In musculoskeletal diseases such as arthritis pain often does not reflect the course and severity of disease. In particular, the persistence of pain or hyperalgesia after spontaneous or therapeutically achieved remission of disease may be a burden.25,41 Why pain is persistent in some patients but transitory in others is difficult to explore solely in patients. Experimental arthritis models are usually believed to be homogeneous and to allow for an analysis of the relationship between disease and pain. It is unclear, however, to which extent pain-related behaviors in experimental models may show either transitory or persistent courses, and whether related mechanisms may be investigated in such models. Here, we addressed this question using the model of antigen-induced arthritis (AIA), an immune-mediated monoarticular joint inflammation.5,16
In AIA, rats or mice are immunized against the antigen methylated bovine serum albumin (mBSA), and the injection of mBSA into one knee joint after immunization causes a very reproducible monoarthritis. Within one day, the joint develops pronounced swelling and polymorphonuclear infiltration, but within 2 weeks, acute inflammation decreases leaving the joints with some synovial hyperplasia and minor cartilage and bone destruction. On average, both inflammation and pain-related behaviours disappear within 21 days.3,13,36,37 However, we recently noticed that, in some rats, mechanical hyperalgesia at the knee came back without obvious inflammation in the knee, whereas other rats did not show such new hyperalgesia.26,37 This is reminiscent of the consistently persisting mechanical hyperalgesia after transient inflammatory joint swelling in systemic polyarthritis models such as K/BxN serum transfer arthritis,7 collagen antibody-induced arthritis,2,39 and glucose 6-phosphate-isomerase (G6PI)-induced arthritis.11 We now systematically monitored pain-related behaviors in AIA up to 84 days and asked whether AIA may evoke different courses of pain, and if so, whether persistence of pain-related behaviors may be predicted from behavioral parameters.
Furthermore, we studied whether some pain-related markers are present at 84 days. We explored the invasion of macrophages into the dorsal root ganglia (DRG), a process seen after nerve injury,17 and also in arthritic conditions,19,30,36 the expression of ATF3 in DRG neurons as an indicator for neuropathic changes,18 the expression of the transcription factor pCREB in DRG neurons (see below), and we quantified the number of IBa1-positive cells in spinal cord sections as an indicator of microglial activation.9
We reported previously an upregulation of pCREB in lumbar DRG neurons up to 42 days in rat AIA, thus outlasting the inflammatory process.37 pCREB regulates genes involved in neuroplasticity such as learning28 and has been upregulated in DRG neurons upon inflammation and electrical C-fiber stimulation.33,40,43 Upon treatment of established AIA with Anakinra, an interleukin-1 receptor (IL-1R) antagonist, pCREB upregulation was not observed.37 Since immunization significantly increased the expression of IL-1R1 before and during AIA13 and since IL-1ß induces pCREB in DRG neurons in vitro,37 we tested whether the prevention of pCREB upregulation by Anakinra treatment in the first 21 days of AIA influences the development of AIA and pain.
This study was performed in strict accordance with EC regulations for the care and use of laboratory animals. The protocol was approved by the Thuringian Government (Thüringer Landesamt für Verbraucherschutz) and performed according to the Protection of Animals Act of the Federal Republic of Germany. All injections for immunization, subcutaneous implantation of osmotic minipumps, and induction of monoarticular AIA were performed under short anesthesia with 2.5% isoflurane. Data sampling, evaluation, and presentation complied with the ARRIVE guidelines.
2.1. Antigen-induced arthritis
We used 103 female Lewis rats (age 6-8 weeks, weighing 170-200 g, Janvier, France) for the studies. Lewis rats are particularly susceptible to AIA42 and also particularly suitable for behavioral experiments because they are quite tame. Animals were housed in groups of 4 to 5 animals per standard cage in a climate-controlled room on a 12:12-hour light:dark cycle with water and standard rodent chow available ad libitum.
For immunization, 500 µg of antigen (methylated bovine serum albumin [mBSA], Sigma, Deisenhofen, Germany) in saline emulsified with 500 µL of Freund's complete adjuvant (Sigma) supplemented with 2-mg/mL Mycobacterium tuberculosis strain H37RA (Difco, Detroit, MI) was injected subcutaneously twice with a 1-week interval between immunizations (Fig. 1). For the first immunization step, the solution was injected bilaterally at 4 sites of the back, for the second at 2 different sites of the back and at one site at the neck. After another 2 weeks, a sterile mBSA solution (500 µg in 50 µL) was injected into the left knee joint cavity to induce monoarticular AIA. The experimental design of the whole experiment is shown in Figure 1.
The mediolateral diameter of each knee was assessed using a vernier calliper (Mitutoyo, Neuss, Germany) on every testing day for each animal. The relative swelling was calculated by subtracting the diameter of the noninflamed knee from that of the inflamed one. All efforts were made to minimize suffering of the animals.
2.2. Experimental groups
Animals were divided into different experimental groups (Fig. 1), and the experiments were performed over a period of 12 months. Rats were purchased and delivered at different time points, and they were from different charges. The observation periods after induction of AIA were 3, 42, or 84 days. The different experiments were not made in series, but from each delivery of rats, the animals were randomly assigned to the different experimental groups, to avoid any bias. Nineteen rats received no treatment (naïve control [NC] group), and 15 animals were immunized, but AIA was not induced (immunized control [IC] group). In 67 rats, AIA was induced; 37 animals received no further treatment (AIA group); in 30 animals, osmotic minipumps (Model Alzet 2ML4) with Anakinra (53 mg/kg/d) (n = 15) or saline (n = 15) were implanted 5 days before AIA induction (Anakinra and saline group). Because of the length of the catheter attached to the pump, the compounds were delivered from day 0. In some rats, Anakinra was applied for 3 days (then the experiment was stopped); in the other rats, the application of Anakinra lasted 21 days (at this time point, the maximally possible amount of Anakinra had been administered); and the rats were observed until day 84 of AIA.
2.3. Pain-related behaviors in the course of antigen-induced arthritis
All pain-related behaviors were tested by the same person (J.L.). As prearthritic baseline values before AIA, we used the values obtained in the last week before injection of mBSA into the knee joint (day-7). Gait abnormalities of the ipsilateral hind limb were scored as follows: 0: no guarding; (1) transient limping after a brief compression of the inflamed knee; (2) persistent, visible spontaneous limping; (3) no use of the ipsilateral hind limb; and (4) no walking.
Weight bearing as a functional measure of pain-related guarding behavior of the inflamed hind limb was assessed with an incapacitance tester (Linton Instrumentation, Norfolk, United Kingdom). Animals were placed with both hindpaws on scales inside a plastic cage. The weight force resting on the 2 scales was consecutively measured for 3 times after animals acclimatized and sat calm in the cage. From these averaged values, the relative weight (in %) resting on the inflamed hind limb was evaluated (weight on inflamed hind limb × 100%/[weight on inflamed hind limb + weight on noninflamed hind limb]) as described previously.27
Mechanical hyperalgesia at the inflamed knee joint (primary mechanical hyperalgesia) and mechanical hyperalgesia at the contralateral knee joint were assessed with a PAM device (digital Pressure Application Measurement device; Ugo Basile, Gemonio, Italy). Using a circular probe of 8-mm diameter, increasing pressure (at 50 g/second) was simultaneously applied in the mediolateral axis to the knee joint at the level of the joint space until the animals attempted to escape or vocalized.27 The applied weight force was read out in grams. Testing was performed once for each animal on each testing day to avoid nociceptive sensitization due to repeated testing (Fig. 1).
Mechanical hyperalgesia at the plantar aspect of the hindpaws (secondary mechanical hyperalgesia) was assessed with the Dynamic Plantar Aesthesiometer (Ugo Basile). A blunt tip was positioned below the plantar surface, and the actually applied force (at 1 g/s) at the time of the paw withdrawal reflex was recorded.
Secondary thermal hyperalgesia at the plantar aspect of the hindpaws was assessed using the Plantar Test (Hargreaves Apparatus) (Ugo Basile). An infrared source was placed below the plantar surface, and latency to heat-evoked paw withdrawal was measured. Both the latter testings were repeated twice for each side in an altered sequence, and values were averaged.
2.4. Preparation of tissues
To identify changes in different tissues up to day 84 after induction of monoarticular arthritis, animals were deeply anesthetized with 120-mg/kg sodium thiopentone injected intraperitoneally (Trapanal; Byk Gulden, Konstanz, Germany) and transcardially perfused with 4% ice-cold phosphate-buffered paraformaldehyde (PFA) (Fig. 1). For knee joint histopathology, whole knees were directly removed and postfixed for 24 hours in 4% PFA in PBS at 4°C. In addition whole spinal columns were directly removed after perfusion and postfixed for 24 hours in 4% PFA. Then, the spinal cords from the lumbar levels were prepared, and the DRG from segments L1–L5 from the inflamed sides were excised. Both tissues were fixed in 2% PFA in PBS for at least 24 hours, embedded in paraffin (Histosec; Merck, Darmstadt, Germany), cut into 5-µm sections, and mounted on slides (Super FrostPlus; Menzel-Gläser).
2.5. Histopathological assessment of arthritis
After washing with phosphate-buffered saline, knee joint samples were dehydrated in increasing alcohol concentrations and embedded in Technovit 9100 methyl methacrylate (Heraeus Kulzer, Wehrheim, Germany). Thin sections of 4 µm were cut from Technovit 9100 blocks using the microtome Polycut S (Reichert-Jung, Heidelberg, Germany). Before staining, polymer was removed with 2-methoxyethylacetate, and sections were rehydrated in decreasing alcohol concentrations.
Synovial tissue sections were stained with Mayer's hemalum solution and Eosin G (Merck). Synovitis was scored using the histopathological scoring system of Krenn et al. (2005).14,24 It quantifies in points the enlargement of the lining cell layer (0: only 1 cell layer; 1: 2-3 cell layers; 2: 4-5 cell layers; and 3: more than 5 cell layers), the cellular density of synovial stroma and pannus formation (0: normal cell density; 1: cell density slightly enhanced; 2: cell density moderately enhanced; and 3: high cell density, multinuclear giant cells, and pannus tissue), and leukocytic infiltrate (0: no infiltration; 1: single lymphocytes or plasma cells; 2: aggregates of lymphocytes; 3: dense infiltration with lymphocytes or lymph follicles). The total score ranges from 0 to 9 (0-1: no synovitis; 2-4: low-grade synovitis; and 5-9: high-grade synovitis). Analysis was performed with the microscope Axioplan2 (Zeiss, Oberkochen, Germany) using a ×20 dry objective.
2.6. Immunohistological labeling for pCREB, ATF3, and CD68/ED1 in dorsal root ganglion sections
The embedded DRG sections were dewaxed and autoclaved for 15 minutes (at 120°C, 1 bar) in 0.1-mol/L citrate buffer (pH 6.0). Chilled and PBS-washed sections were incubated for 30 minutes in PBS containing Triton X-100 and 2% goat serum or 2% rabbit serum (DAKO, Glostrup, Denmark). Thereafter, the sections were incubated overnight at 4°C with the primary antibodies (pCREB; Cell Signaling Technology, Danvers, MA, #9198, 1:100; ATF-3, Santa Cruz Biotechnology, Santa Cruz, CA, #H-90, 1:200; ED1 [Mouse anti-rat CD68, clone ED1], #MCA341R, BioRad, Hercules, CA, 1:100) diluted in PBS plus 1% Triton X-100 containing 1% gelatin from cold water fish skin in a moist chamber. After incubation sections were washed 3 times with PBS. As secondary antibodies, we used a biotinylated goat anti-rabbit antibody (1:200; DAKO) for the labeling of ATF3 or pCREB and a biotinylated rabbit anti-mouse antibody (1:200; DAKO) for detection of ED1. After 3 washes with PBS, avidin–biotin–peroxidase complex (Vectastain-Elite ABC Kit; Vector Laboratories, Inc., Burlingame, CA) was applied for 40 minutes. Sections from different experimental groups were developed simultaneously with Jen-chrom px blue (MoBiTec, Göttingen, Germany). The sections were dehydrated and embedded in Entellan (Merck). In control experiments, primary antibody was omitted.
All sections were analyzed using a light microscope (Axioplan 2; Zeiss) coupled to an image analyzing system (Axiovision; Zeiss). In every second section, the proportion of neuronal profiles with ATF3 or pCREB labeling was counted. For each experimental group, the proportion of neurons with positive labeling of the nucleus from different rats were averaged and expressed as mean ± SD.
It was not possible to reliably count the numbers of ED1-positive cells. Therefore, we measured the gray density of sample fields because staining of ED1 provided a dark signal (see also Ref. 36). Dorsal root ganglia of the inflamed side were measured with a minimum area of at least 0.03 mm2 per animal. Naïve-control animals and primary antibody omission were used as controls. For each group, the gray value from different rats were averaged and expressed as mean ± SD.
2.7. Immunohistochemical labeling of IBa1 in spinal cord sections
The spinal cord sections were dewaxed in xylol, hydrated, and autoclaved for 15 minutes (120°C, 1 bar) in 10-mmol/L citrate buffer (pH 6.0). After cooling down, the sections were washed in TBS (20 mmol/L Tris base, 137 mmol/L NaCl, pH 7.4) 3 times for 10 minutes and incubated 30 minutes in TBS plus 1% Triton-X-100 plus 10% normal goat serum (Rockland, Gilbertsville, PA). After a short wash with TBS, the anti-IBa1 antibody (#019-19741; Wako, Neuss, Germany) was diluted 1:100 in TBS containing 1% Triton X-100 and 1% normal goat serum and incubated over night at 4°C in a moist chamber. After 3 times washing for 10 minutes, an Alexa-488 goat anti-rabbit antibody (#A11008, Thermo Fisher Scientific, Waltham, MA) was administrated to the sections for 3 hours at 20°C. The secondary antibody was diluted 1:100 in TBS containing 1% Triton X-100 and 1% normal goat serum. After 3× washing with TBS, the sections were incubated for 10 minutes with Hoechst 34580 (diluted 1:500 in TBS containing 1% Triton X-100 and 1% normal goat serum). After washing again with TBS, the sections were embedded in Aqua Poly Mount (Polysciences, Hirschberg, Germany). Control experiments for the antibodies were performed with omission of the primary antibodies.
From 3 sections of every animal, photomicrographs of the ipsilateral and contralateral dorsal horn were made using a light microscope (Axioplan 2, Zeiss, Jena, Germany) coupled to an image analyzing system (illumination time 500 ms, Axiovision, Zeiss). In each photomicrograph (area 320 × 240 µm), the number of visible IBa1-positive cells was counted. The proportions of counted cells from different rats and sides were averaged and expressed as mean ± SD.
2.8. Statistical analysis
SPSS software (SPSS, Inc, IBM Company, New York, NY) and GraphPad PRISM 5 (GraphPad Software, Inc) were used for statistical analysis. Data were first tested for normal distribution using the Kolmogorov–Smirnov test.
Hierarchical cluster analysis was performed to detect intragroup differences of the rats at the end of the observation period (day 84). Ward's minimum variance method was used for analysis, and the squared Euclidean distance was selected. All behavioral parameters (hot beam, plantar test, PAM, incapacity test, and difference of knee diameter) of day 84 were selected for analysis to identify clusters in the untreated AIA group. We then traced back the values of the clusters defined at day 84. For comparison of the behavioral data between groups, repeated-measures one-way analysis of variance (ANOVA) was applied with the between-subject factor GROUP (NC, IC, AIA, Anakinra, saline intraperitoneally) and the within-subject factor TIME (days 1, 3, 7, 14, 21, 28, 35, 42, 49, 56, 63, 70, 77, and 84 after induction of AIA).
Correlation analysis was performed between the baseline values of each behavioral parameter and the corresponding delta values on day 84 (d84 value minus baseline value) using Pearson correlation as well as between all assessed parameters on day 84 between groups and clusters using Spearman's rank correlation coefficient. P values <0.05 were considered significant.
Statistical analysis of cell counts for pCREB-, ATF3-, and ED1-positive cells in the DRG was performed using ANOVA followed by post hoc t-tests or Mann–Whitney U-tests. Significance was accepted at P < 0.05.
3.1. Behavioral changes during and after antigen-induced arthritis
The injection of mBSA into the left knee joint in immunized rats caused a typical AIA, similar as in previous studies.3,21,26,29,36,37 The pattern of swelling and of pain-related behaviors is shown in Figure 2. In each panel, the values of the AIA rats (n = 20) are compared with those of NC rats (n = 8), and both groups were tested for 84 days. Antigen-induced arthritis rats developed significant swelling of the injected knee joint [F(1,26) = 35.796; P < 0.001], which disappeared after about 28 days (Fig. 2A). Swelling was accompanied by significant guarding of the left leg [Fig. 2B, F(1,26) = 25.006; P < 0.001] (Fig. 2B). The whole group of AIA rats and NC rats were significantly different in their responses to pressure at the knee (injected and inflamed in AIA rats) [Fig. 2C, F(1,26) = 8.102; P = 0.009], but not in their responses to pressure onto the contralateral knee [Fig. 2D, F(1,26) = 1.457; P = 0.238]. Antigen-induced arthritis and NC rats showed also significant differences of their responses to heat stimulation of the ipsilateral paw [Fig. 2E, F(1,26) = 11.16; P = 0.003], but not to heat stimulation of the contralateral paw [Fig. 2F, F(1,26) = 2.484; P = 0.127]. Antigen-induced arthritis and and NC rats exhibited significant differences in the mechanical withdrawal threshold at the ipsilateral hindpaw [Fig. 2G, F(1,26) = 6.964; P = 0.014], but not to mechanical stimulation of the contralateral hindpaw [Fig. 2H, F(1,26) = 1.474; P = 0.236]. Antigen-induced arthritis and NC rats differed in the symmetry of weight bearing [Fig. 2I, F(1,26) = 17.554; P < 0.001]. Although some parameters went back to the prearthritic values, others remained altered. In addition, for some parameters, the variability was quite substantial. In the next step, we asked, therefore, whether the whole population of rats may be composed of subgroups, which showed a different behavior in the course of AIA.
All behavioral parameters and the difference of knee diameter of day 84 were selected for analysis to identify clusters in the AIA group. This cluster analysis at day 84 revealed 2 major clusters of rats called cluster 1 and cluster 2 (see dendrogram using Ward linkage in supplementary Figure 1, available at http://links.lww.com/PAIN/A976). We then put together the rats of clusters 1 and 2 and traced back their values of swelling and the pain-related behaviors (Fig. 3). The 2 groups of rats did not differ significantly in the swelling at the knee [Fig. 3A, F(1,18) = 0.054; P = 0.819], the guarding score [Fig. 3B, F(1,18) = 3.19; P = 0.091], but they differed in the magnitude and course of the reduction of the mechanical threshold at the injected knee [Fig. 3C, F(1,18) = 16.598; P = 0.001] and at the contralateral knee [Fig. 3D, F(1,18) = 7.691; P = 0.013]. Thus, rats of cluster 1 (n = 7) exhibited persistent mechanical hyperalgesia at the injected knee joint, and at the noninjected contralateral knee, rats of cluster 2 (n = 13) showed reversal of mechanical hyperalgesia at the injected knee joint and no hyperalgesia at the contralateral knee. The rats of clusters 1 and 2 did not differ in the thermal withdrawal thresholds at the ipsilateral [Fig. 3E, F(1,18) = 0.119; P = 0.735] and contralateral hindpaw [Fig. 3F, F(1,18) = 0.108; P = 0.746], in the mechanical withdrawal threshold at the ipsilateral [Fig. 3G, F(1,18) = 0.119; P = 0.734] and contralateral hindpaw [Fig. 3H, F(1,18) = 0.501; P = 0.488], nor did they show significant differences of weight bearing [Fig. 3I, F(1,18) = 1.191; P = 0.29]. The body weight of the rats of cluster 1 and 2 increased similarly (supplementary Figure S2, available at http://links.lww.com/PAIN/A976). Thus, the cluster analysis yielded a group of rats, which exhibited a persistent primary mechanical hyperalgesia at the inflamed and mechanical hyperalgesia at the contralateral knee.
3.2. Exploration of behavioral factors associated with persistent mechanical hyperalgesia at the injected knee
A closer look to Figure 3C shows that rats of cluster 1 exhibited a greater reduction of the mechanical threshold at the injected knee than rats of cluster 2 already at day 3 of AIA, and this difference remained throughout the 84 days. These data suggest that strong initial primary mechanical hyperalgesia is a risk factor for the persistence of mechanical hyperalgesia. Furthermore, as already mentioned, only rats of cluster 1 developed a significant and persistent reduction of mechanical threshold at the contralateral knee (Fig. 3D). By contrast, secondary mechanical and thermal hyperalgesia at the paws were not different in clusters 1 and 2. These data suggest, therefore, that in particular contralateral effects at the same segment as the inflammatory hyperalgesia are associated with persistent primary mechanical hyperalgesia at the injected knee.
Figure 3 shows the normalized behavioral data in which the values after immunization but before induction of knee inflammation were used as the baseline reference. Looking to the absolute values before injection of mBSA into the knee joint, we noticed that rats of cluster 1 had on average significantly higher thresholds at the knee than rats of cluster 2 before arthritis (Fig. 4A). Thus, rats with higher initial “pain tolerance” at the knee showed stronger and persistent primary mechanical hyperalgesia and mechanical hyperalgesia at the contralateral knee than rats with less pain tolerance. In the rats displayed in Figure 4A, we did not measure the withdrawal threshold before immunization, but we compared the withdrawal threshold of NC rats and of immunized rats and found that the average values were similar (about 365-370 g, see supplementary Figure 3, available at http://links.lww.com/PAIN/A976). Thus, we have no evidence that the immunization changed the thresholds. Independent on the cluster, we found an inverse relationship between the initial withdrawal threshold before injection of the knee (x-axis) and the reduction of threshold at day 84 of AIA for pressure at the ipsilateral knee (Fig. 4B) and contralateral knee (Fig. 4C).
Concerning thermal withdrawal thresholds at the paws, rats of clusters 1 and 2 did not significantly differ, but rats with high thresholds were rather found in cluster 1 (Fig. 4D). Again, we observed an inverse relationship between the initial withdrawal threshold before injection of the knee (x-axis) and the reduction of threshold at day 84 of AIA for heat stimulation at the ipsilateral (Fig. 4E) and contralateral hindpaw (Fig. 4F) independent on the cluster. Clusters 1 and 2 did not differ in the mechanical withdrawal threshold at the hindpaws (Fig. 4G), but the inverse relationship between the initial basal withdrawal threshold and the reduction of threshold at day 84 of AIA was also found here (Figs. 4H and I). In general, the reduction of pain threshold at day 84 was greatest in rats with high initial prearthritic thresholds.
3.3. Exploration of molecular factors associated with persistent mechanical hyperalgesia at the injected knee
We analyzed several parameters to find out whether molecular factors are related to cluster 1 and cluster 2 and to the magnitude of primary mechanical hyperalgesia in individual animals. To show the changes during AIA, we included additional rats, which were observed either to until day 3 or day 42 of AIA (Fig. 1). These additional rats were not included in the cluster analysis, which is based on the behavioral parameters at day 84.
3.3.1. Inflammation at the injected knee joint
Figure 5A shows the histology scores of synovial tissue of rats of NC groups for day 3 (n = 7), day 42 (n = 4), and day 84 (n = 8) and of rats with AIA at day 3 (n = 7), day 42 (n = 10), and day 84 (n = 7 for cluster 1 and n = 13 for cluster 2). On day 84, the histology score is also displayed separately for rats of clusters 1 and 2. The histology score showed pronounced inflammation at day 3 and a strong reduction of the scores at days 42 and 84, consistent with the time course of swelling. At day 84, the histology score was similarly low in rats of both cluster 1 and 2, ie, in both groups of rats inflammation had almost disappeared. Figure 5B shows typical examples. In the acute AIA stage (day 3), the synovial tissue was strongly infiltrated by inflammatory cells, and adipocytes were hardly visible. At day 84, the superficial synovial tissue exhibited a layer of connective tissue with very few inflammatory cells but appeared still thickened and different from the synovia of NC rats (see specimens). Thus, the knee had largely recovered from arthritis but still showed residual small inflammatory changes of the synovial tissue. Importantly, however, these persistent alterations were similar in the rats of clusters 1 and 2, and thus, the persistent mechanical hyperalgesia at the knee in rats of cluster 1 could not be attributed to more severe synovitis. Notably, also the initial swelling was not different in rats of clusters 1 and 2 (Fig. 3A).
3.3.2. Invasion of macrophages into the dorsal root ganglia
As reported previously,29,36 macrophages invaded the lumbar DRG at day 3 of AIA (Fig. 5C). Dorsal root ganglia obtained at day 42 of AIA (n = 9 rats) did not show more macrophages than DRG of NCs (n = 4 rats). On day 84, we did not find macrophage invasion, and the data were similar for both cluster 1 and cluster 2. Figure 5D shows typical DRG sections. Thus, persistent mechanical hyperalgesia could not be attributed to persistent invasion of macrophages into the lumber DRG.
3.3.3. Expression of ATF3 in dorsal root ganglion neurons
Because some arthritis models show an upregulation of ATF3 in DRG, we also looked for the expression of ATF3. At no time point (day 3 [n = 4], day 42 [n = 4], day 84 [n = 20]), we did find any expression of ATF3 in DRG sections.
3.3.4. Microglia in the spinal cord
We quantified the mean numbers of IBa1-positive microglial cells in the dorsal horn of different groups of rats (Fig. 6). The number of microglial cells per section (see Fig. 6B for the approximate area of analysis) was similar in sections from NC rats (n = 6), IC rats (n = 5), and AIA rats (n = 7-13). We did neither find a side difference nor a difference for clusters 1 and 2 at day 84 (Fig. 6A). In the overview, a similar expression pattern was observed in the dorsal horn of NC rats (Fig. 6C), IC rats (Fig. 6D), and AIA rats of days 3 (Figs. 6E) and 84 (Fig. 6F). IBa1-positive cells in higher magnification had also similar morphology in NC rats (Fig. 6G), IC rats (Fig. 6H), and AIA rats at day 3 (Fig. 6I) and day 84 (Fig. 6J).
3.3.5. Expression of pCREB in dorsal root ganglia
As previously reported, we observed a significant upregulation of pCREB in lumbar DRG neurons during AIA. This study confirmed the upregulation of pCREB at day 3 of AIA in untreated rats (to 38 ± 5% of the neurons, not shown), and it showed an upregulation of pCREB in rats treated with saline applied from minipumps for 3 days (Fig. 7A, third column, n = 7 rats). The early upregulation of pCREB at day 3 of AIA was prevented by Anakinra, applied from minipumps for 3 days (Fig. 7A, fourth column, n = 8 rats). This study also revealed a persistent and significant upregulation of pCREB at day 42 (not shown) and at day 84 in untreated rats (Fig. 7A, columns 5 and 6). We did not find a difference between rats of clusters 1 (n = 7) and 2 (n = 13) at day 84. In rats treated with Anakinra for 21 days, pCREB upregulation at day 84 was similar as in untreated AIA rats (Fig. 7A, column 7, n = 7). Typical pCREB-positive DRG neurons are displayed in Figure 7B.
We monitored pain-related behaviors in the rats, which were treated for 21 days with either Anakinra or saline using minipumps. Figure 7C shows the mechanical withdrawal threshold at the injected knee. There was no significant difference between the 2 groups of rats [F(1,13) = 2.310; P = 0.152], although there was a trend toward a reduction of hyperalgesia around day 28. Thus mechanical hyperalgesia was not prevented during the application period of Anakinra, although the upregulation of pCREB was prevented (Fig. 7A). Furthermore, the application of Anakinra for 21 days did not prevent pCREB upregulation at day 84 of AIA.
These data indicate that the upregulation of pCREB is not essential for the expression of mechanical hyperalgesia at the inflamed knee, and they raise the question of the functional role of the upregulation of pCREB. Plotting the percentage of pCREB-positive neurons at day 84 of AIA and the withdrawal threshold showed that there was a correlation between the proportion of pCREB-positive neurons and the reduction of the mechanical withdrawal threshold at the knee. Stronger reduction of the threshold (expressing more severe mechanical hyperalgesia) was associated with a lower proportion of pCREB-positive neurons (including cluster 1 and cluster 2). It may be asked, therefore, whether the upregulation of pCREB is rather associated with recovery from mechanical hyperalgesia.
Using Ward's minimum variance method for unbiased hierarchical cluster analysis at day 84 of AIA, we identified 2 clusters of rats. About one-third of the rats (cluster 1) showed persistent primary mechanical hyperalgesia at the injected knee joint at day 84, whereas in about two-thirds of the rats (cluster 2), primary mechanical hyperalgesia at the knee had disappeared. Retrograde analysis of all pain-related behaviors of these 2 clusters with repeated-measures ANOVA showed that rats of clusters 1 and 2 differed in the severity of primary mechanical hyperalgesia (more intense from day 3 until day 84 in cluster 1) and in the occurrence of mechanical hyperalgesia at the contralateral knee (only present in rats of cluster 1). But they did not differ in secondary mechanical and thermal hyperalgesia at the ipsilateral hindpaw, weight bearing, and guarding score in the first 21 days, and in the time course and severity of inflammation of the synovia. At day 84, neither cluster showed an invasion of macrophages in the DRG, nor an activation of microglia, nor an expression of ATF3, and pCREB in the lumbar DRG of the inflamed side was similar in both clusters. During treatment of rats with Anakinra, the upregulation of pCREB was prevented, but mechanical hyperalgesia was not consistently reduced.
4.1. Cluster analysis and neuronal mechanisms
Notably, unbiased cluster analysis at day 84 distinguished the rats according to the disappearance or persistence of primary mechanical hyperalgesia after subsiding AIA. Remarkably, the differences of primary mechanical hyperalgesia in clusters 1 and 2 became already apparent at day 3 of AIA and were present to day 84. Thus, early severe primary mechanical hyperalgesia seems to be a risk factor for the development of persistent hyperalgesia. In line with this, clinical studies showed that severe acute pain in the first week after surgery predisposes patients to develop chronic postoperative pain.6,15 However, whether specific models of postoperative pain yield similar clusters needs to be determined because postoperative pain exhibits some specific mechanisms.4,31,32
Mechanical hyperalgesia at the contralateral noninjected knee during ipsilateral AIA was noted in our previous studies, but this study shows that mechanical hyperalgesia at this site was only present and persistent in rats of cluster 1. Thus, contralateral segmental mechanical hyperalgesia upon unilateral arthritis seems to be a characteristic escorting feature of severe primary hyperalgesia and chronification. The contralateral hyperalgesia in AIA cannot be directly compared with bilateral hyperalgesia in the K/BxN, collagen antibody-induced arthritis, and G6PI models in which arthritis is induced systemically and hence develops bilaterally. Although in AIA the contralateral knee joint is not evidently inflamed, we noted previously some neuronal reactions in the lumbar DRG of both sides. Such bilateral changes were the invasion of macrophages in the lumbar DRG in early AIA,36 and in DRG neurons, the upregulation of IL-1-R1 during immunization and early AIA,13 the upregulation of pCREB in early AIA,37 the upregulation of CGRP12 and bradykinin 2 receptors,38 and the downregulation of somatostatin 2a and 4 receptors.1 Notably, these symmetric bilateral changes in AIA were only observed in lumbar, not in thoracic segments. Bilateral segmental changes in the nervous system after unilateral pathologies have been repeatedly reported, and a role of midline crossing spinal interneurons, the sympathetic nervous system, and/or dorsal root reflexes has been proposed, thus implicating spinal mechanisms.10,22,34
Intriguingly, before AIA induction, rats of cluster 1 showed higher mechanical withdrawal thresholds at the knees than rats of cluster 2, suggesting higher initial “pain tolerance.” It must be noted that these baseline thresholds were obtained before AIA induction but after immunization. But we have no evidence that the mechanical threshold at the knee was altered by immunization (see supplementary Figure 3, available at http://links.lww.com/PAIN/A976). In postsurgical pain patients, it remained unclear whether stronger pain after surgery can be predicted from pain thresholds before surgery.35
Secondary mechanical and thermal hyperalgesia at the hindpaw, typical phenomena of general central sensitization, were present from day 1 of AIA. Similarly, as previously observed, secondary mechanical hyperalgesia at the hindpaw disappeared with time, whereas the rats did not completely recover from thermal hyperalgesia at the hindpaw.13,21 Importantly, however, these symptoms were equally present in rats of clusters 1 and 2 and did not predict persistent primary mechanical hyperalgesia at the knee. The same applies for guarding behavior and asymmetric weight bearing.
The precise segmental neuronal mechanisms leading to persistent mechanical hyperalgesia are not known. Spinal sensitization as well as spinal disinhibition could play a role. Possibly “pain tolerant” cluster 1 rats have stronger spinal baseline inhibition, which subsides during arthritis.
4.2. Exploration of the involvement of molecular mechanisms
Since the monitoring of pain-related behaviors for 84 days precluded the harvest of tissue for histology at earlier time points, we could only test whether rats of clusters 1 and 2 can be distinguished by cellular and molecular changes at day 84. Intriguingly, none of the factors analyzed at this time point distinguished rats of clusters 1 and 2.
Rats of both clusters showed overlapping courses and severities of inflammatory swelling. Although the inflammatory AIA phase is characterized by substantial synovitis (see data of days 3 and 42 of AIA), the synovitis scores at day 84 were very low and not different in the 2 clusters. Thus, residual inflammation does not explain the persistence or absence bilateral mechanical hyperalgesia at the knees. Furthermore, macrophage invasion into the DRG, typical for early AIA, was not any more evident at days 42 and 84 of AIA.
An interesting possibility is the development of a neuropathic component since in K/BxN serum transfer,7 adjuvant,19 collagen-induced,20 and G6PI-induced arthritis,11 a large proportion of DRG neurons showed a persistent expression of ATF3. However, AIA did not evoke ATF3 expression in DRG neurons (see also Ref. 36). Furthermore, the analysis of IBa1-positive cells in the spinal did not reveal a morphologically visible microglia activation, which would be expected in a neuropathic state.8 Morphologically visible microglia activation may be absent in inflammatory models.8 Functional activation of microglial cells cannot be excluded because even acute spinal application of TNFα in normal rats causes microglial cells to release soluble interleukin-6 receptors, a prerequisite of spinal interleukin-6 effects.23 Thus, we have no evidence for the development of a neuropathic component. It should be noted, however, that the chronic pain phenotype consisted mainly in local mechanical hyperalgesia at the joints, whereas weight bearing and guarding score had been normalized, arguing against severe pain.
We can exclude further factors such as age, sex (all rats were female), genetic differences, housing conditions including food, influence of the experimenter (all tests were performed by the same person, but the data were only evaluated after completion of the experiments). However, we cannot exclude the importance of additional factors such as hormones.
Since strong early primary mechanical hyperalgesia predisposed for chronification of hyperalgesia, clustering of rats and correlation with molecular markers at early time points may give more hints, which molecular mechanisms should be tested.
4.3. The significance of pCREB upregulation
Previously, we found in rat DRG neurons an upregulation of pCREB up to 42 days of AIA,37 and in this study, pCREB was even upregulated for 84 days after AIA induction. The functional significance of pCREB upregulation remained unclear. Here, we found that the application of Anakinra prevented the increase of pCREB in the early phase of AIA but did not significantly alter primary mechanical hyperalgesia at the injected knee joint. Thus CREB activation in DRG neurons is not essential for the generation of mechanical hyperalgesia. Interestingly, we found an inverse correlation between the proportion of pCREB-positive neurons at day 84 of AIA and the reduction of the mechanical withdrawal threshold at the knee. Stronger reduction of the threshold (indicating more severe mechanical hyperalgesia) was associated with a lower proportion of pCREB-positive neurons. However, rats of clusters 1 and 2 were not distinguished by differences in the average of pCREB-positive neurons. These data suggest that upregulation of pCREB in sensory neurons is rather associated with recovery from hyperalgesia on an individual basis.
The study was performed on female Lewis rats and may not be representative for male animals. In rats observed for 84 days, transitory changes of cellular markers could not be analyzed.
The study shows that transitory joint inflammation does evoke persistent mechanical hyperalgesia at the joint in a subgroup of rats, reminiscent of clinical observations. Initial severe primary mechanical hyperalgesia and segmental contralateral mechanical hyperalgesia in noninflamed tissue seem to predispose for the development of persistent mechanical hyperalgesia. Transitory general central sensitization outside of the segment of inflammation may contribute to pain severity but may not necessarily lead to chronic hyperalgesia. A neuropathic component does not seem to be imperative for the development of persistent local mechanical hyperalgesia, although it may be important for the persistence of severe chronic pain in other models. Future research should address the mechanisms of bilateral segmental hyperalgesia and the role of pCREB.
Conflict of interest statement
The authors have no conflicts of interest to declare.
Appendix A. Supplemental digital content
Supplemental digital content associated with this article can be found online at http://links.lww.com/PAIN/A976.
The authors thank Mrs. Konstanze Ernst and Mrs. Antje Waller for technical help and the Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung, program Neuroimpa) for funding.
. Bär KJ, Schurigt U, Scholze A, Segond von Banchet G, Stopfel N, Bräuer R, Halbhuber KJ, Schaible HG. The expression and localisation of somatostatin receptors in dorsal root ganglion neurons of normal and monoarthritic rats. Neuroscience 2004;127:197–206.
. Bas DB, Su J, Agalave NM, Lundberg J, Codeluppi S, Baharpoor A, Nandakumar KS, Holmdahl R, Svensson CI. Collagen antibody-induced arthritis
evoked persistent pain with spinal glia involvement and transient prostaglandin dependency. Arthritis
. Boettger MK, Hensellek S, Richter F, Gajda M, Stöckigt R, Segond von Banchet G, Bräuer R, Schaible HG. Antinociceptive effects of TNF-α neutralization in a rat model of antigen-induced arthritis
. Brennan TJ. Pathophysiology of postoperative pain. PAIN 2011;152(3 suppl):S33–46.
. Buchner E, Bräuer R, Schmidt C, Emmrich F, Kinne RW. Induction of flare-up reactions in rat antigen-induced arthritis
. J Autoimmun 1995;8:61–74.
. Chapman CR, Vierck CJ. The transition of acute postoperative pain to chronic pain: an integrative overview of research on mechanisms. J Pain 2017;18:359.e1–359.e38.
. Christianson CA, Corr M, Firestein GS, Mobargha A, Yaksh TL, Svensson CI. Characterization of the acute and persistent pain state present in K/BxN serum transfer arthritis
. PAIN 2010;151:394–403.
. Clark AK, Gentry C, Bradbury EJ, McMahon SB, Malcangio M. Role of spinal microglia
in rat models of peripheral nerve injury and inflammation. Eur J Pain 2007;11:223–30.
. Clark AK, Grist J, Al-Kashi A, Perretti M, Malcangio M. Spinal cathepsin S and fractalkine contribute to chronic pain in the collagen-induced arthritis
. Donaldson LF. Unilateral arthritis
: contralateral effects. Trends Neurosci 1999;22:495–6.
. Ebbinghaus M, Müller S, Segond von Banchet G, Eitner A, Wank I, Hess A, Hilger I, Kamradt T, Schaible HG. The contribution of inflammation and bone destruction to pain in arthritis
—a study in murine glucose-6-phosphate isomerase (G6PI)-induced arthritis
. Ebbinghaus M, Segond von Banchet G, Massier J, Gajda M, Bräuer R, Kress M, Schaible HG. Interleukin-6-dependent influence of nociceptive sensory neurons on antigen-induced arthritis
Res Ther 2015;17:334.
. Ebbinghaus M, Uhlig B, Richter F, Segond von Banchet G, Gajda M, Bräuer R, Schaible HG. The role of interleukin-1ß in arthritic pain: main involvement in thermal but not in mechanical hyperalgesia
in rat antigen-induced arthritis
. Eitner A, Pester J, Vogel F, Marintschev I, Lehmann T, Hofmann GO, Schaible HG. Pain sensation in human osteoarthritic knee joints is strongly enhanced by diabetes mellitus. PAIN 2017;158:1743–53.
. Fletcher DM, Stamer UM, Pogatzki-Zahn E, Zaslansky R, Valentin N, Perruchoud C, Kranke P, Komann M, Lehman T. Chronic postsurgical pain in Europe an observational study. Eur J Anaesthesiol 2015;32:725–34.
. Griffiths RJ. Characterisation and pharmacological sensitivity of antigen arthritis
induced by methylated bovine serum albumin in the rat. Agents Actions 1992;35:88–95.
. Hu P, McLachlan EM. Macrophage and lymphocyte invasion of dorsal root ganglia after peripheral nerve lesions in the rat. Neuroscience 2002;112:23–38.
. Hunt D, Raivich G, Anderson PN. Activating transcription factor 3 and the nervous system. Front Mol Neurosci 2012;5:7.
. Inglis J, Nissim A, Lees DM, Hunt SP, Chernajovsky Y, Kidd BL. The differential contribution of tumour necrosis factor to thermal and mechanical hyperalgesia
during chronic inflammation. Arthritis
Res Ther 2005;7:R8.
. Inglis JJ, Notley CA, Essex D, Wilson AW, Feldmann M, Anand P, Williams R. Collagen-induced arthritis
as a model of hyperalgesia: functional and cellular analysis of the analgesic actions of tumor necrosis factor blockade. Arthritis
. Jochmann E, Boettger MK, Anand P, Schaible HG. Antigen-induced arthritis
in rats is associated with increased growth-associated protein GAP-43-positive intraepidermal nerve fibres remote from the joint. Arthritis
Res Ther 2015;17:299.
. Koltzenburg M, Wall PD, McMahon SB. Does the right side know what the left is doing? Trends Neurosci 1999;22:122–7.
. König C, Morch E, Eitner A, Möller C, Turnquist B, Schaible HG, Ebersberger A. Involvement of spinal Interleukin-6 trans-signaling in the induction of hyperexcitability of deep dorsal horn neurons by spinal tumor necrosis factor-α. J Neurosci 2016;36:9782–91.
. Krenn V, Morawietz L, Burmester GR, Häupl T. Synovialitis-Score: histopathologisches Graduierungsschema rheumatischer und nicht- rheumatischer Synovialitiden [in German]. Z Rheumatol 2005;64:334–42.
. Lee YC, Cui J, Lu B, Frits ML, Iannaccone CK, Shadick NA, Weinblatt ME, Solomon DH. Pain persists in DAS28 rheumatoid arthritis
remission but not in ACR/EULAR remission: a longitudinal observational study. Arthritis
Res Ther 2011;13:R83.
. Leuchtweis J, Boettger MK, Niv F, Redecker C, Schaible HG. Enhanced neurogenesis in the hippocampal dentate gyrus during antigen-induced arthritis
in adult rat—a crucial role of immunization. PLoS One 2014;9:e89258l.
. Leuchtweis J, Imhof AK, Montechiaro F, Schaible HG, Boettger MK. Validation of the digital pressure application measurement (PAM) device for detection of primary hyperalgesia in rat and mouse antigen-induced knee joint arthritis
. Methods Find Exp Clin Pharmacol 2010;32:575–83.
. Lonze BE, Ginty DG. Function and regulation of CREB family transcription factors in the nervous system. Neuron 2002;35:605–23.
. Massier J, Eitner A, Segond von Banchet G, Schaible HG. Effects of differently activated rodent macrophages on sensory neurons. Implications for arthritis
. Miller RE, Tran PB, Das R, Ghoreishi-Haack N, Ren D, Miller RJ, Malfait AM. CCR2 chemokine receptor signaling mediates pain in experimental osteoarthritis. Proc Natl Acad Sci U S A 2012;109:20602–7.
. Pogatzki-Zahn EM, Segelcke D, Schug SA. Postoperative pain—from mechanisms to treatment. Pain Rep 2017;2:e588.
. Pogatzki-Zahn EM, Zahn PK, Brennan TY. Postoperative pain—clinical implications of basic research. Best Pract Res Clin Anaesthesiol 2007;21:3–13.
. Qiao LY, Vizzard MA. Up-regulation of phosphorylated CREB but not c-jun in bladder afferent neurons in dorsal root ganglia after cystitis. J Comp Neurol 2004;469:262–74.
. Rees H, Sluka KA, Westlund KN, Willis WD. Dorsal root reflexes in articular nerves occur bilaterally in a chronic model of arthritis
in rats. J Neurophysiol 1996;76:4190–3.
. Sangesland A, Storen C, Vaegter HB. Are prooperative experimental pain assessments correlated with clincial pain outcomes after surgery? A systematic review. Scand J Pain 2017;15:44–52.
. Segond von Banchet G, Boettger MK, Fischer N, Gajda M, Bräuer R, Schaible HG. Experimental arthritis
causes tumor necrosis factor alpha-dependent infiltration of macrophages into rat dorsal root ganglia which correlates with pain- related behavior. PAIN 2009;145:151–9.
. Segond von Banchet G, König C, Patzer J, Eitner A, Leuchtweis J, Ebbinghaus M, Boettger MK, Schaible HG. Long-lasting activation of the transcription factor CREB in sensory neurons by interleukin-1β during antigen-induced arthritis
in rats: a mechanism of persistent arthritis
. Segond von Banchet G, Petrow PK, Bräuer R, Schaible HG. Monoarticular antigen-induced arthritis
leads to pronounced bilateral upregulation of the expression of neurokinin 1 and bradykinin 2 receptors in dorsal root ganglion neurons of rats. Arthritis
. Su J, Gao T, Shi T, Xiang Q, Xu X, Wiesenfeld-Hallin Z, Hökfelt T, Svensson CI. Phenotypic changes in dorsal root ganglion and spinal cord in the collagen antibody-induced arthritis
mouse model. J Comp Neurol 2015;523:1505–28.
. Tamura S, Morikawa Y, Senba E. Up-regulated phosphorylation of signal transducer and activator of transcription 3 and cyclic AMP-responsive element binding protein by peripheral inflammation in primary afferent neurons possibly through oncostatin M receptor. Neuroscience 2005;133:797–806.
. Taylor P, Manger B, Alvaro-Gracia J, Johnstone R, Gomez-Reino J, Eberhardt E, Wolfe F, Schwartzmann F, Furfaro N, Kavanaugh A. Patient perceptions concerning pain management in the treatment of rheumatoid arthritis
. J Int Med Res 2010;38:1213–24.
. Webster JI, Tonelli L, Sternberg EM. Neuroendocrine regulation of immunity. Annu Rev Immunol 2002;20:125–63.
. Woolf CJ, Costigan M. Transcriptional and posttranslational plasticity and the generation of inflammatory pain. Proc Natl Acad Sci U S A 1999;96:7723–30.