Angiogenic Mechanisms of Human CD34⁺ Stem Cell Exosomes in the Repair of Ischemic Hindlimb

Abstract

Rationale

Paracrine secretions appear to mediate therapeutic effects of human CD34⁺ stem cells locally transplanted in patients with myocardial and critical limb ischemia as well as in animal models. Earlier, we had discovered that paracrine secretion from human CD34⁺ cells contains pro-angiogenic, membrane-bound nano-vesicles called exosomes (CD34Exo).

Objective

Here, we investigated the mechanisms of CD34Exo-mediated ischemic tissue repair and therapeutic angiogenesis by studying their miRNA content and uptake.

Methods and Results

When injected into mouse ischemic hindlimb tissue, CD34Exo, but not the CD34exo-depleted conditioned media, mimicked the beneficial activity of their parent cells by improving ischemic limb perfusion, capillary density, motor function and their amputation. CD34Exo were found to be enriched with pro-angiogenic miRNAs such as miR-126-3p. Knocking down miR-126-3p from CD34exo abolished their angiogenic activity and beneficial function both in vitro and in vivo. Interestingly, injection of CD34Exo increased miR-126-3p levels in mouse ischemic limb, but did not affect the endogenous synthesis of miR-126-3p suggesting a direct transfer of stable and functional exosomal miR-126-3p. miR-126-3p enhanced angiogenesis by suppressing the expression of its known target, SPRED1; simultaneously modulating the expression of genes involved in angiogenic pathways such as VEGF, ANG1, ANG2, MMP9, TSP1 etc. Interestingly, CD34Exo, when treated to ischemic hindlimbs, were most efficiently internalized by endothelial cells relative to smooth muscle cells and fibroblasts demonstrating a direct role of stem cell-derived exosomes on mouse endothelium at the cellular level.

Conclusions

Collectively, our results have demonstrated a novel mechanism by which cell-free CD34Exo mediates ischemic tissue repair via beneficial angiogenesis. Exosome-shuttled angiomiRs may signify amplification of stem cell function and may explain the angiogenic and therapeutic benefits associated with CD34⁺ stem cell therapy.

INTRODUCTION

Stem and progenitor cell-based therapies have emerged as one of the most promising treatment options for patients with cardiovascular disease. Transplantation of autologous human CD34⁺ stem cells has been shown to improve perfusion and function in ischemic tissues and reduce amputation rates in patients with critical limb ischemia^{1, 2}. Laboratory experiments suggest that the benefits of human CD34⁺ cell transplantation occur primarily via increases in vascular angiogenesis³. Although involvement of CD34⁺ cell-secreted paracrine factors in the angiogenic process have been implicated⁴, the specific components and mechanisms by which the paracrine factors induce vessel growth and functional recovery post-ischemia remain largely undefined.

In our earlier study, we have established a novel mechanism that human CD34⁺ cells secrete membrane-bound nano-vesicles called exosomes (i.e. CD34Exo) that mediate most of the pro-angiogenic paracrine activity of the cells⁵. We have shown that the exosomes secreted by CD34⁺ cells were similar to exosomes described in previous reports- in their morphology, in size and shape, in expressing known exosomal protein markers as well as in expressing CD34⁺ cell-specific CD34 protein maker on their surface. Moreover, CD34Exo mimicked the function of their parent cells, at least in part, and induced angiogenic activity both in vitro and in vivo. Exosomes from several different cell types have been shown to carry and transfer selective cytosolic components such as proteins, lipids and nucleic acids⁶ to communicate with cells at the vicinity or at a distance, altering their function^{7, 8}. Interestingly, the unique cargo of exosomes is often distinct from the cell of their origin, although they are also known to carry selective cell-specific signature molecules such as parent cell-specific surface proteins or disease-specific signature proteins originating from the parent cells.

In several recent parallel investigations, role of exosomes as a mediator of cardiac communication among different cell types in the heart has been studied intensively. Both human and mouse stem and progenitor cell-derived exosomes have been shown to augment myocardial function post-ischemia^9–12. Remarkably, cardiac progenitor cell (CPC) -derived exosomes isolated from neonatal patients were found to have higher regenerative potential for cardiac tissue repair compared to CPC exosomes from older children¹³. Moreover, expression of certain exosomal cargo, such as miR-126 was significantly lower under high-glucose or diabetes conditions in human CD34+ exosomes, indicating that the exosomal cargo is dependent on the physiological condition of the cell of their origin. Recent research highlights significant potential of exosomes-derived from pericardial fluid and plasma obtained from human heart failure patients to induce therapeutic angiogenesis^14–16. In contrary to the beneficial effects shown by the stem cell exosomes, cardiac fibroblast-derived exosomes were shown to disseminate the damaging effects of cardiac remodeling by transferring miR-21* as a paracrine signaling mediator of cardiac hypertrophy¹⁷. While most of the current cardiovascular exosomes studies have examined the RNA and miRNA content as well as function, the in vivo uptake mechanisms of exosomes and exosomal miRNAs are largely undefined.

In our current study, we have investigated the role of CD34Exo in the beneficial angiogenesis associated with CD34⁺ cell therapy after ischemic injury by i) exploring whether CD34Exo induce beneficial effects in the absence of cells or other paracrine factors, ii) identifying the molecular components responsible for the angiogenic and therapeutic function of CD34Exo, iii) studying the in vitro and in vivo uptake mechanisms od CD34Exo.

METHODS

Cell culture and exosomes isolation

CD34⁺ cells and the CD34⁺-cell–depleted mononuclear cells (MNCs) were obtained from Baxter Healthcare Corporation (Deerfield, IL, USA). The cells were purified from mobilized peripheral-blood mononuclear cells (AllCells LLC, Emeryville, CA, USA) with an Isolex 300i device (Baxter Healthcare); cell purity (85–95%) was determined via flow cytometry. Both CD34⁺ cells and MNCs (250,000 cells/mL) were cultured in X-VIVO 10 serum-free cell-culture medium (Lonza Inc. Allendale, NJ) and ultra-pure exosomes were isolated from the conditioned media by ultracentrifugation and separated from the soluble protein fraction on a sucrose gradient⁵. The exosomes from both cell types were characterized using dynamic light scattering, flow cytometery analysis, electron microscopy and immunoblotting as described before^{5, 18}. Human umbilical-vein endothelial cells (HUVECs) (Lonza Inc. Allendale, NJ) were maintained in endothelial growth medium-2 (EGM^™-2; Lonza Inc.) and starved in EBM-2 medium containing 0.25% fetal bovine serum for 24 hours before cell assays were performed.

Mouse hind limb ischemia model

All animal protocols were approved by the AAALAC accredited Northwestern University and Icahn School of Medicine at Mount Sinai Animal Care and Use Committee. Immunocompromised BalbC mice (8–10 weeks old) were anesthetized with Isoflurane delivered at approximately 2%. A ligation was made around the femoral artery and all arterial branches were removed. A small segment of the artery was then dissected free. Post ischemia, mice were randomly assigned to receive intramuscular injections of PBS, CD34⁺ cells, CD34⁺ cell conditioned media, CD34⁺ Exosomes, CD34⁺ exosomes-depleted conditioned media, or MNC exosomes (all derived from equal number of cells; treatment dose: 5×10^6 cells/kg). All treatments were injected directly into the ischemic hindlimb in a 20-μl volume and injected at 4 different locations immediately after the surgery.

Laser-Doppler Perfusion Imaging (LDPI) and scoring for limb functional recovery

For laser Doppler measurements of the ischemic and control limbs, animals were anesthetized with Isoflurane (2%) and LDPI measurements were taken at 7, 14, 21, and 28 days following the hind limb ischemic surgery. At each time point, tissue perfusion was measured via LDPI, measuring blood flow in both the ischemic and non-ischemic limb and reporting results as the ratio of these two measurements. Ischemic and non-ischemic tissues were harvested at day 28 for histological analyses. Before sacrifice, the mice were injected with 50 μg of BS-1 lectin to identify the mouse vasculature. In addition, motor function and tissue damage was semiquantitatively assessed on postoperative day 7, 14, 21, and 28 by established scoring systems.

Immunohistology for quantification of capillary density

At day 28, muscle tissue from the ischemic limb was harvested, fixed in methanol, paraffin-embedded, and cross-sectioned (6 μm) for histological immunostaining. Briefly, sections were blocked with 10% donkey serum (30 min, room temperature). Primary antibodies for BS-1 lectin (Vector Laboratories, Burlingame, CA, USA), an endothelial cell marker was applied to tissue slices for 1h at 37 °C and further with AlexaFluor-conjugated secondary antibodies (Invitrogen Corporation, Carlsbad, CA, USA) and nuclei were counterstained with DAPI (Vector Laboratories, Burlingame, CA, USA). BS-1 lectin positive cells were imaged and quantified using fluorescent microscopy (Zeiss).

MicroRNA quantification

Total RNA from the CD34⁺ cells, CD34⁺-depleted MNCs, and their respective exosomes preparations were extracted using the miRNeasy Mini Kit (Qiagen) according to the manufacturer’s protocol (including a DNase step). Taqman small RNA assay and Taqman qRT-PCR were used to assess miRNA and mRNA expression, respectively. miRNA and mRNA expressions were normalized to U6 or 18s rRNA, respectively.

MicroRNA Microarray

MiRNA profiling was completed at Asuragen using Affymetrix miRNA microarrays and data analysis was performed by Asuragen. CD34⁺ cells, CD34Exosomes, MNCs (depleted of CD34⁺ cells), MNCExosomes- all isolated from the same donor were processed and compared for 1100 known human microRNAs. At least three healthy donors were used for the experiments (n=3). RNA was isolated as before. The expression level Signals were scaled in GCOS 1.2 to give a median array intensity of 100. This was done to enable different arrays to be compared. The program Spotfire DecisionSite 8.2 (www.spotfire.com) was used for gene-profiling analysis.

LNA-mediated miRNA knockdown

Anti-miR 126-3p, or, scrambled control LNA (Ambion, Carlsbad, California, U.S) was transfected into CD34⁺ cells using Lipofectamine RNAiMAX (Invitrogen Corporation, Carlsbad, CA, USA) for 16 h. Cells were washed 3 times and re-plated in fresh media to remove any external miRs. Cells were incubated for exosomes secretion for 40h. Exosomes were isolated using the protocol described before and used in an in vitro Matrigel tube formation assay using HUVECs to evaluate their angiogenic properties. Mouse hind limb ischemia model was used to compare the therapeutic efficacy of miR-126 KD exosomes with the control scrambled-miR-exosomes. Pro-angiogenic mir-126-3p expression in both the cells and exosomes were quantified using Taqman qRT-PCR assay as described before.

In-vitro Matrigel tube formation assay

HUVECs (2.0×10⁴, serum-starved overnight) were mixed with either PBS (control), or, with control scrambled CD34Exo or miR 126-3p-KD CD34Exo collected from the conditioned media of 2.0×10⁴ CD34⁺ cells – and seeded into 48-well plates. The plates were previously coated with 150 μl of growth-factor–reduced Matrigel^™ (BD). Tube formation ability of control, or, exosomes-treated HUVECs was examined by phase-contrast microscopy about 2–4 hours later. Each condition in each experiment was assessed at least in duplicates. Tube length was measured as the mean summed length of capillary-like structures (>6 cell long) per well, per high-power field using ImageJ software.

Exosomes treatment and qRT-PCR analysis of the HLI muscle tissue

HLI was induced in mice as described before and randomized to receive treatments of either PBS control, or, CD34⁺ exosomes-scrambled RNA, or, CD34⁺ exosomes-anti miR-126-3p in the ischemic limb; two injections of 10 μl each, flanking the ischemic tissue. A small piece of muscle flanking the treatment areas was dissected after certain time points. Total RNA was isolated and expression of miR-126-3p, mouse pri-126, mouse Spred-1 and mouse VEGF was determined as described before¹⁹.

Live Confocal Imaging and Flow Cytometry Analysis of Cy3-Exo Treated HUVECs and Ischemic tissue

Cy3 tagged miR, or, untagged scrambled control (Ambion, Carlsbad, California, U.S) was transfected into CD34⁺ cells using Lipofectamine RNAiMAX (Invitrogen Corporation, Carlsbad, CA, USA) for 16 h. Cells were washed 3 times and re-plated in fresh media to remove any external miRs. Cells were incubated for exosomes secretion for about 24–40h. Intact exosomes were RNase-treated at 37^oC for 10min to remove any external or, surface-bound Cy3. RNase activity was stopped by RNase-inhibitor. Exosomes were washed with excess PBS, pelleted, re-suspended in PBS, applied to HUVECs (that were plated on coverslips) and imaged immediately using live confocal microscopy.

For analysis of fluorescent-treated ischemic tissue, the exosomes were isolated from a stable dye, Rhodamine-PE-treated CD34⁺ cells. Ischemic tissues, treated with CD34Exo, were harvested at different time points and immediately processed for histology, or, flow cytometry analyses. For flow cytometry analysis, the tissues were digested into single cell suspension using collagenase digestion (0.5mg/ml, 20min at 37^oC), the cells were first stained with a live-dead kit, fixed (with 4% PFA), washed, and stained with cocktail of antibodies/secondaries specific for endothelial cells, smooth muscle cells and fibroblasts and analyzed by flow cytometry (BD LSR Fortessa with six lasers, 16-color analysis, CA, USA). For immunofluorescence studies, mice were injected with lectin (Vector laboratories, CA, USA; injected by tail-vein injections), hindlimb tissues were processed immediately after harvesting, stained for lecin and imaged using Nikon A1R laser scanning confocal microscopy.

Quantified results are presented as mean±SEM; comparisons between groups were evaluated with Students t-test or ANOVA and P<0.05 was considered significant. Detailed methods for all protocols used in the manuscript are provided in the online Supplemental Materials section.

RESULTS

Cell-free CD34Exo enhance therapeutic recovery and vascular angiogenesis in a mouse model of HLI

To evaluate the therapeutic efficacy of CD34Exo, we isolated and purified CD34Exo as described previously⁵. Using dynamic light scattering analysis, we confirmed that isolated CD34Exo fraction is devoid of contaminating protein aggregates as well as larger extra cellular vesicles (Online Figure I). CD34Exo-depleted CM, exosomes derived from MNCs (MNCExo) or PBS was used as negative control. To study beneficial effects of CD34Exo on tissue repair post-HLI, we created mouse models by ligating femoral artery as described before²⁰. We then injected the ischemic limbs with either PBS, CD34⁺ cells, CD34-CM, CD34Exo, CD34Exo-depleted-CM or MNCExo - all isolated from equal number of cells (treatment dose: 5×10^6 cells/kg). Mice treated with CD34⁺ cells, CD34-CM or CD34Exo had robust improvements in tissue necrosis, tissue perfusion, limb motor function and limb salvage compared to controls (Figure 1A–E). Interestingly, we observed that that the ischemic limbs injected with CD34 Exo, but not with MNCExo, were retaining their limbs and overall muscle mass, which were evident as early as day 7-post HLI/treatment (Online Table IA, IB). Quantification of limb tissue microcapillary density indicated significant increase in the mice treated with CD34⁺ cells, CD34Exo or CD34-CM (Figure 2A, 2B) suggesting enhanced angiogenesis. Interestingly, CD34-CM depleted of CD34Exo (CD34Exo-depleted-CM) lost its beneficial effects (Figure 1 & Figure 2) suggesting that the exosomes in CD34-CM provide most, if not all of the beneficial effects. Together these data suggest that cell-free CD34Exo have independent therapeutic activity similar to that of CD34⁺ cells and that CD34Exo induces angiogenesis.

Figure 1.

CD34Exo improve revascularization and functional benefits in mouse HLI. Day28 of mouse ischemic (L) hind limb treated as indicated (A); day28 LDPI images, severely restricted limb perfusion in blue and normal in red (B); quantification of LDPI perfusion (C); limb motor function (D) and salvage (E) scores of the ischemic limb (in a scale of 1–5; 1-poor and 5-strong); n=10–21; *P< 0.05.

Figure 2.

CD34Exo improve revascularization in mouse HLI. Capillaries shown by lectin staining from day28 ischemic tissues in CD34Exo-, CD34⁺ cells-, CD34-CM, CD34Exo-depleted CM-, MNCExo- or PBS-treated ischemic limbs (A); Quantification of ratio of capillary density between ischemic and non-ischemic limbs in each treatment groups. *p< 0.05 vs PBS, MNCExo and CD34Exo-depleted CM (B); n=4–5.

CD34Exo is enriched with angiomiR-126-3p and loss of miR-126-3p attenuates CD34Exo-mediated angiogenesis

To identify factors responsible for CD34Exo-induced beneficial effects and therapeutic angiogenesis, we examined the exosomal protein and RNA content. Exosomes function as mediators of cell-cell communication by selectively carrying biologically active molecules in the form of proteins and miRNAs²¹. To this end, we extracted protein from CD34Exo and blotted using human angiogenesis protein array that contains known proteins with angiogenesis function (Online Figure II). Angiogenic protein profiling indicated no significant enrichment in CD34Exo as compared to MNCExo (Figure 3A, 3B). We then isolated RNA from angiogenic CD34Exo and performed microarray profiling of miRNAs comparing that to control non-angiogenic MNCExo. To avoid contaminating RNAs sticking to outside of exosomal membranes, we performed RNase A digestion of intact exosomes without altering the quantity or quality of the isolated RNA (Online Figure III). Furthermore, analysis of total exosomal RNA confirmed that CD34Exo was mostly enriched with small RNAs and miRNAs as compared to CD34⁺ cells (Online Figure IV, V). MicroRNA microarray analysis revealed that CD34⁺ cells and CD34Exo had strong similarity in their miRNA composition and had significantly greater expression of several angiomiRs as compared to MNCs and MNCExo (Figure 3C & Online Figure XI). We determined the most enriched miRNAs in CD34⁺ cells and MNCs to compare with that of CD34Exo and MNCExo (Online Table II). Interestingly, the exosomes from both cell types were distinctively enriched with certain miRNA species of unknown function compared to the cells and vice versa (Online Figure XII & Online Table II). The exosomes-specific miRNAs may suggest selective packaging of exosomal miRNAs that may be involved in exosomes-specific signaling and function in target cells.

Figure 3.

CD34Exo carry pro-angiogenic miRNA. Human angiogenesis protein array comparing CD34Exo and MNCExo (A); quantification of proteins present in CD34Exo and MNCExo using human angiogenesis protein array (B); miRNA expression profiles comparing CD34⁺ cells with MNCs and CD34Exo with MNCExo, shown using heat map expression analysis (C).

MiR-126-3p was one of the most enriched miRNAs in both CD34⁺ cells and CD34Exo and was one of the most differentially expressed miRNA between CD34⁺ cells/CD34Exo and MNCs/MNCExo (Figure 3C). Further, it was significantly enriched in CD34Exo compared to CD34⁺ cells (Figure 4A). Interestingly, expression of miR-126-3p was lower in CD34Exo from CMI-patient-derived CD34+ cells compared to CD34Exo from healthy individuals although the difference was not statistically significant (Online Figure XIII). To address whether miR-126-3p contributes to CD34Exo function, we knockdown (KD) its expression in CD34⁺ cells using anti-miR-126-3p or a scrambled miR control; the exosomes isolated from miR-126KD cells had reduced expression of miR-126-3p (miR-126-KD-Exo) compared to the exosomes isolated from scrambled control-treated cells (Scrambled-control-Exo) (>99% decrease; Figure 4A). To measure the angiogenic activity of miR-126-KD-Exo, we treated HUVECs with either miR-126-KD-Exo or Scrambled-control-Exo and quantified the tube length in a in vitro Matrigel tube formation assay. We observed a significant decrease in tube length in miR-126-KD-Exo-treated HUVECs compared to Scrambled-control-Exo-treated HUVECs (Figure 4B & Online Figure IX). These results indicate that angiomiR-126-3p present in CD34Exo is a potential mediator of CD34Exo-mediated angiogenic function in vitro.

Figure 4.

Loss of miiR-126 in CD34Exo results in loss of its angiogenic function in vitro and in vivo. MiR-126 expression in CD34⁺ cells or exosomes isolated from CD34⁺ cells, treated as indicated, n=3–6 (A); tube formation of HUVECs on Matrigel, treated as indicated; *P<0.05, n=3 (B); experimental scheme illustrating CD34Exo injection to ischemic (I) hindlimb (C); Day28 of mouse ischemic hind limb treated as indicated and LDPI images, severely restricted limb perfusion in blue and normal in red (D); quantification of LDPI perfusion (E); limb motor function (F) and salvage (G) scores of the ischemic limb (in a scale of 1–5; 1-poor and 5-strong); Quantification of ratio of capillary density between ischemic (I) and non-ischemic (NI) limb (H, I); *p< 0.001, vs. scrambled-control-cells or scrambled-control-Exo, n=6–8.

CD34Exo directly transfers miR-126-3p to regulate gene expression in ischemic hindlimb tissue

To gain further mechanistic insights into the role of exosomal miR-126-3p in CD34Exo-induced therapeutic angiogenesis, we injected mouse ischemic hindlimb with either miR-126-KD-CD34Exo or Scrambled-control-Exo and measured limb function and capillary density (Figure 4C). miR-126-KD-CD34Exo-treated ischemic hindlimbs had significantly lower limb perfusion, limb motor function and limb salvage compared to the Scrambled-control-Exo (Figure 4D–G). Similarly, the capillary density as measured by lectin-positive cells was significantly lower in the miR-126-KD-CD34Exo-treated ischemic hindlimbs (Figure 4H, 4I). This indicates that knocking down exosomal miR-126-3p abrogated the angiogenic and therapeutic potential of CD34Exo.

Next, to investigate whether the observed angiogenesis was mediated directly by the transfer via CD34Exo, or indirectly as a result of de novo synthesis in the mouse ischemic tissue, we quantified mature miR-126-3p (Exo-transferred plus mouse-tissue synthesized) and mouse-specific primary miR-126-3p (pri-miR-126) (only mouse-tissue synthesized) in the ischemic limbs treated with either PBS, miR-126-KD-CD34Exo or Scrambled-control-Exo. We detected significantly high level of miR-126-3p in the Scrambled-control-Exo-treated limbs compared to miR-126-KD-CD34Exo-treated limbs (Figure 5A). On the other hand, there was no difference in the levels of mouse-specific pri-miR-126 levels between hindlimbs treated with either PBS, scrambled-miR-CD34Exo or miR-126-KD-CD34Exo (Figure 5B), suggesting there was no de novo synthesis of miR-126-3p in the mouse limb post-CD34Exo treatment. Therefore, the observed increase in mature miR-126-3p post-CD34Exo treatment was most likely due to the direct transfer of mature miR-126-3p from the cargo of CD34Exo.

Figure 5.

Loss of miiR-126 in CD34Exo results in loss of angiogenic gene activation function. Expression of miR-126 (A) and mouse pri-miRNA (B) normalized to U6; mSPRED1 (C); mVEGF (D); ANG1 (E) and MMP9 (F) mRNAs normalized to 18s RNA, in tissue homogenates of non-ischemic (NI) or ischemic (I) limbs treated as indicated, at 4h; n=4–9; P<0.05, *vs. PBS or, PBS I; †vs. scrambled-control-Exo.

We then tested whether CD34Exo-mediated enhanced angiogenic activity could be due to the regulation of known miR-126-3p target mRNAs in mouse hindlimbs. In line with this hypothesis, we observed a significant reduction in SPRED1 mRNA expression, a direct target of miR-126-3p^{19, 22}, in scrambled-miR-CD34Exo, but not in miR-126-KD-CD34Exo-treated HLI tissues compared to other treatment groups (Figure 5C). Conversely, we observed increased expression of pro-angiogenic mRNAs including VEGF, ANG1 and MMP9 by several folds in the scrambled-miR-CD34Exo treated HLI tissues (Figure 5D–F). Similar gene expression pattern was observed for additional genes involved in angiogenic pathways (Online Figure X). Notably, miR-126-KD-CD34Exo treatment failed to induce angiogenic mRNA expression compared to PBS treatment in these ischemic hindlimbs. Collectively, these data suggest that CD34Exo directly delivers miR-126-3p to ischemic tissues by inducing angiogenic gene expression.

Uptake and transfer of CD34Exo and its miRNA content in vitro and in vivo

To investigate the direct transfer of miRNAs via CD34Exo, we used fluorescently (Cy3)-tagged miRNA and isolated exosomes containing Cy3-miR. We treated CD34⁺ cells with Cy3-miR and isolated Cy3-miR-containing exosomes from the CM of those cells. We eliminated non-specific exosomal surface-bound Cy3-miR by treating the isolated CD34Exo with RNase A. Flow cytometry analysis confirmed that Cy3-miR is contained in CD34Exo (Figure 6A, 6b)²³. Next we treated HUVECs with Cy3-miR-CD34Exo or with a non-fluorescent control-miRNA-CD34Exo prepared in the same way. Live confocal microscope imaging shows presence of Cy3-positive punctate cytosolic vesicles only in the Cy3-miR-CD34Exo-treated HUVECs, but not in the controls (Figure 6C). To study the in vivo uptake of CD34Exo in the ischemic hindlimb, we used PE-CD34Exo isolated from the CM of Rhodamine-PE-tagged CD34⁺ cells and treated PE-CD34Exo or an unstained Exo control to the ischemic hindlimb, digested the tissue to prepare single cell suspensions and analyzed for the presence of PE fluorescence in the cells using flow cytometry (at 2h) and histology at 12h (Online Figure XI). To identify the individual cell types uptaking CD34Exo, we used known cell type-specific antibodies. Our data revealed that PE-CD34Exo were uptaken most efficiently by endothelial cells as compared to smooth muscle cells and fibroblasts (Figure 6D, 6E). Interestingly, we observed the appearance of division in PE-CD34Exo-uptaken lectin-positive endothelial cells in the ischemic hindlimb (Figure 7A). In this line, CD34Exo-positive endothelial cells displayed different scatter properties than control PBS-uptaken endothelial cells (Figure 7B) suggesting that CD34Exo possibly induced cell cycle changes in those cells. To confirm involvement of cell cycle regulators in CD34Exo-injected tissue, we tested the expression of Cyclin genes at day 28 post-CD34Exo injection in the ischemic limbs. We identified increased Cyclin A and Cyclin B mRNA expressions in hind limbs treated with CD34Exo as compared to PBS control (Figure 7C, 7D and Online Figure XII). Collectively, our data suggest that CD34Exo from human CD34⁺ progenitor cells carry and transfer stable and functional miR-126-3p to activate angiogenic and cell cycle-related gene expression in the mouse ischemic hind limb tissue (Figure 8).

Figure 6.

Uptake and transfer of CD34Exo and exosomal miRNAs by endothelial cells in vitro and in vivo. Flow cytometry analysis of CD34⁺ cells (A) or exosomes isolated from CD34⁺ cells (B), transfected as indicated; confocal image of HUVECs treated with Cy3miRNA-CD34Exo (C); flow cytometry analysis of single cell suspensions from post-ischemic hindlimb tissue injected with R-PE-CD34Exo at 2h (D); % of cells uptaking exosomes (PE) from total number of each cell types quantified from Figure 5D (E), (n=3–5, *p<0.05).

Figure 7.

CD34Exo induce proliferation of endothelial cells. Confocal image of post-ischemic hindlimb tissue injected with R-PE-CD34Exo at 12h (A); flow cytometry analysis of back-tracked CD31-positive endothelial cells (in blue) from single cell suspensions of ischemic hindlimbs treated as indicated (B); expression of Cyclin A (C) and Cyclin B (D) mRNA normalized to 18S rRNA in tissue homogenates of ischemic limbs treated as indicated, at 28d; n=5–6; P<0.05.

Figure 8.

Summary of the new knowledge based on our hypothesis. Exosomes secreted by CD34⁺ cells induce tissue repair in mouse ischemic hind limb by delivering exosomal angiomiR-126 to endothelial cells inducing cell cycle changes and angiogenesis.

DISCUSSION

The molecular underpinning of paracrine factor-induced neovascularization that contributes to the therapeutic benefits of human CD34⁺ cell therapy following tissue ischemia is largely unknown. Here we demonstrate for the first time that CD34Exo constitute a key component of the paracrine secretion from CD34⁺ cells that promotes ischemic tissue repair. Cell-free CD34Exo were independently therapeutic and improved microcirculation in the ischemic limb, similar to CD34⁺ cells in a mouse model of HLI. Our results provide further insights into the molecular mechanisms underlying the angiogenic activity of CD34Exo by demonstrating that angiomiRs such as miR-126-3p, that are enriched in CD34Exo have a profound effect on their angiogenic activity. The distinct combination of angiomiRs enriched in CD34Exo may provide a plausible explanation for their superior angiogenic and therapeutic efficacy compared to MNCExo, which is also consistent with enhanced potency of CD34⁺ cells versus MNCs³. MiR-126, known as an essential component of endothelial angiogenic activity¹⁹, was functional and it was one of the most enriched microRNAs transported by CD34Exo. Treatment with CD34Exo suppressed endogenous SPRED-1, one of the known direct targets of miR-126-3p, and induced expression of several pro-angiogenic mRNAs, which are not known to be a direct target of miR-126-3p (Online Figure XIII), in a miR-126-dependent manner. We speculate that additional pathways may play a role in the CD34Exo-mediated angiogenesis. Moreover, the lack of angiogenic activity of miR-126-KD-CD34Exo could also be due to alteration of other unknown moieties (in addition to loss of miR-126-3p) in CD34Exo generated from miR-126-KD-CD34 cells. Thus, miR-126-3p may affect tissue angiogenesis via its direct targets or indirectly by modulating other angiogenic components in exosomes or in the exosomes-treated limbs. We do not rule out contribution from other microRNAs or factors in the repertoire of CD34Exo, alone or in combination, to the angiogenic and therapeutic activity of CD34Exo. In agreement, miR-10a and miR-130, enriched in CD34Exo, have antifibrotic activity suggesting that the overall benefits of CD34⁺ exosomes may exceed beyond angiogenesis. Moreover, our previous work has shown that exosomes from sonic hedgehog-modified CD34⁺ cells activated angiogenic sonic hedgehog signaling in other cell types²⁴. The complete molecular content of CD34Exo and the mechanism of their cell-specific uptake still remains to be fully characterized in future studies.

In patients with coronary artery disease, the plasma levels of miR-126-3p is significantly reduced^{25, 26}, indicating a dysfunctional endothelium in those patients. In patients with acute coronary syndrome, miR-126-3p is down-regulated across the coronary circulation implicating a high level of consumption of miR-126-3p in trans-coronary passage of patients²⁷. These data suggest an essential role of miR-126-3p in regeneration and repair of ischemic tissues. Therefore, CD34Exo-mediated delivery of miR-126-3p or other angiomiRs to the damaged tissues may provide a promising therapeutic strategy²⁸ as miRNAs are likely to be more stable within the exosomes membrane and uptaken efficiently to target cells⁶. Moreover, miRNAs directly supplied in the systemic circulation or injected into the tissues are prone to RNAse digestion, therefore less stable in vivo. Although exosomes are known to play a role in angiogenesis during cancer metastasis²⁹, in immune regulation or in disease pathogenesis⁶, their contribution to vessel growth under ischemic cardiovascular conditions is largely unknown. Our findings have unveiled a unique mechanism of revascularization that highlights the contribution of horizontal transfer of angiomiRs from human CD34⁺ stem cells to endothelial cells that mediates cell-cell communication. Exosome-shuttled angiomiRs may signify efficient amplification of stem cell function and may explain the results of clinical trials demonstrating significant functional improvement from a modest number transplanted CD34⁺ cells that are retained in the ischemic tissues. Moreover, we are beginning to propose CD34Exo therapy as a suitable cell-free alternative to CD34⁺ cell therapy. Exosomes have great potential to overcome certain limitations of cell therapies for cardiovascular diseases, such as- retention, viability and functional impairment of transplanted cells in the unfavorable ischemic environment, while retaining the benefits.

Recent cell therapy approaches have shown promising results to augment cardiac function and quality of life by several means including reduced fibrosis, enhanced vascular angiogenesis, replacement and repair of damaged myocardial tissue^{1, 30, 31, 32, 33}. Emerging evidence highlight a significant contribution from several different stem cell types conferred by paracrine mechanism mediated through exosomes and their miRNA repertoire ^{10, 15, 34–39}. Our results identified CD34Exo as a key therapeutic component in the paracrine secretion of human CD34⁺ stem cells. We have demonstrated a novel mechanism by which cell-free CD34Exo mediates ischemic tissue repair via beneficial angiogenesis. Thus, the benefits of CD34 cell therapy on functional recovery could be induced primarily through CD34Exo-mediated transfer of angiomiRs such as miR-126-3p to endothelial cells in the ischemic tissue. This new knowledge can be used to refine existing treatments and to develop alternative therapeutic approaches that may benefit patients with cardiovascular diseases.

NOVELTY AND SIGNIFICANCE.

What Is Known?

• Human CD34⁺ stem cells have been shown to induce therapeutic neovascularization in preclinical studies and in Phase I, II and III clinical trials.

• One of the key mechanisms by which CD34⁺ cells induce neovascularization appears to include paracrine secretion.

• Exosomes, secreted nanovesicles, are one of the major components of CD34⁺ cell paracrine secretion, which have independent angiogenic activity, both in vitro and in vivo.

What New Information Does This Article Contribute?

• Exosomes from CD34⁺ cells induce therapeutic angiogenesis that are mediated by exosomal angiomiRs such miR-126-3p.

• CD34⁺ exosomes are most effectively internalized by endothelial cells in the ischemic tissue to induce cell cycle induction, angiogenesis and proliferation.

• Exosome-shuttled angiomiRs may signify amplification of stem cell function and may explain the angiogenic and therapeutic benefits associated with CD34⁺ stem cell therapy.

In this study we addressed a key undefined paracrine mechanism by which human CD34⁺ stem cells mediate their beneficial effects. Our data show that exosomes, a major component of the paracrine secretion from CD34⁺ cells stimulate beneficial angiogenesis in a mouse model of hindlimb ischemia. CD34⁺ exosomes were selectively enriched with several angiomiRs that are distinct from miRs in non-therapeutic exosomes from CD34^- MNCs. Another key discovery is the first evidence of an in vivo trafficking mechanism of CD34⁺ exosomes and exosomes-shuttled ‘angiomiRs’. Our findings have both scientific and clinical importance. It will lead to a deeper understanding of the mechanisms of stem cell-exosomes trafficking and function on one hand and therapeutic interventions on the other.

Nonstandard Abbreviations and Acronyms

AngiomiRs

Pro-angiogenic miRNAs

CD34Exo

CD34⁺ cell-derived exosomes

CM

conditioned media

CPC

Cardiac progenitor cells

DLS

dynamic light scattering

Exo

exosomes

HLI

hindlimb ischemia

HUVECs

human umbilical-vein endothelial cells

MNCs

mononuclear cells (human CD34⁺ stem cell-depleted)

PBS

phosphate buffered saline

SPRED1

Sprouty-related, EVH1 domain-containing protein 1

VEGF

vascular endothelial growth factor