Macrophage Plasticity and the Role of Inflammation in Skeletal Muscle Repair

Effective repair of damaged tissues and organs requires the coordinated action of several cell types, including infiltrating inflammatory cells and resident cells. Recent findings have uncovered a central role for macrophages in the repair of skeletal muscle after acute damage. If damage persists, as in skeletal muscle pathologies such as Duchenne muscular dystrophy (DMD), macrophage infiltration perpetuates and leads to progressive fibrosis, thus exacerbating disease severity. Here we discuss how dynamic changes in macrophage populations and activation states in the damaged muscle tissue contribute to its efficient regeneration. We describe how ordered changes in macrophage polarization, from M1 to M2 subtypes, can differently affect muscle stem cell (satellite cell) functions. Finally, we also highlight some of the new mechanisms underlying macrophage plasticity and briefly discuss the emerging implications of lymphocytes and other inflammatory cell types in normal versus pathological muscle repair.


Introduction
Tissue regeneration is an evolutionary conserved process in which interactions between in�ltrating in�ammatory cells and resident cells must be �nely coordinated if homeostasis and functionality are to be restored. Perturbation of these interactions leads to unsuccessful regeneration and oen compromises survival of the individual [1,2]. Skeletal muscle, the most abundant tissue of the body, is essential for breathing, posture maintenance, and locomotion, besides serving important homeostatic and metabolic roles, such as heat production and carbohydrate or amino acid storage. Loss of muscle functionality in acute or chronic conditions results in diminished mobility and strength, in addition to metabolic disorders, which can have potentially lethal consequences. Abnormal muscle repair can occur in the context of persistent myo�ber degeneration and/or in�ammatory in�ltration, such as in Duchenne muscular dystrophy (DMD), or when extracellular matrix (ECM) deposition is excessive or inappropriately timed, eventually leading to the substitution of the normal muscle architecture by �brotic tissue [3]. erefore, preservation of the capacity of skeletal muscle to regenerate in a coordinated manner in response to direct mechanical trauma (acute injury), or following secondary damage as a consequence of genetic neuromuscular alterations, is of utmost importance.

Injury-Induced Skeletal Muscle
Regeneration: A Model for Tissue Repair e capacity of muscle to regenerate relies primarily on a speci�c population of normally quiescent muscle stem cells, named satellite cells due to their particular position and intimate association with muscle �bers [4]. Many additional cell types also play a role in efficient tissue repair, including resident cells within the skeletal muscle niche such as PICs (PW1 + interstitial cells), mesoangioblasts, FAPs (�bro/adipogenic progenitors), and other ECM-associated cells [5]. However, the in�ammatory cells that in�ltrate the injured muscle appear to be the most critical, alongside satellite cells, for successful regeneration. Among these 2 F 1: In�ammation and macrophage polarization in skeletal muscle injury and repair. �atellite cells are muscle-resident stem cells which are located underneath the basal lamina of myo�bers and are normally quiescent (top right). �pon muscle injury, satellite cells get activated, start to proliferate as myoblasts, and subsequently fuse and differentiate into myotubes that later grow thereby replacing damaged muscle. �everal cell types in�uence the outcome of regeneration, in particular in�ammatory cells released from the blood (top le). �roin�ammatory monocytes and neutrophils (not shown) extravasate shortly aer damage, invading the injured areas where they differentiate into proin�ammatory macrophages that phenotypically resemble M1 macrophages. �ese cells clear the damage and release a number of cytokines that stimulate myoblast proliferation. M2-like macrophages are present locally at later stages of regeneration acting as promoters of myoblast differentiation and fusion. �ther cell types such as mast cells and lymphocytes also have less de�ned roles in muscle repair (not shown). in�ammatory cells, it is the monocytes/macrophages which play the greatest role in this repair process ( Figure 1). In response to local vascular damage and signals released by degenerating myo�bers, these cells extravasate from the blood and in�ltrate the injured areas, to phagocytose myo�ber debris. In addition to this critical function, in�ammatory cells produce growth factors, cytokines, in�ammatory mediators, and damage signals that have a profound impact on satellite cell behavior during the repair process [6]. In concert with monocyte/macrophage recruitment, quiescent satellite cells are activated by damage/in�ammationassociated signals and begin to proliferate, thereby providing a sufficient supply of myonuclei for the formation of new myo�bers. �hile most of the proliferating satellite cells will commit to myogenic differentiation, a small population will undergo self-renewal and replenish the pool of quiescent satellite cells, thus maintaining muscle stem cell homeostasis [7].

Mediators of In�ammation
A further critical step in the repair process is the reestablishment of the ECM around the individual �bers and bundles which helps strengthen the muscle and provides additional support for contraction. Correct remodeling and reorganizing of the muscle ECM aer damage is necessary for providing new scaffold structures over which nascent myo�bers will be formed, as well as ensuring correct spatial organization of the new myo�bers [8]. Excessive and persistent ECM deposition (�brosis) leads to failure in restoring the previous structure of myo�bers, thus provoking a defective regenerative outcome. Although several studies have shown that satellite cell-derived myoblasts may synthesize many components of the ECM, the major matrix-producing cell is the �broblast [9]. �ike satellite cells, resident �broblasts proliferate and migrate to the injury site immediately aer muscle damage, where they function in close proximity to satellite cells and regenerating myo�bers. Indeed, recent �ndings have demonstrated the relevance of the interplay between satellite Mediators of In�ammation 3 cells and �broblasts and/or FAPs as a determinant factor for the efficiency of the repair process [10][11][12]. Speci�c deletion of �broblasts using genetic approaches resulted in impaired regeneration due to the lack of proliferation of satellite cells and their premature differentiation, strongly suggesting a paracrine action of �broblasts on muscle cells [10]. An important part of the functional role of ECM in controlling the process of repair is carried out by the basal lamina, a thin layer of non�brillar collagen, noncollagenous glycoproteins, and proteoglycans that is in direct contact with the myo�ber plasma membrane (see [13] for review). e basal lamina also surrounds satellite cells forming part of the niche that is necessary for maintaining the stem-like properties of quiescent satellite cells. Because of this direct satellite cell contact, the basal lamina composition and integrity also in�uence the process of repair, by providing guidance cues for satellite cell migration. In the normal repair process, prevention of excessive accumulation of ECM components and restoration of the original basal lamina integrity are controlled by the balanced activities of extracellular proteases and their inhibitors. Dysregulation of these enzymatic activities may cause unrestricted ECM accumulation and altered basal lamina composition, which eventually could lead to �brosis development and loss of normal muscle architecture [14]. Lastly, proangiogenic factors also need to operate at advanced stages of the repair process to revascularize the newly formed myo�bers, thus restoring the vascular network of the damaged tissue [15]. For a recent and comprehensive review focused on macrophage biology in skeletal muscle injury, muscle disease, and �brosis, see Bosurgi et al. (2011) [16]. In this paper, we focus more speci�cally on the current knowledge of the in�ammatory control of satellite celldependent muscle repair in acute injury and highlight several recent �ndings.

�� In�����t��n �n ��c�ent Muscle �e���r
Just as satellite cells go through a controlled process of activation from quiescence, proliferation, and self-renewal, and �nally differentiation and fusion into new myotubes, the in�ammatory response also undergoes a series of carefully regulated stages to ensure an efficient return to tissue homeostasis. at is, the composition of the in�ammatory in�ltrate is dynamically regulated to facilitate timely initiation of divergent functions, while the duration and intensity of the various in�ammatory components must also be coordinated with the degree of muscle damage and the need to change tissue milieu during repair [3,6,17]. For example, interfering with the in�ammatory response immediately aer acute injury disrupts the phagocytosis of necrotic �bers and impedes seeding of new myo�bers. Just as detrimental is the prolongation of in�ammation which can promote muscle degeneration and �brosis development, as occurs in severe myopathies such as DMD which are characterized by chronic in�ammation [18]. Macrophages have recently been shown to promote survival and proliferation of myogenic precursor cells that were introduced into mdx skeletal muscle [19]. us, a tightly regulated, transient in�ammatory response is required for normal muscle regeneration. Improving our understanding of the different cell subtypes and identifying the factors that regulate their function and the timing of their activity will enable us to improve pharmacological treatment of acute injury and neuromuscular disorders associated with chronic in�ammatory responses.

�� ���ses �� t�e In�����t�ry �es��nse �n Acute Muscle Injury
Most studies of skeletal muscle regeneration use acute models of injury and repair, such as sterile destruction of myo�bers by either injection of toxins, such as cardiotoxin, notexin, or barium chloride, or by performing freeze crush injuries. ese models are useful for synchronizing the repair processes and performing systematic studies, although they do not necessarily re�ect the more physiologic repair associated with contraction injuries or replicate the different kinetics of chronic in�ammation observed in myopathies. Moreover, there are important contributions of mouse strain to the in�ammatory component and kinetics that are brie�y discussed below and elsewhere [20,21]. However, despite these variables, the in�ammatory response to experimentally induced muscle repair follows an ordered pattern. An immediate response to sterile muscle injury is the local activation of the innate immune response via the release of largely unknown factors, but which could include heat shock proteins, high mobility group box 1 (Hmgb1) as well as endogenous myo�ber proteins and nucleic acids that become decompartmentalized as the �ber breaks and act as damageassociated molecular patterns (DAMPs) [22]. One of the earliest subsequent events is the invasion of the damaged site by in�ammatory cells, particularly monocytes and polymorphonuclear leukocytes, which include neutrophils, that secrete proin�ammatory cytokines and phagocytose particles (such as cellular or bacterial debris) [23]. Neutrophils constitute the �rst wave of in�ammatory cells to enter the damaged tissue, reaching elevated numbers as soon as 2 hours aer the initial injury. Neutrophils are, however, short-lived cells, whose number declines rapidly, probably through apoptosis, and they are essentially undetectable 3-4 days aer injury [6]. e exact role of neutrophils in toxin-induced or freezecrush injury is not clearly de�ned. However, several studies on contraction-induced injury show that neutrophils play a key role in repair by causing secondary damage, through the release of reactive oxygen species (ROS) and proteases, as well as facilitating phagocytosis and recruitment of monocytes by the release of cytokines [24,25]. Neutrophils are known to enter into contraction-damaged muscle via a process called diapedesis that requires CD18 (integrin-2) [25]. Interestingly, in contraction-injured CD18-de�cient mice, neutrophil, but not macrophage, recruitment was impaired, while physiological signs of repair such as �ber size and force were more quickly restored compared to wild-type mice [25].
Recent studies have shown that resident macrophages in the muscle epimysium/perimysium connective tissue orchestrate the innate immune response to injury, which is linked to adaptive immunity through in�ammatory dendritic cells (DCs) [26]. In addition to resident macrophages, blood monocytes also enter the damaged tissue and start differentiating into macrophages shortly aer invasion by neutrophils [17]. Other in�ammatory cell types, such as mast cells and T cells, have also been implicated in the repair and �brogenesis of several tissues/organs; however, their role in muscle repair and/or �brosis is generally limited (see also below) [27]. Monocytes originate in the bone marrow and circulate to the blood and the spleen before entering the muscle aer injury [28]. ey are equipped with chemokine and adhesion receptors that allow them to migrate from the blood to the injured tissues, where they produce proin�ammatory cytokines and phagocytose dying or apoptotic cells. In the blood, circulating monocytes can be classi�ed into at least two populations that are distinguishable by their expression levels of Ly-6C (also known as GR1) and of chemokine receptors CCR2 and CX3CR1 [29]. ese two monocyte populations use different mechanisms for extravasation and probably have different functions. e GR1 + monocyte cell pool has been designated as the "in�ammatory� population because they efficiently produce proin�ammatory cytokines [30]. rough the CCR2/CCl2 axis, they are rapidly recruited to, and accumulate at, the site of in�ammation [31,32]. On the other hand, the GR1 − population of monocytes has an "anti-in�ammatory� function, which includes supporting tissue repair and patrolling the vasculature [33]. In contrast to GR1 + cells, GR1 − monocytes enter damaged tissues in a CX3CR1-dependent manner just aer the onset of in�ammation in models of sterile injury [34][35][36]. An important consideration beyond the scope of this paper is the known heterogeneity in the use of cell surface markers between mice and humans, with human monocytes broadly de�ned as expressing different levels of CD14 and CD16 [37]. us, as most studies are performed in mice, care will be needed in trying to extrapolate �ndings to humans and the clinic.
Classically, there are believed to be two waves of tissue-in�ltrating monocytes in most experimental wound healing models� a �rst wave comprising the GR1 + population, endowed with proin�ammatory function and a second wave of GR1 − monocytes with an anti-in�ammatory function. Interestingly, using an acute muscle injury model, Arnold and colleagues showed that the GR1 + monocyte population is the only one recruited to the injury site, switching subsequently within the damaged tissue into an "anti-in�ammatory� macrophage population, thereby dampening the earlier proin�ammatory wave and also supporting myogenesis [38]. Distinct macrophage populations have also been associated with the increased �brosis observed in dystrophic muscle (see also below) [39]. Together these observations suggest that the mechanisms of leukocyte recruitment and maturation could be speci�c for each type of damage, tissue, and microenvironment.

Markers versus Functional Properties
As suggested by the experiments above and additional data from other tissue repair systems [40][41][42], macrophages exist as different functional populations at different times aer injury. Generally, these populations are considered to exhibit opposing activities, being either polarized towards proin�ammatory or anti-in�ammatory activity [38]. Polarized macrophages are currently classi�ed as either M1 or M2, referring to either classical or alternative activation, respectively [40,43]. Proin�ammatory M1 macrophages arise from exposure to the T-helper ()1 cytokines interferon-(IFN) and tumor necrosis factor (TNF) , in addition to lipopolysaccharide (LPS) or endotoxin [43,44]. However, polarization of M2 macrophages is more complex than M1 polarization, with three possible subtypes currently de�ned, each one with diverse physiological roles. Alternatively activated or M2a macrophages are commonly associated with advanced stages of tissue repair and wound healing, arising from exposure to 2 cytokines such as IL-4 and IL-13. As well as promoting the proliferation of nonmyeloid cells, IL-10 is also known to induce M2c macrophages which have an anti-in�ammatory function. Similarly, M2b macrophages also have an anti-in�ammatory role and can release large amounts of IL-10. M2b share many features with tumorassociated macrophages [45]. Like M1 macrophages, M2b macrophages also release proin�ammatory cytokines, such as IL-1 and TNF , but not IL-12. Proin�ammatory macrophages, observed experimentally in the context of muscle repair, are phenotypically similar to classically activated M1 macrophages, and are usually found at early stages aer muscle injury, closely followed by macrophages sharing features with the anti-in�ammatory M2c phenotype, so-called because of their role in deactivating M1 macrophages [38]. Early on, M1 macrophages phagocytose necrotic muscle debris and participate in the processing and presentation of antigens. In addition to producing high levels of proin�ammatory cytokines, M1 macrophages also express inducible nitric oxide synthase (iNOS), which is required to efficiently metabolize L-arginine, a fundamental reaction for producing an abundance of NO for killing intracellular pathogens during infection. Alternatively activated M2a macrophages are more abundant during the �nal phase of tissue repair [46]. Importantly, M2a macrophages have also been linked to �brosis in dystrophic mdx mouse muscles [39,47]. M2b-like macrophages have recently been described in regenerating muscle aer acute injury [48], suggesting that a wide range of M2 macrophage subtypes might be functional during the muscle repair process.
Despite the emergence of compelling evidence for the presence of different macrophage subtypes in muscle repair, a clear understanding of their speci�c functions is still lacking. By analogy with the in vitro cytokine proin�ammatory pro-�le, monocytes entering the muscle at the onset of in�ammation resemble M1 polarized macrophages. Indeed, they produce large amounts of the proin�ammatory cytokines TNF and IL-1 and have an enhanced expression of iNOS. In cell culture models, proin�ammatory macrophages have been shown to exert a positive in�uence on myoblast proliferation while repressing myoblast differentiation [17,38,49]. As the process of muscle regeneration advances, they switch their phenotype to resolve in�ammation and start to express high levels of IL-10, TGF , and other anti-in�ammatory cytokines that dampen the initial cytokine storm. ese cytokines have essential roles in promoting proper wound healing, by supporting myogenesis [50], enhancing angiogenesis, and stimulating the transient deposition of the ECM [51]. Similarly, the later wave of anti-in�ammatory macrophages stimulates both myoblast differentiation and fusion in vitro [17,38,49]. e relevance of these in�ammatory cells in vivo was shown aer depletion of blood monocytes exerted negative effects on the regeneration process [38]. Indeed, it is blood monocytes that are the likely source of M1 and M2 macrophages in injured muscle. It is worth noting that although the in vitro models of macrophage polarization are useful to establish a theoretical classi�cation, these macrophage populations most likely represent the extremities of a continuum of possible activation states. In addition to problems of classifying polarized macrophages in different tissues and repair models, a recent study has also suggested that there is considerable heterogeneity in the gene expression pattern of different resident macrophage populations in different tissues [52]. erefore, caution is suggested when classifying wound healing macrophages in general, and muscle in�ltrating macrophages in particular, especially when comparing them with in vitro polarized macrophages. It is tempting to propose that, rather than belonging to one of these categories, wound healing macrophages could themselves constitute a unique class based on their common characteristics with M1, M2a, M2b, or M2c macrophage subtypes [53]. us, to avoid the confusion that can arise from the mixed phenotypes found in vivo, some authors propose to classify macrophages regarding their function (i.e., host defense, wound healing, or immune regulation) rather than grouping them on the basis of expression of certain markers [54].

Mechanisms of Macrophage Polarization and
Deactivation during Muscle Repair e mechanisms underlying the transition of macrophage phenotypes during muscle repair are poorly understood. However, certain analogies can be established between in vitro macrophage responses to endotoxin and the phenotypic transitions that occur during wound healing. For example, the cAMP response element-binding protein (CREB) plays an important role in generating the anti-in�ammatory macrophage phenotype in response to LPS. is response is mediated by the mitogen-and stress-activated kinases 1 and 2 (MSK1 and MSK2), which are, in turn, activated by p38 MAP kinase (MAPK) [55,56]. In a model of toxic contact eczema induced by phorbol-12-myristate-13acetate, the CREB-induced expression of IL-10, and dual speci�city protein phosphatase 1/MAP kinase phosphatase-1 (��SP1/MKP-1) inhibited the expression of proin�ammatory genes associated with M1 macrophage activation, thus supporting a link between p38/MAPK-1 and CREB in macrophage polarization. An important regulatory function for CREB in macrophage polarization has also been revealed during tissue repair. Indeed, M2, but not M1, macrophage gene expression was impaired by deleting two CREB-binding sites from the C/EBP gene promoter, resulting in abnormal muscle regeneration [57]. Macrophages from the C/EBP promoter mutant mice had a reduced expression of the M2-associated arginase gene aer LPS stimulation. It was hypothesized that this may lead to a switch in arginine metabolism from arginase-mediated polyamine synthesis to iNOS-mediated NO production [57]. Importantly, additional studies showed that shis in macrophage polarization and macrophage competition for arginine metabolism in�uenced the severity of muscle pathology in mdx dystrophic mice [47]. ese studies strongly support the idea that CREB might be a pivotal transcription factor in macrophage polarization that functions by promoting M2-associated genes while repressing M1 activation, with CREB transcriptional activity regulated by balance of p38/MSK1/2-MKP-1 activities.
Although macrophages sustain proper healing by secreting growth factors and cytokines that support myogenesis and promote transient ECM deposition, dysregulation of the expression of cytokines such as TGF or IL-1 can lead to aberrant repair, including �brosis development, especially in muscle pathologies and conditions characterized by chronic in�ammation [21]. Consequently, efficient muscle repair requires resolution of in�ammation, and in particular, deactivation of macrophages, at advanced stages of tissue recovery [58].
e regulatory mechanisms controlling cytokine gene silencing and macrophage deactivation remain largely undeciphered. One recent study investigated the AKT activation status in macrophages of wild-type and MKP-1-de�cient mice during the resolution of in�ammation aer muscle injury. e activity of AKT was higher in MKP-1 −/− than in wild-type macrophages in the late stage of muscle repair, correlating with a loss of pro-and anti-in�ammatory cytokine gene expression, and this effect could be reverted by pharmacologically inhibiting p38 MAPK activity [49]. Conversely, macrophages from wild-type mice treated with the PI3K/AKT inhibitor wortmannin showed a prolonged activation status presumably by preventing deactivation. Furthermore, levels of the phosphatase PTEN, which functions as a tumor suppressor by negatively regulating the AKT/PKB signaling pathway, were lower in macrophages in the absence of MKP-1 during the later stage of muscle repair. PTEN is also a direct target of miR-21 [59], a miR classically associated with cancer and �brosis [60,61], and its expression was previously reported in RAW 264.7 macrophages [62]. Because miR-21 expression was shown to increase in deactivated macrophages in a p38-dependent manner, [49] it is possible that the loss of MKP-1 (through regulation of the miR-21/PTEN/AKT pathway) extends macrophage cell persistence at the site of injury while at the same time provoking their premature deactivation during the tissue repair process. Taken together, these results strongly support a role for MKP-1 in neutralizing p38 MAPK and thereby controlling sequential macrophage activation-deactivation transitions during tissue repair by restraining AKT activation.

Additional Immune Cell Types, Such as
Lymphocytes, Are Also Implicated in Muscle Repair and Fibrosis Macrophages are the predominant in�ammatory cells in skeletal muscle regeneration, yet other immune cells, in particular T lymphocytes, have also been proposed to in�uence repair and �brosis. Like macrophages, T lymphocytes can also differentiate into distinct functional subsets. e two major types are termed 1 and 2, which have distinct roles in orchestrating the host response by generating distinct cytokine pro�les [44], whereas more newly characterized including 17 and Treg subtypes have a less de�ned role in muscle regeneration, although growing evidence suggests they may become important [63]. Cytokines produced in T cells also regulate muscle degeneration and repair. CD4 + 1 cells promote cell-mediated immunity and are able to produce cytokines with anti�brotic properties such as IFN , TNF , IL-12, and IL-2. By contrast, CD4 + 2 cells produce IL-4, IL-5, IL-6, and IL-13, which are cytokines whose primary role is to promote humoral immunity in addition to having pro�brotic roles. Importantly, 1 cytokines inhibit the development of 2 cells, and conversely, 2 cytokines inhibit the development of 1 cells. Clearly, alterations or imbalances in these pathways have the potential to skew repair towards anti-or pro�brotic pathways, as witnessed by the importance of 2 cytokines in the development of liver �brogenesis [46]. Moreover, T-cell-derived cytokines have a clear role in maintaining the polarized state of macrophages in vivo, at least in other models of injury and repair such as aer parasite infection [64]. However, the contribution of T cells to macrophage polarization in sterile injury models where T cells are less abundant remains to be explored. Several studies have suggested roles for T lymphocytes in muscle regeneration. For example, knockout mice lacking the proteolytic activity of the serine protease uPA, and its downstream proteolytic cleavage enzyme plasmin, displayed reduced macrophage and T-lymphocyte in�ltration of injured muscle and persistent myo�ber degeneration [65][66][67]. Another study in mice de�cient for the Cbl-b ubiquitin ligase tumor suppressor gene showed increased in�ltration of CD8 + T cells into injured muscles with a subsequent delay in muscle regeneration [68]. De�ciency of Cbl-b also signi�cantly increased production of the chemokine CCL5 (RANTES) from macrophages during muscle regeneration, whereas neutralization of CCL5 improved the defective muscle regeneration in Cbl-b-de�cient mice. All together, these results suggest that Cbl-b is an important regulator of CD8 + T-cell in�ltration into regenerating muscle, an effect mediated via CCL5 production in macrophages [68]. In another example, athymic BALB/c nude mice, which are T cell de�cient, showed signi�cant increase in central nucleation and increased MMP-9 activity in comparison to wild-type BALB/c [21].
Lymphocytes have also been implicated in the de�cient regeneration and development of �brosis observed in some degenerative myopathies. Early studies identi�ed the presence of T cells and several other in�ammatory cell subtypes in biopsies of human DMD patients and other myopathies [69,70]. However, T-cell-mediated cytotoxicity appeared to be limited in DMD patients, despite the appearance of major histocompatibility complex I (MHC I) on regenerating �bers [71]. Dystrophic scid/mdx mice, which are de�cient in functional T and B lymphocytes, develop much less diaphragm �brosis with age compared with normal mdx mice, concomitant with a decrease in activated TGF in skeletal muscle, [72]. In nu/nu/mdx mice (immunode�cient nude mice in the mdx background), the lack of functional T cells alone was associated with less diaphragm �brosis at 3 months, supporting the pathogenic role for T cells in mdx muscle and revealing this lymphocyte subclass to be an important source of TGF [73]. A speci�c subpopulation of T cells expressing the Vb8.1/8.2 T-cell receptor (TCR) was recently identi�ed and shown to be enriched in mdx muscle. ese T cells produce high levels of osteopontin, a cytokine that promotes immune-cell migration and survival [74]. Intriguingly, osteopontin levels are increased in patients with DMD and in mdx mice aer disease onset. Importantly, loss of osteopontin in mdx double-mutant mice diminishes the in�ltration of natural killer T-cell-(NKT-)like cells, which express both T and NK cell markers and neutrophils. ese mice also show reduced levels of TGF . ese results correlate well with improvements in muscle strength and reduced �brosis in the diaphragm and heart [74]. ymectomy at one month of age induces near complete postnatal depletion of circulating T cells in mdx mice. When this was followed by anti-CD4 and/or anti-CD8 antibody treatment, it failed to improve diaphragm �brosis at six months of age [72,75,76]. Finally, a recent study investigated the role of lymphocytes in muscle dysferlinopathy using Scid/A/J transgenic mice and showed that the absence of T and B lymphocytes resulted in an improvement of muscle regeneration [77].
Several studies have also shown that mast cells may play a role in normal skeletal muscle repair. Mast cells were shown to accumulate in injured muscles from around 8 hours aer saline injection of the gastrocnemius muscle, most of which were recruited from the circulation as very few mast cells are resident in the tissue [78]. Interestingly, mast cells have been linked to development of �brosis and were shown to be persistently present in mdx muscle tissue close to major vessels [78,79]. Several studies in mdx mice and human clinical trials have explored the use of mast cell stabilizers like Oxatomide (Tinset) or Cromolyn (sodium cromoglycate) on the ability to improve muscle repair [80,81]. Although mast cells are known to release many proin�ammatory cytokines such as TNF and interleukins such as IL-1 and IL-6, consistent with their early appearance in muscle aer acute damage, in addition to histamine and proteases such as chymase, their role in macrophage polarization is unknown and their overall contribution to efficient repair requires further investigation.
e above data serves to demonstrate the complexity of the mechanisms that regulate in�ammation, muscle repair, and �brosis development. It is still not clear whether distinct types of  responses and macrophage subtypes operate in dystrophic muscle, and how they mediate their interactions. us, despite our increasing understanding of these immune cells, the implication of the presence of lymphocytes and their subtypes in muscle repair clearly requires further study.

Concluding Remarks and Future Directions
Numerous recent studies have expanded our knowledge of the function of macrophages, which extends far beyond their role in host defense against bacteria or parasites. e progress in this �eld has led to the discovery of an increasing number of macrophage activation states, rendering their classi�cation more difficult. If the in vitro studies on macrophages activation and their subsequent classi�cation in M1 and M2 macrophages have been useful to mirror the 1 and 2 polarization of T cells, the M2 designation has expanded to include all of the non-M1 macrophages. Consequently, a growing number of immunologists now classify them in the extended family of M2-like macrophages. However, the plasticity of these cells makes it difficult to assign speci�c markers to each population, especially since phenotypic changes are temporally dynamic and depend on changes in the microenvironment and on cell intrinsic mechanisms, like in endotoxin tolerance, which represents a switch from a proin�ammatory M1 phenotype to an M2-like anti-in�ammatory phenotype. e discovery of new markers, together with progress in �ow cytometry techniques, will probably increase even more the complexity of classifying macrophages, rendering it essential to rethink the way we create categories of macrophages and forcing us to focus on their function in order to de�ne these di�erent populations more precisely. Finally, in addition to more precisely de�ning and evaluating macrophage functions in tissue repair, future research should also focus on identifying in greater detail the function of alternative immune cell types, such as lymphocytes, in the correct resolution of tissue injury or, conversely, in facilitating �brosis development.

Author's Contribution
Y. Kharraz and J. Guerra contributed equally to the paper.