Chemotaxis and Immunoregulatory Function of Cardiac γδ T Cells in Dystrophin-Deficient Mice

Duchenne muscular dystrophy (DMD) is the most common form of muscular dystrophy, with a frequency of 1 in 3500 male births. This neuromuscular disorder is characterized by the absence of functional dystrophin, a protein that provides a link between intracellular F-actin and the dystrophin glycoprotein complex, which binds to laminin in the extracellular matrix (1). The lack of functional dystrophin destabilizes the muscle fiber, leading to progressive cell damage through membrane leakage (2). Usually, patients are wheelchair bound by ∼6–12 y of age owing to skeletal muscle damage, and most patients die in their 20s, usually because of cardiac-respiratory failure (3, 4). In recent years, life expectancy has increased because of improved treatments, leading to a higher incidence of late cardiac symptoms. Although skeletal muscles are the ones primarily affected, there is a progressive increase in myocardial fibrosis with electric and contractive dysfunction (5).

Mdx mice are widely used as an experimental animal model to study DMD. A naturally occurring mutation in the gene encoding dystrophin leads to skeletal muscle weakness, inflammatory infiltration, damage, fibrosis, and cardiac dysfunction, which can be monitored via echocardiography in 11- to 12-mo-old animals (6, 7).

In both humans and mice, muscle damage and inflammation are secondary to dystrophin mutations, exacerbating the disease (8, 9). Muscle inflammatory infiltration in DMD is characterized by high numbers of myeloid cells, mostly macrophages, and few T cells (10–12). Several studies have demonstrated that experimental interventions that reduce the inflammatory response in vivo can diminish muscle disease in mdx mice. Early depletion of CD4⁺ or CD8⁺ T lymphocytes (from 6 to 8 d of age) led to reduced collagen deposition in skeletal muscles (13), and experiments using null mutants for both dystrophin and perforin, a cytotoxic molecule mostly produced by CD8⁺ T cells and NK cells, showed decreased dystrophic muscle pathologic changes (14). Furthermore, neutralization of TNF and treatment with immunosuppressants, such as cyclosporine and prednisolone, also reduced muscle disease (15–17).

Although γδ T cells have not yet been described in DMD, this lymphoid subpopulation has important proinflammatory roles in other diseases. For example, in polymyositis and in idiopathic inflammatory myopathies, muscle fibers are destroyed by monoclonally expanded γδ T cells expressing a Vγ1 TCR (18). γδ T cells represent a unique subset of lymphocytes, constituting <10% of circulating lymphocytes in humans and <5% in mice (19). Whereas αβ T cells recognize proteins that are processed into peptides and presented on MHC molecules on APCs, γδ T cells can recognize Ags directly (20). Some murine γδ T cell subpopulations recognize MHC-II molecules regardless of the type of bound peptide (21). Other subpopulations recognize T10 and T22 molecules, nonclassical MHC-I closely related molecules that assume atypical MHC-like conformations in the absence of an Ag (21). Apparently, no Ag processing and presentation is required in the context of MHC molecules, as is the case for αβ T cells, and this process for γδ T cells remains largely unknown.

γδ T cells are able to produce a number of cytokines and chemokines and are involved in many aspects of the host immune response, including tissue repair, regulation of immune responses, tumor surveillance, and adaptive immune responses against pathogens (22). γδ T cells have been conventionally subdivided into different subsets according to their TCR Vγ expression and tissue distribution, where they contribute to immune surveillance (23). Among these subsets, Vγ1⁺ and Vγ5⁺ cells preferentially produce IFN-γ, whereas Vγ4⁺ cells can produce either IFN-γ or IL-17 upon activation in the periphery (23). It has been demonstrated that distinct functional γδ T cell subsets that produce IFN-γ or IL-17 are developed in the thymus and are characterized by the expression of the surface markers CD27 or CCR6 (or NK1.1), respectively (24–27). In addition to cytokine-mediated responses, γδ T cells also mount cytolytic responses upon activation that are very similar to those of cytotoxic αβ T cells, ordinarily using the perforin/granzyme–based pathway (28, 29). The involvement of γδ T cells in a wide variety of diseases indicates the importance of this T cell subset in both innate and acquired immunity. Several studies have demonstrated a role for γδ T cells in inflammatory processes in the absence of infection and autoimmune diseases, such as chronic obstructive pulmonary disease (30), allergic airway inflammation (31), celiac disease (32), multiple sclerosis (33), and type 1 diabetes (34). It was also reported that γδ T cells can be activated by danger/damage–associated molecular pattern molecules (DAMPs) that are generated by injured or dying cells (35).

Despite the extensive muscle damage observed in DMD, this T cell subset has not been studied in mdx mice or DMD patients. Considering the importance of γδ T cells in some models of cardiomyopathy, the aim of this study was to evaluate the chemotactic factors and receptors that lead to γδ T cell migration to the hearts of mdx mice and the role played by this subpopulation in cardiac immunoregulation.

We used 6-, 8-, 12-, and 14-wk-old male mdx mice (C57BL/10ScSn-Dmdmdx/J) and age-matched C57BL/10 control mice obtained from the Center for Breeding of Laboratory Animals at FIOCRUZ. Mice were housed for 7–10 d at the Center for Animal Experimentation under environmental factors and sanitation according to the Guide for the Care and Use of Laboratory Animals (8th edition, 2011). This project was approved by the Fiocruz Committee of Ethics in Animal Research (LW13/11), according to resolution 196/96 of the National Health Council of Brazilian Ministry of Health.

The ventricles from mdx mice were collected and cut into fragments 1–2 mm thick in ice-cold PBS. All fragments were transferred to a 0.1% solution of collagenase type IV (powder 300 U/mg) (Sigma-Aldrich, St. Louis, MO) and submitted to seven or eight cycles of enzymatic digestion (15 min each) under gentle agitation at 37°C. Isolated cells were centrifuged (400 × g/10 min) and immediately transferred to ice-cold RPMI 1640 medium supplemented with 10% FCS (Sigma-Aldrich) and maintained on ice until use. The frequency of the γδ T cells in the cardiac tissue is shown as a percentage, rather than as an absolute number, because, even after many cycles of collagenase dissociation, 10–15% of the tissue fragments remain undissociated, which would result in an underestimate of the number of total cells. Skeletal muscle dissociation was accomplished as described in this section; we used, however, collagenase type IA 90 U/ml.

The following mAbs were used: allophycocyanin-conjugated anti-TCRγδ, PE-conjugated anti-CD27, PerCP-conjugated anti-CCR6, PECy7-conjugated anti-CCR9, PerCP-conjugated anti-CXCR2, FITC-conjugated anti-CCR5, PerCP-conjugated anti-CCR3, FITC-conjugated anti-CCR8, FITC-conjugated anti-CCR2, PE-conjugated anti-CD122 (all from BioLegend, San Diego, CA), FITC-conjugated anti-Vγ1, FITC-conjugated anti-Vγ4, FITC-conjugated anti-Vγ5, FITC-conjugated anti-Vδ4, FITC-conjugated anti-Vδ6.3, PE-conjugated anti–Fas-L, FITC-conjugated anti-CD103, PE-Cy7–conjugated anti–IFN-γ and PE-conjugated anti–IL-17 (all purchased from BD, San Jose, CA). Of note, throughout this study, we used the nomenclature described by Heilig and Tonegawa (36) for murine Vγ genes.

For phenotypic labeling by flow cytometry, all cells were incubated for 30 min at 4°C in RPMI 1640 medium supplemented with 10% FCS and 10% inactivated normal sheep serum to block FcγR. All samples were incubated for 30 min at 4°C with previously titrated mAbs, washed twice in RPMI 1640 medium (Sigma-Aldrich) and maintained in 2% formaldehyde (EMD Chemicals, Gibbstown, NJ) until acquisition using a FACSCalibur (BD) or a Cyan ADP flow cytometer (Beckman Coulter, Houston, TX). Data analysis was performed using Summit software version 4.3 (Beckman Coulter).

After cell surface staining, intracellular labeling was performed using a Cytofix/Cytoperm Kit (BD), according to the manufacturer’s instructions, followed by 30 min of incubation with the appropriate concentrations of mAbs at 4°C. All samples were washed twice with Perm/Wash buffer (BD) and immediately acquired using a Cyan ADP flow cytometer (Beckman Coulter).

Cytokines and chemokines were evaluated by flow cytometry in cardiac and spleen extracts from 6- and 12-wk-old mdx and C57BL/10 mice and purified cardiac γδ T cell extracts from 12-wk-old mdx mice. γδ T cells were purified by flow cytometry using a MoFlow Astrios (Beckman Coulter), and samples of ≥95% purity were used. For cardiac extracts, the hearts were reperfused through the aorta with PBS for complete blood removal, and the ventricles were cut into fragments of 1–2 mm thick. Cardiac fragments and purified γδ T cells were incubated in ice-cold extraction buffer (1% Nonidet P-40, leupeptin 1 mmol/L, PMSF 100 mmol/L, pepstatin A 1 mmol/L, EDTA 100 mmol/L; all purchased from Sigma-Aldrich) and centrifuged (500 × g), and the supernatants were frozen until use. TNF, IL-6, IL-10, IFN-γ, and IL-12 were detected using a Cytometric Bead Array inflammation kit (BD), according to the manufacturer’s instructions. These cytokines were also measured in plasma samples, obtained by cardiac puncture, using heparin from 6- and 12-wk-old mdx and C57BL/10 mice. Samples were acquired using a FACSCalibur flow cytometer (BD), and data analysis was performed using the Cytometric Bead Array analysis FCAP software (BD). CCL2, CCL7, CCL3, CCL4, CCL5, and GM-CSF were detected using Mouse Chemokine FlowCytomix Multiplex (eBioscience, San Diego, CA), also according to the manufacturer’s instructions. Samples were acquired using a FACSCalibur flow cytometer (BD), and data analysis was performed using FlowCytomix software (eBioscience). All results are expressed as picogram per milligram of total proteins per sample. Protein concentrations were determined using a bicinchoninic acid–based kit (Thermo Fisher, Rockford, IL), according to the manufacturer’s instructions.

Purified cardiac γδ T cells harvested from 12-wk-old mdx mice were used in cytotoxic assays as effector cells against adherent peritoneal or adherent cardiac cells (not cardiomyocytes) obtained from 25-wk-old mdx mice. In brief, endomysial cardiac cells were isolated from fragmented ventricles using collagenase, as described earlier in this section. Peritoneal cells were collected from the same mice and left to adhere to 24-well plates for 1 h at 37°C. After two consecutive washes to remove cell debris, the adhered cells from both cultures were detached from the plates using TrypLE Express (Life Technologies, Grand Island, NY), counted in a Neubauer chamber and replated separately in 96-well plates (1 × 10⁴/well). After washing the wells again, purified cardiac γδ T cells were added in triplicate as effector cells in an E:T ratio of 15:1 and incubated for 18 h at 37°C in an atmosphere of 5% CO₂. We evaluated cell cytotoxicity by measuring lactate dehydrogenase (LDH) activity (Wienner, Rosario, Argentina) and granzyme B release using a Granzyme B Activity Fluorometric Assay Kit (Biovision, San Francisco, CA) both in culture supernatants. We also evaluated CD8⁺ T lymphocytes (freshly collected) as control effector cells in cytotoxic assays by flow cytometry. In brief, we purified spleen CD8⁺ T cells from 12-wk-old mdx mice and cardiac γδ T cells, and incubated the cells for 18 h with endomysial adherent cardiac cells. We then detached the cardiac cells using TrypLE Express, washed them in PBS, and incubated all samples with 7-aminoactinomycin D (7AAD) (1 μg/ml), as recommended by the supplier (BioLegend). Data analysis was performed for the myeloid cell gate.

Purified spleen γδ T lymphocytes (5 × 10⁴) were collected from 12-wk-old mdx mice and placed in the upper chamber of Transwell tissue culture inserts (5.0-μm pore diameter) (Corning, New York, NY) under sterile conditions. Transwell inserts were placed over individual wells of a 96-well plate containing the following stimuli: apyrogenic BSA 2 mg/ml (Life Technologies, Grand Island, NY, 99% purity), recombinant mouse (rm) CCL2 (10 ng/ml), rmCCL5 (10 ng/ml), rmCCL2 plus rmCCL5, rmCCL2 plus anti-CCL2 mAb (6.5 ng/ml), rmCCL5 plus anti-CCL5 mAb (50 ng/ml), the total protein cardiac extracts from the ventricles obtained from 6- or 12-wk-old mdx mice (2 mg/ml) or 12-wk-old C57BL/10 mice (2 mg/ml), anti-CCL2 mAb or anti-CCL5 plus the total cardiac extracts from 12-wk-old mdx mice, or anti-CCL2 and anti-CCL5 mAbs plus the total cardiac extracts from 12-wk-old mdx mice. Recombinant chemokines were purchased from Sigma-Aldrich, and neutralizing Abs were purchased from BD. The plate was incubated for 3 h (37°C, 5% CO₂), and transmigrated cells were collected from the lower chambers and counted in a Neubauer chamber.

The creatine kinase cardiac isoform (CK-MB) was used as a marker of cardiac damage in individual blood samples collected from tail snips in heparinized capillary tubes. All samples were centrifuged and analyzed using commercially available kits according to the manufacturer’s instructions (LabTest Laboratory, Minas Gerais, Brazil).

Eight-week-old mdx and C57BL/10 mice were i.p. treated on alternating days with low-endotoxin azide-free–anti-γδ mAb (clone UC7-13D5) or low-endotoxin azide-free–isotype control (BioLegend) at 10 mg/kg for 4 wk. Immediately after the treatment (12-wk-old mice), two mice per group were euthanized to evaluate the treatment efficacy and the profile of lymphoid cell migration to the heart. We then followed up the levels of blood CK-MB in the remaining mice every 15 d for 6 wk (18-wk-old mice), when the mice were euthanized. This time point was chosen because the levels of CK-MB were slightly increasing in the isotype-treated mdx mice, following the natural progression of cardiac damage in DMD, and we wanted to evaluate the difference between both groups of mdx mice before they reached similar levels of cardiac damage.

Five-micrometer-thick frozen sections of cardiac tissues (left ventricles) from 18-wk-old mice were fixed using a 4% solution of paraformaldehyde for 20 min at room temperature. Samples were extensively washed using PBS/Tween 20 0.1%, incubated with a 3% solution of H₂O₂ for 20 min, washed again, and incubated with goat serum (to block Fcγ receptors) for 20 min. Samples were labeled with anti-F4/80 mAb (BioLegend) for 3 h at room temperature and washed using PBS/Tween 20 0.1%. For polymer HRP-conjugated secondary mAb and revelation, we used a DAB (3,3′-diaminobenzidine))/nickel kit according to the manufacturer’s instructions (Vector Laboratories, Burlingame, CA).

The data were analyzed using Student t tests or ANOVA (GraphPad Prism version 4.00 software; GraphPad Software, La Jolla, CA). A p value ≤ 0.05 was considered significant.

Given that previous studies demonstrated that γδ T cells are very important in regulating cardiac inflammation and damage (37), we evaluated by flow cytometry the presence of these cells in cardiac muscles from 6-, 8-, 10-, 12-, and 14-wk-old mdx mice (Fig. 1). Fig. 1A shows the FSC × SSC profile of dissociated hearts and the gate used in all analyses (R1). Fig. 1B and 1C show representative histograms of cardiac γδ T cells harvested from 12- and 14-wk-old mice, respectively. γδ T cells progressively increased in mdx cardiac muscle, reaching a peak in 12-wk-old mice and decreasing to <10% in 14-wk-old mice (Fig. 1D). Therefore, all other experiments were performed using 12-wk-old mice, unless indicated. We did not observe γδ T cells in diaphragm or gastrocnemius (Fig. 1E, 1F).

The phenotypic depiction of these cells showed that almost all cardiac γδ T cells are CD27⁺/CCR6⁻ (Fig. 2A) and that 27.2% of these cells express CD103 (Fig. 2B). We also observed that these cells do not express CD122 (Fig. 2B), NK1.1, or CD62-L (Fig. 2C). Because γδ T cell function can be inferred by the type of TCR that is expressed, we characterized these cells based on the expression of Vγ and Vδ chains. We observed that almost all cardiac γδ T lymphocytes from mdx mice are Vγ1⁺ (Fig. 2D). The expression of Vγ4 and Vγ5 subtypes was not observed (Fig. 2E, 2F, respectively), nor was the expression of Vδ6.3 and Vδ4 (Fig. 2G, 2H).

Distinct subtypes of γδ T cells can be identified by the production of some leading cytokines, such as IFN-γ or IL-17. One representative result showed that 52.4% of cardiac γδ T cells isolated from 12-wk-old mdx mice produced IFN-γ (Fig. 3A) but did not produce IL-17 (Fig. 3B), which is consistent with the expression of CD27.

As we observed that γδ T cells have very particular migration kinetics to mdx mouse hearts, we decided to evaluate possible cytokines that could be produced by the heart and could act as chemotactic factors. We used 6- and 12-wk-old mdx and C57BL/10 mice. We observed higher levels of IFN-γ in the cardiac muscle of 12-wk-old mdx mice (Fig. 3D) compared with 6-wk-old mdx mice (Fig. 3C). This finding is in agreement with the presence of IFN-γ–secreting γδ T cells in the hearts of 12-wk-old mdx mice. Furthermore, we observed that TNF levels are higher in the hearts of 6-wk-old mdx mice (Fig. 3C) compared with 12-wk-old mdx mice (Fig. 3D). No significant variations were found in IL-6, IL-10, or IL-12 between both groups of mdx mice. Only basal levels of all cytokines tested were found in either 6-wk-old (Fig. 3E) or 12-wk-old (Fig. 3F) C57BL/10 mice. Regarding the plasma, 6-wk-old mdx mice showed an increase of TNF and IL-12, and 12-wk-old mdx mice showed an increase of IL-6, IFN-γ, TNF, and IL-12, all compared with age-matched C57BL/10. We also observed higher levels of IL-6 and IL-12 in 12-wk-old mdx mice when compared with 6-wk-old mdx mice (Fig. 4). We found only basal levels of all cytokines measured in 6- and 12-wk-old C57BL/10 mice (Fig. 4).

As the cytokines evaluated showed no clear chemotactic attractors, we decided to investigate some chemokines that could be involved in the migration of these cells. First, we evaluated some chemokine receptors expressed by γδ T cells and observed no expression of CXCR2, CCR8, CCR9, or CCR5 by the cardiac γδ T cells (Fig. 5A–D, 5G). In contrast, we observed 22.8% of γδ⁺ CCR3⁺ in the lymphocyte gate (Fig. 5E) and calculated that 79.7% of the TCRγδ⁺ cells are CCR3⁺ (Fig. 5G). Regarding CCR2 expression, we found 10.2% of γδ⁺ CCR2⁺ in the lymphocyte gate (Fig. 5F) and calculated that 32.9% of the TCRγδ⁺ cells are CCR2⁺ (Fig. 5G). Regarding some γδ T cell–recruiting chemokines that could be produced by cardiac tissue, we evaluated CCL7, CCL2, CCL3, CCL4, CCL5 (Fig. 6A–E), and GM-CSF (Fig. 6F) in the ventricle extracts. There were no differences in C57BL/10 between the two ages evaluated, considering all molecules tested (Fig. 6A–F). Moreover, most chemokines were increased in the mdx mice of both ages compared with C57BL/10, except CCL3, CCL4, and GM-CSF in 6-wk-old mice, which were produced at similar levels in both mouse lineages. When we compared only mdx in both ages, we observed statistically similar levels of CCL7 and CCL3 and increases in CCL2, CCL4, CCL5, and GM-CSF in 12-wk-old mice (Fig. 6A–F). As a control, we evaluated these chemokines in spleens from 6- and 12-wk-old mdx mice and observed a very similar profile of chemokines produced by both groups of mice, with only CCL3 increased in the spleens of 12-wk-old mice (Fig. 6G).

Because γδ T cells also secrete chemotactic factors and recruit leukocytes to inflammatory foci, we also evaluated the chemokines that were produced by freshly collected and purified cardiac γδ T cells (Fig. 6H). We observed high levels of CCL2, CCL4, and CCL5 in these cells.

The combination of chemokines secreted by the cardiac tissue and the profile of chemokine receptors expressed by γδ T cells shed light on what factors could be involved in this process. On the basis of our results, we performed in vitro migration assays using recombinant CCL2 and CCL5 ligands and CCR2 and CCR3 ligands. We therefore used spleen purified γδ T cells from 12-wk-old mdx mice, considering that these cells circulate in the blood and would respond to chemotactic factors secreted by the heart. These cells were exposed to recombinant CCL2 and CCL5 individually or in combination (Fig. 7A), and we observed similar high levels of γδ T cell migration in all situations. Moreover, the blockage of each chemokine using neutralizing Abs reduced cell migration (Fig. 7A). When we used fresh cardiac extract from retroperfused hearts, we observed no migration induced by 6-wk-old mdx or 12-wk-old C57BL/10 mice, but extract from 12-wk-old mdx mice induced a significant migration of γδ T cells (Fig. 7B). To further explore the importance of these chemokines in the recruitment of γδ T cells to the heart, we neutralized CCL2 and/or CCL5 in the cardiac extract of 12-wk-old mdx mice (Fig. 7B). We observed a significant reduction in γδ T cell migration only when both chemokines were neutralized compared with the total cardiac extract harvested from 12-wk-old mdx mice.

γδ T cells play important regulatory roles in inflammatory sites, which are performed by two main nonexcluding pathways: the secretion of cytokines and cytotoxic activity. We then evaluated whether purified cardiac γδ T cells from 12-wk-old mdx mice (Fig. 8A, 8B) would exert cytotoxic activity and used an LDH-based assay (Fig. 8C) or 7AAD (Fig. 8D–F, 8J). Regardless of the cytotoxic pathway used, we observed that γδ T cells induced cell death of endomysial cardiac adherent cells (90 ± 3% of F4/80⁺ cells, data not shown) (Fig. 8D–F, 8J), but not peritoneal cells (96 ± 2% of F4/80⁺ cells, data not shown) (Fig. 8C). This result suggests that the repertoire of costimulatory and adhesion molecules affects the cytotoxic capacity of these lymphocytes. As an additional control, we purified spleen CD8⁺ T cells from 12-wk-old mdx mice, and these unstimulated cells were assayed against adherent cardiac cells (Fig. 8G–I, 8K). We observed no significant cytotoxic activity using 7AAD.

To discern which cytotoxic pathway was mainly responsible for the lysis of cardiac target cells, we evaluated the release of granzyme B (perforin-based pathway) by γδ T cells after the interaction with cardiac or peritoneal target cells (Fig. 9A) and the expression of Fas-L (Fig. 9B). Our data point to a cytotoxic function primarily based on perforin/granzymes because we observed a high granzyme B activity in the supernatant of the assayed effector:cardiac target cells (Fig. 9A). Regarding Fas-L expression, in a representative result, we observed only 4.8% of γδ⁺ Fas-L⁺ in the lymphocyte gate (Fig. 9B). Considering three independent experiments, we found 4.0% ± 3.0 (average ± SD) (data not shown). This means ∼16.0% of γδ T cells express Fas-L (in γδ T cell gate) (data not shown).

It was necessary to target the in vivo relevance of γδ T cells in the progression of cardiac damage and myocarditis in mdx mice, and therefore, we depleted these cells by treating 8-wk-old C57BL/10 and mdx mice with anti-γδ mAb or isotype control for 4 wk. Twelve-week-old mdx mice treated with isotype Ab showed ∼20.0% of cardiac γδ T cells (Fig. 10A, 10C) in the gate of lymphoid cells (FSC × SSC) compared with 2.5% in mdx mice treated with anti-γδ mAb (Fig. 10B, 10C). The analysis of CD4⁺ and CD8⁺ αβ T cells in the gate of lymphoid cells (excluding γδ T cells) showed an enrichment of both subpopulations after γδ T cell depletion (Fig. 10D–F). Regarding cardiac damage, we observed basal levels of CK-MB at all time points in both groups of C57BL/10 mice (Fig. 10G). In contrast, isotype-treated mdx mice showed an increase in CK-MB levels from the fourth week of treatment onward, and in anti-γδ–treated mdx mice, we observed the highest levels of CK-MB throughout the period evaluated (Fig. 10G). We then labeled F4/80⁺ cells (mostly macrophages) in frozen sections obtained from the ventricles of 18-wk-old mice and observed no myeloid infiltration in isotype-treated (Fig. 11A, 11E) or anti-γδ–treated (Fig. 11B, 11E) C57BL/10 mice. Isotype-treated mdx mice showed few and small diffuse F4/80⁺ cells (Fig. 11C, 11E, 11F), in contrast to anti-γδ–treated mdx mice, which showed more myeloid infiltration but generally larger inflammatory foci (Fig. 11D–F). We observed no lymphoid cardiac infiltration by flow cytometry or H&E staining in any of the groups studied at this age (data not shown).

Since the discovery of γδ T lymphocytes in 1986 (36), these cells have been described as pivotal players in the immune response. Despite being components of the innate response, these cells directly affect the adaptive response and are capable of altering the course of inflammatory responses and the susceptibility or resistance to infections, sterile inflammation, and tumors, for example. In the particular case of myositis and myocarditis, it has been reported that αβ and γδ T cells act as proinflammatory cells, while playing different roles in the progression of diseases (18, 38). For example, in the particular case of myocarditis induced by coxsackievirus B3, it has been shown that γδ T cells induce the death of protective CD4⁺ Th2 cells, enriching the cardiac tissue in pathogenic Th1 CD4⁺ T cells (37). Moreover, γδ T cells have also been described as important inflammatory cells in a rare variant of myositis (18). Despite the importance of γδ T cells in immunomodulation and proinflammatory activity in muscle diseases, the role played by these cells in the pathogenesis of DMD has not been described. The selective balance of lymphoid and myeloid cell migration and function in muscles in DMD is certainly the result of multiple immunomodulatory pathways, which ultimately define the outcome of the disease. We have been puzzled by the relatively few lymphoid cells found in the muscles of mdx mice and DMD patients, despite the disruption of the sarcolemma. Previous data published by our group (10) showed that CD4⁺ and CD8⁺ αβ T cells are primarily found in the hearts of young (6-wk-old) mdx mice and decline by the age of 12 wk owing to a P2 × 7–dependent shedding of CD62-L (10). This finding coincides with the peak of γδ T cells, as observed in the present work. As long as at this age (12 wk) there is no detectable cardiac damage, with very few inflammatory foci and no functional alterations, we questioned what would be the function of these γδ T cells in the heart. In this work, our goal was to characterize the phenotype, chemotactic factors secreted by the cardiac tissue, and a possible immunoregulatory activity of these γδ T cells in vitro and in vivo.

Because the membrane phenotype of γδ T cells may shed light on the functional activity of these cell subpopulations, we first evaluated the expression of CCR6 and CD27 as main producers of IL-17 and IFN-γ, respectively (25, 26). Cardiac γδ T cells isolated from mdx mice were only CD27⁺, and the production of IFN-γ, but not IL-17, was confirmed by intracellular labeling via flow cytometry. Regarding CD122, despite the fact that the expression of this molecule is related to increased production of IFN-γ in αβ T cells (39), we did not observe this correlation in cardiac γδ T cells as long as these cells were CD122⁻ IFN-γ⁺. In contrast, CD103 (usually expressed by mucosal γδ T cells) was observed in ∼30% of the cardiac γδ T cells. CD103 is expressed in different cell types, including subpopulations of αβ T lymphocytes, playing a role in cell motility (40), intercellular adhesion (41), and immune regulation (42). It has been published that CD8⁺CD103⁺ T cells proliferate less and have reduced cytotoxic activity after in vitro stimulation with alloantigens, acquiring a suppressive function over PBMCs (42). However, it remains to be clarified whether cardiac γδ T cells have a suppressive activity over other inflammatory cells in the cardiac foci of mdx mice. It has been shown previously that Vγ1⁺ T cells obtained from uninfected mice are cytotoxic against macrophages harvested from mice infected with Listeria monocytogenes (43). These data are in agreement with our results, in which we found that γδ T cells are cytotoxic against cardiac adherent cells. Although there is no infection in our model, muscle damage in mdx mice leads to inflammatory infiltration and secondary damage to the tissue, inducing cycles of proinflammatory DAMPs released by the heart. In mdx mice, this cytotoxic function must be transitory because γδ T cells are reduced in the cardiac tissue of 14-wk-old mdx mice despite the continuous release of DAMPs. Interestingly, cardiac γδ T cells were cytotoxic against cardiac macrophages but exhibited no cytolytic function toward peritoneal macrophages. Although both target cells were from mdx mice, the cardiac cells were harvested from an active inflammatory site, which is not observed in the peritoneal cavity. The required molecular repertoire for E:T cell recognition, adhesion, and target cell death is likely fulfilled only when these cells are under the influence of local proinflammatory factors, such as chemokines and cytokines, that alter membrane phenotype and cell function. It is important to note that we obtained as cardiac adherent cells mainly macrophages; cardiomyocytes were not present in the cell culture. Therefore, this result does not suggest that cardiac γδ T cells are directly inducing cardiac damage. On the basis of this figure, we can conclude only that cardiac γδ T cells are cytotoxic against cardiac endomysial cells in a perforin-dependent manner. On the basis of this result and the in vivo data, we suggest that cardiac γδ T cells could play a protective role by selectively inducing the death of inflammatory macrophages, the main inflammatory cell type responsible for the death of myofibers (44).

The evaluation of cytokines in the blood of 6 wk-old mdx mice showed the onset of inflammation in the organism with higher levels of TNF and IL-12 compared with counterpart C57BL/10 mice. The increases in IL-6, TNF, IFN-γ, and IL-12 in 12-wk-old mdx mice arrested the progression of the inflammatory response in these mice. The evaluation of cytokines in the cardiac tissue of 6-wk-old mdx mice showed a different pattern compared with blood, with high levels of IFN-γ, but no production of IL-12. In addition, we found the production of only IFN-γ and TNF in the cardiac tissue of 12-wk-old mdx mice. Of interest, IL-6 was not observed in the hearts of 12-wk-old mdx mice, although high levels were observed in the blood. It is known that IL-6 is an inflammatory cytokine associated with stimulation of hypertrophic skeletal muscle growth and myogenesis through regulation of the proliferative capacity of muscle stem cells (45). However, our data indicate that any given IL-6–dependent functions would be at least markedly reduced in the cardiac tissue at this time point.

We did not observe γδ T cells in skeletal muscles (diaphragm, gastrocnemius, soleus) of mdx mice of any age, and this attests to the specific and polarized recruitment of these cells to the cardiac tissue during a specific window of time. A number of chemotactic factors can play a role in the activation and recruitment of blood γδ T cells to the heart, such as a variety of DAMPs released from damaged muscle fibers, including mitochondria molecules (35), CCL2 (46), CCL5 (47), CCL3 (48), leukotriene B₄ (49), and others. Our data were inconclusive regarding the cytokines secreted by the cardiac tissue as recruiters of γδ T cells. However, we observed that CCL2, CCL4, CCL5, and GM-CSF were increased in the hearts of 12-wk-old mdx mice compared with 6-wk-old mice, despite the fact that few inflammatory macrophages are found in the hearts of these mice (50). Therefore, CCL2 and CCL5, previously described as γδ T cell recruiters in a number of inflammatory conditions (46, 47), could mediate the migration of these lymphocytes to the heart. The evaluation of cognate chemokine receptors of cardiac γδ T cells showed the expression of CCR2 (CCL2 receptor) and CCR3 (CCL5 receptor), suggesting that CCL2/CCR2 and CCL5/CCR3 counterparts could mediate γδ T cell migration. It is noteworthy that when we added both recombinant chemokines, there was no additive effect for cell transmigration, and this may be a result of the intracellular saturation of the signal transduction machinery. According to our calculations, in the extract of 12-wk-old mdx mice, there was 27.5 pg/well of CCL2 and 14 pg/well of CCL5 (in 500 μg of total cardiac proteins), which are ∼120 (CCL2) and 235 (CCL5) times less than the concentration of recombinant chemokines added per well.

Regarding the heart, we observed in vivo that γδ T cells play a central role in inflammatory regulation and cardiac damage, as we depleted these cells and observed the following: enrichment in αβ T lymphocytes in 12-wk-old mdx mice, greater infiltration of F4/80⁺ cells in 18-wk-old mdx mice, increased cardiomyocyte damage at all time points, but no increased fibrosis (data not shown) all in mdx mice. It is important to emphasize that we did not associate γδ T cell function with cardiac damage in 8- or 12-wk-old mdx mice, in neither the in vitro nor the in vivo assays. In contrast, we indicate an early protective immunomodulatory function of these cardiac lymphocytes, possibly based on cytotoxic activity against pathogenic macrophages in the heart. This function could play a role in the late onset of cardiac damage in older mice. We believe that the late establishment of myocarditis in mdx mice is mostly based on CD62-L depletion in T lymphocytes (10) and the cytotoxic activity of local γδ T cells selectively killing pathogenic macrophages. All these data place γδ T cells in a central position in cardiac immunoregulation and damage, and γδ T cells seem to be a pivotal player contributing to the late onset of heart failure in mdx mice.

We thank the Multiuser Research Facilities of Flow Cytometry, Unities of Cell Sorting and Multiparametric Analysis of Instituto Oswaldo Cruz, FIOCRUZ, Rio de Janeiro, Brazil.

The authors have no financial conflicts of interest.