Background

Macrophages are pleotropic immune cells capable of adapting to changes in the local microenvironment. Over the last several years, research has shown that macrophage phenotype is dynamic, existing on a continuum (Mosser and Edwards, 2008; Italiani and Boraschi, 2014; Martinez and Gordon, 2014). However, to date macrophage populations continue to be described using the restrictive M1 and M2 classifications. It is commonly accepted that these designations are an oversimplification of macrophage phenotype and represent opposite extremes of a continuum (Mosser and Edwards, 2008; Gordon et al., 2014; Italiani and Boraschi, 2014; Martinez and Gordon, 2014; Murray et al., 2014). M1 macrophages are classically activated, have pro-inflammatory functions and are involved in host responses to pathogens and tissue injury. M2 macrophages are alternatively activated, exhibit anti-inflammatory functions and are involved in wound healing and tissue repair. In addition to the functional definition of M1 and M2, cell surface markers have been identified to distinguish between these populations. Surface markers associated with M1 macrophages include CD40, CD64 and the co-stimulatory molecules CD80/CD86 (Lolmede et al., 2009; Ambarus et al., 2012), whereas M2 macrophages have been shown to express high levels of CD163, CD206 and galactose receptors (Lolmede et al., 2009; Ambarus et al., 2012; Roszer, 2015).

From tissue to tissue, macrophage populations are heterogeneous adopting different functional roles depending on the local environment (Gordon et al., 2014; Italiani and Boraschi, 2014). This, coupled with the macrophage continuum, has led to inconsistencies with regard to identification and nomenclature across fields and across species (Murray et al., 2014). Skeletal muscle macrophages have primarily been studied in rodent models of injury, where the M1 versus M2 macrophage classification has proven useful (Smith et al., 2008; Chazaud et al., 2009; Tidball and Villalta, 2010; Kharraz et al., 2013; Novak and Koh, 2013; Saclier et al., 2013b; Rigamonti et al., 2014; Tidball et al., 2014; Wang et al., 2014; Sciorati et al., 2016; Varga et al., 2016; Mackey and Kjaer, 2017). The skeletal muscle response to injury is characterized by highly orchestrated temporal processes. Initially, M1 macrophages phagocytize damaged skeletal muscle fibers and debris, followed by M2 macrophage-facilitated repair and regeneration (Chazaud et al., 2009; Tidball and Villalta, 2010; Kharraz et al., 2013; Saclier et al., 2013a; Tidball et al., 2014; Sciorati et al., 2016). Recent work nicely details fiber repair in human skeletal muscle in vivo, showing the presence of macrophages, using a pan-macrophage intracellular marker (CD68+), in regenerating zones along injured fibers (Mackey and Kjaer, 2017). Direct interaction between macrophages and satellite cells (Dumont and Frenette, 2013; Ceafalan et al., 2017; Du et al., 2017; Wehling-Henricks et al., 2018), and defects in skeletal muscle regeneration in the absence of macrophage participation (Arnold et al., 2007; Melton et al., 2016), highlight the necessity of these cells for skeletal muscle repair. In vitro, M1 macrophages promote skeletal muscle cell proliferation and M2 macrophages promote differentiation, suggesting that macrophages may play a role in skeletal muscle growth adaptations, as well as repair (Arnold et al., 2007; Saclier et al., 2013b).

Skeletal muscle is a highly adaptable tissue, able to respond to a wide range of external stimuli, such as exercise, inactivity, hormones and nutritional signals. In contrast to the clearly defined, strongly polarizing responses elicited by acute skeletal muscle injury, the role of tissue resident macrophages during less polarizing processes, such as responses to the aforementioned stimuli, is relatively unknown. Under non-damaging exercise conditions, animal studies report an increase in macrophage populations following aerobic and resistance exercise, linked to both metabolic and growth adaptations (DiPasquale et al., 2007; Ikeda et al., 2013). However, the mechanisms by which macrophages in skeletal muscle influence training adaptations remain to be explored. It has also been reported that resident human skeletal muscle macrophage abundance is affected by aging, obesity and diabetes (Przybyla et al., 2006; Hong et al., 2009; Varma et al., 2009; Tam et al., 2012; Fink et al., 2014; Reidy et al., 2017); however, the inconsistent use of macrophage markers across studies has made the interpretation of these findings difficult. Further, the applicability of the distinctive M1/M2 markers of polarized macrophages to tissue resident macrophages is unclear, as surface markers may not be mutually exclusive on resident macrophages under non-polarizing conditions (Italiani and Boraschi, 2014). Thus, there is a need in the field for a standardized, validated method for identifying and quantifying macrophages in human skeletal muscle. Establishing a simple and reproducible protocol for studying muscle macrophages will aide in the characterization of their role in muscle adaptations to various stimuli, independent of injury.

Although results from studies of skeletal muscle macrophages in animal models are informative, these studies often use macrophage markers that are not directly translatable for use in humans. Even when human homologs do exist, the same surface markers in mouse and rat skeletal muscles often identify different populations in human skeletal muscles, complicating the extrapolation of findings from rodent models to human studies (Murray et al., 2014). For example, CD68 is used as a pan-macrophage marker in humans and an M1 marker in mice. There is a need in the field for a standardized, validated method for identifying and quantifying macrophages in human skeletal muscle. Taking into account the limited mass of frozen muscle tissue available for analyses from human skeletal muscle biopsies (normally in the range of 100 mg), an immunohistochemical method is the most feasible approach to identifying and quantifying human skeletal muscle macrophage populations.

A variety of markers have been used to characterize human macrophages by flow cytometry. The most detailed studies have been performed utilizing peripheral blood mononuclear cells (PBMCs), artificially polarized to an M1 or M2 phenotype (Martinez et al., 2006; Ambarus et al., 2012; Iqbal, 2015). These in vitro studies characterize the expression of various marker combinations on M1 and M2 macrophages and provide a good starting point for choosing markers to identify macrophage populations in frozen human skeletal muscle tissue. CD14 has been identified as a monocyte marker, expressed mainly by macrophages but also neutrophils and dendritic cells (Table 1). In blood, CD14 co-staining with CD16 is used to stratify monocytes into three subsets: classical (CD14++/CD16-), intermediate (CD14++/CD16+) and non-classical (CD14+/CD16++) (Sprangers et al., 2016; Boyette et al., 2017). It is thought that classical monocytes give rise to tissue macrophages under homeostatic conditions; however, during an inflammatory insult all monocyte populations differentiate into macrophages (Italiani and Boraschi, 2014; Sprangers et al., 2016). In tissue, CD16 is predominantly used to identify NK cells, but is also expressed on neutrophils, granulocytes, dendritic cells and some macrophage populations (Table 1). CD11b is a commonly used marker and is expressed on subsets of lymphocytes and monocytes, these include natural killer (NK) cells, granulocytes and macrophages (Table 1). CD68 is expressed by cells in the monocyte lineage, including macrophages, and is the most commonly used macrophage marker in human skeletal muscle tissue (Table 1) (Stupka et al., 2001; Beaton et al., 2002; Peterson et al., 2003; Crameri et al., 2004; Przybyla et al., 2006; Crameri et al., 2007; Mahoney et al., 2008; Mikkelsen et al., 2009; Varma et al., 2009; Paulsen et al., 2010a; Paulsen et al., 2010b; MacNeil et al., 2011; Tam et al., 2012; Chistiakov et al., 2017; Mackey and Kjaer, 2017; Reidy et al., 2017). CD68 is a member of the lysosomal/endosomal-associated membrane glycoprotein (LAMP) family of proteins, which are mainly associated with the endosomal/lysosomal compartment. Though largely intracellular, CD68 can traffic to the cell surface. Of note, other cell types have been reported to express CD68, including hematopoietic cells, fibroblasts and endothelial cells (Table 1) (Kunisch et al., 2004; Gottfried et al., 2008; Paulsen et al., 2013; Chistiakov et al., 2017). CD206, the mannose receptor, is a well-accepted macrophage marker in skeletal muscle and is widely used to identify M2 macrophage subsets (Lolmede et al., 2009; Ambarus et al., 2012; Italiani and Boraschi, 2014; Roszer, 2015), although CD206 expression by other cell types (including satellite cells) has been reported (Table 1) (Jansen and Pavlath, 2006). M2 macrophages also express CD163 (Table 1) (Lolmede et al., 2009; Ambarus et al., 2012; Roszer, 2015). CD80 and CD86 are co-stimulatory molecules expressed by antigen presenting cells upon activation and have been used to identify M1 macrophage populations (Table 1) (Mosser and Edwards, 2008; Lolmede et al., 2009; Ambarus et al., 2012). Using three grams of discarded human hamstring muscle from patients undergoing anterior cruciate ligament (ACL) reconstruction surgery, we isolated and labeled mononuclear cells with antibodies against some of the markers described above (CD11b, CD14, CD16, CD86 and CD206) and performed multichannel flow cytometry. Due to the intracellular expression of CD68, we were not able to include CD68 in flow cytometry analyses. Mononuclear cells from skeletal muscle did not express CD16, but co-expressed the other 4 markers tested (Figures 1A-1E). Thus, human skeletal muscle macrophages have a ‘mixed’ phenotype, co-expressing both M1 (CD86) and M2 (CD206) cell surface markers (Figure 1D).

Table 1. Overview of monocyte and macrophage markers



Figure 1. Flow cytometry from discarded human hamstring muscle showing co-expression of both M1 and M2 macrophage markers. Mononuclear cells isolated from human skeletal muscle express A. Both pan-monocyte markers CD11b and CD14; B. Both the M2 macrophage marker, CD206, and the pan marker, CD11b; C. Both the M1 marker, CD86, and the pan marker CD11b; D. Both the M2 marker, CD206, and the M1 marker, CD86; E. Overlay of CD206+/CD86+ populations from panel D onto CD14/CD11b flow plot shown in panel A. This overlay shows that CD206+/CD86+ macrophages (denoted in dark blue) also express the pan-monocyte markers CD11b and CD14. Red boxes indicate cells that are double positive for the markers shown.

Using these 4 cell surface antibodies against CD11b, CD14, CD86 and CD206 that label skeletal muscle resident macrophages, we sought to develop a simple and reproducible immunohistochemical method for identification of macrophages in fresh frozen human skeletal muscle sections. CD68 is a commonly used pan-macrophage marker in human skeletal muscle. However, CD68 expression is predominantly intracellular, requiring permeabilization steps to perform immunohistochemistry (IHC). These permeabilization steps lead to inconsistent results and compromise staining with additional antibodies for cell surface markers. The cell surface localization of CD11b results in better morphological definition of macrophages and more consistent staining across samples than intracellular CD68 staining. Moreover, CD11b can readily be combined with CD206 and other cell surface markers for IHC. For these reasons, we compared CD11b and CD68 staining, and found CD11b to be comparable to CD68 as a pan-macrophage marker in human skeletal muscle (Figures 2A-2E; CD11b and CD68 antibodies cannot be used to label sections simultaneously for technical reasons, see General Note 12). Furthermore, distinct morphological features of muscle macrophages that may be related to functional status was revealed through CD11b and CD206 double IHC (Figures 8A-8C) (Durafourt et al., 2012; McWhorter et al., 2013). We describe here a detailed method for combined IHC using CD11b and CD206 antibodies on frozen human skeletal muscle sections. We also describe in detail our approach to quantifying macrophage subsets in non-injured skeletal muscle. We find the majority of human muscle macrophages are CD11b+/CD206+, whereas a small subset are CD11b+/CD206- (Figures 4A-4C). We were unable to obtain IHC results with antibodies against M1 cell surface markers (CD80 or CD86). Of note, anti-CD163 works well on frozen human skeletal muscle and can be used in place of CD11b and in combination with CD206 with this protocol (see General Note 13). Moreover, combining Pax7 (a satellite cell marker) IHC with CD206 shows very little co-expression, supporting the conclusion that CD206 co-localizes with CD11b in human skeletal muscle and is a valid macrophage marker. This protocol allows reproducible quantification of the relative abundance of CD11b+/CD206+ and CD11b+/CD206- macrophage populations in human skeletal muscle, which can be extended to human skeletal muscle adaptations, aging and disease, enabling comparison of results across studies, across labs and across diverse human populations.