Identification and Function of Fibrocytes in Skeletal Muscle Injury Repair and Muscular Dystrophy

Abstract

We identified and characterized the function of CD45⁺/Collagen I⁺ fibrocytes in acutely injured skeletal muscle of wild-type and Ccr2−/− mice, and in quadriceps and diaphragm muscles of mdx^5cv mice, a mouse model for Duchenne muscular dystrophy. Fibrocytes were not detected in peripheral blood of wild-type mice after acute muscle injury or mdx^5cv mice. Fibrocytes were detected in acutely injured muscles and in mdx^5cv quadriceps and diaphragm muscles. These cells expressed F4/80 and CCR2, and they were mostly Ly6C^low. They expressed a low level of collagens but a high level of pro-fibrotic growth factors as compared with intramuscular fibroblasts. Fibrocyte expression of collagens and pro-fibrotic growth factors was not increased in Ccr2−/− mice as compared with wild-type controls. Fibrocyte expression of both pro-inflammatory and pro-fibrotic cytokines was significantly higher in mdx^5cv diaphragm than in mdx^5cv quadriceps. In co-cultures, fibrocytes from the mdx^5cv diaphragm stimulated a higher level of fibroblast expression of extracellular matrix genes than those from the mdx^5cv quadriceps. Our findings suggest that intramuscular fibrocytes most likely originate from infiltrating monocytes/macrophages and differentiate within injured muscles. They likely contribute to the normal muscle injury repair by producing growth factors. They do not appear to contribute to the persistent muscle fibrosis associated with poor injury repair in Ccr2−/− mice. But they likely contribute to the persistent inflammation and progressive fibrosis in mdx^5cv diaphragm.

INTRODUCTION

Skeletal muscle injury is a commonly encountered clinical condition, which can be acute or chronic, depending on causes. Acute muscle injury can be caused by trauma, ischemia, freeze, toxins, or chemicals. Chronic muscle injury can be associated with inflammatory myopathies, toxic myopathies, or muscular dystrophies. Skeletal muscle has excellent regenerative capacity, and the injury repair involves an inflammatory response, myogenic cell (satellite cell) activation and differentiation, revascularization, and extracellular matrix (ECM) remodelling. Previous studies by our group and others demonstrate that acute muscle injury can repair within a few weeks, and the repairing process requires an adequate inflammatory response, which is predominated by monocyte/macrophage (MO/MP) infiltration (1–8). MOs/MPs are recruited into injured muscles mainly via the CC chemokine ligand 2 (CCL2)/CC chemokine receptor type 2 (CCR2) (2–4, 6). The infiltrating MOs/MPs not only phagocytose damaged muscle fibers to allow tissue repair, but also produce myotrophic growth factors to promote muscle regeneration (3, 4). Mice deficient in CCL2 or CCR2 showed poor muscle regeneration with persistent muscle fibrosis following acute injury (2–4, 6). ECM mainly consists of collagens and fibronectin. ECM remodelling with a transient reactive increase of ECM protein deposition in endomysium is important as it provides a structural support to the muscle injury repair. However, persistent excessive ECM protein deposition leads to fibrosis.

Muscular dystrophy refers to a group of genetically-determined progressive muscle diseases. Duchenne muscular dystrophy (DMD) is the most common muscular dystrophy, which affects 1 in 3,500 live male births (9). The disease is X-linked recessive and caused by a defective dystrophin gene, which leads to muscle membrane instability, muscle necrosis, secondary muscle inflammation and fibrosis, muscle dysfunction, and premature death. The disease is devastating with no cure at this point (10). The most commonly used animal model for studying DMD is the mdx mice. Muscle pathology of DMD patients and mdx mice features chronic inflammation and fibrosis, which can directly cause muscle dysfunction and weakness (11–14). In this regard, DMD is also a fibrotic disease. Studies from our group and others have shown that ameliorating muscle fibrosis represents a viable therapeutic approach for DMD, because it can improve muscle function and clinical phenotype (15–19). Inflammation in skeletal muscles of the mdx mice is also predominated by macrophages. It starts around 3 weeks of age. A high level of inflammation persists for 2–3 months, and then subsides spontaneously in limb muscles. Progressive muscle fibrosis predominantly occurs in diaphragm, which correlates with an impaired respiratory function (14, 20, 21). It remains largely unclear, however, why the limb and diaphragm muscles in the mdx mice display different phenotypes.

Fibrocytes are circulation-derived monocytes with features of both fibroblasts and leukocytes (22). They express fibroblast proteins as well as some monocyte markers. Co-expression of collagen I and the pan-leukocyte antigen CD45 (CD45⁺/ColI⁺) has been conventionally used to identify circulating and tissue fibrocytes (23). Fibrocytes can rapidly enter injured tissues to contribute to wound repair (22). They can also play a pathogenic role in tissue fibrosis by producing a low level of ECM proteins, secreting pro-fibrotic growth factors to activate tissue fibroblasts, and may differentiating into effector fibroblasts with an increase of ECM protein expression and loss of CD45 cell marker (23, 24). The contribution of fibrocytes to tissue fibrogenesis has been shown in many fibrotic disease models, including pulmonary fibrosis, liver cirrhosis, and renal fibrosis, among others (25–32). In human patients, the number of circulating fibrocytes correlates with the disease severity of idiopathic pulmonary fibrosis, systemic sclerosis, and liver cirrhosis induced by hepatitis C infection (33–35), which suggests that enumerating circulating fibrocytes may serve as a biomarker for these fibrotic diseases, and targeting fibrocytes may become a useful therapeutic approach for fibrotic diseases.

Based on the important roles of fibrocytes in wound repair and fibrosis in a number of non-muscle tissues, we hypothesized that fibrocytes might also be present in injured muscles to play a physiological role in normal acute skeletal muscle injury repair, and a pathological role in skeletal muscle fibrosis associated with poor acute muscle injury repair in Ccr2−/− mice and chronic muscular dystrophy in mdx mice. Fibrocytes might also become a potential target for treating muscle fibrosis in various settings. To address this hypothesis, we identified and characterized the function of fibrocytes in acutely injured muscles of wild-type (WT) and Ccr2−/− mice, and in quadriceps and diaphragm muscles of mdx^5cv mice.

MATERIALS AND METHODS

Animals

C57BL/6J (CD45.2) mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). Ccr2−/− mice were originally kindly provided by Dr. Israel Charo (36), which were backcrossed nine times with the C57BL/6J mice as previously described (37). ColI-GFP transgenic mice with the GFP expression driven by the collagen I (α1) gene promoter to label with high sensitivity and specificity collagen I α1 producing cells were kindly provided by Dr. David Brenner (38–40). Ccr2−/−/ColI-GFP mice were generated by crossbreeding Ccr2−/− mice with ColI-GFP mice. Mdx^5cv mice, instead of mdx mice, were used because the mdx^5cv mice are in the C57BL/6J background, same as the ColI-GFP and Ccr2−/− mice. Mdx^5cv mice showed muscle pathology comparable to mdx mice (41). Mdx^5cv mice were derived from the Jackson laboratory. Mdx^5cv/ColI-GFP mice were generated by crossbreeding ColI-GFP mice with mdx^5cv mice. C57BL/6J (CD45.1) mice were purchased from the Jackson Laboratory. Our study protocols were approved by the Institutional Animal Care and Use Committee at Icahn School of Medicine at Mount Sinai (New York, NY).

Acute muscle injury

To induce acute skeletal muscle injury, 100 μl 1.2% barium chloride (BaCl₂) was injected into the right quadriceps muscle of each mouse (age 10–14 weeks).

Histopathological analysis

Male mice with acute muscle injury were sacrificed at days 1, 3, 7, and 14 after BaCl₂ injections. Male mdx^5cv and mdx^5cv/ColI-GFP mice were sacrificed at 4 weeks, 3 months, and 6 months. Quadriceps and diaphragm muscles were collected, fresh frozen in liquid nitrogen-cooled isopentane, sectioned at 8 μm, stained with hematoxylin and eosin, and viewed under a bright field microscope.

Immunostaining

Frozen sections of injured muscles from WT and Ccr2−/− mice and of diaphragm and quadriceps muscles from mdx^5cv mice were fixed in 4% paraformaldehyde for 30 minutes and blocked in 5% serum for 2 hours, followed by overnight incubation with the primary goat anti-human collagen III antibody (Southern Biotech, Birmingham, AL, USA, 1:10) for 2 hours. The sections without primary antibody incubation were used as negative controls. The sections were then washed with PBS and incubated with Alexa 488-conjugated donkey anti-goat secondary body (Invitrogen, Carlsbad, California, USA, 1:500) for 1 hour. After stained with DAPI, slides were mounted with mounting medium and the antibody binding was visualized under a fluorescent microscope.

Isolation of bone marrow-derived macrophages (BMMac) and activation of M1 and M2 phenotypes

To obtain macrophages, bone marrow cells were cultured in 6-well plates (1 million cells per well). Each well was filled with 2 ml of media (DMEM, 20% of fetal bovine serum, 30% of the supernatant from L929 cells containing M-CSF, 100 μM of 2-ME, and antibiotics). At day 6, non-adherent cells were discarded along with the conditioned media. The adherent cells were cultured in fresh DMEM media with 10% of fetal bovine serum. The M1 phenotype was induced by treating the BMMac with LPS (100 ng/ml, Sigma-Aldrich) + IFNγ (10 ng/ml, R&D systems) for 6 hours, while the M2 phenotype was induced by the treatment with IL-4 (20 ng/ml, R&D systems) and IL-13 (10 ng/ml, R&D systems) for 24 hours.

Single-cell suspension preparation, flow Cytometry analysis, cell sorting, and cell quantification

Muscle single-cell suspension was prepared by collagenase/dispase digestion (Zhao et al., 2016). Briefly, each quadriceps muscle was minced in 2.5 ml of digestion solution containing 1U/ml of collagenase B and 1U/ml of dispase II (Roche Diagnostics, Indianapolis, IN, USA) in PBS and incubated at 37°C for 1 hour. The reaction was terminated by addition of 10 ml of PBS with 10% of FBS. The mixture was then filtered through 70 μm cell strainer and centrifuged at 250 g for 5 minutes. The pellet was collected and the supernatant was centrifuged again at 250 g for 5 minutes. The pellet was combined with the pellet from the first centrifugation, washed with PBS, and centrifuged at 670 g for 10 minutes. The pellet was resuspended in 3 ml of PBS, filtered through 40 μm cell strainer, layered on the equal volume of Lymphocyte-M solution (Cedarlane, Burlington, NC, USA), and centrifuged at 2095 g for 45 minutes. Cells at the interface were collected, centrifuged at 670 g for 10 minutes, and resuspended in FACS staining buffer (PBS with 2% of normal mouse serum (Invitrogen, Frederick, MA, USA) and 2% of BSA (Sigma, St. Louis, MO, USA)). Blood samples were collected from retro-orbital venous plexus, lysed using red blood cell lysis buffer (BD Bioscience, San Jose, CA, USA), centrifuged at 400 g, and resuspended in FACS staining buffer.

The following antibodies were used for flow cytometry. PerCP- Cy5.5- anti -CD45 was purchased from BD Bioscience (San Jose, CA, USA). APC-anti-CD115, PerCP-Cy5.5-anti-CD45.2, APC-anti-F4/80, and PE-Cy7-anti-Ly-6C were purchased from Biolegend (San Diego, CA, USA). Rat IgG2b against mouse CCR2 (MC-21) was kindly provided by Dr. Matthias Mack (Mack et al., 2001). Biotin-anti-rat IgG2b and Streptavidin-APC-Cy7 were purchased from BD Biosciences. PE-anti-CCR5, PE-anti-CCR7, PE-anti-CD31, and PE-anti-Sca-1 were purchased from eBiosciences (San Diego, CA, USA). Alexa-700-anti-α7 integrin was purchased from R&D Systems (Minneapolis, MN, USA). All flows were done using LSR II (BD Bioscience, San Jose, CA, USA). We used fluorescence-labelled corresponding normal IgG isotypes as negative controls for gating. Data were analysed using Flowjo 8.2.6 (Tree Star, Inc., Ashland, OR, USA).

Cell sorting was performed by the Flow Cytometry Core of the Icahn School of Medicine at Mount Sinai. We used CD45⁺/GFP (Col1)⁺ as a marker for identifying and sorting fibrocytes. We used CD45⁻/GFP (Col1)⁺ as a marker for tissue fibroblasts, as this cell population predominantly contains fibroblasts with a few satellite cells (CD45⁻CD31⁻Sca-1⁻a7-integrin⁺, <10%) and fibro/adipogenic progenitor cells (FAP) (CD45⁻CD31⁻Sca-1⁺CD34⁺, <5%) but no endothelial cells at the time of isolation. We used CD45⁺/F4/80⁺ as a marker for intramuscular macrophages. We used the mice of the same genotype (WT, Ccr2−/−, or mdx^5cv) but without the GFP-Col1 transgene as negative controls to determine the cut-off for the GFP (Col1)⁺ cell population.

For quantification of intramuscular fibrocytes and fibroblasts, we weighed each injured muscle, counted and calculated total cell number within single cell suspension of each injured muscle, and measured the percentages of fibrocytes and fibroblasts in each injured muscle by flow cytometry. The number of fibrocytes or fibroblasts was then calculated by total cell number of each muscle×percentage of fibrocytes or fibroblasts/muscle weight (mg).

Quantitative RT-PCR

Sorted cells were lysed in TRIzol reagent (Ambion, Grand Island, NY, USA). Total RNA was then purified and further cleaned up using RNeasy Micro Kit (Qiagen, Hilden, Germany). Reverse transcription was performed using SuperScriptTM II kit (Invitrogen, Frederick, MA, USA) following manufacturer’s instructions. The cDNA samples were then subjected to real-time PCR using Sybr-green reagent and an Eppendort Realpix4 cycler. The gapdh expression was used as an internal control. Reaction specificity was determined by product melting curves. PCR products were verified by running 3% agarose gels. Data were analysed by ΔΔCt method and presented as Fold Changes. The primer sequences were listed in Table 1.

Table 1.

Primer Sequences for qRT-PCR

Gene name	Primer	Sequence
tnfa	F	5′ - CTTCTGTCTACTGAACTTCGGG - 3′
	R	5′ - CACTTGGTGGTTTGCTACGAC - 3′
il1b	F	5′ - GACGGCACACCCACCCT - 3′
	R	5′ - AAACCGTTTTTCCATCTTCTTCTTT - 3′
il6	F	5′ - GAACAACGATGATGCACTTGC - 3′
	R	5′ - CTTCATGTACTCCAGGTAGCTATGGT - 3′
il10	F	5′ - CAGGACTTTAAGGGTTACTTG - 3′
	R	5′ - ATTTTCACAGGGGAGAAATC - 3′
ifng	F	5′ - TGAGTATTGCCAAGTTTGAG - 3′
	R	5′ - CTTATTGGGACAATCTCTTCC - 3′
tgfb1	F	5′ - GGATACCAACTATTGCTTCAG - 3′
	R	5′ - TGTCCAGGCTCCAAATATAG - 3′
pdgfa	F	5′ - GTCCAGGTGAGGTTAGAGG - 3′
	R	5′ - CACGGAGGAGAACAAAGAC - 3′
pdgfb	F	5′ - CCACTCCATCCGCTCCTTT - 3′
	R	5′ - AAGTCCAGCTCAGCCCCAT - 3′
ctgf	F	5′ - AGGACTGCAGCGCGCAATGT - 3′
	R	5′ - GAGGCCCTTGTGTGGGTCGC -3′
col1a	F	5′ - GCTCCTCTTAGGGGCCACT - 3′
	R	5′ - CCACGTCTCACCATTGGGG - 3′
col3a	F	5′ - AACCTGGTTTCTTCTCACCCTTC - 3′
	R	5′ - ACTCATAGGACTGACCAAGGTGG - 3′
col6a	F	5′ - CGCCCTTCCCACTGACAA - 3′
	R	5′ - GCGTTCCCTTTAAGACAGTTGAG - 3′
fibronectin	F	5′ - AAACTTGCATCTGGAGGCAAACCC - 3′
	R	5′ - AGCTCTGATCAGCATGGACCACTT - 3′
arginase-1	F	5′ - CAATGAAGAGCTGGCTGGTGT - 3′
	R	5′ - GTGTGAGCATCCACCCAAATG - 3′
cd68	F	5′ - CAATTCAGGGTGGAAGAAAG - 3′
	R	5′ - TCTGATGTAGGTCCTGTTTG - 3′
cd163	F	5′ - TCCACACGTCCAGAACAGTC - 3′
	R	5′ - CCTTGGAAACAGAGACAGGC - 3′
cd206	F	5′ - GGTGTGGGCTCAGGTAGT - 3′
	R	5′ - TGTGGTGAGCTGAAAGGT - 3′
fizz1	F	5′ - GATGAAGACTACAACTTGTTCC - 3′
	R	5′ - AGGGATAGTTAGCTGGATTG - 3′
ym1	F	5′ - CTTCTAAGACTGGAATTGGTG - 3′
	R	5′ - GTACAAACCTCATAGTAAGCC - 3′

Cell transfer

Fibrocytes (CD45⁺GFP⁺) and fibroblasts (CD45⁻GFP⁺) from day 7 injured quadriceps of the GFP-Col1 transgenic mice (CD45.2) were sorted and injected into day 6 injured quadriceps of the WT C57 BL/6 (CD45.1) recipient mice (10⁶ cells in100μl PBS/mouse). Over 80% donor cells were viable before injections. The recovery rates ranged from 15% to 20% at the time point analysed. For the no transfer control, an equal volume of PBS was injected instead of cells. After 24 hours, single cell suspensions of injured quadriceps from recipients were subjected to FACS analysis for the expression of GFP and CD45.2.

Fibrocytes and fibroblasts co-culture

Fibrocytes (CD45⁺GFP⁺) sorted from single cell suspensions of quadriceps or diaphragm muscles of 3-month-old mdx^5cv mice were plated together with NIH3T3 fibroblasts at a ratio of 1:1 in DMEM supplemented with 10% of FBS in 6-well culture plates (1×10⁵cells in 2 ml of medium/well). 48 hours after the co-cultures, cells were trypsinized and fibrocytes were depleted by negative selection using the anti-CD45 MACS system (Miltenyi Biotec, San Diego, CA, USA) following the manufacturer’s instruction. The purified NIH3T3 cells were then counted and lysed in TRIzol Reagent. The ECM gene expression by the NIH3T3 cells was analysed by quantitative RT-PCR as described above.

Statistical analyses

GraphPad (GraphPad software, Inc., La Jolla, CA, USA) was used for statistical analyses. All data were presented as mean ± SEM. Two-tailed Students t test was used when comparing two groups, and analysis of variance was performed with Bonferroni correction for multiple comparisons. A p value of <0.05 was considered statistically significant.

RESULTS

CD45⁺/GFP (ColI)⁺ fibrocytes were not detectable in peripheral blood of WT/ColI-GFP mice at any stages after acute muscle injury

We used intramuscular BaCl₂ injection to induce acute muscle injury. BaCl₂ injection usually causes massive muscle fiber necrosis with infiltration of inflammatory cells, which peaks at day 3. Endomysial collagen deposition gradually increases, starting from day 1 and peaking at day 7. After that, clearance of necrotic fibers and resolution of inflammation are accompanied by gradual thinning of endomysium and reduction of ECM protein synthesis. Therefore, there is a transient inflammatory response and a transient increase of ECM protein synthesis and deposition during normal acute muscle injury repair (4). We used the transgenic reporter collagen-α(I)-GFP (ColI-GFP) mice to identify, track, and isolate the collagen I (ColI)-producing cells. We first addressed whether CD45⁺/GFP (ColI)⁺ fibrocytes were present in peripheral blood after acute muscle injury. We performed flow cytometry analysis in the WT/ColI-GFP mice at days 1, 3, 7, and 14 after BaCl₂ injections. The CD45⁺/GFP (ColI)⁺ cells were not detectable in the peripheral blood at any stages after injury (Fig. 1A). Likewise, no CD115⁺/GFP (ColI)⁺ cells were detected in the peripheral blood to indicate collagen I expression by blood monocytes (data not shown).

Figure 1. CD45⁺GFP⁺ Fibrocytes were not detected in peripheral blood but were detected in acutely injured muscles of WT/ColI-GFP mice, they expressed F4/80 and CCR2, and they were Ly6C^low.

A) FACS analysis of CD45⁺GFP⁺ fibrocytes in peripheral blood after acute skeletal muscle injury. B) FACS analysis of CD45⁺GFP⁺ fibrocytes in injured muscles at different time points. C) Changes of fibrocyte and fibroblast numbers (×10³/mg muscle tissue) in injured muscles at different time points after acute muscle injury. D) and E) FACS analysis of the day 7 intramuscular CD45⁺GFP⁺ fibrocytes for the expression of F4/80 (D), Ly6C (D), and CCR2 (E). Histogram shows IgG control (shaded) and CCR2 antibody staining. n=15–20 mice/time point.

CD45⁺/GFP (ColI)⁺ fibrocytes were detected in injured muscles at late stages of injury repair, and they expressed F4/80 and CCR2

To address whether CD45⁺/GFP (ColI)⁺ fibrocytes were present in injured muscles of WT/ColI-GFP mice, we performed flow analysis using muscle single cell suspensions. The CD45⁺/GFP (ColI)⁺ cells were not detected at day 1 but were detected at a very low number at day 3 (Fig. 1B). A significant portion of CD45⁺ cells were positive for GFP (ColI) at days 7 and 14 (Fig. 1B). The number of fibrocytes per mg muscle was markedly reduced from day 7 to day 14 (Fig. 1C). To further determine the nature of the intramuscular CD45⁺/GFP (ColI)⁺ cells, we did flow analysis and found that most of these cells were positive for F4/80 (Fig. 1D) and CCR2 (Fig. 1E) but not CCR5 or CCR7 (data not shown), and most of them were Ly6C^low (Fig. 1D). Given the CCR2⁺/Ly6C^low/F4/80^hi phenotype of the intramuscular fibrocytes, we think that these cells are a subset of macrophages. In the absence of detectable fibrocytes in peripheral blood, we think that the intramuscular fibrocytes are most likely derived from blood via CCR2, initially as CCR2⁺ inflammatory monocytes, and then differentiate into macrophages and collagen-expressing macrophages (fibrocytes) within injured muscle.

Intramuscular fibrocytes expressed a low level of collagen genes but a high level of pro-fibrotic growth factor genes as compared with intramuscular fibroblasts

To address whether intramuscular fibrocytes expressed collagen genes and pro-fibrotic growth factor genes, we performed qRT-PCR of collagen I, collagen III, collagen VI, PDGFα, PDGFβ, TGF-β1, and CTGF genes using the sorted CD45⁺/GFP (ColI)⁺ intramuscular fibrocytes and CD45⁻/GFP (ColI)⁺ intramuscular fibroblasts at day 7, as well as the F4/80⁺ intramuscular macrophages at day 3 (Fig. 2). We used in vitro cultured bone marrow macrophages treated with IFN-γ and LPS (M1) or Il-4 and Il-13 (M2), and NIH 3T3 fibroblasts as controls. As expected, the mRNA expression of the collagen genes was detected in the 3T3 fibroblasts but not in the M1 or M2 cells. The mRNA expression of the collagen genes was barely detectable in the day 3 F4/80⁺ macrophages but was readily detected in the day 7 fibrocytes, although the expression level was much lower than that in the 3T3 fibroblasts and intramuscular fibroblasts (Fig. 2A). The mRNA expression of PDGFα and PDGFβ genes but not TGF-β1 gene was significantly higher in the day 7 fibrocytes than in the day 3 F4/80⁺ macrophages. While the PDGFα gene expression was lower in the fibrocytes than in the fibroblasts, the PDGFβ and TGF-β1 gene expression was significantly higher in the fibrocytes than in the fibroblasts (Fig. 2B). The mRNA expression of CTGF gene was very low in the intramuscular fibrocytes (data not shown). Taken together, the intramuscular fibrocytes expressed a low level of collagen genes but a high level of some pro-fibrotic growth factor genes, including TGF-β1 and PDGFβ.

Figure 2. Intramuscular CD45⁺GFP⁺ fibrocytes expressed a low level of collagens but a high level of pro-fibrotic growth factors.

Quantitative RT-PCR was performed to assess the gene expression by the sorted intramuscular CD45⁺/F4/80⁺ macrophages at day 3 (day 3 MP), intramuscular CD45⁺GFP⁺ fibrocytes at day 7 (day 7 FC), and intramuscular CD45⁻GFP⁺ fibroblasts at day 7 (day 7 FB). NIH3T3 fibroblast cell line and M1 and M2 macrophages were used as controls. A) Expression of the collagen genes. B) Expression of the pro-fibrotic growth factor genes. Fold change refers to the comparison to the NIH3T3 cells, except for PDGFβ gene. Fold change of PDGFβ gene refers to the day 7 fibroblasts, because the level of this gene expression in NIH3T3 cells was extremely low. Each experiment was independently performed twice. Each time, 5 mice/time point were used for cell isolation and RNA preparation. *p<0.05; **p<0.01

Differentiation of intramuscular fibrocytes into fibroblasts was not detected in injured muscles at day 7

To address whether the CD45⁺/GFP (ColI)⁺ fibrocytes could differentiate into CD45⁻/GFP (ColI)⁺ fibroblasts in injured muscles, we did cell transfer experiments by injecting CD45⁺/GFP (ColI)⁺ fibrocytes (Fig. 3A) isolated from day 7 injured muscles of WT/GFP (ColI) mice (CD45.2 background) into day 6 injured muscles of WT mice (CD45.1 background). We used day 7 CD45⁻/GFP (ColI)⁺ fibroblasts (Fig. 3A) as a control. We chose day 7 to examine the transferred cells because the cell numbers of both fibrocytes and fibroblasts peaked at this time point (Fig. 1C), and both numbers reduced quickly after this point along with inflammation resolution and ECM remodeling. While CD45.2⁻/GFP (Col1)⁺ fibroblasts were detected in the injured muscles of the WT recipient mice at day 7 after receiving the donor CD45⁻/GFP (ColI)⁺ fibroblasts at day 6, they were not detected in the injured muscles of the WT recipient mice at day 7 after receiving the donor CD45⁺/GFP (ColI)⁺ fibrocytes at day 6 (Fig. 3B). Only fibrocytes, but not fibroblasts, were detected in the injured muscles of the WT recipient mice when fibrocytes were transferred (Fig. 3B). Therefore, there was no detectable differentiation of CD45⁺/GFP (ColI)⁺ fibrocytes into CD45⁻/GFP (ColI)⁺ fibroblasts within acutely injured muscles from day 6 to day 7.

Figure 3. Intramuscular differentiation of fibrocytes to fibroblasts was not detected in injured muscles at day 7.

A) FACS analysis identifying donor CD45.2⁺GFP⁺ fibrocytes and CD45.2⁻GFP⁺ fibroblasts for sorting. B) FACS analysis for CD45.2⁻/GFP⁺ fibroblasts in WT (CD45.1 background) injured muscles at day 7 after receiving the intramuscular transfer of the CD45.2⁺GFP⁺ fibrocytes (FC transfer) or CD45.2⁻GFP⁺ fibroblasts (FB transfer) at day 6. The control recipients (No transfer) only received intramuscular injections of PBS at day 6. n=10–15 recipient mice/group.

CD45⁺/ColI⁺ fibrocytes were also detected in injured muscles of Ccr2−/−/ColI-GFP mice, and they expressed a similar level of collagen and pro-fibrotic growth factor genes as compared with WT controls

Acute muscle injury cannot be repaired normally in Ccr2−/− mice due to the lack of an adequate inflammatory response. In contrast with the WT mice whose injured muscles were largely repaired at day 14, the injured muscles from the Ccr2−/− mice showed very small regenerated muscle fibers with low grade inflammation and persistent endomysial fibrosis at day 14 (Fig. 4A) (4). We addressed whether fibrocytes were present in injured muscles and contributed to the persistent muscle fibrosis in this setting. By flow cytometry, the fibrocytes were not detected at day 1, but were detected at days 3, 7, and 14, with the number being very few at day 3 and greatly increased at days 7 and 14 (Fig. 4B–C). Compared with the WT/ColI-GFP mice, the number of the intramuscular fibrocytes was markedly reduced in the Ccr2−/−/ColI-GFP mice at day 7, but the low number of the fibrocytes remained at day 14 (Fig. 4C). Quantitative RT-PCR showed that the fibrocyte expression of the ECM protein genes, including collagen I, collagen III, collagen VI, and fibronectin, and the pro-fibrotic growth factor genes, including TGF-β, PDGFα, and PDGFβ, was similar in the Ccr2−/−/ColI-GFP mice and WT/ColI-GFP controls at day 7 (Fig. 4D). Therefore, the intramuscular fibrocytes did not express a higher level of ECM protein or pro-fibrotic growth factor genes to contribute to the persistent muscle fibrosis in the CCR2 deficient mice following acute injury. To further address whether the fibroblasts in injured muscles of CCR2 deficient mice were functionally more active than those in WT mice to correlate with the fibrosis, we compared the ECM and pro-fibrotic growth factor gene expression in day 7 fibroblasts isolated from injured muscles of Ccr2−/−/ColI-GFP and WT/ColI-GFP mice (Fig. 4E). The mRNA expression of collagen III and fibronectin genes was significantly increased in the intramuscular fibroblasts from the Ccr2−/−/ColI-GFP mice, but the mRNA expression of collagen I, collagen VI, TGF-β1, PDGFα, and CTGF genes was not significantly different (Fig. 4E). The PDGFβ mRNA expression was very low in the intramuscular fibroblasts (data not shown). Therefore, compared with the WT controls, the CCR2 deficient intramuscular fibroblasts did appear functionally more active, as they expressed a higher level of collagen and fibronectin genes, but they did not express a higher level of pro-fibrotic growth factor genes.

Figure 4. Intramuscular fibrocytes did not contribute to the persistent muscle fibrosis in Ccr2−/− mice following acute injury.

A) H&E staining and collagen III immunostaining of injured muscles at days 7 and 14 in WT (upper panel) and Ccr2−/− mice (lower panel). B) FACS analysis of intramuscular fibrocytes (CD45⁺GFP⁺) in injured muscles of Ccr2−/−/ColI-GFP mice at different time points. C) Changes of the fibrocyte number (×10³/mg muscle tissue) in injured muscles of WT/ColI-GFP (WT) and Ccr2−/−/ColI-GFP (CCR2KO) mice during acute muscle injury repair. D) & E) Quantitative RT-PCR analysis of ECM and pro-fibrotic growth factor gene expression by intramuscular fibrocytes (D) and fibroblasts (E) sorted from day 7 injured muscles in Ccr2−/−/ColI-GFP (CCR2KO) and WT/ColI-GFP (WT) mice. Fold change refers to the comparison to the wild-type controls. For H&E staining and collagen III immunostaining, n= 5 mice/group/time point. For FACS, n=15–20 mice/time point. For qRT-PCR, each experiment was independently performed twice. Each time, 5 mice/group were used for fibrocytes and fibroblasts isolation and RNA preparation. *p<0.05; **p<0.01; ***p<0.001. Bar = 50μm

CD45⁺/ColI⁺ fibrocytes were detected in diaphragm and quadriceps muscles but not in peripheral blood of mdx^5cv/ColI-GFP mice, and they expressed F4/80 and CCR2

Skeletal muscle inflammation was readily detected in the quadriceps and diaphragm muscles of the mdx^5cv mice at age 4 weeks (Fig. 5A). The muscle necrosis and inflammation persisted at age 3 months with thickened endomysium deposited by collagens, which was more noticeable in the diaphragm (Fig. 5A–B). The inflammation gradually subsided with no progressive fibrosis seen in the quadriceps at 6 months, but the inflammation was persistent with more severe fibrosis seen in the diaphragm at 6 months (Fig. 5A–B). To address whether fibrocytes are present and contributed to the diaphragm fibrosis, we did flow cytometry and gene expression analyses. CD45⁺/GFP (ColI)⁺ fibrocytes were not detected in the peripheral blood of the mdx^5cv/ColI-GFP mice at 4 weeks, 3 months, or 6 months (Fig. 5C), but they were detected in the quadriceps and diaphragm muscles at these stages (Fig. 5D). At 4 weeks, the number of the fibrocytes was higher in the quadriceps and the number of the fibroblasts was higher in the diaphragm. At 3 months, while the number of the fibrocytes was similar in the quadriceps and diaphragm, the number of the fibroblasts was significantly higher in the diaphragm than in the quadriceps. At 6 months, while the numbers of both fibrocytes and fibroblasts were greatly reduced in the quadriceps as compared with the earlier stages, the numbers of the fibrocytes and fibroblasts in the diaphragm were also reduced but to a lesser degree, and both numbers were significantly higher in the diaphragm than in the quadriceps (Fig. 5D). The findings are consistent with the persistent chronic inflammation and progressive fibrosis seen in the diaphragm but not the quadriceps. The intramuscular fibrocytes were positive for F4/80, and the majority of them were Ly6C^low (Fig. 5E). They also expressed CCR2 (Fig. 5F) but not CCR5 or CCR7 (data not shown). The findings suggest that the intramuscular fibrocytes in the mdx^5cv mice are most likely a subset of macrophages, which are originally recruited from circulation as CCR2⁺ inflammatory monocytes, and then differentiate within injured muscles.

Figure 5. CD45⁺/ColI⁺ fibrocytes were detected in mdx^5cv diaphragm and quadriceps muscles but not peripheral blood.

A) H&E staining of quadriceps and diaphragm muscles of mdx^5cv mice at 4 weeks, 3 months, and 6 months. Arrows and asterisks indicate areas of inflammation and fibrosis, respectively. B) Collagen III immunostaining of the quadriceps and diaphragm muscles of mdx^5cv mice at 4 weeks, 3 months, and 6 months. C) FACS analysis of fibrocyte (CD45⁺GFP⁺) in peripheral blood of mdx^5cv/ColI-GFP mice at 4 weeks (4W), 3 months (3M), and 6 months (6M). D) FACS analysis of fibrocytes (CD45⁺GFP⁺) and fibroblasts (CD45⁺GFP⁻) in the quadriceps and diaphragm muscles of mdx^5cv/ColI-GFP mice at 4 weeks (4W), 3 months (3M), and 6 months (6M). E) & F) FACS analysis of intramuscular fibrocytes (CD45⁺GFP⁺) sorted from the mdx^5cv/ColI-GFP quadriceps and diaphragm at 3 months for expression of F4/80 (E), Ly6C (E), and CCR2 (F). Histogram shows IgG control (shaded) and CCR2 antibody staining. For H&E and collagen III immunostaining, n=5 mice/group. For FACS, n=30 mice/group/time point. *p<0.05; ***p<0.001. Bar = 50μm

Fibrocytes in the mdx^5cv diaphragm expressed a higher level of pro-inflammatory and pro-fibrotic cytokines than those in the mdx^5cv quadriceps

The intramuscular fibrocytes are most likely a subset of macrophages, and macrophages can be functionally heterogeneous depending on the tissue environment. We next addressed whether the diaphragm fibrocytes exerted more pro-inflammatory and pro-fibrotic functions than the quadriceps fibrocytes to contribute to the persistent inflammation and progressive fibrosis seen in the mdx^5cv diaphragm but not the quadriceps. We performed qRT-PCR to study the gene expression of M1 and M2 cytokines and cell markers using the fibrocytes sorted from the mdx^5cv/ColI-GFP diaphragm and quadriceps at 3 months. The mRNA expression of the pro-inflammatory cytokines, including TNF-α, IL-1β, and Il-6, was significantly higher in the diaphragm fibrocytes than in the quadriceps fibrocytes (Fig. 6A). The mRNA expression of IFNγ, IL-10, CD163, and Ym1 genes was also significantly increased in the diaphragm fibrocytes. However, the mRNA expression of arginase 1, CD206, Fizz1, and CD68 genes was not significantly different (Fig. 6B). We next addressed whether the diaphragm fibrocytes expressed a high level of the ECM and pro-fibrotic growth factor genes as compared with the quadriceps fibrocytes. qRT-PCR results showed that the mRNA expression of collagen I, collagen III, and collagen VI gene but not fibronectin gene was significantly higher in the diaphragm fibrocytes than in the quadriceps fibrocytes (Fig. 6C). The mRNA expression of TGF-β1 and PDGFα genes was also significantly high in the diaphragm fibrocytes. Likewise, the mRNA expression of collagen and TGF-β1 genes was significantly higher in the diaphragm fibroblasts than in the quadriceps fibroblasts (Fig. 6D–E). While the collagen gene expression was much higher in the fibroblasts, the TGF-β1 gene expression was much higher in the fibrocytes (Fig. 6D). Taken together, compared with the quadriceps fibrocytes in the mdx^5cv/ColI-GFP mice, the diaphragm fibrocytes expressed a high level of both pro-inflammatory and pro-fibrotic cytokines at 3 months. The diaphragm fibroblasts also appeared more functionally active than the quadriceps fibroblasts in terms of the ECM gene expression.

Figure 6. Mdx^5cv diaphragm fibrocytes expressed a high level of pro-inflammatory and pro-fibrotic cytokine genes as compared with mdx^5cv quadriceps fibrocytes.

A) to E) Quantitative RT-PCR of sorted intramuscular fibrocytes (CD45⁺GFP⁺) and fibroblasts (CD45⁻GFP⁺) from the mdx^5cv/ColI-GFP quadriceps and diaphragm at 3 months for the expression of the pro-inflammatory cytokine genes by fibrocytes (A), M1/M2 cell marker and cytokine genes by fibrocytes (B), ECM genes by fibrocytes (C), pro-fibrotic growth factor genes by fibrocytes (FC) and fibroblasts (FB) (D), and ECM genes by fibroblasts (E). Fold change refers to the comparison to the quadriceps fibrocytes or fibroblasts. Each experiment was independently performed twice. Each time, 10 mice were used for fibrocytes and fibroblasts isolation and RNA preparation. *p<0.05; **p<0.01.

Fibrocytes from the mdx^5cv/ColI-GFP diaphragm but not the quadriceps stimulated a high level of ECM gene expression by co-cultured fibroblasts

To further address whether the high level of pro-fibrotic growth factor gene expression by the mdx^5cv diaphragm fibrocytes promoted the ECM gene expression by fibroblasts, we performed fibrocytes and fibroblasts co-culture experiments (Fig. 7). We used the fibrocytes sorted from the mdx^5cv/ColI-GFP diaphragm and quadriceps at 3 months, and cultured with NIH3T3 fibroblasts. The growth rate of the NIH3T3 cells was not changed by the co-cultures (data not shown). However, the mRNA expression of collagen I, collagen III, and fibronectin genes by the NIH3T3 cells was significantly increased by co-culturing with the mdx^5cv/ColI-GFP diaphragm fibrocytes but not the quadriceps fibrocytes. The findings support a pro-fibrotic role for the mdx^5cv diaphragm fibrocytes.

Figure 7. Mdx^5cv diaphragm fibrocytes stimulated a high level of ECM gene expression by co-cultured fibroblasts.

Quantitative RT-PCR analysis of the ECM gene expression by the NIH3T3 fibroblasts cultured with or without the fibrocytes sorted from the mdx^5cv/ColI-GFP diaphragm or quadriceps at 3 months. Fold change refers to the comparison to NIH3T3 cells without co-culture. The experiment was independently performed three times. Each time, 5–7 mice were used for fibrocytes isolation. *p<0.05.

DISCUSSION

Fibrosis with excessive ECM protein synthesis and deposition directly causes tissue dysfunction. Targeting fibrogenesis effectors, the activated tissue fibroblasts, represents an important approach to develop anti-fibrotic therapies. It has been shown in the fibrotic non-muscle tissues that the tissue effector fibroblasts can be derived from a variety of cellular sources, including resident tissue fibroblasts, epithelial-mesenchymal transition, and circulating fibrocytes. Fibrocytes are circulation-derived collagen-producing monocytes with features of both fibroblasts and monocytes. They play both physiological and pathological roles in non-muscle tissues to participate in wound repair and tissue fibrosis (22–24). Our present study demonstrates that fibrocytes are also present in injured skeletal muscles. They likely participate in normal acute skeletal muscle injury repair by producing growth factors to activate the fibroblast function for the reparative ECM remodelling. They also likely contribute to the muscle dystrophy in the mdx^5cv diaphragm by producing a high level of both pro-inflammatory and pro-fibrotic cytokines.

It has been shown that circulating fibrocytes quickly enter injured tissues to participate in wound repair (22). The number of fibrocytes recruited varies depending on the local tissue environment and the type of injury (42, 43). In our skeletal muscle injury models, the differentiation of fibrocytes from monocytes does not appear to occur in circulation; it most likely occurs within injured muscle based on the following findings. First, fibrocytes were not detected in the peripheral blood at any stages after acute muscle injury, or in the peripheral blood of the mdx^5cv mice at any stages studied. Second, the intramuscular fibrocytes were not detected at the early stage (day 1) after acute injury, and the number of these cells markedly increased after day 3, at which point the active muscle recruitment of MOs/MPs was near complete (1). It remains elusive which signals within injured muscle drive the fibrocyte differentiation. Previous studies established a critical role for T cells in fibrocyte differentiation, as the T cell activation profile influenced the fibrocyte differentiation in vitro (44, 45). While the Th1 cytokines, INF-γ and IL-12, inhibited the fibrocyte differentiation, the Th2 cytokines, IL-4, IL-13, and TGF-β1, promoted the fibrocyte differentiation (44, 46). During acute skeletal muscle injury repair, intramuscular macrophages and fibroblasts express a high level of TGF-β1, which may participate in the stimulation of fibrocyte differentiation. It has been reported that fibroblasts can also influence the fibrocyte differentiation by producing regulatory soluble factors (47, 48). Moreover, the fibrocyte differentiation can be regulated by the FCγ receptors (49, 50). Whether these regulatory mechanisms play a role in the fibrocyte differentiation within injured muscle needs to be further studied.

The majority of the intramuscular fibrocytes appear to be macrophages as they express the tissue macrophage marker F4/80, and the majority of them are Ly6C^low. It has been shown that fibrocytes can be recruited into injured tissues via CCR2, CCR5, CCR7, or CXCR4 (26–28, 30). The expression of CCR2 but not CCR5 or CCR7 by the intramuscular fibrocytes suggests that these cells most likely originate from the CCR2⁺/Ly6C^hi inflammatory monocytes. This is consistent with the notion that fibrocytes originate from the CD11b⁺/CD115⁺/Gr1⁺ inflammatory monocytes (45). Murine blood monocytes mainly consist of two subsets, CX3CR1^lowCCR2⁺Ly6C^hi and CX3CR1^hiCCR2⁻Ly6C^low cells (51). CX3CR1^lowCCR2⁺Ly6C^hi cells are inflammatory monocytes, and CX3CR1^hiCCR2⁻Ly6C^low cells are patrolling monocytes and may contribute to resident macrophages (51, 52). It has been shown that acutely injured muscle recruits only the inflammatory CCR2⁺/Ly6C^hi/F4/80^low monocytes, and within injured muscle, they switch their phenotype into the Ly6C^low/F4/80^hi macrophages after phagocytosing damaged tissues and cells (1). Given the CCR2⁺/F4/80⁺/Ly6C^low phenotype of the intramuscular fibrocytes, and no detectable fibrocytes in peripheral blood at any stages after acute muscle injury, we think that the intramuscular fibrocytes are most likely derived from circulation via CCR2, initially recruited as CCR2⁺ inflammatory monocytes, and then differentiate into macrophages and collagen-expressing macrophages (fibrocytes) in muscle. One alternative interpretation of the intramuscular CD45⁺/GFP (Col1)⁺ cells is that they may not be fibrocytes but tissue fibroblasts which express macrophage markers in injured muscles. However, this possibility is remote, as to the best of our knowledge, the expression of macrophage markers by tissue fibroblasts has never been reported. In addition, unlike tissue fibroblasts, the intramuscular CD45⁺/GFP (Col1)⁺ cells express a very low level of ECM genes, and they also express CCR2, which fit the characteristics of fibrocytes. Fibro/adipogenic progenitor cells (FAPs) have been shown to play an important role in muscle fibrosis and adipose tissue deposition (53–55). They express the progenitor cell marker CD34 and the mesenchymal progenitor marker PDGFRα. The intramuscular fibrocytes do not overlap with FAPs as they do not express CD34 as assessed by flow cytometry or PDGFRα as assessed by qRT-PCR (data not shown).

It has been shown in non-muscle tissues that fibrocytes can participate in tissue wound repair and fibrosis by producing a low level of collagens (26, 56), secreting fibrogenic growth factors TGF-β1 and PDGF (57), and may differentiating into tissue effector fibroblasts. Although the differentiation of fibrocytes into α-smooth muscle actin positive myofibroblasts can be easily achieved in vitro (28, 44, 58), the ability and extent of such differentiation remains unclear and controversial in the disease mouse models in vivo (23, 28–31, 39, 44, 58–63). By bone marrow transplant studies in the lung, liver, and renal fibrosis mouse models, the contribution of fibrocytes to myofibroblasts appeared minimal (39, 58, 60). In a mouse model of renal fibrosis induced by unilateral ureter obstruction, fibrocytes were not detected in peripheral blood, they were detected at a very low number in kidneys at the late stages but not the early stage after the injury, and they did not appear to contribute to myofibroblasts (39). Likewise, our present study showed that fibrocytes were not detectable in peripheral blood after acute muscle injury, they were detected in injured muscles at late stages but not an early stage, and no differentiation of fibrocytes into fibroblasts was detected in injured muscles from day 6 to day 7. However, the differentiation cannot be definitively excluded by the cell transfer experiments, because the number of fibrocyte engraftment was low, and a significant number of these cells might die after the transplantation. We focused on day 7 for the cell transfer experiment, because the cell numbers of both fibrocytes and fibroblasts peaked at this time point, and both numbers reduced quickly after this point along with inflammation resolution and ECM remodelling. We did try longer exposures, but the recovery rates of the donor cells were too low to allow us to make any conclusions. Given the low level of collagen expression, the intramuscular fibrocytes do not appear to make a significant direct contribution to the ECM production during the acute skeletal muscle injury repair. However, the intramuscular fibrocytes did express a high level of TGF-β1 and PDGFβ genes as compared with the intramuscular fibroblasts, suggesting that they might enhance the reparative function of muscle fibroblasts, as TGF-β1 and PDGF are potent growth factors for fibroblast activation and function. This might in turn contribute to the transient increase of ECM production and deposition to support the injury repair.

Fibrocytes contribute to tissue fibrogenesis in many fibrotic disease models (25–32). However, in the skeletal muscle fibrosis associated with poor acute injury repair due to an inadequate inflammatory response caused by the CCR2 deficiency, the intramuscular fibrocytes do not appear to contribute to the persistent muscle fibrosis. They expressed neither a high level of collagen genes nor a high level of pro-fibrotic growth factor genes, and their number was also low at the stage of fibrogenesis (day 7) in Ccr2−/− mice as compared with WT controls, which was most likely due to the markedly impaired monocyte recruitment in the Ccr2−/− mice. The intramuscular fibroblasts from the Ccr2−/− mice at day 7, however, expressed a higher level of collagen genes than the WT controls. The findings indicate that the fibroblasts are the major effectors driving muscle fibrosis, and they are functionally more active in Ccr2−/− mice than in WT controls following acute muscle injury. The findings also suggest that cells other than fibrocytes and fibroblasts within injured muscles may serve as an important cellular source of TGF-β1 and PDGF to activate fibroblasts and drive fibrosis in injured muscles of Ccr2−/− mice. Taken together, the persistent muscle fibrosis in Ccr2−/− mice following acute injury appears to be driven by poor muscle regeneration rather than inflammation or fibrocytes.

In contrast to the low function of fibrocytes in acutely injured muscles of Ccr2−/− mice, the function of the fibrocytes in the mdx^5cv diaphragm is high as compared with the function of the fibrocytes in the mdx^5cv leg muscle quadriceps. Endomysial inflammation and mild endomysial fibrosis are present in both diaphragm and quadriceps muscles in mdx^5cv mice at 3 months, and the number of the fibrocytes is similar in these two muscles. However, after this time point, both inflammation and fibrosis gradually subside in the quadriceps but not the diaphragm, and the diaphragm undergoes persistent inflammation and progressive fibrosis (14, 20, 21). It is largely unclear why limb muscles and diaphragm display different phenotypes. It has been speculated that diaphragm has poor regenerative capacity than quadriceps, which may explain the efficient muscle repair of the mdx^5cv quadriceps but not the diaphragm. It has also been shown that the intramuscular macrophages switch their phenotype from pro-inflammatory at 4 weeks to pro-regenerative at 12 weeks in the mdx leg muscles, which may contribute to the spontaneous improvement of the leg muscle pathology in mdx mice at the late stages (64). Based on our current findings, we think that the different functional activities of fibrocytes may also contribute to the different phenotypes seen in the mdx^5cv diaphragm and limb muscles. Although the fibrocyte number is similar in the mdx^5cv diaphragm and quadriceps at 3 months, the fibrocyte function is more active in the diaphragm than in the quadriceps. The diaphragm fibrocytes expressed a higher level of both pro-inflammatory and pro-fibrotic cytokines than the quadriceps fibrocytes. The fibrocytes from the mdx^5cv diaphragm but not the quadriceps stimulated a high level of ECM gene expression by the co-cultured fibroblasts. Therefore, the high level of the TGF-β1 and PDGF gene expression by the mdx^5cv diaphragm fibrocytes likely contribute to the high fibrotic function of the mdx^5cv diaphragm fibroblasts. In addition, the diaphragm fibrocytes also expressed a higher level of some M1 and M2 cytokines and cell markers than the quadriceps fibrocytes. The findings support the notion that fibrocytes can produce not only pro-fibrotic cytokines but also pro-inflammatory cytokines, and they play a pathological role not only in chronic fibrotic diseases but also in chronic inflammatory diseases (57, 65, 66). Chronic inflammation and fibrosis are likely linked in muscular dystrophy. Our findings support the hypothesis that intramuscular fibrocytes, a subset of macrophages which are most likely recruited from circulation and differentiate in situ, may play a pathological role in maintaining chronic inflammation and driving progressive fibrosis in muscular dystrophy.

In summary, our present study demonstrates that fibrocytes are present in both acutely and chronically injured skeletal muscles caused by chemical-induced injury and muscular dystrophy, respectively. Our findings suggest that intramuscular fibrocytes most likely originate from infiltrating macrophages and differentiate within injured muscles. Intramuscular fibrocytes are functionally heterogeneous, different in acute and chronic injury settings, and different in limb and diaphragm muscles of the mdx^5cv mice. They may promote acute muscle injury repair by producing growth factors for reparative ECM remodelling. While they do not appear to contribute to the persistent muscle fibrosis associated with impaired inflammation and poor regeneration following acute injury in the Ccr2−/− mice, they do appear to play a pathological role in maintaining chronic inflammation and driving progressive fibrosis in the mdx^5cv diaphragm muscle. Although gene therapy and cell therapy may ultimately cure DMD, anti-fibrotic therapies remain very much needed for promoting muscle regeneration and improving gene and cell engraftment efficiency (67, 68). Targeting and modulating the fibrocyte function may potentially be therapeutically useful. Future studies should further elucidate the in vivo functions of intramuscular fibrocytes and uncover the detailed mechanisms of intramuscular fibrocyte differentiation and functional regulation.

ABBREVIATIONS LIST

BMMac

Bone marrow-derived macrophages

CCL2

CC chemokine ligand 2

CCR2

CC chemokine receptor type 2

CTGF

Connective tissue growth factor

ECM

Extracellular matrix

FAP

Fibro/adipogenic progenitor cells

MOs

Monocytes

MPs

Macrophages

PDGF

Platelet derived growth factor

PDGFRα

Platelet derived growth factor receptor alpha

TGF-β1

Transforming growth factor-beta1

WT

Wild-type