To examine the relationship between inflammation and posttraumatic arthritis (PTA) in a murine intraarticular fracture model.
To examine the relationship between inflammation and posttraumatic arthritis (PTA) in a murine intraarticular fracture model.
Male C57BL/6 and MRL/MpJ “superhealer” mice received tibial plateau fractures using a previously established method. Mice were killed on day 0 (within 4 hours of fracture) and days 1, 3, 5, 7, 28, and 56 after fracture. Synovial tissue samples, obtained prior to fracture and on days 0, 1, 3, 5, and 7 after fracture, were examined by reverse transcription–polymerase chain reaction for gene expression of proinflammatory cytokines and chemokines. Synovial fluid and serum samples were collected to measure cytokine concentrations, using enzyme-linked immunosorbent assay. Whole joints were examined histologically for the extent of synovitis and cartilage degradation, and joint tissue samples from all time points were analyzed immunohistochemically to evaluate the distribution of interleukin-1 (IL-1).
Compared to C57BL/6 mice, MRL/MpJ mice had less severe intraarticular and systemic inflammation following joint injury, as evidenced by lower gene expression of tumor necrosis factor α and IL-1β in the synovial tissue and lower protein levels of IL-1α and IL-1β in the synovial fluid, serum, and joint tissues. Furthermore, after joint injury, MRL/MpJ mice had lower gene expression of macrophage inflammatory proteins and macrophage-derived chemokine (CCL22) in the synovial tissue, and also had reduced acute and late-stage infiltration of synovial macrophages.
C57BL/6 mice exhibited higher levels of inflammation than MRL/MpJ mice, indicating that MRL/MpJ mice are protected from PTA in this model. These data thus suggest an association between joint tissue inflammation and the development and progression of PTA in mice.
Joint trauma resulting in posttraumatic arthritis (PTA) is estimated to account for 12% of the 27 million Americans with symptomatic osteoarthritis (OA) (1–3). Even with optimal treatment, displaced articular fractures in the lower extremity are associated with a 10–20% incidence of PTA (4). Despite the impact of PTA, the sequence of events leading to arthritis following an articular fracture is not fully understood.
An articular fracture is a complex event with several injurious aspects, including mechanical insult to the joint tissues, release of blood and marrow contents into the joint space, and potentially systemic polytrauma (5). The inflammatory response resulting from articular fracture may be a significant factor in the progression of PTA, but its effect remains incompletely characterized (6). Proinflammatory cytokines such as interleukin-1 (IL-1) and tumor necrosis factor α (TNFα) are up-regulated in injured and degenerative joints and may play an important role in the pathogenesis of PTA, similar to their role in primary OA of the joints in patients without antecedent injury (7, 8). The manner in which these cytokines and mediators influence sustained joint tissue inflammation and cartilage degeneration following articular trauma, however, remains unclear.
In order to further characterize these mechanisms, we developed a murine model of intraarticular fracture of the tibial plateau that results in progressive arthritic changes in the bone, articular cartilage, and other joint tissues in the C57BL/6 mouse (9). However, the inbred MRL/MpJ strain, known as the “superhealer” mouse strain, is protected from PTA and does not develop degenerative joint changes following articular fracture (10). The MRL/MpJ strain is of particular interest because of its demonstrated ability to regenerate a wide array of tissues, ranging from fibrocartilage in the ear to myocardium to articular cartilage (11–14). While the exact genetic differences responsible for the enhanced regenerative capability of this strain are not known, the MRL/MpJ mouse exhibits decreased levels of inducible proinflammatory cytokines, such as IL-1 and TNFα, in response to lipopolysaccharide stimulation (15).
We hypothesized that the superhealer MRL/MpJ strain of mice responds to joint injury (i.e., a closed articular fracture) with a reduced inflammatory response when compared to C57BL/6 mice. To assess this hypothesis, we measured synovial gene expression of inflammatory cytokines and chemokines, synovial fluid and serum levels of inflammatory cytokines, and the extent of infiltrating inflammatory cells in the synovium, at a variety of early and late time points following closed articular fracture, in MRL/MpJ mice compared to C57BL/6 mice.
All procedures were performed in accordance with an Institutional Animal Care and Use Committee–approved protocol. Male C57BL/6 mice (n = 42; Charles River Laboratories) and male MRL/MpJ mice (n = 42; The Jackson Laboratory) at 8 weeks of age were obtained, and were thereafter housed until 16 weeks of age, which represents the age of peak bone mass (16). At this point, 6 mice from each strain were killed as prefracture controls.
The remaining experimental mice received a moderately severe closed articular fracture of the left lateral tibial plateau. Animals were anesthetized and placed in a custom cradle with the left hind limb in 90° of flexion. As previously described (9), a custom indenter was attached to a materials testing system (ElectroForce ELF3200; Bose), and a 10N compressive preload was applied to the anterior aspect of the proximal tibial plateau of the left hind limb. The tibia was then loaded in compression at a rate of 20 N/second in load-control mode to induce fracture. The total displacement of the indenter was scaled to the relative size of the tibial plateau in the mouse strains (2.5 mm for C57BL/6 and 3.2 mm for MRL/MpJ), a technique that, as we have previously shown, allows for equivalent severity of injury (10). High-resolution digital radiographs (MX-20; Faxitron) were obtained to confirm the fracture. The right hind limb, which was left unfractured, was used as a contralateral control. No immobilization or surgical intervention was utilized. Animals were allowed immediate, full weight-bearing after fracture.
For each strain, 6 mice were killed on days 0, 1, 3, 5, 7, 28, and 56 after fracture. Prior to fracture and on days 0 (within 4 hours of fracture), 1, 3, 5, and 7 after fracture, the following samples were obtained postmortem. Serum was collected via retroorbital puncture, followed by cardiac stick. Synovial fluid was collected from both knees of each animal using a calcium sodium alginate compound, as previously described (17). Joint capsule tissue was obtained from both knees for RNA isolation. Prior to fracture and on days 0, 1, 3, 5, 7, 28, and 56 after fracture, both hind limbs of all animals were formalin-fixed and paraffin-embedded for histology and immunohistochemistry.
Gene expression was measured in synovium from joint capsule tissue, which was harvested with a 3-mm biopsy punch as previously described (18). Tissue samples from each knee were pooled from a total of 6 animals on days 0, 1, 3, 5, and 7 after fracture. RNA was isolated using a 2-step TRIzol protocol. Reverse transcription–polymerase chain reaction (RT-PCR) was run, in duplicate, on 1 μg of RNA using SYBR Green Master Mix and the commercially available RT2 Profiler PCR array, to assess the synovium for messenger RNA (mRNA) expression levels of 84 mouse inflammatory cytokine and receptor genes (SABiosciences). The relative expression of mRNA for each gene in each sample was first normalized to the geometric mean value for 3 housekeeping genes (GAPDH, HPRT1, and HSP90ab1) in that sample, and then compared to mRNA expression prior to fracture using the 2 method (19). In accordance with the commercial analysis software provided by the array manufacturer, results were considered significant when mRNA expression levels were 3-fold different from that in mice prior to fracture.
Concentrations of IL-1α, IL-1β, and TNFα were measured in the synovial fluid and serum samples using commercially available enzyme-linked immunosorbent assay (ELISA) kits (MLA00, MLB00B, and MTA00B; R&D Systems) and run as directed by the manufacturer, with the following exceptions: 1) the 2-fold dilution series of the standard was extended to more accurately quantify samples containing protein levels between 0 and 7.8 pg/ml, and 2) the assay required a 1:5 dilution of the synovial fluid samples and 50 μl of neat serum.
Both hind limbs from each animal at all time points were placed in 10% neutral buffered formalin for 72 hours. The limbs were then decalcified (Cal-Ex Decalcification Solution; Fisher Scientific) for 72 hours, dehydrated in ethanol, infiltrated with xylene, and paraffin-embedded using a commercially available automated tissue processor (ASP300S; Leica Microsystems).
Immunohistochemistry was used to localize IL-1α and IL-1β cytokine proteins within the joint tissue. Serial sections (8 μm) of fixed limbs from both strains at all time points were pretreated by heating the tissue for 20 minutes at 95°C for antigen retrieval. Endogenous peroxidase was quenched for 45 minutes in the dark with 3% H2O2 in methanol with 0.1% weight/volume saponin, for membrane permeabilization. Sections were then stained with a polyclonal antibody against murine IL-1β (anti-mouse IL-1β/IL-1F2, AF-401-NA; R&D Systems) or murine IL-1α (anti-mouse IL-1α, AF-400-NA; R&D Systems) at 1:10 dilution at room temperature for 18 hours. During every staining protocol, negative controls were created simultaneously by application of 1–2% horse blocking serum, instead of the primary antibody, to tissue sections. Chromogenic detection was achieved with diaminobenzidine (DAB) substrate (Vectastain), and sections were digitally photographed. Regions of the joint evaluated included the articular cartilage, calcified cartilage, subchondral bone, and synovium. The intensity of positive staining for cytokines was assessed qualitatively using a discrete 4-point scoring system.
Synovitis was assessed using histologic analysis and semiquantitative grading. Serial sections (8 μm) of fixed limbs from both strains at all time points were stained with standard Harris' hematoxylin and eosin, and digital images of the joints were obtained. All limbs were evaluated at the synovial insertion of the lateral tibia, lateral femur, medial tibia, and medial femur, using a modified form of an established standardized synovitis score (20, 21) in which the total possible site score was assigned on a scale of 0–6. The mean synovitis score, obtained from the readings of 3 independent blinded graders (JSL, BDF, and EZ), was reported for each limb. The score for all 4 sites was summed to yield the total joint synovitis score.
Immunohistochemistry was used to assess synovial infiltration of activated macrophages. Serial sections (8 μm) of fixed limbs from both strains at all time points were pretreated with 0.05% proteinase K (Sigma-Aldrich) for 5 minutes at 37°C for antigen retrieval, as previously described (22). Endogenous peroxidase was quenched with 3% H2O2 in methanol for 30 minutes. Sections were stained with a monoclonal antibody against a surface marker of activated macrophages (anti-F4/80, Clone CI:A3-1; Serotec) at 1:100 dilution at room temperature for 1 hour. During every staining protocol, negative controls were created simultaneously by application of 1–2% rabbit blocking serum, instead of the primary antibody, to tissue sections. Chromogenic detection was achieved with DAB substrate (Vectastain), and digital images of the joint tissue were obtained.
All statistical analyses were performed using Statistica software (version 7; StatSoft). Nonparametric statistical analyses were utilized, with significance reported at the 95% confidence level. For serum levels of cytokines, the Mann-Whitney U test was used to compare differences between C57BL/6 and MRL/MpJ mouse strains at each time point, and the Kruskal-Wallis test was used to compare differences between the baseline prefracture condition and conditions at all time points postfracture within a mouse strain. For synovial fluid levels of cytokines and joint synovitis scores, Wilcoxon's matched pairs test was used to compare differences between fractured and unfractured limbs at each time point in each strain independently. The Mann-Whitney U test was used to compare differences between C57BL/6 and MRL/MpJ mouse strains at each time point for fractured and unfractured limbs independently, and the Kruskal-Wallis test was used to compare differences between the baseline prefracture condition and conditions at all time points postfracture within a mouse strain.
In C57BL/6 mice, 54 of the 84 inflammatory cytokine and receptor genes that were evaluated by RT-PCR were up-regulated in the synovial tissue after lateral plateau fracture, and 0 of 84 genes were down-regulated. In contrast, in MRL/MpJ mice, only 33 of the 84 inflammatory genes evaluated were up-regulated in the synovial tissue after lateral plateau fracture, and 7 of 84 genes were down-regulated.
Several genes of interest were highly up-regulated after fracture. TNFα gene expression in C57BL/6 mice was significantly up-regulated on day 1 (increase of 13-fold) and remained significantly elevated through day 7 after fracture (Figure 1A). In contrast, MRL/MpJ mice showed no significant increase in TNFα gene expression (increase of <1.5-fold) in synovial tissue after fracture at any time point.
IL-1α gene expression was up-regulated after fracture in both strains (Figure 1A). In C57BL/6 mice, IL-1α gene expression was significantly up-regulated on day 1 after fracture (increase of 7-fold) and remained significantly elevated through day 5, before returning to baseline levels on day 7 (Figure 1A). In MRL/MpJ mice, IL-1α gene expression was intermittently up-regulated after fracture (7-fold increase on day 0, 4-fold increase on day 3, 3-fold increase on day 7).
IL-1β gene expression was up-regulated after fracture, to much higher levels than those of IL-1α (Figure 1A). In C57BL/6 mice, IL-1β gene expression was up-regulated 720-fold on day 0, within 4 hours of fracture, compared to a 74-fold increase in MRL/MpJ mice (Figure 1A). Furthermore, IL-1β gene expression remained elevated more than 200-fold in C57BL/6 mice on days 1 and 3 postfracture, with significant elevation extending to days 5 and 7 postfracture. In contrast, IL-1β gene expression in MRL/MpJ mice returned to prefracture levels by day 3.
Serum IL-1α levels were significantly increased following fracture in both the C57BL/6 mice and the MRL/MpJ mice. In C57BL/6 mice, peak concentrations of serum IL-1α occurred on days 1 and 3 postfracture compared to the serum concentrations prior to fracture (P < 0.05). In MRL/MpJ mice, IL-1α concentrations on days 0 and 7 postfracture were significantly elevated compared to prefracture levels (P < 0.05) (Figure 1B). The effect of strain was found to be significant, as C57BL/6 mice showed significantly higher total serum concentrations of IL-1α in comparison with MRL/MpJ mice on days 1 and 3 postfracture (each P < 0.05) (Figure 1B). It should be noted that in synovial fluid samples from 22 of 72 C57BL/6 mice and 24 of 72 MRL/MpJ mice, the IL-1α protein concentration was below the detection limit of the immunoassay. The distribution of samples with IL-1α levels below the detection limit was concentrated at the prefracture time point and on days 0, 1, and 7 postfracture in both mouse strains, and this distribution was similar between the right and left limbs. For purposes of statistical analysis, these samples were assigned a value of one-half the lower limit of detection of the assay.
Serum IL-1β levels were also increased following fracture in both C57BL/6 and MRL/MpJ mice, with higher concentrations being observed in C57BL/6 mice compared to MRL/MpJ mice (P < 0.001). C57BL/6 mice had significantly higher levels of IL-1β compared to MRL/MpJ mice at the time of prefracture and at 1, 3, and 5 days after fracture. The highest concentrations of IL-1β in both strains occurred in the serum on day 1 postfracture (Figure 1C). The elevation in serum concentrations of IL-1β on day 1 was significantly higher than that prior to fracture or on days 5 and 7 postfracture in C57BL/6 mice, and was significantly higher than that prior to fracture and on day 3 postfracture in MRL/MpJ mice (each P < 0.05) (Figure 1C).
TNFα was undetectable in the majority of synovial fluid samples analyzed (47 of 72 from C57BL/6 mice and 43 of 72 from MRL/MpJ mice). Among the samples that had sufficient protein concentrations to analyze, no effect of strain or time on synovial fluid TNFα levels was observed (Figures 2A and B, left). Given that protein concentrations were expected to be higher in the synovial fluid than in the systemic serum, subsequent analysis of serum levels of TNFα was not performed.
IL-1α protein concentrations in the synovial fluid of C57BL/6 and MRL/MpJ mice showed nonsignificant increases on days 3 and 5 postfracture, in both the fractured limb and the contralateral (unfractured) limb (Figures 2A and B, middle). Interestingly, concentrations of IL-1α in the synovial fluid were not different between fractured and unfractured limbs in either strain at any time point. Thus, these data, in combination with the findings in the serum, suggest that a systemic response of IL-1α may occur following articular fracture.
IL-1β protein concentrations in the synovial fluid from fractured limbs of both C57BL/6 mice and MRL/MpJ mice were significantly elevated compared to the concentrations in the synovial fluid from unfractured limbs on days 0–7 after fracture (each P < 0.05) (Figure 2, right). Peak concentrations of IL-1β in both strains were comparable and occurred on day 0 (within 4 hours of fracture). Interestingly, baseline and unfractured limb synovial fluid levels of IL-1β were significantly higher in MRL/MpJ mice compared to C57BL/6 mice. However, the fold change was greater in the C57BL/6 mice postfracture. C57BL/6 mice showed a 7-fold increase in IL-1β concentration in the synovial fluid of the fractured limb relative to the unfractured limb, while MRL/MpJ mice showed only a 2-fold increase. Taken together, these data suggest that a local intraarticular response of IL-1β occurs in synovial fluid after articular fracture, followed by a systemic increase in serum levels.
The distribution of both isoforms of IL-1 in various joint tissues was assessed via immunohistochemistry. Regions of the joint evaluated included the articular cartilage, calcified cartilage, subchondral bone, and synovium (Figure 3A). In the fractured limbs of C57BL/6 mice, there was a peak in IL-1α expression, as revealed by immunohistochemical staining, in the articular cartilage, calcified cartilage, and synovium on day 1 postfracture (Figure 3B). In the unfractured limbs of C57BL/6 mice, there was a peak in IL-1α staining in the same 3 joint tissues (articular cartilage, calcified cartilage, and synovium) on day 3 postfracture (Figure 3B). In the fractured limbs of MRL/MpJ mice, there was an increase in IL-1α staining in only the synovium, on days 1 and 7 postfracture, while in the unfractured limbs of MRL/MpJ mice, there was minimal variation in IL-1α expression in the joint tissue in any region of the joint at any time point (Figure 3B).
In the fractured limbs of C57BL/6 mice, there was a peak in IL-1β levels, identified by immunohistochemical staining, in both the subchondral bone and synovial tissue on day 3 after fracture (Figure 3C), while in the unfractured limbs, there was a peak in IL-1β staining in the synovium only, on day 3 postfracture. In the fractured limbs of MRL/MpJ mice, there was a peak in IL-1β staining in the synovium only, observed on day 3 postfracture, whereas in the unfractured limbs, there was no significant change in IL-1β expression in the joint tissue in any region of the joint at any time point (Figure 3C).
Grading of the synovium using a modified synovitis score (mean ± SD) revealed no significant effect of mouse strain at any of the time points examined. Synovitis progressed similarly in both strains until day 7 postfracture, when C57BL/6 mice showed a significant increase in the synovitis score (12.8 ± 3.7) compared to that at prefracture (2.0 ± 3.1) (P < 0.05), whereas MRL/MpJ mice demonstrated a nonsignificant elevation in the synovitis score from pre- to postfracture (from 3.7 ± 1.5 to 10.8 ± 2.8). On days 28 and 56 postfracture, synovitis decreased in both mouse strains, returning to near-prefracture levels (3.7 ± 0.7 in C57BL/6 mice and 3.7 ± 2.9 in MRL/MpJ mice).
Gene expression of multiple proteins associated with macrophage chemoattraction was up-regulated following fracture in both strains, but to a lesser extent in MRL/MpJ mice (Figure 4). Expression of CCL3 (macrophage inflammatory protein 1α [MIP-1α]) and CCL4 (MIP-1β) was significantly elevated in both strains on all days tested, although more so in the C57BL/6 mice than in the MRL/MpJ mice. CCL3 expression peaked in both strains on day 3 after fracture, with a 274-fold increase in C57BL/6 mice and only a 42-fold increase in the MRL/MpJ mouse strain. CCL4 expression peaked on day 3 after fracture in C57BL/6 mice, showing a 140-fold increase from pre- to postfracture, while in MRL/MpJ mice, CCL4 expression peaked on day 1, showing only a 36-fold increase from pre- to postfracture.
CCL9 (MIP-1γ) exhibited significant increases in gene expression on days 0–5 in both strains, and also on day 7 in C57BL/6 mice (Figure 4). CCL9 expression peaked on day 1 after fracture in C57BL/6 mice, with a 25-fold increase, while in MRL/MpJ mice, CCL9 expression peaked at only an 8-fold increase on day 0. CCL20 (MIP-3α) gene expression was significantly elevated in C57BL/6 mice on days 1, 3, and 5, peaking on day 5 (up to 12-fold increase). MRL/MpJ mice showed elevated gene expression of CCL20 on days 1 and 5, with a peak on day 1 (35-fold increase).
CCL19 (MIP-3β) expression was significantly elevated in C57BL/6 mice on all days tested, but none of the changes in CCL19 expression were significant in MRL/MpJ mice on any day after fracture. In addition, CCL22 (macrophage-derived chemokine) gene expression was elevated following fracture in both strains (Figure 4). Significant changes in CCL22 expression were seen in both strains on days 0, 3, 5, and 7 after fracture. The peak occurred in both strains on day 7, with a 27-fold change in C57BL/6 mice and only a 6-fold change in MRL/MpJ mice.
Immunohistochemistry was used to identify activated macrophages present within the synovial tissue. Starting on day 3 postfracture, C57BL/6 mice showed markedly increased infiltration of activated macrophages in the lateral synovial tissue of fractured joints when compared to MRL/MpJ mice (Figure 5A). Furthermore, starting on day 3 postfracture, C57BL/6 mice showed increased macrophage infiltration in the medial synovium, distant from the lateral fracture site, when compared to MRL/MpJ mice (results not shown).
Postmortem joint sections from mice obtained on days 28 and 56 after fracture were also examined. At these time points following lateral plateau fracture, C57BL/6 mice showed sustained inflammation in the lateral synovial tissue, as evidenced by persistent infiltration of the synovium with activated macrophages (Figure 5B). In contrast, at these time points, MRL/MpJ mice demonstrated a paucity of macrophages. Furthermore, C57BL/6 mice showed sustained global synovitis and joint-wide inflammation, as evidenced by persistent macrophage infiltration in the medial joint tissue, whereas MRL/MpJ mice showed minimal macrophage infiltration in the medial aspect of the joint at these later time points (results not shown).
Within 4 hours of fracture, MRL/MpJ mice showed evidence of decreased local intraarticular and systemic inflammation when compared to C57BL/6 mice, and this attenuated inflammatory response may help explain how MRL/MpJ mice are protected from the development of PTA after intraarticular fracture in our model (10). The findings of this study support the hypothesis that, in comparison with C57BL/6 mice, MRL/MpJ mice exhibit a significantly diminished inflammatory response following closed articular fracture. This observation suggests that inhibition of the inflammatory response after acute fracture may provide a novel therapeutic approach for PTA.
A number of methods were utilized to assess local and systemic inflammation after articular fracture. RT-PCR analysis was used to evaluate inflammatory gene expression, particularly that of IL-1α, IL-1β, TNFα, and a variety of associated chemokines, in the synovial joint tissue, to assess changes in inflammatory response to injury in the C57BL/6 mouse strain compared to the MRL/MpJ mouse strain. Serum and synovial fluid levels of these proinflammatory cytokines were assessed as well. Specifically, we found (by gene expression analysis and immunohistochemistry) that IL-1β levels were differentially elevated in the synovial fluid, serum, and synovium, while IL-1α expression was elevated in the serum and synovium. These data suggest that there is a very early inflammatory response to articular fracture that is greater in the C57BL/6 mouse strain than in the MRL/MpJ mouse strain.
The total synovial cellular infiltration following articular fracture was not different between the 2 strains. However, the C57BL/6 mice had significantly more activated macrophages present in the synovium near the time of fracture, and this was sustained over time following articular fracture. As revealed by immunohistochemical staining, IL-1β levels were increased in the synovium and subchondral bone of the C57BL/6 mice and in the synovium of the MRL/MpJ mice. IL-1α levels were increased in the synovium, cartilage, and calcified cartilage of the C57BL/6 mice, whereas no increases were observed in the MRL/MpJ mouse strain. These trends in IL-1α staining are consistent with the observed increases in synovial fluid protein levels in both limbs. The elevated serum levels followed by elevated local protein expression in the joint tissues from both limbs suggest that there is a systemic response of IL-1α following the occurrence of articular fracture in this mouse model.
Articular fracture led to increased inflammatory gene expression in the synovium and inflammatory protein expression in the synovium, serum, and synovial fluid, and subsequently led to infiltration of the synovium with activated macrophages in association with progression to arthritis. Taken together, these findings suggest both a systemic response and a local intraarticular organ–level response of the joint to injury.
TNFα levels did not increase in the synovial fluid after fracture, although there was evidence of increased gene expression in the C57BL/6 mice. It is possible that this is a reflection of the very short half-life of TNFα or may indicate that there is a significant amount of binding of the available TNFα, and this will lead to such a low level of detection in the synovial fluid. TNFα is the primary agent in the pathophysiologic processes of inflammatory arthritis, although the levels of TNFα are generally increased transiently and may rapidly return to undetectable levels in vivo (23). Other methods of selective inhibition of TNFα may be needed to assess its role in PTA in this model.
Many chemokines are implicated in the pathogenesis of arthritis because of their increased expression levels in the synovial fluid of rheumatoid arthritis (RA) patients and in animal models of arthritis (24–36). The proposed function of chemokines in arthritis development is through the regulation of inflammation, including infiltration of monocytes, macrophages, neutrophils, and lymphocytes, among others (37, 38), but their exact contributions have not been fully elucidated.
Synovial inflammation associated with arthritis and joint injury is likely driven, in part, by up-regulation of chemokines and infiltration of inflammatory cells. The mediators and targets in the cascade of cytokine and chemokine up-regulation are unclear, but previous studies of isolated synovial fibroblasts have demonstrated that IL-1 and TNFα up-regulate the MIP chemokines CCL3, CCL4, CCL19, and CCL20 (32, 34, 35, 39–41). Our data also suggest that IL-1 and TNFα are up-regulated first, with the highest levels observed within 1 day of injury, and then maximal up-regulation of the MIP chemokines occurs from day 1 to day 3 postinjury, followed by the simultaneous up-regulation of macrophage-derived chemokine (CCL22) and the infiltration of macrophages into the synovial tissue on days 5–7 postinjury (as illustrated in Figure 6). Furthermore, increased chemokine gene expression has been observed clinically in the synovium of patients undergoing surgery for meniscal injury, in whom synovial inflammation has been linked to increased pain and dysfunction (42).
The fracture in this model is displaced at the time of injury, and there is no attempt to treat the fracture, similar to that in other animal models of joint instability or injury (43, 44). The scale of the murine knee limits the possibilities of surgical reduction and fixation. An additional limitation of mouse models is the limited yield of biosamples, particularly from joint tissues. Pooling of tissue samples for RNA isolation in small-animal models has been previously reported when specific joint tissues are of interest, such as the synovium and joint capsule tissue (18, 45, 46). An alternative strategy for isolating RNA from individual mice is to homogenize entire knee joints or paws, but this approach involves significant heterogeneity in the tissue sample (47–49). In spite of this limitation, we demonstrated that protein expression patterns in the synovial fluid and joint tissues in individual animals followed the same trends as the changes in mRNA levels from pooled samples.
The primary clinical scenario that would be analogous to our findings is the early events that have been observed after articular fracture in humans. To date, very few studies have characterized the acute local and systemic response to articular fracture in humans. Future studies may examine the expression patterns of local and systemic cytokines and chemokines in patients with articular fracture. The time from fracture to reduction and fixation in a human joint can vary from <24 hours to longer than 3 weeks; the events that occur after fracture are potentially important in the subsequent response to injury. The pronounced infiltration of synovial macrophages observed in C57BL/6 mice starting at 3 days after fracture and continuing to 4 and 8 weeks postfracture is characteristically seen in chronic inflammatory arthropathies such as RA. It is widely accepted that both the inflammatory processes and destructive features of RA are driven through synovitis (50).
Our findings thus suggest that inhibition of the inflammatory response may provide a novel therapeutic approach for PTA after articular fracture. Targeted blockade of specific cytokines has been the focus of several therapies for RA. This has led to the development of specific inhibitors of TNFα and IL-1, such as etanercept, a soluble form of TNFα receptor type II, and anakinra, a recombinant form of IL-1 receptor antagonist (IL-1Ra). Endogenous IL-1Ra is a specific receptor antagonist that competitively inhibits the binding of both IL-1α and IL-1β to their active receptor (51). Previous research has shown that administration of either TNFα inhibitors (52–54) or IL-1Ra (55) in mouse models of collagen-induced arthritis will ameliorate joint inflammation and cartilage destruction. Given the findings of the current study, future studies are warranted to better characterize the inflammatory response to articular injury in humans and to test, in animals and humans, antiinflammatory agents for their ability to block the injury-related inflammatory response and the development of PTA.
In summary, MRL/MpJ mice had reduced severity of intraarticular and systemic inflammation following joint injury when compared to C57BL/6 mice, as evidenced by lower gene expression of TNFα and IL-1β in the synovial tissue and lower protein levels of IL-1α and IL-1β in the synovial fluid, serum, and joint tissues. Furthermore, C57BL/6 mice had increased gene expression of MIPs and macrophage-derived chemokine in the synovial tissue and reduced acute and late-stage infiltration of the synovium with activated macrophages after joint injury. Collectively, these data suggest an association between an increased and prolonged inflammatory response to articular fracture and the development and progression of PTA in mice. These results provide a basis for novel pharmacologic approaches that would target early inflammation of the joint tissues following injury, which may halt the progression of PTA.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Olson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Furman, Guilak, Olson.
Acquisition of data. Lewis, Furman, Zeitler, Huebner, Kraus.
Analysis and interpretation of data. Lewis, Furman, Zeitler, Huebner, Kraus, Guilak, Olson.
We would like to thank Steve Johnson for his excellent technical support, and Drs. Amy McNulty and Beverley Fermor for their assistance.