Age dependent increase in the levels of osteopontin inhibits skeletal muscle regeneration

Osteopontin expression is increased in old macrophages upon muscle injury

To analyze age specific changes in muscle regeneration, Young and old Tibialis Anterior (TA) and Gastrocnemius (Gastroc) muscles were injured by cardiotoxin (CTX) and compared by tissue cryosections at different time points after injury. As seen by H&E staining, pronounced myofiber disintegration (Figure 1 A) and infiltration of mononucleated CD11b+ inflammatory cells (Figure S1 A) into the injury site were observed at 3 days after injury irrespective of animal age. CD11b has been extensively used as macrophage marker [7, 15]. In complete agreement with previously published work [30] while young muscle was robustly repaired at 5 days post injury (5DPI) with new regenerating myofibers, tissue repair was significantly worse in the old mice, with fewer newly formed myofibers and evident fibrosis (Figure 1 A).

Macrophage population is predominant in the injured muscle [31] and OPN is known to be expressed in a variety of cell types including macrophages [32-34]. Hence, we hypothesize that OPN, which becomes elevated in a muscle of MDX mice [24] and in the injured muscle of wild type mice [26], might play an important role in physiological muscle aging and compared the levels of this cytokine in the intramuscular macrophages of young versus old mice. We isolated CD11b+ macrophages from young and old muscle (injured for 3 and 5 days), using magnetic bead column (MACS). This method yielded ~90% cell purity (Figure S2 A). Using Western blotting, we found approximately 8-fold increase in ~30 KDa OPN in CD11b+ macrophages that have been isolated from old, as compared to young muscle (Figure 1 B and C). OPN undergoes extensive posttranslational modifications and enzymatic cleavage at different sites, which gives rise to various cleaved forms ranging from ~25-75 KDa; such OPN variants regulate diverse cell behavior and expression of multiple genes [35, 36]. To confirm the age-specific increase in OPN, we also performed qRT PCR studies and found that OPN mRNA was significantly elevated in CD11b+ cells that were isolated from old injured muscle, as compared to young (Figure 1 D). FACS analysis quantitatively determined that the numbers of CD11b+ inflammatory leukocytes are similar in the injured muscles of young and old mice (Figure S1 B).

These data suggest that while the numbers of CD11b+ cells that infiltrate injured skeletal muscle do not significantly change with age, OPN expression becomes elevated in the intramuscular inflammatory macrophages of old mice, as compared to young.

OPN is increased in the serum and myofiber niche of injured old mice

OPN is a secreted molecule that is present in soluble form in many bodily fluids including milk, blood and urine [21, 22]. To analyze the age specific systemic changes in OPN, ELISA was performed on serum obtained from young and old mice (resting or after muscle injury). In correlation with the high levels of OPN produced by the old intramuscular CD11b+ cells, serum levels of OPN were also found to be increased with age and interestingly, only when old mice were injured (Figure 2 A). These results suggest either that CD11b+ cells, which infiltrate injured muscle of old mice secrete elevated levels of OPN into blood circulation or that other cells in the old injured animals contribute high levels of soluble OPN into blood sera.

To confirm and extrapolate these results and to determine whether the production of OPN by CD11b+ cells and / or elevated levels in blood serum translate into higher levels of this cytokine in the local niches of old muscle stem cells, we performed co-immunostaining of OPN with laminin (a key component of myofiber basement membrane) in 10 μm muscle cryosections. As shown in Figure 2 B, we found more OPN in the ECM of old injured muscle, as compared to young (particularly, at 5 DPI). Resting muscle that does not have infiltrating inflammatory leukocytes did not have detectable OPN in either young or old animals (Figure 2 B).

To analyze whether old muscle cells (themselves) up-regulate production of OPN, we performed western blotting on the lysates of injured myofibers and muscle stem cells. As shown in Figure 2 C (quantified in Figure 2 E), no statistically significant age-specific elevation of OPN was detected in muscle fibers; however, as compared to young, higher levels of OPN were observed in the old satellite cells (Figure 2 D, quantified in Figure 2E). As previously published, our satellite / myogenic cell isolation method yields high cell purity [37] which we also confirmed here for both young and old cells by Pax7 immunostaining (Figure S2 B). Further, ~1.5 fold increase in OPN mRNA levels was seen in old myogenic stem cells when compared to young (Figure 2 F).

These data demonstrate that OPN becomes persistently elevated in the local and systemic niches of old muscle stem cells after tissue injury.

Age-specific elevation in OPN inhibits myogenic responses

To test the functional significance of the uncovered age-specific up-regulation of OPN and to analyze the specific role of macrophages, we performed co-cultures of myofibers with their associated muscle stem cells and CD11b+ macrophages in transwell plates. This system allows the transport of secreted molecules from two different compartments but does not permit direct cell-cell contacts between macrophages and either myofibers or the myogenic cells (Figure 3 A). In this experiment, young and old satellite cells were cultured with their own local niche (myofiber) and in their own serum, but the age of the CD11b+ cells was either isochronic (young for young and old for old) or heterochronic (young for old and old for young) (Figure 3 A). Importantly, OPN was neutralized in some of the co-cultures and exogenous OPN was added to some of the co-cultures (Figure 3 A). The regenerative capacity of satellite cells was determined based on their ability to quickly (overnight) generate colonies of desmin+/BrdU+ myogenic cells progenitor cells [30].

As shown in Figure 3 B and C, young satellite cells cultured with young serum and young macrophages (YYY) produced many more myogenic colonies than old satellite cells cultured with old serum and old macrophages (OOO). Remarkably, presence of young CD11b+ cells rejuvenated myogenic responses of old satellite cells, even when old serum and old myofibers were present (OOY); and neutralization of OPN enhanced myogenicity of old satellite cells in co-cultures with old myofibers, old serum and old CD11b+ cells (OOO+ nOPN) (Figure 3 B and C). Confirming the negative role of OPN on myogenicity, recombinant OPN reduced the regenerative capacity of young satellite cells, which were cultured with young myofibers and young serum (YYY+ rOPN) (Figure 3 B and C). In agreement with the data on the elevated levels of OPN in the circulation of old injured mice, neutralization of OPN enhanced myogenesis in the cultures of old satellite cells, old myofibers and old sera (even when CD11b+ cells were not present) (Figure 3 C). Interestingly, transwell co-cultures where old macrophages were co-cultured with young myofiber explants in young serum did not show significant decrease in myogenicity of young satellite cells (YYO) (Figure 3 C). These results indicate that young myofiber niche and serum counteract the negative effect of old macrophages, which is consistent with previous report on heterochronic parabiosis [29]. Further, administration of nOPN antibody to young myofiber explants that were cultured with young serum and young macrophages (YYY+ nOPN), did not show a significant increase in myogenicity of young satellite cells (Figure 3 C), indicating that myogenic responses might be already maximal for the young muscle stem cells in context of their young local and systemic niches.

To validate these data further, we adsorbed a range of OPN concentrations into matrigel substrate of young and old myofiber explant cultures (Figure S3 A). The diminished generation of desmin+/BrdU+ colonies followed by the decrease in the numbers of eMyHC+ differentiated myotubes was observed in both young and old myofiber explants cultures that were exposed to ectopic OPN, as compared to control vehicle (Figure S3 B, C, D and E). These results substantiate the negative effects of OPN on myogenic cell proliferation and subsequent differentiation. This data also confirm that desmin/BrdU assay correctly detects proliferating fusion-competent myoblasts, which are the progeny of activated satellite cells and are committed to form eMyHC+ myotubes [30].

Neutralization of OPN in vivo improves muscle regeneration and up regulates myogenic markers in old mice

The anti-myogenic effects of OPN that manifested in cultured satellite cells prompted us to test if neutralization of OPN would enhance and exogenous OPN would inhibit muscle regeneration in vivo. TA muscle of young and old mice were injured with CTX and 24-30 hours later neutralizing OPN antibody was injected intramuscularly into left TA of old mice, recombinant OPN was injected into left TA of young mice, while control Goat IgG and PBS were injected into right TAs of these animals daily for 4 days (Figure 4 A). TA muscles were dissected from young and old mice on day 5 post injury and assayed for the numbers of newly formed myofibers by H&E staining and by the immunofluorescence for eMyHC in 10 μm cryosections (Figure 4 B and D). As compared to control injections, neutralization of OPN in old mice enhanced muscle regeneration by 25%, while exogenous OPN decreased the regenerative capacity of young muscle by 20%, (Figure 4 C and D). Consistently with these results, neutralization of OPN significantly increased the levels of MyoD, Myogenin and eMyHC in the injured / regenerating muscle of old mice whereas exogenous OPN moderately decreased MyoD, Myogenin and eMyHC gene expression in the muscle of young mice (Figure 4 D). These results suggest that attenuation of OPN can enhance regeneration of aged muscle, while exogenous OPN inhibits repair of young muscle in vivo.

Animal strains

C57BL6/J (2-3 months old) and C57BL/6J (22-24 month old) male mice were obtained from pathogen free breeding colonies at The Jackson Laboratories and NIH respectively. Animals were housed at the Northwest Animal Facility, University of California, Berkeley and procedures were performed in accordance to administrative panel on the Office of Laboratory Animal Care, UC Berkeley.

Reagents and antibodies

BrdU, ECM and DNase1 were from Sigma while recombinant mouse osteopontin was purchased from R&D. Antibodies to BrdU (ab6326; rat monoclonal), osteopontin for western blotting (ab8848; rabbit polyclonal), GAPDH (ab 9483; goat polyclonal) and Desmin (ab 15200; rabbit polyclonal) were from Abcam. Laminin (Rat monoclonal) and Actin (rabbit polyclonal) antibody was from Sigma. CD11b (rat monoclonal), goat normal IgG and osteopontin (goat polyclonal) for immunostaining and in vivo neutralization were from R&D, Pax7 and eMyHC was from DSHB; secondary HRP antibodies were purchased from Santa Cruz. For flow cytometry, Cd11b-FITC and rat IgG2b-FITC was from eBiosciences while secondary fluorophore antibodies were from Molecular Probes, Invitrogen.

Western blotting and ELISA

For whole muscle lysate preparation, Tibialis Anterior (TA) or Gastrocnemius (Gastroc) muscle was homogenized in Miltenyi Tissue dissociator instrument in Tissue lysate buffer (Thermo Scientific) containing 1X phosphatase inhibitor (Roche) and 1X Protease inhibitor (Roche). For cells, lysis was performed in RIPA buffer (50mM Tris-Cl pH7.6, 150mM NaCl, 0.1% SDS, 1% NP-40 containing phosphatase inhibitor and protease inhibitors). The protein concentration was determined by Bradford Assay. Lysates were re-suspended in 1X Laemmli buffer (BioRad) containing beta-Mercaptoethanol, boiled for 5 minutes at 1000C, and either run on precast 4-20% Gels or 10% Criterion gels from BioRad. The proteins were then transferred to nitrocellulose membrane (GE) for 2 hours and blocked in 5% Milk in PBS containing 0.05% Tween-20. Protein expression was visualized using ECL-Prime reagent (Amersham), detected by BioRad Gel Doc/Chemi Doc imaging system and analyzed by Image J.

For blood serum collection, young and old mice blood was collected and incubated at 370C for 15 minutes with gentle agitation. The tubes were spun down for 3 minutes at 5000 rpm and supernatant collected. For osteopontin ELISA, R&D kit was used according to manufacturer's recommendation.

Immunofluorescence and Histological analysis

Immunofluorescence histological analysis was carried as described earlier [42]. In brief, cells were pulsed with 0.01mM BrdU for 4-5 hours, fixed with 4% PFA for 15 minutes at room temperature followed by permeabilization with 0.25% Triton-X 100 for 12 minutes. This was followed by 0.2 units/ul DNase1 treatment for 25 minutes, washed with 1XPBS and blocked for one hour in blocking buffer (1%BGS + 1% BSA + 0.1% Na-Azide in 1XPBS) followed by primary antibody incubation in blocking buffer for 2 hours or overnight and secondary antibody incubation for 1 hour in blocking buffer with Alexa fluorophore conjugated species specific secondary antibody (Invitrogen). For muscle sections, harvested TA muscle was soaked in 20% sucrose in PBS for few hours, embedded in OCT (Tissue-Tek), frozen in isopentane cooled liquid nitrogen and cryosectioned at 10 μm (Shandon cryostat). Sections were aligned on super frost plus slides and subsequently stained with Hematoxylin and Eosin to visualize areas of injury and determine muscle regenerative potential. The number of regenerative myofibers/mm2 throughout the muscle was calculated by Image J software. For immunostaining, muscle sections were post fixed with 4% PFA for 10 minutes and permeabilized with 0.25% TritonX-100 for 12 minutes and blocked with CAS (Invitrogen) for 30 minutes. The primary antibody incubation in CAS block was carried either at room temperature for 2 hours or overnight at 40C followed by respective fluorophore conjugated secondary antibodies. Cells were mounted with mounting media containing DAPI (Prolong Gold Antifade, Invitrogen) to visualize nuclei in all immunostaining experiments.

CD11b cells isolation and Transwell culture

For CD11b+ cell isolation, 10 and 30 ul of cardiotoxin (CTX, 0.25mg/ml) was injected in TA and Gastroc of young and old mice and muscles were harvested after 3 days and 5 days post injury. Mononuclear cells were released from injured muscles by trituration after incubation in Collagenase type I (600 units/ml). This was passed through 100 and 40 μm strainer respectively and spun at 500g for 5 minutes at 40C. The pellet was resuspended in ice cold 2% BSA in PBS and blocked with Fc receptor blocking solution (Miltenyi Biotec) for 10 minutes followed by incubation with CD11b magnetic beads (Miltenyi Biotec) for another 15 minutes at 40C. The CD11b+ cells were sorted using Mini MACS column (Miltenyi Biotec) according to manufacturer recommendations. For transwell culture, CD11b+ immune cells isolated by MACS were plated isochronically or heterochronically on transwell insert of 0.3 μm PE membrane in OptiMEM medium whereas myofiber explants (triturated myofibers with associated satellite cells) were cultured on lower chamber coated with ECM in 5% isochronic mouse serum in OptiMEM with and without rOPN and or nOPN for 24 hours. The cultures were pulsed with BrdU (0.01mM) for 4 hours followed by removal of transwell insert and myofiber washing while adherent myofiber associated stem cells were fixed with 4% PFA and co-stained with Desmin and BrdU antibodies.

In vivo muscle experiments. For in vivo muscle experiments, TA muscle was injured with CTX in young and old mice and 30-40 hours after injury, recombinant osteopontin (4ug/ml) was injected in one leg of injured TA of young mice whereas control PBS was injected in other leg of same mice while neutralizing osteopontin antibody (5ug/ml) in one leg and control goat IgG in other leg was injected in old mice each day for a total period of 3 days followed by muscle dissection which were processed for RNA and protein lysate or frozen down for cryosectioning. For 10 μm sections, whole TA muscle was sectioned and stained for H&E to visualize the site of injury and newly formed regenerating myofibers were counted. All the experiments were done in 3 independent young and old mice.

qRT-PCR assay

For qRT-PCR, RNA was isolated from MACS sorted CD11b+ cells using RNeasy purification kit from Qiagen according to manufacturer instructions. For whole muscle, tissues were kept in RNA later and homogenized using Miltenyi Biotec Tissue dissociator and further RNA was isolated using Qiagen kit. 500ug of RNA was reverse transcribed using SA Biosciences RT Kit. For qRT-PCR, cDNA obtained was diluted to 1:5 and reaction was set using BioRad iQ5 SybrMix in iQ5 BioRad cycler. All reactions were run in triplicate from at least three independent experiments and gene expression analyzed using delta Ct values. The list of primers used for qRT-PCR is provided as supplementary table S1 in supplementary section.

Statistical analysis

At least 3 independent set of experiments were performed and student t-test was used to calculate significance *** p≤0.01 and ** p≤0.05. Data was plotted as mean ±S.E.M.