Proteomic screens of chemokines in the skeletal muscles of CMD patients
Proteome analysis consisted of skeletal muscle biopsies from patients with confirmed diagnosis of Bethlem Myopathy (BM, n=5), Ulrich Congenital Muscular Dystrophy (UCMD, n=8) and Merosin-deficient congenital muscular dystrophy type 1A (MDC1A, n=5). Patients were from 2 to 67 years of age and presented a range of clinical symptoms from mild to severe. The clinical characteristics of CMD patients are summarized in Table S1. Assay controls included skeletal muscle biopsies collected from 18, 43 and 58-years old healthy individuals with no CMD history. The human chemokine array was used to simultaneously survey 38 known human chemokines. The data were analyzed by grouping the patients by disease type (BM, UCMD and MDC1A) and represented based on a comparison to muscle samples of healthy donors. The ratio of mean pixel densities of individual chemokine in CMD samples to that in control group is presented as the fold-difference (Table 1, Fig.S1A). Analysis of biopsies collected from the healthy group demonstrated that homeostatic muscles contain constitutively low levels and limited repertoire of chemokines, with CTAK (CCL27), GRO (CXCL1,2,3), IL-8 and IL-10, MCP-1 (CCL2), MIP-1 beta (CCL4), NAP-2 (CXCL7), and RANTES (CCL5) being above the minimal value of the detection. Proteome profile of CMD-derived biopsies reveled disease-related changes (Fig. 1). Cross analysis of patient data showed that all examined muscle biopsies share similar chemokine profile regardless of genetic abnormality or severity of disease. Moreover, most dominant molecules were found to be a pro-inflammatory chemokines and common for all three CMD types, those included NAP-2 (CXCL7), GCP (CXCL6), GRO (CXCL1,2,3), RANTES2 (CCL5), and MCP-1 (CCL2) (Fig. 1, Table 1). On special note, the level of CCL5 was associated with CMD severity, showing modest increase (1.7-fold) in BM-derived muscles and considerably higher presence in muscles of UCMD (5.7-fold) and MDC1A (11.0-fold) patients (Table 1). However, prognostic value of this molecule will require further statistical analysis in a larger cohort of patients with careful phenotypic evaluation. Collectively, robust analysis of CMD muscle-associated chemokines revealed a distinct subset of chemokines that may contribute to the pathology of the major CMD types. These data also revealed several chemotactic pathways that could be further exploited for the effective recruitment and homing of the systemically transplanted therapeutic stem cells to CMD-affected muscles and the improvement of cell-based therapies.
Proteome analysis of chemokines in LAMA2-deficient mice
The Lama2 (dyW) mice recapitulate the clinical manifestations of MDC1A patients and display severe muscular dystrophy, with death around 5-7 weeks of age. Proteomic screens of chemotactic molecules in the gastrocnemius muscle (GCM) were conducted at different stages of disease progression: initial (7-14 days), acute (3-4 weeks), and terminal (5-6 weeks). Assay controls were consisted of proteomic profiles utilizing the GCM tissues from age-matched wild-type counterparts. Densitometry data analysis of individual chemokine is reported as the fold-difference change of dyW compared to wild-type samples. Evaluation of 25 known mouse chemotactic molecules in the GCM of wild-type mice showed constitutive presence of few chemokines at low levels, mainly produced as a part of normal homeostasis of muscle tissue (Fig. 2A, Fig. S1B). Contrary, analysis of the GCM from dyW mice showed abundant presence of several chemotactic molecules at all stages of disease (Fig. 2B). Pairwise comparison of the GCM sampled from dyW and normal mice revealed the significant induction of seven distinct CC and CXC class chemokine ligands, including CCL6, C5/C5a, RARRES2, CCL27, IL-16, CCL2, CXCL1, CCL8, CCL12, CCL9/CCL10, and CXCL12 (Table 2). Patterns of most identified chemokines showed early induction as soon as 1 week of age and maintained expression until termination point. Analysis of the diaphragm muscle (DM) sampled from dyW and wild-type mice showed considerable presence of CCL6 and RARRES2 and modest de-regulation of IL16, CCL2, CXCL1 and CCL12 (data not shown). Interestingly, the levels of some identified molecules were variable depending on the tissue and did not necessarily coincide with stage of disease. Specifically, CCL6, RARRES2 and CCL9/10 showed the highest level at initial stage of disease (2 weeks) in the DM but reached the maximum only in acute phase (3 weeks) in the GCM. For the same token, IL6 level had modest increase at 2 weeks in the DM and was maintained at that level as disease progressed; however, its level did not change significantly and was steady in the GCM. CCL8 and CCL12 levels dis not vary much and demonstrated consistent pattern associated with all disease stages. Contrary, analysis of serum samples did not produce any noticeable changes in chemokine profile despite disease progression; however, the serum collected from 1 week-old animals showed temporal increase in CCL6 (5.6 -fold), CXCL5 (4.9-fold) and CCL9/CCL10 (9.0-fold), which reached the basal level by second week of life (data not shown). Together, these data suggest that rapid induction and sustained expression of several chemotactic molecules in the GCM and DM of LAMA2-deficient mice may provide a selective mechanism for inflammatory cell recruitment and, thus, play a role in disease pathology. Also, identified chemotactic signatures suggest that therapeutic stem cells can be recruited to the affected tissues by the similar chemotactic mechanism as immune cell trafficking.
Proteome analysis of chemokines in COL6-deficient mice under physiological and pro-inflammatory conditions
The homozygous Col6a1-/- mutants completely lack COL6 in the tissues but have normal life span. Skeletal muscles of adult Col6a1-/- mice display a myopathic histology, including fiber necrosis, phagocytosis, a pronounced variation in the fiber diameter, and signs of stimulated regeneration of fibers with necrotic fibers particularly frequent in the diaphragm [20]. Since the muscle phenotype of the Col6a1-/- mice is much milder than that of the human UCMD patients, CTX was used to exacerbate the temporal muscle abnormality, as previously shown in this model [22]. CTX selectively injures myofibers but leaves nerves, blood vessels and satellite cells morphologically intact. Proteome profile was undertaken to define a signature of chemokine release corresponding to sequential stages of skeletal muscle injury and regeneration. Following CTX treatment, the GCM biopsies were harvested at defined periods (0/uninjured, 1, 3, 7, 14, and 21 days). Assay controls consisted of data collected from the age-matched wild-type mice under similar treatment conditions. Data is presented as the fold-change difference of densitometry reads between the Col6a1-/- and control groups. Comparative proteome profiling of the Col6a1+/+ and Col6a1-/- GCMs under uninjured conditions showed considerable presence of several chemokines in Col6a1-/--derived muscles, including Ccl21 (4.0-fold), C5/C5a (18.1-fold), RARRES2 (8.4-fold), IL16 (11.9-fold), Ccl2 (29.3-fold), Ccl8 (8.7-fold), Ccl12 (22.3-fold), Cxcl1 (5-fold) and Cxcl12 (42.6-fold) (Fig. 3A and Table 3). Twenty-four hours after CTX injury, a substantial release of several pro-inflammatory chemokines was evident in the wild-type GCM, as judged by increased levels of Ccl6, C5/C5a, RARRES2, IL16, Cxcl5, Ccl8, Ccl12, and Ccl9/Ccl10 (Fig. 3B). Interestingly, chemokine profile of Col6a1-/- mice showed modest response to CTX, affecting only few molecules, including C5/C5a, Cxcl1, Cxcl10, Ccl8 and Ccl9/Ccl10 (Fig. 3C). Moreover, the level of induced chemokines was comparable between wild-type and Col6a1-/- muscles, with the exception of Ccl21, C5/C5a, Cxcl10, and Cxcl12 (Table 3). Further analysis of regenerating muscles showed that the levels of CTX-induced chemokines in wild-type GCM were reestablished to the baseline values by 2 weeks after injury. In contrast, high levels of CTX-induced chemokines in Col6a1-/- mice persisted for 2 weeks until they restored to the basal level by week 3 (Table 3).
Chemotaxis-mediated recruitment of ADSC to skeletal muscles in vivo
Our previous FACS-based assessment of receptor activity in minimally cultured primary stem cells showed that ADSC have very limited percentage of cells with functional chemokine receptors, accounting for their ineffective recruitment into muscles [21]. Because homing of systemically infused ADSC to the muscle could be enhanced via the use of cells uniformly expressing principal receptor, ADSC engineered to express Ccr2 were systemically transplanted into NCr nude mice followed by localized intramuscular administration of the mouse recombinant Ccl2 chemokine into the left GCM. As expected, mice transplanted with heterogeneous ADSC (less than 6% of cells positive for Ccr2 receptor) showed significant cell entrapment in lungs during first 24 hours, with no detectable engraftment into chemokine-treated or untreated limbs (Fig. 4A). In the next 48-72 hours, control mice did not show any appreciable signals, indicating exiting of cells along unspecific axes within the body (Fig. 4B, C). In sharp contrast, as early as 24 hours after transplantation, migration of Ccr2-postive ADSC along created chemotactic gradient was observed in the chemokine-treated GCM but not in untreated right limb (Fig. 4D). Additional 48 hours led to a more robust migration of the receptor-expressing cells from the circulation to the chemokine-treated muscle, as revealed by the marked increase in red fluorescent signal (Fig. 4E, F). Direct immunofluorescence analysis of muscle biopsies showed preferential colonization of Ccr2-positive ADSC around blood vessels (data not shown). Taken together, these findings strongly support our hypothesis that engagement of stem cell receptor with tissue-derived chemokine is a critical step in cell recruitment into muscle tissue.
Chemotactic recruitment of ADSC to CMD-affected muscles
Comparative proteome analysis of human and mouse CMD samples as well as pairwise matching of identified chemokines and their cognate receptors allowed us to identify few common chemotactic axes, including CXCL7/CXCL1,2,3-CXCR2 and CCL2-CCR2, as potential gradients for targeting of ADSC into CMD muscles. Chemotactic activity of ADSC uniformly expressing CMD muscle-specific chemokine receptors was assessed in a pre-clinical setting using the COL6 deficient mouse model. Reinstatement of COL6 deposition in skeletal muscles was examined after systemic transplantation of selected Ccr2- and Cxcr2-positive ADSC (1.5-2x106/injection/mouse) fluorescently labeled with DiOC18 under physiological and CTX-induced inflammatory conditions, respectively. The right GCM was injected with CTX and left GCM was injected with PBS and served as control for natural homing followed by cell transplant 24 hours after injury. To compare potential of selected ADSC in normal and muscle-damaged condition, mice were also treated with unselected ADSC in a similar fashion. Live imaging of both CTX-treated and control limbs was performed at 4 (dorsal view), 14 and 21 (ventral view) days after cell transplantation. The experimental animals were concluded 3 weeks after a single transplant. The mice transplanted with heterogeneous ADSC did not show any appreciable signals (data not shown). Contrary, the red fluorescent signals of DiOC18/Ccr2- and DiOC18/Cxcr2-positive ADSC were detectable in all transplanted animals and persisted for the entire study, suggesting the efficient homing and engraftment of cells in the muscle environment under the influence of chemokines in both physiological and CTX-injured conditions (Fig. 5A). Moreover, the detected fluorescent signals were significantly higher in the limbs treated with CTX. Interestingly, the transplantation with Cxcr2-positive ADSC (Cxcl1/Cxcr2, Cxcl2/Cxcr2 and Cxcl5/Cxcr2 axis) resulted in more efficient migration than Ccr2-positive ADSC (Ccl2/Ccr2 axis) (Fig. 5A, B). This is in good agreement with our chemokine protein array profiles showing the significant induction of Cxcl1, Cxcl2 and Cxcl5 chemokines after CTX injury, which are all potent chemoattractants for Cxcr2-positive inflammatory cells. Direct immunofluorescence analysis of muscle biopsies at 21 days after transplant revealed significant presence and wide distribution of DiOC18-ADSC in muscle interstitium (Fig. 5C). Co-immunostaining with antibodies against COL6 and LAMA2, which is an integral component of the muscle basement membrane, showed overlapping fluorescent signals of both proteins at the basement membrane of individual myofibers in both Ccr2- and Cxcr2-positive transplants. Morphometric analysis showed that the number of COL6-labeled myofibers was greater in Cxcr2-positive ADSC transplanted mice (Fig. 5D). Together, these findings strongly suggest that stem cells uniformly expressing specific receptors can be efficiently targeted to the CMD-affected skeletal muscle under influence of the CMD muscle-derived chemotactic signals.