arms between control rats and MA4D or MA14D rats. However, MA28D rats spent significantly lesser in the open arms than by their corresponding controls (P < 0.001 by one-way ANOVA followed by Dunnett post hoc test), again indicative of the development of anxiety-like behavior. No changes in locomotor activity were detected between groups in the EZM, ruling out a motor component in the effects observed.

The FST was performed to evaluate the possible development of depressive-like behaviors (fig. 2B), in which MA28D but not MA4D or MA14D rats exhibited significantly higher immobility scores compared with the scores of their corresponding controls (P < 0.05 by one-way ANOVA followed by Dunnett post hoc test). Moreover, this effect was accompanied by a significant decrease in climbing behavior (P < 0.05 by one-way ANOVA followed by Dunnett post hoc test), but there were no significant changes in swimming behavior. As expected, desipramine treatment diminished the rat’s immobility (P < 0.01 vs. corresponding control, Student t test) and significantly increased their climbing behavior (P < 0.01 vs. corresponding control, Student t test). Descriptive statistics is shown in table 2.

The expression of pERK_1/2 was evaluated as an indicator of neuronal activity in the SC, PGi, LC, and PFC in the early, middle, and late phases of monoarthritis development (fig. 3). In the ipsilateral side of the SC, significant increases in the expression of pERK_1/2 were observed at 14 (P < 0.05 by one-way ANOVA followed by Dunnett post hoc test) and 28 (P < 0.001 by one-way ANOVA followed by Dunnett post hoc test) days after monoarthritis induction when comparing with that in control groups. No significant changes were observed at 4 days of monoarthritis disease. In addition, no significant changes in pERK_1/2 were observed at 4 or 14 days after monoarthritis induction in any of the brain areas analyzed. By contrast, significantly more pERK_1/2 was observed in the PGi (P < 0.05 by one-way ANOVA followed by Dunnett post hoc test), LC (P < 0.01 by one-way ANOVA followed by Dunnett post hoc test), and PFC (P < 0.05 by one-way ANOVA followed by Dunnett post hoc test) at 28 days after monoarthritis induction compared with that in the corresponding controls. Descriptive statistics is shown in table 2.

As expected, topical application of diclofenac to the ipsilateral paw decreased the inflammation and increased the pain threshold of that paw, without producing any significant change in body weight or body temperature (fig. 4). Although a small edema was observed in monoarthritic rats that received diclofenac (P < 0.01, control + diclofenac vs. monoarthritis + diclofenac by two-way ANOVA followed by Student–Newman–Keuls post hoc test), this was substantially smaller than that seen in those that received vaseline (P < 0.001, monoarthritis + vaseline vs. monoarthritis + diclofenac by two-way ANOVA followed by Student–Newman–Keuls post hoc test: fig. 4B). A similar effect on the inflammation score was observed (fig. 4B) and the ankle-bend score of monoarthritic rats that received a topical application of diclofenac was significantly lower (P < 0.001, monoarthritis + vaseline vs. monoarthritis + diclofenac by two-way ANOVA followed by Student–Newman–Keuls post hoc test), reflecting a higher nociceptive threshold, although this score remained significantly higher than that observed in control animals (P < 0.001 for the inflammation score and P < 0.01 for the ankle-bend score, control + diclofenac vs. monoarthritis + diclofenac by two-way ANOVA followed by Student–Newman–Keuls post hoc test: fig. 4C). Diclofenac treatment also increased the force supported by the ipsilateral paw of monoarthritic rats (P < 0.001, monoarthritis + vaseline vs. monoarthritis + diclofenac by two-way ANOVA followed by Student–Newman–Keuls post hoc test), completely abolishing mechanical hyperalgesia (fig. 4C). Although not quantified, photographs and footprints revealed a clear improvement in paw posture and weight loading in the affected paw after diclofenac treatment (fig. 5). Descriptive statistics is shown in table 3.

Diclofenac administration had significant effects on monoarthritis-induced anxiety- and depression-like behaviors. In the EZM, monoarthritis-induced anxiety-like behavior was reversed by diclofenac treatment (P < 0.05, monoarthritis + vaseline vs. monoarthritis + diclofenac by two-way ANOVA followed by Student–Newman–Keuls post hoc test; fig. 5A), yet no changes in locomotor activity were produced in this paradigm (fig. 5A). Similarly, monoarthritis-induced depression-like behavior was successfully reversed by topical diclofenac administration, as revealed by the significant differences in immobility time between groups (P < 0.05, monoarthritis + vaseline vs. monoarthritis + diclofenac by two-way ANOVA followed by Student–Newman–Keuls post hoc test; fig. 5B). The climbing behavior displayed by the monoarthritic rats receiving vaseline showed a tendency to decrease compared with the respective control group (P = 0.056 control + vaseline vs. monoarthritis + vaseline, by two-way ANOVA followed by Student–Newman–Keuls post hoc test; fig. 5B). Descriptive statistics is shown in table 3.

Local inflammation produced a significant increase in the expression of pERK_1/2 levels in the ipsilateral SC as witnessed in monoarthritic rats treated with vaseline as compared with that in the corresponding controls (P < 0.001, monoarthritis + vaseline vs. control + vaseline by two-way ANOVA followed by Student–Newman–Keuls post hoc test). Analgesia through application of diclofenac to the inflamed paw significantly reduced pERK_1/2 levels (P < 0.001, monoarthritis + vaseline vs. monoarthritis + diclofenac by two-way ANOVA followed by Student–Newman–Keuls post hoc test; fig. 6, A and B) and the significant increase in pERK_1/2 levels in the PGi–LC–PFC pathway of monoarthritic rats (P < 0.05, monoarthritis + vaseline vs. control + vaseline in the PGi; P < 0.01 monoarthritis + vaseline vs. control + vaseline in the LC and PFC by two-way ANOVA followed by Student–Newman–Keuls post hoc test; fig. 6, A and B) was also successfully restored to control levels after topical treatment with diclofenac (P < 0.05, monoarthritis + vaseline vs. monoarthritis + diclofenac in the PGi and PFC; P < 0.01, monoarthritis + vaseline vs. monoarthritis + diclofenac in the LC by two-way ANOVA followed by Student–Newman–Keuls post hoc test; fig. 6, A and B). Descriptive statistics is shown in table 3.

The results point to the absence of an effect of the contralateral administration of diclofenac (fig. 7) on the paw volume (fig. 7A), and paw-pinch threshold (fig. 7C) of the inflamed paw (ipsilateral paw), as can be observed by the lack of significant changes between the values obtained before the treatment and after the treatment, in the group of monoarthritic animals receiving diclofenac (paired Student t test). However, in the ankle-bend score (fig. 7B), contralateral treatment with diclofenac caused a decrease in the values obtained after the treatment when comparing with those obtained before the treatment (P < 0.01; paired Student t test; fig. 7B). Descriptive statistics is shown in table 4.

In addition, quantification of the immunoblots (fig. 8) showed that the increased ERK_1/2 activation observed in MA + VAS group was maintained in the MA + DIC group in the SC (P < 0.01 for MA + VAS and MA + DIC vs. control, unpaired Student t test), PGi (P < 0.01 for MA + VAS vs. control and P < 0.05 for MA + DIC vs. control unpaired Student t test), LC (P < 0.05 for MA + VAS vs. control and P < 0.01 for MA + DIC vs. control, unpaired Student t test), and PFC (P < 0.05 for MA + VAS and MA + DIC vs. control, unpaired Student t test). Descriptive statistics is shown in table 4.

The current findings demonstrate that chronic joint inflammation interferes with sensitivity to noxious stimulation and with physiological movement of the inflamed joint, as well as profoundly affecting emotional states when the painful condition is prolonged. Furthermore, we demonstrated that chronic pain enhances ERK_1/2 activation in certain central nervous system regions, specifically in the SC, PGi, LC, and PFC. Promoting analgesia through the topical application of anti-inflammatory cream successfully attenuated nociception, as well the anxiety- and depressive-like behaviors observed. Interestingly, this effect was accompanied by the normalization ERK_1/2 activation in the SC, PGi, LC, and PFC.

We initially analyzed here the nociceptive behavior produced by joint inflammatory pain with the use of the monoarthritis model. As expected, unilateral arthritic inflammation induced constant allodynia and hyperalgesia in the ipsilateral paw at all time points studied. The induction of monoarthritis was associated with the development of anxiety-like symptoms in the EZM and marble-burying tests. The EZM is based on the natural aversion of rodents to bright and exposed places, whereas marble-burying behavior gauges the level of anxiety-like behavior of a rodent on encountering unfamiliar and bright objects. In both tests, anxiety-like behavior was evident at late disease stages, and in the case of the marble-burying test, the anxiety-like behavior was already evident at 14 days after monoarthritis induction. We also evaluated the time course of depression-like behavior by using the FST. This test induces “behavioral despair” in animals and abandoning the struggle to escape a stressful environment may resemble the psychological concept of “entrapment” described in clinical populations.²³  MA28D rats displayed greater immobility and milder escape behavior (climbing) in the FST test. This reduction in climbing behavior was not accompanied by changes in swimming behavior, suggesting noradrenergic dysfunction. This is mainly due to the fact that noradrenaline-selective reuptake inhibitors produce significant increases in this behavior in the FST²³  indicating that a decrease in the climbing behavior suggests a dysfunction in the noradrenergic system. More studies are underway to clarify whether this impairment is really occurring. These data are in consistent with previous findings,^25–27  and with clinical data obtained from patients with osteoarthritis,²⁸  indicating that chronic inflammatory pain conditions profoundly affect emotional states, as also described previously in other rheumatic diseases.²⁹  Moreover, the time course for development of anxiety and then depression may indicate a certain chronological development of pain-related mood disorders, which was already observed in other pain models.³⁰  Overall, this implies that the affective consequences of chronic inflammatory pain evolve over time, probably mediated by long-term molecular and neural plastic changes at different brain areas.

Because ERK_1/2 have been proposed as promising target molecules in the regulation of both pain^31,32  and affective disorders,^33–35  we investigated whether monoarthritis-induced affective disorders were associated with altered ERK_1/2 activation. In the SC, pERK_1/2 was not significantly increased at 4 days after monoarthritis induction but was significantly increased at both 14 and 28 days after monoarthritis. Indeed, other authors showed that 2 and 4 days of inflammation were not accompanied by increased ERK_1/2 activation in the SC.³⁶  The significant increase of pERK_1/2 observed at 14 and 28 days may be related with increased metabolic activity which was already demonstrated by Schadrack et al.,³⁷  at least for the 14 days of monoarthritic time point. The onset of affective symptoms was accompanied by increased pERK_1/2 levels in the PGi, LC, and PFC of monoarthritic rats, in agreement with previous reports of increased pERK_1/2 expression in the rat PGi 7 days after the induction of neuropathic pain.^38,39  We observed increased ERK_1/2 activation in the LC after chronic monoarthritis, possibly due to increased excitatory input from the PGi. Interestingly, CFA injection has been shown to induce a sharp increase in LC pERK_1/2 after 5 min that disappears 7 h later,⁴⁰  and indeed, we observed no changes in pERK_1/2 in the early stages after CFA administration. By contrast, pERK_1/2 expression was consistently increased 28 days postadministration, coinciding with altered nociceptive behavior and the onset of anxiety and, especially, depressive states. These findings suggest that the LC is involved in both acute pain and the subsequent development of pain-induced affective disorders. Finally, ERK_1/2 activation was also enhanced in the PFC region 28 days after CFA administration. The PFC is one of the most important projection targets of LC noradrenergic terminals, and increased pERK_1/2 expression in this area has been correlated with anxiety- and depressive-like behaviors.^9,10,13,35  Another interesting observation was that the pattern of ERK_1/2 activation at supraspinal level was bilateral, which means that unilateral inflammation of the ankle produced both ipsilateral and contralateral phosphorylation of ERK_1/2 in the PGi–LC–PFC pathway. This was already shown in other pain studies,^38,39  and particularly, such lack of lateralization of ERK_1/2 phosphorylation was common in studies regarding pain-related anxiety paradigms.^9,10,13  Taken together, these findings suggest that alterations in the levels of pERK_1/2 may be one of the molecular mechanisms that underlie the onset of chronic pain–induced affective disorders.

One of the most remarkable findings of the current study was that the blockade of nociceptive inputs affected monoarthritis-induced anxiety- and depression-like behavior. As indicated elsewhere, administration of sodium diclofenac ointment attenuated inflammation and pain,^41,42  as well as reducing pERK_1/2 levels in the lumbar SC. Significantly, the induction of analgesia was accompanied by the disappearance of anxiety- and depression-like behaviors, suggesting that sensorial pain inputs are the source of affective alterations in chronic monoarthritis. Importantly, the increased ERK_1/2 activation in the PGi–LC–PFC pathway on monoarthritis induction was successfully restored to control levels by diclofenac treatment, indicating that the increased ERK_1/2 activity in these regions is a consequence of the nociceptive inputs. Based on its increased expression only when anxiety- and depression-like behaviors are present, we propose that sustained altered pERK1/2 expression in the PGi–LC–PFC pathway in chronic monoarthritis is probably more closely related to the development of pain-related affective disorders than with nociception itself. However, further studies in other areas widely involved in facilitating or inhibiting nociception, such as the periaqueductal gray, rostral ventromedial medulla, and dorsal reticular nucleus, will be necessary to confirm this hypothesis.

To study the possible site of action of diclofenac, we evaluated the effect of administering vaseline/diclofenac in the contralateral paw. Behavioral studies showed that diclofenac administered into the contralateral paw modify neither the paw volume nor the hyperalgesia level, but a slight reduction was observed in the ankle-bend score. Nevertheless, when studying the effect of contralateral application of diclofenac on the expression of pERK_1/2 in the PGi, LC, and PFC, no significant changes were observed with respect to its control group of monoarthritic rats receiving vaseline on the contralateral paw. Hence, these data suggest that topical application has a low systemic effect that would explain the small effect in the ankle-bend test. Indeed, previous data have shown that local administration (cream, gel, or dermal patch) produces a very low systemic effect when compared with that in oral administration.^43–45  However and of relevance for our study, pERK_1/2 expression is not modified in any of the brain areas studied; so, it seems unlikely that a central effect produced may be involved in all the pain-related features restored by diclofenac treatment. Overall data suggest that diclofenac might be peripherally acting and that the blockade of the nociceptive inputs originated in the inflamed paw is able to restore pERK_1/2 expression levels in the brain areas studied. However, it is important to note that further studies will be necessary to discern the peripheral and/or central effect of diclofenac in the reversal of all the monoarthritis-related features.

Although the translation of findings from animal models to humans must always be approached with caution, we believe that our findings have interesting parallels in the clinical setting. Thus, we propose that achieving effective pain relief reverses the molecular changes induced by chronic pain. This hypothesis is consistent with the data from patients with osteoarthritis in whom successful arthroplasty reverses thalamic atrophy.⁴⁶  Furthermore, the current findings suggest that effective analgesia also benefits other symptoms (anxio-depressive behaviors) that, although not directly related to sensorial pain, have been identified as major contributors to a worse patient outcome. Finally, it is important to note that neuroelasticity (i.e., reversal of an effect on removal of the stimulus) is observed at the onset of anxio-depressive symptoms. It is possible that a critical window exists after which inducing such reversal will be more difficult, or no longer feasible, due to additional changes in neuronal architecture. These findings have important implications for the ongoing debate regarding the optimal therapeutic approaches to treating patients with arthritis (pharmacological vs. surgical strategies).

The delayed onset of monoarthritis-induced affective pathologies suggests that peripheral pain inputs induce some reorganization in the central nervous system. Our study demonstrates that affective behavioral changes are accompanied by ERK_1/2 activation in the PGi–LC–PFC pathway. Moreover, these findings indicate that successful analgesia can reverse sensorial and affective pain-induced changes.

The authors thank Raquel Rey-Brea, A.S. (Department of Neuroscience, University of Cádiz, Cádiz, Spain), José Antonio García Partida, B.S. (Department of Neuroscience, University of Cádiz, Cádiz, Spain), Jesus Gallego-Gamo, A.S. (Department of Neuroscience, University of Cádiz, Cádiz, Spain), and Elisa Galvão, A.S. (Departamento de Biologia Experimental, Universidade do Porto, Porto, Portugal), for assistance throughout this study. The authors also thank Joaquin Berrocoso Domínguez, Diploma in Computer Science (Consejería de Economía, Innovación, Ciencia y Empleo de la Junta de Andalucía, Seville, Spain), for the Red Mice software.

The main funding sources of this study are: “Cátedra Externa del Dolor Fundación Grünenthal-Universidad de Cádiz,” Cádiz, Spain (to Borges); “Cátedra em Medicina da Dor Fundação Grünenthal-Faculdade de Medicina da Universidade do Porto,” Porto, Portugal; “Fondo de Investigación Sanitaria,” Madrid, Spain (PI13/02659, PI12/00915); “Centro de Investigación Biomédica en Red de Salud Mental,” Madrid, Spain (G18); and “Consejería de Economía, Innovación, Ciencia y Empleo de la Junta de Andalucía,” Seville, Spain (CTS-510, CTS-4303 and CTS-7748).

The authors declare no competing interests.