Growth and repair factors, osteoactivin, matrix metalloproteinase and heat shock protein 72, increase with resolution of inflammation in musculotendinous tissues in a rat model of repetitive grasping

All experiments were approved by the Temple University Institutional Animal Care and Use Committee in compliance with NIH guidelines for the humane care and use of laboratory animals. Rats were housed individually in the central animal facility in transparent plastic cages in a 12 h light: 12 h dark cycle with free access to water. Studies were conducted on a total of 34 young adult (3.5 months of age at onset of experiments) Sprague-Dawley, female rats. All rats were handled for 1 week, and then food restricted to 80–90 % of their naïve full body weights for another week before onset of the experiment to motivate interest in 45 mg food pellets used for reward and retrieval (Banana flavored 45 mg food pellets; Bio-Serve, Flemington, NJ). All rats were weighed once to twice per week, provided regular rat chow daily and allowed to gain weight over the course of the experiments since they were young adult rats. All rats were maintained within 5–10 % of the weights of age-matched normal control rats (used only for weight matching purposes and therefore not included in the study). Rats were then randomly divided into food restricted control rats (n = 12) and high repetition negligible force (HRNF) task rats (n = 22) performing this task for 3 weeks (n = 12) and 6 weeks (n = 10).

The HRNF task has been described in detail previously [9]. Briefly, rats were trained for 5 min/day, 5 days/week for 10–12 days to learn to retrieve a 45-mg pellet of food from a shoulder-height portal at a reach rate of 4 reaches/min, making this a high repetition task (i.e. faster than 30 s/cycle) [33]. After this initial training period, task rats then performed the HRNF task for 2 h/day in four 30 min sessions/day on 3 days/week for 3 to 6 weeks in customized operant behavioral apparati (Med Associates, St Albans, VT), described in detail previously [9]. Retrieval of the 45 mg food pellet from the portal was estimated as <5 % maximum pulling force, making this a negligible force task [34]. Rats were allowed to use their preferred reach limb to reach and retrieve the food pellet, hereafter referred to as the reach limb. Tissues used in this study were from the reach limbs.

Twelve animals were euthanized with an overdose of sodium pentobarbital (120 mg/kg body weight) and forearm flexor muscles were collected the flexor digitorum mass from FRC, and 3- and 6-week HRNF rats (n = 3/group). Each muscle belly was divided into two longitudinal parts; half was used for RNA extraction and half for protein extraction. The half for RNA extraction was put into RNAlater RNA Stabilization Reagent (QIAGEN, Valencia, CA) for 2 h at room temperature and then stored at -80 °C. Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA). The concentration of each RNA sample was determined using a spectrophotometer and the integrity was monitored on 1 % formaldehyde denatured gels. After confirming RNA integrity on an agarose gel, cDNA was prepared from mRNA extracts from the above tissue samples (n = 3/group) using a High Capacity cDNA Reverse Transcription kit (Applied Biosystems™, Foster City, CA). PCR primer sets for the following were used: OA (Cat# PPR45879B, SABiosciences, Frederick, MD) and GAPDH (Cat# PPR51520A, SABiosciences, Frederick, MD). Quantitative real time PCT (qPCR) was then performed in duplicate for 20 μl reactions each containing 1 μl cDNA reaction mix, 100nM of each primer and 10 μl 2x SYBR® Green PCR Master Mix method (Applied Biosystems™, Foster City, CA), on an ABI 7500 Fast Real-Time PCR system (Applied Biosystems™). PCR cycles consisted of an initial cycle of 50 °C for 2 min and the second cycle of 95 °C for 10 min, followed by a two-step program of 95 °C for 15 s and 60 °C for 1 min for 40 cycles. Using GAPDH as the internal control, relative gene expression among samples was calculated by comparison of C_t (threshold cycle) values. A dissociation curve was checked for each qPCR run to confirm specific amplification of target RNA. The PCR primer set (OA) used was purchased from SABiosciences.

Animals were euthanized with an overdose of sodium pentobarbital (120 mg/kg body weight) before muscle and tendon collection from the flexor digitorum mass from FRC rats (n = 8) and from rats that performed the HRNF task for 3 weeks (n = 4) or 6 weeks (n = 6). As described above, the muscles were divided into half longitudinally. The half for protein analysis was snap frozen and stored at -80 °C until homogenization and preparation for protein analysis using previously described methods [35, 36]. Total protein was determined using BCA-200 protein assays (Bicinchoninic Acid, Pierce, Rockford, IL). Then, 30 μg of protein sample was mixed with 5X Laemmli sample buffer (Bio-Rad, Hercules, CA) mixed with denaturing buffer, 25 % β-mercaptoethanol, heated to 100 °C for 5 min, subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred at 100 V for 1 h at 4 °C to an nitrocellulose membrane (Bio-Rad) using semi-dry transfer apparatus (Bio-Rad). The membranes were blocked with 5 % non-fat milk in Tris-buffered saline (TBS)-0.1 % Tween-20 (TBST) for 1 h at room temperature and then incubated with the following primary antibodies: anti-rabbit beta Actin (1:1000; Sigma, St Louis, MO), anti-mouse GAPDH (1:500; Santa Cruz Biotechnology, Delaware Avenue, CA), anti-mouse Hsp72 (1:300; Stressgen, Ann Arbor, MI), anti-rabbit MMP-1 (1:500; Abbiotec, San Diego, CA), anti-mouse MMP-2 (1:200; Abcam, Cambridge, MA), anti-rabbit MMP3 (1:200; Novus Biologicals, Littleton, CO), anti-mouse MMP-13 (1:200; Abcam), anti-rabbit OA/GPNMB (1:500; Bioss Inc. Woburn, MA) and custom made anti-chicken OA (1:250), which was produced as previously described in [37], in blocking buffer (same as above) overnight at 4 °C. The blot was then washed with 1X TBS-0.1 % Tween-20 (TBST) and incubated with the following horseradish peroxidase-conjugated secondary antibodies: donkey anti-chicken, anti-mouse or anti-rabbit (1:5000; Jackson Immunoresearch, West Grove, PA), in blocking buffer for 1 h at room temperature. The blot was washed again with TBST, incubated with SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, Rockford, IL) and exposed to film. Quantification of the bands was performed using either Image J or myImageAnalysis v2.0 software (Thermo Scientific).

For protein analysis, tissues from above animals that had been prepared for protein analysis were used in ELISAs to determine HSP72 and inflammatory cytokine levels. All muscle and tendon samples were homogenized separately with 0.5 to 1.0 ml RIPA buffer ((NaCl, KCl, NaH₂PO₄, KH₂PO₄, and DDH₂O + NaOH) plus EDTA free complete protease inhibitor cocktail tablets (Roche Diagnostics, GMPH, Germany) using a PowerGen 125 Homogenizer. Tissue homogenates were centrifuged at 14,000 rpm for 15 min. at 4 °C. Total protein was determined using BCA-200 protein assays (Bicinchoninic Acid). For ELISA, tissue lysates (50 microliter aliquots) were analyzed using a commercially available single-plex ELISA kit for HSP72 (EKS-700, Stressgen, purchased before its purchase by Enzo Life Sciences, Inc, Farmingdale, NY), according to the manufacturers’ protocol. Tissue lysates were also analyzed for IL-1alpha, IL-1beta, TNF-alpha and IL-10 using commercially available ELISA kits according to manufacturer’s protocols (each from BioSource International). Each sample was run in duplicate. Data (ng protein of HSP72 and pg protein for the cytokines) were normalized to μg total protein.

Animals received an overdose of sodium pentobarbital (120 mg/kg body weight) before being perfused transcardially with first sterile saline and then 4 % paraformaldehyde in 0.1 M phosphate buffer (pH 7.4): FRC (n = 6), 3-week HRNF (n = 6) and 6-week HRNF (n = 4). Tissues were collected and postfixed “en bloc” by immersion overnight. A proximal portion of the muscle mass was then removed with a scalpel for cross-sectional sectioning while the remaining flexor digitorum muscles and tendons were separated as a flexor mass from the bones for longitudinal sectioning, as previously depicted [38]. All tissues were cryoprotected in 30 % sucrose in phosphate buffered saline (PBS) before frozen-sectioning using a cryostat into 15 μm cross sectional or longitudinal slices. Sections were then placed onto charged slides (Fisher, Super Frost Plus) and allowed to dry overnight before storage at -80 °C.

Sections on slides were treated with 3 % H₂O₂ in methanol to block for endogenous peroxidase for 30 min (if the tissues were to be used for HRP-DAP visualization), washed in PBS, then permeabilized with 0.05 % pepsin in 0.01 N HCL prior to blocking with 10 % goat serum in PBS for 20 min at room temperature. Sections on slides were then probed in batches with the following primary antibodies: anti-HSP72 (1:1000, C92F3A-5, Enzo Life Sciences, Inc) or anti-osteoactivin (1:350, custom made anti-chicken OA) produced and verified, as previously described [37]. On the 2^nd day after washing, one set of sections from all groups that had been incubated with anti-OA antibody was incubated with goat anti-chicken secondary IgG antibody conjugated to HRP (Jackson Immunoresearch Laboratories, West Grove, PA; diluted 1:100 in PBS, and incubated 2 h at room temperature before washing in PBS) and visualized using DAB (Fast DAB, Sigma). These sections were counterstained lightly with eosin, then dehydrated and coverslipped with DPX mounting medium for bright field microscopy (for HRP-DAB). A second set of sections was incubated with anti-HSP72 antibody and then with a donkey anti-mouse secondary IgG antibody conjugated to Cy3 (red fluorescent tag) diluted as above (Jackson Immunoresearch Laboratories). A third set of slides was incubated with both anti-OA antibody and anti-HSP72 and then with appropriate secondary antibodies conjugated to Cy2 (green fluorescent tag) or Cy3 (red fluorescent tag) (Jackson Immunoresearch Laboratories). The fluorescent tag-labeled sections were washed with PBS and coverslipped with 80 % glycerol in PBS for epifluorescence microscopy.

The specificity of the OA and HSP72 antibodies are shown in western blots that are part of this study. In addition, negative control staining was performed by omitting either the primary antibody or the secondary antibody. No labeling was observed as a result of incubation of tissues with serum and then secondary antibodies alone (data not shown). Preabsorption controls were performed to demonstrate if the antibodies bound specifically to the antigen of interest. Specifically, specificity of the OA and HSP72 antibodies was determined via the use of recombinant human Osteoactivin/GPNMB/Fc Chimera (Cat# 2550-AC, R&D Systems, Minneapolis, MN) and recombinant human HSP70 protein (Product# SPP-755, StressGen, Farmingdale, NY; purchased from StressGen before its purchase by Enzo Life Sciences, Inc.), respectively. A ten-fold excess of purified protein was pre-incubated with the matching antibody overnight at 4 °C, the mixture centrifuged and then the pre-absorbed antibody supernatant was incubated with the tissues (after pepsin and goat serum treatments) similar to that described above before washing and incubation with secondary antibodies. No labeling was observed in the tissues for either pre-absorbed antibody (data not shown), matching previously published specificity of the custom made OA antibody in bone tissues [37].

The percent area with OA and HSP72 immunostaining in muscles and tendons was quantified from HRP-DAB stained and fluorescent stained slides, respectively. This quantification was performed using an image analysis program (Bioquant Osteo II, Nashville, TN) and previously described thresholded pixel count methods [39]. The person performing the quantification was blinded to group assignment.

Statistical analyses were performed using Prism 4 and 5 (GraphPad Software, La Jolla, CA). Univariate ANOVAs were used to analyze quantitative PCR, Western blot densitometry and ELISA results for differences in analytes between groups. Then, post hoc analyses were carried out using the Bonferroni test for multiple comparisons with FRC used as the control group and adjusted p values are reported. An adjusted p value of < 0.05 was considered significant for all analyses. Data are presented as the mean ± standard error of the mean (SEM).

Utilizing quantitative real time PCR (qPCR) analysis, we observed a significant up-regulation of OA gene expression in muscles of 6-week HRNF animals (p < 0.01) compared to food restricted only control (FRC) rats (Fig. 1a). These findings were confirmed using Western blot analysis, which showed progressive increases in OA protein levels with HRNF task performance compared to FRC rats (Fig. 1b). The 65 kDa and 115 kDa molecular weight bands of OA were increased (p < 0.01) while the 47 kDa band was not increased with HRNF task performance (Fig. 1b-d). To further examine the expression of OA in the muscle tissue of our rat model, we performed immunohistochemistry and subsequent quantification of OA in the flexor digitorum muscles of HRNF rats. OA immunostaining was increased in 6-week HRNF muscles (p < 0.01) compared to controls (Fig. 1e). OA was localized to the sarcolemma and to macrophage-like cells located between individual muscle fibers in 6-week HRNF muscles yet was absent in control muscles (Fig. 1f and g).

Western blotting detection showed that MMP-1, -2 and -3 were significantly increased in 3- and 6-week HRNF rat muscles compared to FRC rats (Figs. 2 and 3). Specifically, the 54 kDa molecular weight band of MMP-1 (Fig. 2a and b) and the 66 kDa active form of MMP-2 (Fig. 2c and e) were increased in HRNF rat muscles (p < 0.05 each), particularly by 6 weeks of task performance. The 72 kDa latent form of MMP-2 did not change significantly across groups (Fig. 2c and d). Expression levels of the active forms of MMP-3 (45 kDa and 28 kDa) also increased significantly in 6-week HRNF rats (p < 0.05 each) compared to 3-week HRNF and FRC rat muscles (Fig. 3a, c and d). In contrast, the latent form of MMP-3 (57–59 kDa) did not change (Fig. 3a and c), and the level of active MMP-13 (48 kDa) decreased with HRNF task performance (p < 0.01) compared to controls (Fig. 3e and f).

ELISA analysis showed significant increased HSP72 protein in 6-week HRNF muscles (p < 0.05) compared to controls (Fig. 4a). This result was supported by Western blot analysis that confirmed the expression and increase of the HSP72 protein in 6-wk HRNF rat muscles compared to controls (Fig 4b). We also evaluated the expression of OA and HSP72 using immunohistochemical methods. Little to no HSP72 or osteoactivin was observed in FRC rat muscles (representative examples shown in Fig. 4c and d). Double labeling immunohistochemistry in 6-week flexor digitorum muscles demonstrated increased expression of HSP72 (in red; Fig. 4e, g) and OA (in green; Fig. 4f, g), with co-localization in the sarcolemma surrounding the myofibers (Fig. 4g). Quantification of HSP72 immunostaining confirmed its increased immunoexpression in 6 week HRNF rat muscles compared to controls (Fig. 4h). No DAPI labeled cell bodies were visualized within the center of any of the myofibers at this 6-week HRNF time point (data not shown), indicating that the myofibers were not injured.

Next, we examined the expression of OA and HSP72 in flexor digitorum tendons. Western blot analysis showed that only the 47 kDa form of OA was increased in tendons of 6-week HRNF task performance (p < 0.01) compared to FRC rats (Fig. 5a and b). Immunoexpression of OA and its quantification showed significantly increased OA immunostaining in 6-week HRNF tendons (p < 0.05) compared to control rat tendons (Fig. 5c-f). OA was localized to tenocytes within 6-week HRNF tendons and to fibroblast- and macrophage-like cells located within the surrounding connective tissue (epitendon), but was absent in control tendons and epitendon (Fig. 5d-f ). Similarly, HSP72 immunostaining was significantly increased in tenocytes located within 6-week HRNF tendons compared to controls (Fig. 5g-i).

Western blot analysis showed the level of MMP-1 (54 kDa) was increased in 6-week HRNF rat tendons (p < 0.05), compared to 3-week HRNF and FRC rat tendons (Fig. 6a and b). Note the variability of MMP-1 in the 3-week HRNF tendons (Fig. 6a). This was a consistent finding. We next found that the latent (57–59 kDa) and active (45 kDa) forms of MMP-3 were increased significantly in 6-week HRNF rat tendons (p < 0.05) compared to 3-week HRNF and FRC rats (Fig. 6c-e). Variable expression of the latent forms of MMP-3 was observed in FRC and 3-week HRNF tendons, while the active form of MMP-3 (45 kDa) was consistently increased in the 6-week HRNF tendons.

Lastly, we also observed transient yet significant increases of IL-1alpha in 3-week HRNF tendons (Fig. 7b) and IL-1beta in 3-week HRNF muscles (Fig. 7c). We also observed a very small but significant decrease in IL-1alpha in 6-week HRNF muscles compared to controls (Fig 7a). No increase was observed for IL-1beta in tendons (Fig. 7d). TNF-alpha and IL-10 did not alter significantly with task performance in either tissue compared to controls (data not shown).