Mesenchymal stem cells promote type 2 macrophage polarization to ameliorate the myocardial injury caused by diabetic cardiomyopathy

Thirty 8-week-old male Sprague–Dawley (SD) rats were purchased from the Chinese People’s Liberation Army (PLA) General Hospital and fed either a high-fat diet (HFD) or a normal chow diet (NCD) for 8 weeks. To establish the DCM rat model, a single intraperitoneal injection of streptozotocin (STZ, 25 mg/kg, Sigma-Aldrich) was administered to the HFD-fed rats. Then rats of all groups were fed a normal chow diet until the end of experiment. The DCM rat model was evaluated and confirmed by echocardiography.

MSCs between passage 3 and 4 were used for tail-vein infusion. MSC infusion (2 × 10⁶ MSCs suspended in 0.3 mL sterile physiological saline) was performed every 2 weeks, 4 times in all. The normal rats and DCM rats were infused with 0.3 mL sterile physiological saline.

Rats were weighed at baseline, 8 weeks, 12 weeks, 20 weeks and calculated ratio of heart weight to body weight at the end of the experiment. Blood was collected from the tail vein at baseline, 8 weeks, 12 weeks, and 19 weeks within 24–48 h after MSC infusion. Fasting blood glucose (FBG), fasting insulin (FINS), insulin sensitivity index (ISI), total cholesterol (TC), triglyceride (TG) and intraperitoneal glucose tolerance test (IPGTT) levels were determined by ELISA or biochemical analyzer at different time points throughout the experiment. To perform the IPGTT, rats from each group were fasted for approximately 6 h and injected intraperitoneally with glucose (2 g/kg). Tail-vein blood samples were collected before injection and 30, 60 and 120 min after injection to measure the glucose levels. Data are presented as the mean ± SD. Differences between two groups were analyzed using analysis of Independent-Sample T Test. Differences between multiple groups were analyzed using analysis of variance (ANOVA).

Rats were anesthetized with isoflurane, and echocardiography was performed using a Vevo 770 ultrasound system. Left ventricular end-diastolic internal diameters (LVIDD), LV end-systolic internal diameters (LVIDS), left ventricular end-diastolic anterior wall (LVAWD), left ventricular end-systolic anterior wall (LVAWS), left ventricular end-diastolic posterior wall thickness (LVPWD), and left ventricular end-systolic posterior wall thickness (LVPWS) were measured by parasternal short axis scans at the papillary muscle level, and the fractional shortening (FS) was calculated simultaneously. The left ventricular ejection fraction (LVEF) was measured by parasternal long axis scans. Pulsed-wave Doppler imaging and tissue Doppler were used to assess early to late diastolic transmitral flow velocity (E/A) ratio and early to late diastolic mitral annular velocity (E’/A’) ratio, respectively.

Rat hearts were isolated, imaged, weighed and measured. Heart samples from each group were fixed in 10% formalin after washing with phosphate-buffered saline (PBS), processed, embedded in paraffin, and cut into 4-μm sections at the papillary muscle level. Heart sections were evaluated by hematoxylin–eosin (H&E) and Masson’s trichrome staining. Five randomly selected photomicrographs of cardiomyocyte and cardiac fibrosis were chose for quantitative microscopic analysis, respectively. Myocytes with round nuclei and clearly defined sarcolemmal borders were selected for analysis of myocyte cross-sectional area (CSA) in H&E staining. And fibrosis was determined on the stained heart tissue sections highlighting the collagen fibers blue in Masson’s trichrome staining. Image J software was used to analyze myocyte CSA and quantification of myocardial fibrosis. Data are presented as the mean ± SD. Differences between multiple groups were analyzed using analysis of variance (ANOVA).

Peritoneal macrophages were obtained as previously described [10]. Then, macrophages were cultured in a 12-well culture plate, and after a 2-h incubation, the culture medium was replaced with a high-glucose medium (33 mmol/L) containing 1 μg/mL LPS and incubated for 12 h. MSCs were then cocultured with treated macrophages at a ratio of 1:10 MSC:macrophage for another 48 h in a Transwell system (Corning). For the in vitro experiments, the cells were divided into the following groups: (1) normal, (2) damage group (high glucose + LPS), (3) damage + MSCs, (4) damage + MSCs + COX-2 inhibitor (10 μM, Abcam), and (5) damage + MSCs + COX-2 inhibitor + prostaglandin E2 (PGE2; 0.1 μM, Sigma). For groups 4 and 5, MSCs were pretreated with COX-2 inhibitor for 2 h before coculturing with macrophages either with or without adding PGE2 to the coculture system.

For the flow cytometry analysis, heart tissue was digested by type II collagenase (Sigma) at 37 °C for 40 to 50 min while rocking. Next, 10% FBS was added to terminate enzyme activity. The cell suspension was filtered through a 75 µm cell strainer and centrifuged at 1500 rpm for 10 min. For the in vitro experiments, macrophages from different treatment groups were collected using sterile PBS. Cells from the heart tissue or the in vitro experiment were incubated with antibodies against CD68 (Alexa Fluor 647, AbD Serotec), CD11c (FITC, AbD Serotec) and CD163 (RPE, AbD Serotec) at room temperature and kept in the dark for 20 min. Cells were then washed with PBS and analyzed using a FACSCalibur flow cytometer.

Heart tissue was weighed and homogenized in ice-cold physiological saline at a 1:9 ratio (weight/volume). The homogenate was then centrifuged at 3000 rpm at 4 °C for 10 min, and the supernatant was collected for analysis. Total protein was quantified via Coomassie blue staining. The concentrations of IL-6, IL-10, TNF- α and PGE2 in the cell culture, serum and heart tissue were detected by ELISA following the manufacturer’s instructions (R&D).

Data are presented as the mean ± SD and were analyzed using SPSS 17.0. Differences between two groups were analyzed using analysis of Independent-Sample T Test. Differences between multiple groups were analyzed using analysis of variance (ANOVA). P < 0.05 was considered statistically significant.

AD-MSCs were characterized based on their morphology, phenotype and differentiation capacity after three passages. Flow cytometry showed that more than 90% of these cells were positive for CD29 and CD90 but negative for CD34, CD45, and CD11b (Fig. 1a). The AD-MSC morphology exhibited a typical spindle-like shape and formed a swirl-like monolayer. AD-MSCs cultured in lipid-inducing solution showed many lipid droplets after Oil Red O staining, and cells cultured with osteogenic-inducing fluid stained orange (Fig. 1b).

Macrophages were also characterized based on morphology and phenotype. Figure 1c showed that the macrophages were highly refractive, large and round. Flow cytometry results showed a highly pure cell population (> 99%) positive for specific surface antibodies CD11b and CD68 (Fig. 1c).

To evaluate the rat DCM model, normal and DCM rats were evaluated for biochemical indicators including fasting blood glucose, fasting serum insulin, blood total cholesterol and triglyceride levels, calculated insulin sensitivity indexes and echocardiography at 12 weeks. Echocardiography results showed that contractile function (LVEF, FS) and diastolic function (E/A ratio) were obviously decreased in the morbid rats (Fig. 2a). Additionally, the body weights of the DCM rats were significantly higher than those of the control rats after consuming the HFD, but at late stages of induced DCM, the weights of the DCM rats decreased gradually until they were lower than those of the normal rats (Fig. 2b). The hematological indicators FBG, FINS, TC and TG increased markedly in the DCM rats compared with the normal rats (Fig. 2c, d, f, g). The ISI calculated by FBG and FINS declined significantly after HFD and STZ treatment (Fig. 2e). These data are consistent with previous reports, thus indicating that the DCM model was successfully established in these rats [11].

To assess the effect of MSCs on glucose and lipid metabolism and cardiac function in DCM rats, we performed hematological assessments and echocardiography at 19 and 20 weeks, respectively. Echocardiography data (Fig. 3a, Table 1) demonstrated that the LVAWD, LVAWS, LVPWD and LVPWS of the DCM and MSC-treated groups were obviously thinner than those of the normal group. The LVIDD and LVIDS of the DCM and MSC-treated groups were significantly larger than those of the normal group. Compared with the DCM group, MSC administration alleviated ventricular wall thinning and ventricular chamber enlargement. The LVAWS in the MSC group was significantly thicker than that in the DCM group. LVEF and FS, which represent cardiac systolic function, were reduced in the DCM and MSC-treated rats, and MSCs slightly improved cardiac systolic function. The E’/A’ ratio, which reflects cardiac diastolic function, decreased dramatically in the DCM hearts. However, injecting MSC increased the E’/A’ ratio and ameliorated diastolic function significantly (Fig. 3a, Table 1). Blood sample results indicated that the fasting blood glucose levels of the DCM rats decreased significantly after MSC treatment (Fig. 3b). Moreover, as shown in Fig. 3c–g, glucose levels and lipid metabolism were improved in the MSC-treated rats. These data indicate that administering MSCs may improve diastolic function in DCM rats and partially ameliorate systolic function. In addition, MSC infusion may reverse hyperglycemia, blood lipid metabolic abnormalities and insulin resistance in DCM rats.

To demonstrate the ability of MSCs to ameliorate inflammation and myocardial injury in DCM rats, we measured heart weight, calculated the ratio of heart weight to body weight (HW/BW) and stained heart sections with H&E and Masson’s trichrome. At the end of experiment, the heart weights of DCM rats were significantly lower than those of the normal rats (Fig. 4a), HW/BW of DCM group was larger than that of normal group (Fig. 4b). And staining the DCM hearts exhibited obvious myocardial thinning, chamber dilation, extensive myocardial disorganization, hypertrophic and atrophic cardiomyocytes, cell degeneration, inflammatory cell infiltration (Fig. 4c, d). Figure 4e showed that myocyte CSA of DCM hearts were significantly larger than that of normal hearts. Otherwise, Masson staining and quantitative analysis showed that DCM hearts existed abundant collagen deposition (Fig. 4f–g). However, the heart weights of the MSC-treated rats increased significantly and HW/BW was lower than that of DCM rats (Fig. 4a, b). Moreover, according to Fig. 4c, d, f, MSC-treated hearts presented decreased ventricular chamber, thicker wall, alleviated injury of myocardial and cardiomyocytes, reduced myocardial inflammation and interstitial fibrosis compared to DCM hearts. Myocyte CSA and quantification of myocardial fibrosis were also decreased in DCM + MSCs group (Fig. 4e, g). These data indicated that DCM rats had severely abnormal basal metabolism, abnormal myocardial structure and cardiomyocyte morphology, cardiac inflammation and interstitial fibrosis, all of which were ameliorated by MSC infusion.

To analyze the influence of MSC treatment on macrophage polarization, we analyzed the phenotypes and ratios of cardiac macrophages by flow cytometry at 20 weeks. The results showed that the total number of cardiac macrophages increased significantly in both the DCM and MSC-treated rats (Fig. 5a). Notably, M1 macrophages among the total cells were predominant in the DCM rat hearts, while M2 macrophages were increased in the MSC-treated rats (Fig. 5b, c). Consistently, the proportion of M1 in total macrophages and M1/M2 of DCM hearts were significantly higher than those of normal and MSC-treated hearts. The percentage of M2 in total macrophages was decreased in DCM hearts, while it was increased obviously in MSC-treated hearts (Fig. 5d–f, g–i). There data demonstrated that increased macrophages in DCM hearts were M1 macrophages, and MSC treatment might increase the quantity and proportion of M2 macrophages significantly.

To further evaluate the MSC treatment effects, we analyzed chronic inflammation in the hearts and blood of the rats. As shown in Fig. 5, the proinflammatory cytokines IL-6 and TNF-α were significantly increased in the DCM group, whereas MSC treatment markedly reversed the cytokine elevation (Fig. 5j, k, m, n). Moreover, MSC treatment also significantly increased the anti-inflammatory cytokine IL-10 levels compared with those in the DCM group (Fig. 5l, o). These data indicated that M1 macrophages and proinflammatory cytokines play important roles in chronic and persistent inflammation in DCM rats. Furthermore, MSC infusion increased the M2 macrophage polarization, inhibited the proinflammatory response and increased anti-inflammatory cytokines to relieve inflammation in DCM hearts.

To further verify the effects of MSCs on ameliorating cardiac chronic inflammation via M2 macrophage polarization, we used an in vitro high-glucose (HG) and LPS-induced cell culture system to mimic the rat DCM model. Induced macrophages were cocultured with MSCs, and the ratios of different macrophage phenotypes were analyzed. The results showed that HG + LPS stimulation increased the number of CD11c-positive macrophages (M1) and reduced the number of CD163-positive macrophages (M2) compared with those of the normal group. Furthermore, the results indicated that coculturing with MSCs promoted polarization of M1 to M2 macrophages, as indicated by a significant increase in M2 macrophages ratios and a decrease in M1 macrophages ratios (Fig. 6a, b). Additionally, the secreted inflammatory cytokines, IL-6, IL-10 and TNF-α, were analyzed. The results showed that the IL-6 and TNF-α levels were significantly increased, but the IL-10 concentration significantly decreased in the HG + LPS group supernatant. Compared with the HG + LPS group, MSCs cocultured with macrophages inhibited IL-6 and TNF-α secretion and increased the IL-10 concentration (Fig. 6c). These data indicate that high glucose combined with LPS stimulation polarized macrophages to the M1 phenotype and stimulated the release of the proinflammatory cytokines IL-6 and TNF-α. MSCs could promote M2 macrophage polarization and stimulated IL-10 secretion.

Previous studies have demonstrated that PGE2 is involved in the immunoregulatory effects of MSCs [12]. Our results also showed a high concentration of PGE2 in the supernatant of MSCs stimulated with high glucose and LPS. Moreover, inhibiting COX-2 significantly decreased PGE2 secretion by MSCs (Fig. 7a). We further hypothesized that PGE2 might be involved in MSC-induced M2 macrophage polarization. To verify this, we pretreated MSCs with 10 μM COX-2 inhibitor to specifically suppress PGE2 expression, then analyzed the macrophage polarization. The results showed that the M1 macrophages increased significantly in the COX-2-inhibited group, whereas the M2 macrophages decreased significantly compared with those in the MSC group. Exogenous PGE2 was then added to the COX-2-inhibited group, and the M1 and M2 macrophage polarizations were significantly reversed (Fig. 7b, c). Moreover, the IL-6 and TNF-α levels were significantly increased, but the IL-10 level was significantly decreased in the COX-2-inhibited group. Adding exogenous PGE2 also inhibited the IL-6 and TNF-α levels and increased IL-10 levels (Fig. 7d). These results indicated that MSCs might promote M2 macrophage polarization by secreting PGE2, which can alleviate the chronic inflammation induced by high glucose and LPS.