Activation of the Pro-Oxidant PKCβII-p66^Shc Signaling Pathway Contributes to Pericyte Dysfunction in Skeletal Muscles of Patients With Diabetes With Critical Limb Ischemia

Critical limb ischemia (CLI) represents the most severe manifestation of peripheral arterial disease and the major cause of foot amputation in the U.S. (1). Most amputations are performed on people with diabetes, who have rampant atherosclerosis and poor angiomyogenesis (2–5).

Satellite cells and pericytes, which reside on opposite sides of the myofiber basement membrane, represent the main myogenic stem/progenitor cells in postnatal skeletal muscle (6). Satellite cells are well-acknowledged targets of diabetes-induced damage and are contributors of diabetic vascular myopathy (7–10). Recent evidence indicates that pericytes play a key role in vascular and muscular regeneration (11–13) and have the potential to become favorite candidates for cell therapy for peripheral arterial disease and myocardial ischemia (14–17). However, to the best of our knowledge, no investigation has assessed the impact of diabetes on the functional and molecular makeup of muscular pericytes (MPs).

The mitochondrial adaptor protein Shc1, isoform p66 (p66^Shc), a redox enzyme that triggers mitochondrial apoptosis, is implicated in the pathophysiology of aging and cardiovascular disease (18–20). Studies in experimental models suggest that modulation of p66^Shc expression and activity may be a novel and effective target for the treatment of cardiovascular complications. For instance, abrogation of p66^Shc results in protection from angiotensin II–induced cardiomyocyte damage (21), improvement of neovascularization and muscle fiber survival after induction of limb ischemia, and preservation of proliferation and differentiation of satellite cells exposed to high oxidative stress (22,23). Upstream modulators of the p66^Shc signaling pathways, such as the protein kinase C isoform β-II (PKCβII), which activates p66^Shc through phosphorylation of its serine 36 residue at the COOH-terminal moiety, also represent an attractive therapeutic target to halt ischemic complications (20,24). After successful use in diabetic animals, the PKCβII inhibitor LY333531, also known as Ruboxistaurin, has recently shown a clear therapeutic benefit in clinical trials of diabetic microvascular complications (25,26). However, whether PKCβII inhibitors exert positive effects on myogenic stem cells from patients with diabetes with vasculopathy currently remains unknown.

The current study investigates molecular therapeutic targets to improve angiomyogenesis in patients with complicated diabetes. We assessed functional and molecular features of MPs harvested at the rim of tissue above the plane of limb amputation, which is a rescuable area similar to the peri-infarct border zone. We show for the first time that diabetic MPs (D-MPs) are dysfunctional in many respects. They have a reduced capacity to expand in culture, differentiate into multinucleated myotubes, and support endothelial cells (ECs) network formation. Furthermore, an increased oxidative stress is recurrent in D-MPs, due to the downregulation of superoxide dismutase 1 (SOD-1) and catalase, and upregulation and activation of p66^Shc. Importantly, blockade of p66^Shc phosphorylation by PKCβII inhibition restores D-MP functionality, providing a novel indication for molecular treatment of the angiomyogenic pathology in patients with diabetes.

Skeletal muscle biopsy samples were obtained from control subjects and patients with diabetes after informed consent was given, in line with the guidelines on human rights of the Declaration of Helsinki. MPs were isolated from the following: 1) different anatomic locations of the lower extremities from control subjects, referring to our institutions for investigations/therapeutic interventions related to leg varicosity or suspected bone-related pathologies that then resulted in negative findings (n = 14); or 2) sartorius muscles from patients with type 2 diabetes at the occasion of major amputation for CLI (n = 18). CLI was diagnosed according to Trans-Atlantic Inter-Society Consensus Document on Management of Peripheral Arterial Disease (TASC) 2007 guidelines (i.e., rest for pain and/or ulcer or gangrene, transcutaneous oximetry at the dorsum of the foot of <30 mmHg, and/or ankle pressure of <70 mmHg). Patient characteristics are listed in Table 1. We attempted to perform all of the analyses on the same sample. When this was not possible because of the small size of the harvested tissue, priority was given to histological analyses and then to the proliferation assay. Supplementary Table 1 summarizes the attribution of samples to different assays.

Morphometric analysis was performed on hematoxylin-eosin–stained sections. For capillary analysis, sections were boiled in sodium citrate buffer, pH 6.0, incubated with mouse monoclonal anti-CD31 antibody clone CJ70A (1:200; Dako), then washed and incubated with secondary antibody followed by DAB+ substrate 1:50 (REAL EnVision Kit; Dako). Analysis of inflammatory infiltrates used rabbit monoclonal anti-CD3 (clone 2GV6) and mouse monoclonal anti-CD68 (clone PG-M1) antibodies (both from Ventana). To localize MPs within the muscle structure, cryosections were treated with an alkaline phosphatase (ALP) staining solution containing nitro-blue tetrazolium chloride (160 μg/mL) and 5-bromo-4-chloro-3′-indolyphosphate p-toluidine salt (300 μg/mL; both from Sigma-Aldrich) in ALP buffer (100 mmol/L TrisHCl, 150 mmol/L NaCl, and 1 mmol/L MgCl₂, pH 9.0). The same solution was used to stain isolated cells previously fixed in 2% formalin. ALP quantification in cryosections was performed using Cell Profiler open source software distributed under a GPLv2 public license.

Cryosections from diabetic and control muscle biopsy samples were stained with antilaminin antibody and analyzed using an ImageJ macro. For each sample (n = 3), an area of >2,000 single fibers was measured.

Human MPs were isolated after well-established procedures (12,27). Briefly, muscle biopsy samples were finely minced and digested with collagenase II (100 units/mL) for 45 min at 37°C with shaking. The digestion mixture was centrifuged and resuspended in growth medium (α-minimum essential medium supplemented with 20% FBS). The cell suspension was filtered through a 70-μm cell strainer, dispensed in plastic dishes at clonal density (1,000 cell/cm²), and incubated at 37°C with 5% CO₂ in the growth medium. MPs were selected by plastic adherence in culture for at least 10 days when they form colonies positive for ALP, neural/glial antigen 2 (NG2), and CD146 (27).

For ultrastructural analysis, a pellet of MPs was fixed for 2 h at 4°C in a mixture of 2% paraformaldehyde (PFA) and 2% glutaraldehyde in 0.05 mol/L, pH 7.3, cacodylate buffer; postfixed in 1% osmium tetroxide; and embedded in epon-Araldite. Thin sections were counterstained with uranyl acetate and lead citrate, and examined with a Philips/FEI Morgagni electron microscope.

After fixation with 4% PFA, cells were permeabilized with 0.3% Triton X-100 in PBS 1× plus 1% BSA, then blocked with goat serum 10% in PBS 1×; and incubated with polyclonal rabbit anti-NG2 (Merck Millipore), mouse monoclonal anti-CD146 (clone OJ79c; Abcam), mouse monoclonal antidesmin (clone D33; Dako), mouse monoclonal anti-MyHC (clone MF20; Developmental Studies Hybridoma Bank), and mouse monoclonal anti-Ser36-phospho-p66^Shc (clone 6E10; Abcam); and diluted following the manufacturer instructions. Cells were further incubated with goat anti-mouse or anti-rabbit fluorescent secondary antibodies (Alexa Fluor 488 or 555; Life Technologies). Nuclei were counterstained with DAPI (PanreacAppliChem). Microphotographs were acquired using the imaging software AxioVision Imaging System (Zeiss, Jena, Germany). When required for three-dimensional image acquisition, an Olympus FV 1000 confocal laser scanning microscope with a 60× oil-immersion lens was used.

MPs were stained for surface antigen expression using the following antibodies: CD44-APC, CD90-APC, and ALP-PerCP Cy5.5 (all from BD Biosciences). For Ser36-phospho-p66^Shc quantification, cells were fixed and permeabilized using BD Citofix/Citoperm kit (BD Biosciences). They were then incubated with anti-Ser36-phospho-p66^Shc (Abcam), followed by Alexa Fluor 633–conjugated secondary antibody. Fluorescence was analyzed on a FACSCanto flow cytometer using the FACSDiva software (BD Biosciences), setting a nonlabeled population as a negative control.

MPs were seeded at 10⁵ cells/well in two-well glass chamber slides and were expanded in culture for 7 days to allow cell fusion and myotube formation. Cells were fixed with 4% PFA and permeabilized with 0.3% Triton X-100 in 1% BSA. Myogenic differentiation and myotube formation were assessed by staining cells with anti–myosin heavy chain (MyHC) antibody (clone MF20; Developmental Studies Hybridoma Bank), followed by Alexa Fluor 555–conjugated secondary antibody.

MPs were induced to differentiate into adipocytes using an inductive medium for 4 days (i.e., DMEM GlutaMAX high-glucose, 1% penicillin and streptomycin, 1% sodium pyruvate supplemented with 20% FBS, and insulin [1 µg/mL]). Adipogenic differentiation was assessed by Oil Red O staining (Sigma-Aldrich) after PFA fixation.

MP proliferation was assessed by ELISA BrdU colorimetric assay kit (Roche). Briefly, cells (10³ cells/well in a 96-well microplate in triplicate) were grown for 24 h and then treated with BrdU (10 μmol/L final concentration) for an additional 24 h. The colorimetric reaction was stopped by adding H₂SO₄ (250 mmol/L final concentration) and read immediately at 450 nm.

Human umbilical vein ECs (HUVECs) and MPs were seeded in an eight-well Permanox chamber slide (Nunc) coated with Matrigel (3D; BD Biosciences) alone (3.75 × 10⁴ cells/well) or cocultured at 1:4 ratio (MPs to HUVECs) in EBM medium, supplemented with 0.1% BSA. Cells were incubated for 5 h postseeding to allow network formation. Pictures of the network were taken on an inverted phase-contrast microscope (AxioObserverA; Zeiss, Jena, Germany), and network formation was assessed by counting the number of branches per field. Similar angiogenesis assays were performed to assess the effect of MP-derived paracrine factors, by adding conditioned culture media (CCMs) to HUVECs seeded onto Matrigel. In a subset of these studies, the influence of MPs-CCM on in vitro angiogenesis was assessed in the presence of high oxidative stress by adding 300 μmol/L H₂O₂ to the media. All experiments were performed in duplicate.

Levels of vascular endothelial growth factor vascular endothelial growth factor-A, IGF-I, MCP-1, SPARC (secreted protein, acidic, and rich in cysteine), hepatocyte growth factor, fibroblast growth factor, and angiopoietin (Ang)-2 in MPs-CCMs were measured by ELISA using Duo Set kits (R&D Systems/Bio-Techne). Levels of Ang-1 were analyzed using the Human Angiopoietin-1 ELISA Kit (Sigma-Aldrich). All analyses were performed following the manufacturer instructions in triplicate (25 μL of sample per replicate).

Exosomes were isolated by sequential ultracentrifugation from MPs-CCMs collected after 40 h of growth in a medium added with 20% exosome-free FBS, as previously described (28,29). Exosome-like microparticle abundance was revealed through nanoparticle tracking analysis, and aliquots of the same preparation were added to HUVECs in the Matrigel assay.

The HUVEC intracellular redox state after treatment with MPs-CCMs was investigated by using a lentiviral vector encoding for the redox-sensing green fluorescent protein (roGFP), which reports the reduced glutathione/oxidized glutathione balance (30). After overnight treatment of HUVECs with MPs-CCMs, H₂O₂ (100 and 300 μmol/L) was added. Fluorescence measurements were performed in clear 24-well plates (Corning, Lowell, MA) on a fluorescence plate reader GENios plus (Tecan, Männedorf, Switzerland) using a 535-nm emission filter. The degree of oxidation of the roGFP was estimated from the ratios of light intensities obtained during 1-min intervals under 400- and 485-nm excitation.

The production of reactive oxygen species (ROS) by MPs was evaluated by labeling cells with MitoSox (Invitrogen) according to the manufacturer instructions. Briefly, semiconfluent MPs were treated for 10 min with 5 μmol/L MitoSox dye in culture. Fluorescence was analyzed using a FACSCanto flow cytometer and the FACSDiva software (BD Biosciences). In inhibitory studies, the scavenger N-acetyl-cysteine (NAC; 1 mmol/L; Sigma-Aldrich) was added to MPs to infer the impact of ROS on the functions of MPs.

Functional assays were performed on MPs treated with the PKCβII inhibitors LY333531 (200 nmol/L) or CGP53353 (2 µmol/L), or with vehicle. Pharmacological inhibition was extended for 7 days in the myogenic differentiation assays, and for 24 h in the proliferation and angiogenesis assays.

RNA was extracted using the miRNeasy Mini Kit (Qiagen) according to the manufacturer instructions. cDNA was synthesized using the TaqMan Reverse Transcription kit and then was analyzed using the QuantStudio 6 Flex Real-Time PCR System, normalizing the data to the 18S rRNA. Primers for the CH2 domain of the p66 isoform were designed as follows:

• 5′GATCCCGAATGAGTCTCTGTCATCGTTGATATCCGCGATGACAGAGACTCATTCTTTTTTCCAAA-3′

• and 3′GGCTTACTCAGAGACAGTAGCAACTATAGGCGCTCTGTCTCTGAGTAAGAAAAAAGGTTTTCGA-5′.

Cells were harvested after overnight serum starvation in lysis buffer (TRIS-HCl 20 mmol/L, NaCl 150 mmol/L, EDTA 1 mmol/L, EGTA 1 mmol/L, Triton X-100 1%) supplemented with 1× COMPLETE protease inhibitor cocktail (Roche), 1× phosphatase inhibitor cocktail type II and III (Sigma-Aldrich), and 1× Benzonase (Novagen). Twenty-five micrograms of total cell lysates were loaded onto a NuPAGE 10% Bis-Tris precast gel (Life Technologies). Proteins were then transferred onto an Amersham Hybond polyvinylidene fluoride membrane (0.45 µm; GE Healthcare). Membranes blocked with 5% skim milk in 0.1% PBS-TWEEN 20 were decorated with mouse monoclonal primary antibodies diluted according to the manufacturer instructions. Membranes were subsequently incubated with horseradish peroxidase–conjugated anti-mouse secondary antibody (1:1,000 in 5% skim milk in 0.1% PBS-TWEEN 20). Bands were revealed using the C-Digit blot scanner (LI-COR Biosciences).

Continuous variables were expressed as the mean ± SEM and compared using parametric tests (t test and ANOVA), unless not normally distributed. A P value <0.05 was considered statistically significant. The GraphPad Prism 5 software package was used for these analyses.

A comparative analysis of muscle specimens from control and patients with diabetes was performed to assess the organization of myofibers and vascular cell structure. As illustrated in Fig. 1A, left panel, muscle sections from patients with diabetes show abundant and enlarged adipocytes (black star), and myofibers featuring degeneration (arrow) and nuclear centralization (arrowhead). Also, the average myofiber cross-section area was shifted to lower values in the diabetic samples compared with controls (Fig. 1A, right panel). Moreover, we observed an inflammatory infiltrate in diabetic samples, as evidenced by the abundance of CD3⁺ lymphocytes and, particularly, CD68⁺ macrophages (Fig. 1B), which is in line with findings of a previous investigation of ischemic muscles (31). Additionally, the count of CD31⁺ capillaries revealed a significant decrease in muscle vascularization (Fig. 1C).

MPs embracing CD31⁺ capillary ECs were identified by immunocytochemistry for NG2 (Fig. 1D) and ALP (Fig. 1D and Supplementary Fig. 1A and B). An unbiased profiler analysis performed using the Cell Profiler software (32) indicates a remarkable reduction in the density of ALP⁺ MPs in diabetic muscles (Fig. 1D, bottom panel).

Transmission electron microscopy analyses indicate ultrastructural alterations of D-MPs, consisting of blebbing and vacuolation (Fig. 2A). Immunocytochemistry studies confirm the pericyte identity of freshly isolated cells, based on the expression of ALP, CD146, and NG2 (Fig. 2B and C). Expanded MPs express the same markers, although they are negative for the satellite cell marker PAX7 (Fig. 2D). Moreover, expanded MPs coexpress NG2 and ALP (Fig. 2E). Flow cytometry analyses indicate the statement of mesenchymal markers CD44 and CD90 (Fig. 2F). MPs from control subjects or patients with diabetes did not differ from each other regarding antigen expression, except for desmin, which was downregulated in D-MPs (Supplementary Fig. 2A–C).

High-confluence MPs differentiate and fuse forming MyHC⁺ syncytial myotubes (33,34). As shown in Fig. 3A, D-MPs have reduced capacity to generate MyHC⁺ myotubes compared with control MPs (C-MPs), which may be ascribed in part to decreased proliferation, resulting in delayed cell confluence (Fig. 3B). Additionally, we found that D-MPs have an increased propensity to differentiate into adipocytes (1.7 ± 0.8-fold increase compared with C-MPs, n = 3; Supplementary Fig. 2D). In contrast, diabetes did not affect apoptosis, as assessed by caspase activity (6.7 ± 2.2 vs. 7.3 ± 2.4 relative fluorescence units at 499 and 521 nm in control subjects) or cell motility in a scratch assay at both 5 h postscratch (8.7 ± 3.7% vs. 3.6 ± 1.7% wound closure in control subjects) and 24 h postscratch (60.0 ± 13.8 vs. 43.4 ± 17.4% in control subjects).

We next analyzed the ability of MPs to form networks on Matrigel. Both C-MPs and D-MPs were able to form branched structures. In comparison with networks formed by HUVECs, the MP structures consisted of longer tubes and wider meshwork (Fig. 3C). When cocultured with HUVECs, MPs locate near the intersection points or around the endothelial tubes (Fig. 3D, white arrow and arrowhead). Importantly, D-MPs have detrimental effects on network formation, resulting in a less reticulated system (Fig. 3E).

To investigate whether paracrine mechanisms are responsible for the negative impact of D-MPs on angiogenesis, we next tested the effect of MP-CCM on HUVECs in the Matrigel assay. As shown in Fig. 4A, the D-MP-CCM induces a decrease in network formation, suggesting an alteration of the secretome. To verify this possibility, we interrogated secreted angiogenic myokines. ELISA showed significantly decreased levels of proangiogenic factors (i.e., vascular endothelial growth factor-A, IGF-I, and Ang-1, in D-MP-CCM (Fig. 4B).

Specific micro-RNAs (miRs) have recently been shown to regulate angiogenesis, and their deregulation may contribute to vascular complications and ischemia (29,35–37). Data from PCR analysis did not show any difference between D-MPs and C-MPs in the expression of classic angio-miRs, miR-16, miR-503, and miR-27b (Supplementary Fig. 3A). Also, the analysis of CCMs confirmed that D-MPs and C-MPs secrete similar amounts of miR-16, whereas miR-503 and miR-27b were undetectable (data not shown). Likewise, no difference was observed in the expression of the miR-27b targets thrombospondin-1, sprouty-2, and semaphorin 6A (Supplementary Fig. 3Aii and B), which are negative modulators of angiogenesis (36).

Secreted exosomes mediate the functional cross-talk between pericytes and ECs (29,37,38). Therefore, we next investigated whether diabetes has an impact on the capacity of MPs to release exosomes or modifies exosome activity upon angiogenesis. Results indicate no difference in exosome-sized vesicles shed by C-MPs and D-MPs (7.4 ± 2.5 × 10⁻⁴ vs. 6.9 ± 2 × 10⁻⁴ exosome concentration/cell, respectively). Also, the addition of exosomes from D-MPs to HUVECs in a Matrigel assay does not alter the network formation (10.7 ± 1.9 vs. 10.8 ± 0.4 intersections/field in HUVECs plus C-MPs), although exosome-depleted D-MP-CCM still inhibits the process (data not shown).

We next investigated the role of ROS as paracrine mediators of the negative cross-talk between D-MPs and ECs. To this end, HUVECs were transduced with lentiviral vectors carrying the cytosolic or the mitochondrial isoforms of roGFP, which act as fluorescent indicators of the intracellular redox status (30), and were then incubated with C-MP-CCM or D-MP-CCM before being exposed to increasing doses of the pro-oxidant H₂O₂. We found that C-MP-CCM protects HUVECs from the pro-oxidant action of H₂O₂, whereas D-MP-CCM does not (Fig. 4C). In additional experiments, we test whether ROS per se impairs angiogenesis in vitro. To this end, HUVECs were exposed to a fixed dose of H₂O₂ (300 µmol/L) before being seeded into the Matrigel assay. As expected, H₂O₂ inhibited the HUVEC network formation capacity in a way that was comparable to the inhibition caused by D-MPs or their CCM. Interestingly, the addition of C-MP-CCM contrasted the inhibitory effect of H₂O₂ on network formation, whereas the D-MP-CCM were ineffective (Fig. 4D).

We next investigated the redox status of C-MPs and D-MPs by staining cells with MitoSox, a dye that selectively reacts with superoxide anion O₂^• in mitochondria. The results of flow cytometry analyses indicate that D-MPs have remarkably increased oxidative stress levels compared with C-MPs, considering either the average MitoSox dye intensity or the number of MitoSox⁺ cells (Fig. 5A). Moreover, the ROS scavenger system involved in the maintenance of cellular redox balance was depressed, as indicated by the downregulation of SOD-1 and catalase (Fig. 5B). To confirm that the ROS imbalance contributes to the dysfunction of D-MPs, we investigated whether the addition of NAC, a generic antioxidant, restores inherent and paracrine activities of the diabetic cells. In line, NAC partially amended the myogenic differentiation capacity of D-MPs (Fig. 6A), restored the proliferation of D-MPs (Fig. 6B), and abrogated the inhibitory effect of D-MP-CCM on HUVEC network formation in vitro (Fig. 6C). Since the excess of ROS activates the p66^Shc signaling pathway by its selective phosphorylation, the next step was to compare the levels and phosphorylation state of this protein in C-MPs and D-MPs.

Using primers specifically designed for the p66 isoform of Shc1 protein, we found that p66^Shc mRNA expression is significantly upregulated in D-MPs compared with C-MPs (Fig. 7A). Under unbalanced redox conditions, p66^Shc is phosphorylated at the Ser36 of its unique CH2 domain and translocates to the mitochondrial transmembrane space, where it fuels additional ROS production, causing the formation of a permeability-transition pore and apoptosis. We analyzed the phosphorylation state of p66^Shc in MPs by different methods. Using a specific Western blot Ser36-p-p66^Shc antibody, we observed an increase in the band of activated p66^Shc with no difference in the total protein content (Fig. 7B). Flow cytometry (Fig. 7C) and immunocytochemistry (Fig. 7D) confirmed the increased statement of phospho-p66^Shc in D-MPs. Nonetheless, treatment of D-MPs with NAC was ineffective in restoring p66^Shc phosphorylation to control levels (Supplementary Fig. 4A and B). This result has different keys of interpretation. First, the thiol groups of the NAC molecule may undergo auto-oxidation processes, especially in culture, which reduces their activity. Thus, a possible explanation is that NAC did not reach biologically relevant concentrations in mitochondria, where active p66^Shc accumulates. Alternatively, p66^Shc activation by PKCβII may be partially independent of ROS or resistant to temporary ROS reduction. The latter possibility is compatible with a reported epigenetic activation of the p66^Shc promoter in the context of diabetes (39). Also, ROS scavenging by NAC may be insufficient to contrast the activity of protein kinases responsible for p66^Shc phosphorylation, namely PKCβII (40). Accordingly, we found that NAC supplementation does not affect PKCβII localization or expression (Supplementary Fig. 4C and D).

To assess the direct involvement of p66^Shc in the dysfunction of D-MPs, p66^Shc activation was blocked by interfering with its phosphorylation. To this purpose, we inhibited the PKCβII kinase, which is responsible for p66^Shc activation (39). Two compounds were used to inhibit PKCβII in D-MPs: LY333531, also known as Ruboxistaurin, which has already been tested for therapeutic efficacy in clinical trials of diabetic retinopathy, and CGP53353 (41). Flow cytometry confirmed that both LY333531 and CGP53353 reduce the levels of Ser36-phospho-p66^Shc (Fig. 8A). We next reassessed the features of the dysfunction of D-MPs after treatment with PKCβII inhibitors. Results indicate an overall functional improvement upon PKCβII inhibition, including myogenic differentiation (Fig. 8B), proliferation (Fig. 8C), and interference with network formation (Fig. 8D) Those data indicate that blocking the PKCβII-p66^Shc pathway may have important therapeutic implications for total restoration of the functions of D-MPs.

The current study is the first to investigate the anatomical, functional, and molecular diversity of MPs from skeletal muscles of patients with diabetes who have CLI. Importantly, MPs were isolated from the tissue immediately at the rim of normal tissue surrounding the plane of dissection, a critical zone where it is mandatory to concentrate efforts for limb salvage. Results show for the first time specific alterations consisting of ultrastructural modifications, proliferative impairment leading to blunted myogenic potential, and acquisition of an antiangiogenic activity. Altogether, these deficits could contribute to the extension and severity of peripheral complications. Importantly, we discovered that the diabetes-associated incompetence of MPs is attributable to increased ROS levels, weakened antioxidative protection, and an activated PKCβII-p66^Shc signaling pathway. Also, excessive ROS production and release transmit negative signals to adjacent vascular cells. Muscle regeneration via trans-differentiation of MPs and satellite cells into myoblasts is crucial for the recovery of damaged muscles, whereas the overgrowth of adipogenic cells may be deleterious. Some studies (42) suggest that a subset of MPs contribute to fat accumulation. Our investigation shows structural data that are compatible with a differentiation bias of MPs favoring adipogenesis at the detriment of myogenesis. Concurrent mechanisms may participate in this adverse remodeling, including poor metabolic control, ischemia, and lack of exercise (43).

Unraveling the molecular mechanisms underpinning the dysfunction of MPs could help develop new strategies to maintain tissue integrity and improve clinical outcomes, especially after major limb amputation, a condition associated with high mortality. Current mechanistic understanding of diabetic vasculopathy is mainly inferred from animal models (44). However, the clinical transferability of data from rodents mimicking human diabetes/CLI is limited and often equivocal. Therefore, the successful isolation of MPs from the muscles of patients with diabetes undergoing major limb amputation for CLI discloses excellent opportunities for disease modeling and therapeutics. Since there is a potential overlapping between MPs and satellite cells, it was essential to ascertain that isolated cells express high levels of the canonical pericyte marker NG2 and ALP, along with mesenchymal markers (CD90 and CD44), and the staminal marker CD146, although being negative for PAX7. In fact, ALP positivity and PAX7 negativity are determinants to distinguish pericytes from ALP⁻/PAX7⁻ mature myocytes or ALP⁻/PAX7⁺ satellite cells. When grown to overconfluence, MPs spontaneously fuse together to form syncytial myotubes expressing MyHC. Interestingly, the ability to form syncytial myotubes was drastically reduced in D-MPs, and this could be one of the reasons (together with satellite cell dysfunction) of the aberrant tissue repair of diabetic skeletal muscles.

Skeletal muscle physiology is maintained by mutual trophic influences among MPs, ECs, and myocytes (45). However, this cross-talk is perversely modified by diabetes. The present investigation integrates the results of our recent study showing that p75NTR expression in ECs exposed to high glucose activates the transcription of miR-503, which negatively affects pericyte function (37,46). Here, we newly show that D-MPs exert a drastic reduction of the ability of ECs to form in vitro networks through an alteration in the secretion of angiogenic growth factors. Instead, the screening of several classic angio-miRs or exosomes did not provide any supplementary clue.

An additionally accountable mechanism emerged from studies of the MP redox state and was confirmed by the recognition of a reduced scavenging capacity of D-MPs, consisting of the downregulation of SOD-1 and catalase gene expression. Accordingly, the restoration of the physiological redox state by NAC supplementation led to the correction or attenuation of D-MP dysfunctions. A more in-depth analysis of the original culprit directed us to recognize the intracellular stress sensor as p66^Shc protein. Several lines of evidence support this possibility, as follows: 1) p66^Shc is upregulated at the mRNA level, and 2) was hyperphosphorylated at the Ser36 residue, which corresponds to an activated pro-oxidant state of p66^Shc. However, NAC administration did not revert p66^Shc activation so other targets are needed to suppress p66^Shc activation. Previous studies in animal models used the abrogation or silencing of p66^Shc to infer its pathophysiological importance. However, this approach has translational limitations because p66^Shc is thought to have bivalent actions (47). Thus, its total suppression may not be therapeutically desirable. Therefore, we decided to interfere with the excessive phosphorylation of p66^Shc by inhibiting PKCβII (39,41). Importantly, the use of PKCβII-specific inhibitors, namely LY333531 or CGP53353, reverted p66^Shc phosphorylation and restored D-MP functions. Although NAC has been efficiently used to inhibit PKCβII expression in other experimental settings, dosages and cell type specificity may explain the observed differences (48,49). Moreover, NAC is susceptible to auto-oxidation in culture conditions, reducing its effectiveness. Importantly, PKCβII inhibitors lead, at least in vitro, to a more effective rescue of D-MP myogenic differentiation capacity compared with NAC treatment, thus warranting additional investigation in patients with limb ischemia. This may be clinically relevant since generic antioxidant therapy has given disappointing results in trials assessing vitamins supplementation efficacy for cardiovascular diseases (50).

In conclusion, we demonstrate that 1) MPs can be efficiently isolated from skeletal muscles of patients without diabetes and patients with diabetes; 2) D-MPs are dysfunctional in terms of reduced myogenic ability, decreased proliferation, and antiangiogenic properties; 3) those alterations are strictly related to an increased oxidative status driven by p66^Shc overexpression and activation; and 4) antioxidant treatment as well as p66^Shc phosphorylation blockade by inhibition of PKCβII can rescue D-MP functional competence.

These results are important since one of the used PKCβII inhibitors, LY333531, is currently at the final step of clinical trial investigation (phase 3) in patients with diabetic retinopathy. This opens invaluable opportunities for the treatment of life-threatening complications of diabetes. We hypothesize that the use of PKCβII inhibitors may halt the progression of CLI and allow surgeons to decide to perform less extensive amputations, when these become necessary, ultimately improving the quality and duration of life of patients with complicated diabetes.

Acknowledgments. The authors thank Dr. Laura Cantone from EPIGET (Epidemiologia, Epigenetica e Tossicologia) Laboratory, Dipartimento di Scienze Cliniche e di Comunità Università degli Studi di Milano, Milan, Italy, for nanoparticle tracking analysis of exosomes and the Centre of Advanced Microscopy “P. Albertano,” in the Department of Biology, University of Rome Tor Vergata, for confocal images.

Funding. This work has been supported by British Heart Foundation grants RJ5905 and RM/13/2/30158 and Italian Ministry of Health grant RF-2011-02346867 to P.M., Cariplo Foundation grant 2013-0887 to G.S., European Research Council grant N322749 DEPTH to G.C., and Uncovering Excellence Grant 2014 MDESMPLAT to C.G.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. R.V. participated in the study design, researched and interpreted the data, and wrote the manuscript. C.F. and E.S. isolated and characterized human pericytes. S.T. and D.M. performed immunofluorescence and histochemistry analysis. S.P. set and performed Cell Profiler analysis for ALP staining area quantification and statistical analysis. D.F.M. performed measurements of secreted factors by ELISA. G.P. and R.G. performed redox analyses of endothelial cells. G.F. and F.S. conducted and analyzed the electron microscopy studies. R.C., A.G., and S.L. are responsible for patient enrollment and muscle sample collection. G.C. and S.C. helped with the study design, data analysis interpretation, and writing of the manuscript. R.R. and C.B. helped with data interpretation. G.S., C.G., and P.M. designed the study, helped with data analysis and interpretation, and wrote the manuscript. G.S., C.G., and P.M. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the American Heart Association Arteriosclerosis, Thrombosis and Vascular Biology Scientific Sessions, Orlando, FL, 7–11 November 2015; at European Society of Cardiology Congress 2015, London, U.K., 29 August–2 September 2015; and at the 2015 Joint Meeting of the European Society for Microcirculation and European Vascular Biology Organisation, Pisa, Italy, 3–6 June 2015.