Transplanted allogeneic cardiac progenitor cells secrete GDF-15 and stimulate an active immune remodeling process in the ischemic myocardium

Rat c-kit-positive cardiac cells (rCPCs) were isolated from male Wistar-Kyoto (WKY) rats (6–8 weeks of age) as described previously [16]. Briefly, rat hearts were isolated and perfused via the aortic root with phosphate-buffered saline (PBS), followed by a solution of collagenase (128 units/ml) and hyaluronidase (300 units/ml) for 10 min. Perfused hearts were sliced into approximately 1–2-mm pieces, and cardiomyocytes were removed by centrifugation at 300×g. The small cell fraction in the supernatant was collected and expanded until 70% confluence. Cells were detached and c-kit⁺ rCPCs were isolated using Miltenyi Biotech rat anti-mouse IgG microbeads (#130-048-402) after incubating with anti-CD117 antibody (SC-5535), as per the manufacturer’s instructions.

The immunogenicity of rCPCs was evaluated by co-culturing rCPCs with rat splenocytes. Splenocytes were isolated from spleens harvested from WKY and Brown Norway (BN) rats. Briefly, spleens were removed aseptically and placed in tissue culture dishes containing 5 ml media and minced with scalpels. Splenocytes were isolated by mechanical dissociation, and the cell suspension was filtered through a cell strainer (100 μm). Erythrocytes were lysed by incubating the cell suspension with red blood cell lysis solution for 3 min at room temperature. To evaluate immune responses to allogeneic and syngeneic rCPCs, mitotically inactivated WKY rat CPCs were cultured with CFSE-labeled WKY or BN lymphocytes in a 1:5 ratio. hCPCs were co-cultured with BN rat lymphocytes to investigate the xenogeneic response. After co-culturing for 5 days, T cells were isolated, and proliferation was assessed by measuring CFSE intensity after gating for CD3⁺ cells. The stimulation indexes for proliferation of human and rat lymphocytes were calculated by comparing the fold change of individual alloreactive and xeno-reactive lymphocyte proliferation to the mean of syngeneic lymphocyte proliferation.

To evaluate the in vivo immunogenicity and cardiac repair potential of allogeneic CPCs, CPCs were transplanted into the ischemic region of the heart, as described previously [8]. Briefly, immunologically divergent rats (6–8 weeks old) underwent left thoracotomy under isoflurane (2%) anesthesia and myocardial infarction (MI) was induced by permanent ligation of the left anterior descending (LAD) coronary artery using 6–0 sutures. After confirming ischemia by visual inspection, 1 million rat or human CPCs per animal (5 million cells/kg, corresponding to 1 million cells on an average of 200 g rat) suspended in basal medium (IMDM) were injected in the peri-infarct region in 4 equal doses. Injection of IMDM alone served as a vehicle control. To determine the extent of rCPC engraftment, rCPCs overexpressing GFP were used in some experiments. A total of 6 experimental groups were used to achieve desired combinations: for allogeneic evaluation, BN female rats received: (I) IMDM, or (II) WKY rCPCs; for syngeneic evaluation, WKY female rats received, (III) IMDM, or (IV) WKY rCPCs; for xenogeneic evaluation, BN Female rats received, (V) IMDM, or (VI) hCPCs.

At the end of the reaction, supernatant was collected and precleared of cellular debris and particulate matter by centrifugation at 1000g for 30 min, followed by 20,000g for 30 min. Protein content was quantified using the bicinchoninic assay (BCA) method (Thermofisher, Waltham, MA). To normalize the protein content, we used the following formula: (concentration factor) × (total volume of medium)/(total protein content of supernatant medium). ELISA was performed for in the core facility at the University of Maryland School of Medicine using ELISA kits (Millipore and R&D systems Inc. Billerica, MA), according to the manufacturer's protocol.

Two-dimensional and M-mode echocardiography was performed using a VisualSonics Vevo 2100 ultrasound unit (VisualSonics, Toronto, Canada) under isoflurane (2%) anesthesia. Baseline echocardiograms were acquired on day 1 to confirm the expected reduction in cardiac function by measuring left ventricular (LV) ejection fraction (EF) and fractional shortening (FS) by blinded cardiologist. Animals with an EF of 45 ± 2% on post-operative day 1 were included in the study in order to maintain uniformity with respect to the severity of post-MI dysfunction. The similarity of post-MI dysfunction was verified in each treatment group by performing echocardiography at baseline and 24 h after MI. Following ligation of the left anterior descending artery (LAD), one million cells were injected into the LV myocardium near the ischemic area. Transthoracic M-mode images of the left ventricle in the parasternal short-axis view were obtained at the level of the papillary muscles using high-resolution M mode echocardiography. Data were calculated from 5 cardiac cycles according to the generally accepted formulas [8]. Progressive improvement in cardiac function was evaluated by repeating the echocardiography on days 7 and 28.

The humoral immune response to allogenic, syngeneic, and xenogenic CPCs was evaluated by reactivity of rCPC-treated rat serum for alloreactive and xenoreactive anti-donor antibodies using flow cytometry. Rat and human CPCs were incubated with 50 μl of recipient or naïve serum for 30 min at 4 °C. After washing, cells were further incubated with rat anti-IgM and anti-IgG for 30 additional min followed by quantitative analysis using flow cytometry.

Heart tissue from treated animals was harvested at day 5 post-MI, minced, and then digested by collagenase D (Roche) at 37 °C for 50 min on a rocking platform (180–200 rpm). After enzymatic digestion, the cell suspension was filtered through a 70-µm cell strainer (Fisher Scientific #22363548) and centrifuged at 500×g for 10 min. To lyse red blood cells, the cell pellet was incubated in ammonium-chloride-potassium lysing buffer (Gibco # A10492-01) at room temperature for 3–5 min, then washed with iced cold fluorescence activated cell sorter washing buffer (2.5% fetal bovine serum in PBS without calcium and magnesium). Cells were resuspended in the washing buffer, and samples were incubated with Fc-Block (anti-rat CD16/CD32, 0.5 μg per 1 million cells) before incubation with isotype controls or primary antibodies, according to the manufacturer’s instructions. Cells were then washed with washing buffer. Approximately 2 × 10⁵ events (cells) were analyzed by flow cytometry (BD-LSRFortessa) and populations gated as detailed below and analyzed by FlowJo software. The antibodies are described in Table 1 (Data Supplement). T cells and Tregs were first gated (FSC-A vs SSC-A) as lymphocytes. For total T cells, the CD3 cells were gated and further analyzed for CD4 and CD8. For Tregs, CD4 cells were gated, from this gate, CD25⁺ and Fox-P3⁺ double-positive cells were determined. For macrophages, CD45 cells were gated. From CD45-positive cells, CD68 (Total macrophages) were gated and further analyzed for CD163 for M2 macrophages [22]. To see the uptake of rCPCs secreted GDF15 by CD48 receptor on T- cell, allogeneic rCPCs were co-cultured with spleenocytes with or without GDF15 KD for 5 days. At day 5, spleenocytes were collected and flow was performed for CD4, CD48 and GDF15.

Tissues were processed as described previously [8]^. Briefly, rat hearts were excised under anesthesia after collection of echocardiographic data and perfused with 10% formalin solution, (Sigma Aldrich #HT501128). Tissues were cryopreserved using 30% sucrose (prepared in 1× PBS) and embedded in optimal cutting temperature compound (Fisher Scientific, TissueTek #NC1029572). A commercial cryostat was used to cut 7-μm sections, which were stained for different antibodies according to the manufacturers’ instructions. Tissue sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) nuclear stain (Sigma #F6057) together with other required stains such as Foxp3 and pP65. All images were obtained with an EVOS microscope and quantified using Image J software.

Manipulation of gene expression was performed by lentiviral transduction. All lentiviruses were produced in HEK293T. HEK293T cells (American Type Culture Collection, Manassas, VA) were cultured in DMEM media (CellGro) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific #A38402-02). The lentivirus expression system from Origin (Catalog #TL710232) has 4 unique shRNA 29-mers for knockdown of GDF15. The titer of each lentivirus preparation was calculated based on the amount of virus required to yield 50% GFP⁺ cells following transduction of 100,000 rCPCs. We calculated the MOI according to the company (Origene, Inc) manufacturing protocol. Briefly, we calculated MOI using the following formula:

Our desired MOI was 2. Cells were transduced in 12-well dishes with increasing amounts of lentivirus in media supplemented with 8 μg/ml polybrene (Sigma Aldrich #TR-1003). Three days after transduction, the percentage of GFP+ cells in each well were determined by flow cytometry using the LSR Fortessa. GDF15 KD was confirmed by western blot.

Unpaired non-parametric tests with Mann–Whitney’s correction were performed to compare two groups. For comparisons between more than two groups, a one-way ANOVA with Tukey’s post hoc test was performed. Grouped echocardiography data was analyzed by 2-Way ANOVA with Bonferroni correction. Continuous data is plotted as box-and-whiskers plots. The middle horizontal line represents the median. The upper and lower whiskers represent the maximum and minimum values of non-outliers. The number of subjects is numerically expressed under each box and whisker column. Extra dots represent outliers. Data were analyzed using GraphPad Prism 9 software. P values ranging from 0.01 to 0.05, 0.01 to 0.001, 0.001 to 0.0001 or < 0.0001 is represent as significant *, very significant **, extremely significant ***/**** respectively.

Consistent with our published results on human CPCs, flow cytometry also demonstrated that both rCPCs and rMSCs expressed high levels of mesenchymal markers (CD105⁺, CD90⁺), but low levels of endothelial cell markers (CD45⁻, CD31⁻) [8]. Regarding immune antigens, both cell types expressed major histocompatibility complex class I (MHC I⁺) and the co-stimulatory immune markers (CD80⁺), but not MHC II or CD86 (Fig. 1A). In addition, rCPCs expressed high levels of c-kit⁺ compared with rMSCs. These results show that CPCs display a hypoimmunogenic baseline profile, suggesting their potential to evade immune recognition in vivo. To confirm the biological relevance of the favorable baseline immunogenic profile of allogeneic rCPCs, we performed one-way mixed lymphocyte reactions with rat CD4⁺ T cells isolated from male rats of 2 different strains, WKY and BN. Rat CD4⁺ T cells were then co-cultured with rCPCs obtained from WKY rats. The combination of WKY CD4⁺ T cells and WKY rCPCs is syngeneic, whereas the combination of BN CD4⁺ T cells and WKY rCPCs is allogeneic. We also evaluated a xenogeneic combination: CD4⁺ T cells from a BN female rat with human adult CPCs. Syngeneic or allogeneic rCPCs elicited negligible T cell proliferation (Fig. 1B). Although Syngeneic CPCs didn’t elicit total T cell proliferation, allogeneic CPCs showed low (statistically insignificant) T-cell proliferation. Consistent with this low T cell proliferative response, the levels of the cytokines like transforming growth factor beta (TGF-β) was significantly increased in cell-free media of syngeneic (Fig. 1C) and interleukin 2 (IL-2) [23‐26], TGF-β in allogeneic (Fig. 1D) combinations [27] as identified by multiplexed ELISA. In contrast, significant T-cell proliferation and decreased levels of TGF-β, IL-2, and IL-10 were observed in the xenogeneic co-culture conditions (Fig. 1E. Forkhead box protein 3 (FOXP3) expressing CD4+ T cells (Treg cells) are accumulated in the injured myocardium after MI injury and plays an important role to recover heart function [28‐30]. To further explore the immunomodulatory properties of rCPCs, we performed in vitro co-culture studies combining rCPCs with CD4⁺ T cells isolated from WKY (syngeneic) and BN (allogeneic) rats. After 5 days in co-culture, the syngeneic (Fig. 1F, G) and allogeneic (Fig. 1H, I) combinations significantly increased the percentage of CD25⁺ and Foxp3⁺ Treg cells when compared to control. Thus, Syngenic or Allogenic CPCs do not induce an immunogenic reaction but have an immunomodulatory property by increasing a subset of CD4+ cells (T-regs). These results show that consistent with their hypoimmunogenic baseline profile, rCPCs are not immunogenic and exert immunomodulatory effects in vitro.

To determine the immunomodulatory effects of rCPCs in vivo, one million allogeneic rCPCs were injected intramyocardially in BN rats after MI and, 5 days later, the presence of T-regs in single-cell suspensions from the hearts was determined by flow cytometric analysis. Transplantation of allogeneic rCPCs induced a significant increase of CD4⁺/CD25⁺/FoxP3⁺ Tregs (Fig. 2A, B) in the injured myocardium. Since T-regs stimulate M2 polarization of monocytes/macrophages [31], we also evaluated M2 macrophages (Fig. 2C, D) in the same biological replicates. Helper T-cell (CD4⁺) and killer T-cells (CD8⁺) cells (Fig. 2E–H) or B-cells (CD45R⁺) and total T-cells (CD3⁺)cells (Fig. 2I–L) remain unchanged. Collectively, these data show that the baseline immunomodulatory profile of rCPCs correlates with immunomodulatory biological effects both in vitro and in vivo.

To determine the effects of CPCs on post-MI LV dysfunction, allogeneic, syngeneic, or xenogeneic CPCs were transplanted after MI and LV function was assessed after 28 days. Transplantation of one million allogeneic or syngeneic rCPCs significantly improved LV function after 4 weeks compared with controls (IMDM), as measured by an increase in LVEF (IMDM vs allogeneic: 42 ± 2 vs 49 ± 1%, P < 0.05, n = 9–12; IMDM vs syngeneic: 41 ± 2 vs 55 ± 3%, P < 0.05, n = 4–6) (Fig. 3A) and LV FS (IMDM vs allogeneic: 21 ± 1 vs 26 ± 1%, P < 0.05, n = 9–12 and IMDM vs syngeneic: 21 ± 1 vs 30 ± 2%, P < 0.05, n = 4–6) (Fig. 3B). No significant functional improvement of LV was observed with xenogeneic CPCs (Fig. 3A, B). In contrast to xenogeneic treatment group, LV end-systolic volume was significantly decreased after syngeneic and allogeneic rCPCs compared with IMDM and was associated with a significant increase in cardiac output after syngeneic and allogeneic rCPCs (Fig. 3C, D).

The extent of post-MI fibrosis after 4 weeks was determined by measuring the area of fibrosis (blue) relative to the total myocardial area (pink and blue) after Masson’s trichrome staining of heart sections in different groups. In contrast to xenogeneic group, heart tissues from either syngeneic or allogeneic CPCs group had significantly smaller fibrotic areas compared with placebo (Fig. 3E , F). These results indicate that, consistent with their phenotype and in vitro and in vivo immunomodulatory effects both syngeneic and allogeneic, but not xenogeneic CPCs, reduce scar size following MI in the rat.

We evaluated neovessels, arterioles formation and cardiomyocyte proliferation at day 28 post-transplantation. Transplanted allogeneic and syngeneic rCPCs significantly increased neovessel and arteriole density; confirmed by co-staining for Isolectin B4 (IB4) and smooth muscle actin (α-SMA) whereas xenogeneic rCPCs did not show any increase in neovessel and arteriole density as compared to placebo (Fig. 3G–I). Similarly, transplanted allogeneic and syngeneic rCPCs, but not xenogeneic rCPCs, significantly increased cardiomyocyte proliferation in the border zone of the infarcted area as identified by co-staining for the mitotic marker phospho-histone H3 (pHH3) and alpha-sarcomeric actin (Fig. 3J, K). The spatiotemporal development of immune rejection in the scar, border zone, and remote myocardium was examined by assessing inflammatory cells in hematoxylin and eosin-stained sections. The allogeneic and syngeneic treatment groups showed very mild cell-mediated rejection, in contrast to the xenogeneic group, which was characterized by significant mononuclear infiltration of the infarct and peri-infarct area with both interstitial and perivascular distributions (Fig. 3L, M). This interpretation was performed by a blinded cardiac pathologist. The immune scoring system applied was a modified descriptive grading system, which is a based upon the established International Society of Heart and Lung Transplantation grading system performed for heart transplantation rejection [32]. To assess the humoral response, recipient rat serum was collected at day 28 after transplantation and circulating anti-donor antibodies were detected by incubating rCPCs with the collected serum. In the allogeneic and syngeneic groups there were no significant differences compared to control, whereas in the xenogeneic group the titers of IgG and IgM antibodies were significantly increased at day 28 (Fig. 3N).

Previous studies with liquid chromatography coupled with mass spectrometry (LC–MS/MS) of human CPCs have identified GDF15 as one of the highly expressed proteins in its secretome and playing an important role in modulating cell survival pathways on target cells [8]. Immunoblot analysis of the secretome of rCPCs confirmed that these cells produce significant amounts of GDF15 (Fig. 4A, B). To explore the direct role of GDF15 in myocardial functional recovery, we knocked down GDF15 expression in rCPCs (rCPCs^GDF15KD) and transplanted these cells in the rat MI model. Our results showed that LVEF and LVFS were significantly decreased in rats that received rCPCs^GDF15KD compared to control rCPCs (Fig. 4C, D). Hearts treated with rCPCs^GDF15KD had significantly larger scars relative to rCPCs, and scar size was significantly reduced in rCPC-treated hearts when compared with the placebo-treated group (Fig. 4E, F). Further, the levels of circulating anti-inflammatory cytokines IL-2, IL-10, were significantly lower in the rCPC^GDF15KD group but inflammatory cytokine Tumor Necrosis factor (TNF-a) in the rCPC^GDF15KD group when compared to the rCPC group (Fig. 4G). No significant difference was observed with TGF-B. To delineate the mechanism of action of GDF15, we injected 1 million rCPCs or rCPCs^GDF15KD intramyocardially in rats undergoing MI. To Assess the T-regs and M2 macrophages cells, single-cell suspensions from freshly obtained leukocyte-enriched fractions of the entire heart at 5 days post-MI and flow cytometry was performed. Data showed that CD4⁺/CD25⁺/FOXP3⁺ T-regs and CD45⁺/CD68⁺/CD163⁺ M2 macrphages cells were significantly decreased in rCPCs^GDF15KD transplanted group compared to rCPCs (Fig. 4H–K). Increased GDF15 secretion was observed with the transplantation of rCPCs as compared to rCPCs^GDF15KD (Additional file 1: Fig. S2), confirmed by immunohistochemistry at day 5. These data show that GDF15 expression by rCPCs plays a key role in their immunomodulatory and cardioprotective effects in vivo. Activation of NF-κB in Treg cells were observe by flow cytometry of single-cell mixtures from treated hearts (Fig. 4L) and quantify (Fig. 4M) NF-κB activation in Treg cells (CD4⁺/FOXP3⁺/pP65⁺). These results were further confirmed by immunohistochemistry by co-staining for FOXP3⁺ and pP65 for NF-κB activation in treg cells (Fig. 4N, O). NF-κB activation was decreased with rCPC treatment compared with placebo or rCPCs^GDF15KD treatment in the T-reg cells. There was no expression of pP65 in the cardiomyocyte (Fig. 4P). These data suggest that rCPC secretome contain GDF15, which modulates NF-kB activity in FOXP3 positive T-reg cells modulates M2 levels in the myocardium contributing to the improved myocardial recovery in rat MI model. To explore the receptor activity and downstream signaling pathways involved in the GDF15-mediated inhibition of NF-kB, we performed an invitro co-culture experiment with spleenocytes and rCPCs with or without GDF15. As the expression of Glial-cell- derived neurotrophic factor family receptor α-like (GFRAL), is highly restricted to neuronal cells of the hindbrain and is virtually absent in all of peripheral tissues [33‐35]. Recently, CD48 is the first discovered receptor of GDF15 in the immune system and is exclusively expressed on immune cells [21]. In co-culture experiment we looked for the uptake of rCPCs derived GDF15 by CD48 receptor on T-cells by flow. Data showed that all CD4 cells has CD48 receptor and the uptake of GDF15 by T-cells was significantly higher in rCPCs group as compared to control and rCPC-GDF15KD cells (Additional file 1: Fig. S1). This data suggest that rCPC secretome that contain GDF15, which modulates NF-kB activity in FOXP3 positive T-reg cells by CD48 receptor present on T-cells.

To explore the possible role of Tregs in the recovery of LV function after MI, we used the nude rat which is T cell deficient. We injected 1 million cells [rCPCs, Treg⁻ cells (CD4⁺/CD25⁻), Treg⁺ cells (CD4⁺/CD25⁺), and rCPCs withTreg⁺ cells] intramyocardially following LAD ligation. Echocardiograms performed 4 weeks later showed significant increases in LV ejection fraction (Fig. 5A) and fractional shortening (Fig. 5B) in rats treated with combination cell (rCPCs + Tregs) therapy compared with rats treated with rCPCs alone or with placebo. Our data showed that rCPCs alone failed to recover the myocardium as shown in Fig. 5A and B, but when we injected rCPCs and Treg together, that significantly improved the ejection fraction and fraction shortening as compared to placebo, rCPCs, Treg⁻ cells and Treg⁺ cells. This data suggest that rCPCs derived GDF15 and Treg⁺ cells work synergistically. We also assessed M2 macrophages in single-cell suspensions of whole hearts treated with rCPCs or with the combination cell therapy (rCPCs + Treg) by flow cytometric analysis. To assess the M2 macrophages we isolated CD11b/c cells using magnetic microbeads, CD11b/c cells were co-stained for CD163 (Fig. 5C–E). CD11b/c⁺/CD163⁺ were significantly higher with the combination cell therapy compared with rCPCs therapy alone (Fig. 5F). These results suggest that Tregs augment M2 macrophage polarization, providing cardioprotective effects to the ischemic heart [18]. This concept is further supported by the gene expression of ARG1 (Fig. 5G), which is a hallmark for M2 macrophages, and the gene expression of the anti-inflammatory cytokine IL10 (Fig. 5H) in hearts that received the combination cell therapy (rCPCs + Treg). Since previous studies have shown that elevated GDF15 levels functionally protect the myocardium through an antihypertrophy response during left ventricle pressure overload, the protein levels of GDF15 and its known downstream effector, SMAD 2/3, were determined in RV myocardium [36]. We also isolated the protein from nude rat treated with placebo, rCPCs and rCPCs + Treg after MI from the infarcted zone at day 5 and determined the GDF15 expression by western blot analysis. Data showed that relative to placebo and rCPCs, rCPCs + Treg treated group has increased protein expression of GDF15 (Fig. 5I and J). This data suggests that alone rCPCs in T cell deficient rat do not secret GDF15, but in combination with Treg cells allogenic rCPCs secrete significantly higher GDF15, that co-relate to our invivo data Fig. 5A and B.