OUP user menu

The 106b∼25 microRNA cluster is essential for neovascularization after hindlimb ischaemia in mice

Jonathan Semo, Rinat Sharir, Arnon Afek, Camila Avivi, Iris Barshack, Sofia Maysel-Auslender, Yakov Krelin, David Kain, Michal Entin-Meer, Gad Keren, Jacob George
DOI: http://dx.doi.org/10.1093/eurheartj/eht041 First published online: 18 February 2013

Abstract

Aims MicroRNAs (miRNAs, miR) are endogenous short RNA sequences that regulate a wide range of physiological and pathophysiological processes. Several miRNAs control the formation of new blood vessels either by increasing or by inhibiting angiogenesis. Here, we investigated the possible role of the miR-106b∼25 cluster in postnatal neovascularization and in regulation of the angiogenic properties of adult bone marrow-derived stromal cells.

Methods and results To study the effect of miR-106b∼25 deletion on neovascularization, we used a miR-106b∼25 knockout (KO) mouse model. After inducing hindlimb ischaemia, we showed that vascularization in ischaemic mice devoid of miR-106b∼25 is impaired, as evident from the reduced blood flow on laser Doppler perfusion imaging. The miR-106b∼25 cluster was also shown here to be an essential player in the proper functioning of bone marrow-derived stromal cells through its regulation of apoptosis, matrigel tube formation capacity, cytokine secretion, and expression of the stem-cell marker Sca-1. In addition, we showed that capillary sprouting from miR-106b∼25 KO aortic rings is diminished.

Conclusion These results show that the miR-106b∼25 cluster regulates post-ischaemic neovascularization in mice, and that it does so in part by regulating the function of angiogenic bone marrow-derived stromal cells and of endothelial cells.

  • MicroRNAs
  • Angiogenesis

Introduction

Critical limb ischaemia has been estimated to develop in 500–1000 individuals per million per year.1 Many of the cells required to form neovasculature or remodel existing collaterals are derived from local proliferation of endothelial and smooth muscle cells.2,3 However, several studies suggest an additional contribution, derived from bone marrow progenitors, which are mobilized in the event of limb or myocardial ischaemia, migrate to ischaemic tissue, and become actively incorporated into new vessels.47

MicroRNAs (miRNAs, miR) are short RNA sequences (∼22 nucleotides) that regulate the expression of target genes by base-pairing with specific binding sites located in the 3′-untranslated region of target mRNAs.8 Several specific miRNAs were shown to control angiogenesis, including miR-126, which is highly expressed in endothelial cells and is essential for blood vessel growth in mice.9 In addition, angiogenesis is blocked by miR-221 and miR-222,10 but is increased by miR-27b11 and members of the let-7 family.11

The miR-106b∼25 cluster consists of three mature miRNAs, namely miR-106b, miR-93, and miR-25.12 This cluster is a paralogue of the miR-17∼92 and miR-106a∼363 clusters.13 Ablation of miR-106b∼25 or miR-106a∼363 has no obvious phenotypic consequences, although compound mutant embryos lacking both miR-106b∼25 and miR-17∼92 die at midgestation.12

The miR-106b∼25 cluster was reported to be overexpressed in human cancer and to act as an oncogene by targeting p21 and the pro-apoptotic member of the BCl2 family, Bim14,15 In addition, members of the miR-106b∼25 cluster modulate adult neural stem-cell proliferation and differentiation.16 High expression of miR-93 was reported in human umbilical vein endothelial cells,17 and it was reported to promote tumour angiogenesis.18

The aim of this study was to gain an understanding of the role of the miR-106b∼25 cluster in postnatal neovascularization and in regulation of the angiogenic properties of adult bone marrow-derived stromal cells (BMSCs).

Methods

Animals

Animal studies were approved by The Animal Care and Use Committee of Tel Aviv Sourasky Medical-Center (see Supplementary material online).

Hindlimb ischaemia model and laser-Doppler perfusion imaging

Female, 8-week-old mice underwent ligation of the femoral artery. Shortly after surgery, as well as 7 and 14 days post-operatively, the blood flow in both hindlimbs was determined by laser-Doppler perfusion imaging (LDPI) (Moor Instruments, Axminster, UK).

RNA extraction and quantitative real-time PCR

For miRNA detection, 25 ng of total RNA was transcribed to cDNA, qRT–PCR was performed with Sybr Green and primers for miR-106b, miR-93, and miR-25 (Exiqon, Denmark).

In situ hybridization

To detect expression of miR-93 and miR-25, skeletal muscle was stained according to the method of Obernosterer19 (see Supplementary material online).

Migration assay

Vascular endothelial growth factor (VEGF)(10 ng/mL) or SDF1 (Peprotech, Rehovot, Israel) was loaded into the lower well of a transwell insert (Nunc, Denmark). Migrating cells were counted in a BD Biosciences FACSCanto II flow cytometer.

Proliferation assay

Cells were plated in 96-well plates and incubated overnight in DMEM serum-free medium. Fetal calf serum (20%) was then added to the medium for 16 h. Proliferation was assayed with XTT (Biological Industries).

In vitro Matrigel tube formation assay

Bone marrow-derived stromal cells s, or BMSCs mixed with H5V cells, or H5V cells mixed with 200 µL of medium conditioned by BMSCs [from miR-106b∼25wild-type (WT) or KO mice] were seeded in each well of 24-well plates coated with 200 μL Matrigel (BD Biosciences).

Flow cytometry

Bone marrow cells were incubated with anti Sca-1, anti-FLK-1, and anti-C-kit antibodies (eBioscience, San Diego, CA, USA) or their corresponding isotype controls, for 30 min at 4°C in the dark, and analysed by flow cytometry (BD Biosciences FACSCanto II).

Mouse angiogenesis cytokine array

Proteome Profiler was used (R&D Systems, Minneapolis, MN, USA) according to the manufacturer's instructions (see Supplementary material online).

Retroviral infection

pMSCV-miR retroviral plasmids were transfected into Phoenix packaging cells (ATCC), using a standard Ca2PO4 protocol. H5V endothelial cells were incubated with viral supernatant for 24 h in the presence of 5 µg/mL polybrene (Sigma-Aldrich).

ELISA for cleaved caspase-3

Cleaved caspase-3 was assayed, in 40 µg of protein, using a commercially available Cleaved Caspase-3 kit (R&D Systems).

Aortic rings angiogenesis assay

Assay was performed according to Baker et al.20

Statistical analysis

Groups were compared using Student's t-test or one-way ANOVA. Repeat measure ANOVA was applied for the murine studies. The Bonferroni correction was taken to account for multiple testing. Significance was set at P < 0.05 (*P < 0.05; **P < 0.005). Results are expressed as means ± SEM.

Results

Expression of microRNA-106b∼25 cluster in ischaemic muscles

To examine possible changes in the expression of individual members of the miR-106b∼25 cluster in response to ischaemia, we used a mouse hindlimb ischaemia model. Levels of expression were determined in ischaemic muscles, in the contralateral non-ischaemic muscles, and in muscles of sham-operated animals. No change in expression levels was detectable 1 day after ischaemia was induced (Figure 1A). At 7 and at 21 days after the operation; however, expression of the entire miRNA cluster in the ischaemic limbs of the operated mice, but not in the non-ischaemic limbs, was increased (Figure 1B and C), with maximal effect seen on Day 7 for miR-106b and on Day 21 for miR-93 and miR-25 (fold changes of 3.14 ± 0.44, 8.12 ± 0.29, and 3.36 ± 0.3, respectively).

Figure 1

Relative expression of the miR-106b∼25 cluster after induction of hindlimb ischaemia. RNA was extracted from muscles of ischaemic, non-ischaemic, and sham-operated mice (n = 4 for each group) 1 day (A), 7 days (B), and 21 days (C) after induction of hindlimb ischaemia. Relative expression was determined by qRT–PCR with specific primers for miR-25, miR-106b, and miR-93. Results were derived by the ΔΔCt method and were normalized to the expression of U6. Relative expression is expressed as fold of sham-operated mice (*, P < 0.05). (D) Relative expression of the mir-106b∼25 cluster after induction of myocardial infarction.

The expression of the miR-106b∼25 cluster was also determined after experimental murine myocardial infarction (MI) in the infarct and in the remote myocardium, After induction of MI the expression of the miR-106b∼25 was up-regulated in the remote myocardium, (when compared with sham) (Figure 1D).

We further assessed the expression levels of miR-25 and miR-93 by LNA-based in situ hybridization in ischaemic muscles and in sham-operated muscles 7 days after ischaemia induction, using specific probes for miR-25 and miR-93. We also used a U6 probe as a positive control and a scrambled probe as a negative control. Expression of miR-25 was increased only slightly in response to ischaemia (Figure 2C and D), whereas miR-93 expression was significantly increased (Figure 2A and B). Expression of miR-93 was detected in muscle fibres as well as in capillaries.

Figure 2

Up-regulation of mir-93 after induction of hindlimb ischaemia. In situ hybridization in muscles of ischaemic (n = 3) and sham-operated mice (n = 3), 7 days after the induction of hindlimb ischaemia, was performed with double-DIG-labelled LNA probes specific for miR-93 (A and B) or miR-25 (C and D).U6 probe was used as a positive control (E and F) and a scramble probe as a negative control (G and H). Haematoxylin and eosin staining of sham (I) and ischaemic (J) is shown.

MicroRNA-106b∼25 is essential for an improvement of the blood flow after hindlimb ischaemia

To further characterize the functional role of the miR-106b∼25 cluster in the post-ischaemic improvement of the blood flow, we induced hindlimb ischaemia in miR-106b∼25 KO mice. Laser-Doppler perfusion imaging demonstrated (all values are expressed as the percentage of perfusion in the non-ischaemic limb) a significant increase in the blood flow in the WT compared with the KO mice on Day 7 (53.15 ± 4.6 and 28.21 ± 7.69%, respectively, P = 0.02) and Day 14 (65.24 ± 10.2 and 33.23 ± 2.7%, respectively, P = 0.019) after ischaemia induction (Figure 3A and B). In addition, the number of lectin-positive capillaries was increased in WT compared with KO mice (Figure 3E–G).

Figure 3

Measurement of the blood blow after hindlimb ischaemia. Ischaemia was induced in miR-106b∼25 wild-type (n = 5) or knockout mice (n = 5). The blood flow was measured by laser-Doppler perfusion imaging shortly after ischaemia induction, and at 7 and 14 days after ischaemia, and is expressed as a percentage of the blood flow in the non-ischaemic limb (A). Representative laser-Doppler images are shown at each time point (B). Mdh1-106b∼25 plasmid (30 µg) or Mdh1-GFP (30 µg; control) was injected into the ischaemic limb of miR-106b∼25 knockout (C) or wild-type (D) mice (n = 8 in each group). The blood flow is expressed as in (A). Representative images of lectin-positive capillaries in (E) wild-type and (F) knockout mice. (G) The number of capillaries (per high power field) was counted in four random images for each mouse.

Next, we examined whether re-expression of the miR-106b∼25 cluster in ischaemic muscles could rescue the mutant phenotype of miR-106b∼25 KO mice. A plasmid encoding the miR-106b∼25 cluster was injected into the ischaemic muscle shortly after induction of ischaemia. Re-expression of the miR-106b∼25 cluster was confirmed by real-time PCR (data not shown).

Compared with injection of the control vector Mdh1-GFP, injection of the Mdh1-106b∼25 plasmid significantly improved perfusion in KO mice 7 days after ischaemia induction (41.4 ± 1.9 and 58.7 ± 5.12% and, respectively, P = 0.003; Figure 3C), but at Day 14 this improvement was found not to have persisted (42.25 ± 7.8 compared with 52.69 ± 10.84, P = 0.1). Post-ischaemia, injection of the Mdh1–106b∼25 plasmid did not further improve the blood flow in WT mice (P = 0.17 and P = 0.77 for Days 7 and 14, respectively).

MicroRNA-106b∼25 regulates apoptosis of bone marrow-derived stromal cells

The miR-106b∼25 cluster has been previously shown to regulate apoptosis in cancer cells.14 That finding prompted us to examine the effect of miR-106b∼25 deletion on H2O2-induced apoptosis of cultured BMSCs. AnnexinV/propidium iodide (PI) staining followed by flow cytometric analysis demonstrated increased apoptosis in miR-106b∼25 KO BMSC after treatment with H2O2 (400 µM) for 8 h (39.1 ± 4.2% of AnnexinV-positive cells in the KO cells compared with 14.2 ± 6.1% in the WT), as well as 24 h after treatment (34.9 ± 2.2 and 11.2 ± 1.5%, respectively) (Figure 4A). Also, PTEN inhibition with VO-OHPIC (1 µM for 24 h) partially abolished the difference between WT and KO BMSCs (Figure 4A).

Figure 4

Apoptosis of wild-type and knockout cultured BMCS after H2O2 treatment. H2O2 (400 µM) was added to the culture medium for 8 or 24 h and the cells were stained with AnnexinV-FITC and PI. (A) Representative flow cytometry images of miR-106b∼25 wild-type or knockout bone marrow-derived stromal cells after induction of apoptosis. The percentage of AnnexinV-positive cells is indicated. (B) ELISA of cleaved caspase-3, performed with 40 µg protein extracted from bone marrow-derived stromal cells of miR-106b∼25 wild-type or knockout cells, with or without apoptosis induction (400 µM H2O2). Results are expressed as fold of wild-type. (C) Proliferation of miR-106b∼25 wild-type or knockout bone marrow-derived stromal cells as detected by the XTT assay. Results are expressed as fold of wild-type. (D) Number of miR-106b∼25 wild-type or knockout bone marrow-derived stromal cells that migrated towards 10 ng/mL of VEGF or SDF1 for 4 h or after PTEN inhibition using VO-OHpic in 1 µM for 24 h.

In addition, levels of cleaved caspase-3 (Figure 4B) were significantly higher in the KO cells than in the WT after treatment with 400 µM H2O2 for 24 h (2.17 ± 0.03 fold of WT compared with 1.27 ± 0.06 fold, respectively; P = 0.01).

MicroRNA-106b∼25 deletion reduces tube formation capacity of bone marrow-derived stromal cells

We further investigated the outcome of miR-106b∼25 deletion on neovascularization by assaying Matrigel tube formation in vitro. Bone marrow-derived stromal cells were plated on Matrigel-coated dishes, either alone or mixed with an equal number of H5V endothelial cells. Compared with miR-106b∼25 WT BMSCs, the tube formation capacity of miR-106b∼25 KO BMSCs was reduced when plated either alone or with endothelial cells (Figure 5A and D).

Figure 5

The Matrigel tube formation assay of endothelial cells and bone marrow-derived stromal cells. Tube formation was assayed after plating of miR-106b∼25 bone marrow-derived stromal cells on Matrigel-coated dishes. Wild-type or knockout bone marrow-derived stromal cells (A and B) plated 1:1 with H5V endothelial cells, or (C and D) plated alone, and (E and F) H5V endothelial cells plated with conditioned media obtained from miR-106b∼25 wild-type or knockout bone marrow-derived stromal cells. (G and H) Wild-type or knockout bone marrow-derived stromal cells after PTEN inhibition. (I) Cumulative tube length was measured with NIS-Elements software from three random images. Results are expressed in arbitrary units. (J and K) Dot-blot images and relative expression of 53 cytokines detected in 200 µL of conditioned media obtained from miR-106b∼25 wild-type or knockout bone marrow-derived stromal cells that were subjected to hypoxia (1% O2) for 24 h. (L) Relative PTEN protein levels as detected by ELISA in 293HEK cells transfected with either MDH1–106b∼25 or MDH1-GFP (M) western blot for total-AKT, phosphorylated-AKT and GAPDH performed with cell lysates obtained from 293HEK cells transfected with either MDH1–106b∼25 or MDH1-GFP (with or without serum activation).

Plating of H5V endothelial cells on Matrigel-coated dishes with conditioned media collected from miR-106b∼25 WT or KO BMSCs yielded no detectable effect on tube formation (Figure 5E and F), suggesting that the miR-106b∼25 cluster induces tube formation in H5V endothelial cells in a contact-dependent manner. Examination of the effect of miR-106b∼25 deletion on proliferation of BMSCs also yielded no difference between the proliferation rates of miR-106b∼25 WT and KO BMSCs (Figure 4C).

We next tested the effect of miR-106b∼25 deletion on two important migratory pathways of progenitor cells: the VEGF/Flk-1 and the SDF1/Cxcr4 pathways. Whereas deletion of the miR-106b∼25 cluster did not alter the migration of miR-106b∼25 KO BMSC towards SDF1, migration towards VEGF was significantly reduced in the mir-106b∼25 KO BMSCs compared with the WT cells (P = 0.15 and 0.05 respectively; Figure 4D). Also, PTEN inhibition reversed the mutant phenotype of miR-106b∼25 KO BMSCs both for migration towards VEGF and tube formation capacity (Figures 4D and 5G and H).

MicroRNA-106b∼25 deletion reduces paracrine activity of bone marrow-derived stromal cells

After migrating and homing to hypoxic sites, BMSCs reportedly induce angiogenesis by locally releasing paracrine factors that activate resident endothelial cells.6,7 We, therefore, assayed the secretion of angiogenic cytokines in conditioned media of BMSCs subjected to hypoxic treatment. Of the 53 proteins tested, 11 pro-angiogenic cytokines were found to be up-regulated in conditioned medium from WT BMSCs (fold change of 1.5 or higher in WT than in KO BMSCs, P < 0.05; Figure 5H and I). Among these angiogenic cytokines were CXCL1, CXCL10, proliferin, MMP-3, and members of the IGFBP family. Only two cytokines, endostatin and PDGF, were up-regulated in the conditioned medium from KO BMSCs (fold change of 1.5 or higher in KO than in WT BMSCs, P < 0.05), suggesting that the miR-106b∼25 cluster is important for the release of pro-angiogenic cytokines under hypoxic conditions.

Transfection of the microRNA-106b∼25 cluster reduces PTEN protein levels and increases AKT phosphorylation

We sought to characterize possible signalling pathways that could be regulated by the miR-106b∼25 cluster and influence cytokine expression. We assayed PTEN levels which is a known target for the miR-106b∼25 cluster.21 Upon transfection of the miR-106b∼25 cluster into 293HEK cells, PTEN protein levels were reduced (31 ± 4% reduction P < 0.05; Figure 5J). In addition, transfection resulted in an increase in AKT phosphorylation after serum activation (Figure 5K).

Deletion of the microRNA-106b∼25 cluster reduces expression of the stem-cell markers SCA-1 and CD133 in vitro

The miR-106b∼25 cluster was recently identified as a regulator of neural stem-cell proliferation and differentiation.16 This finding prompted us to investigate the effect of its deletion on the expression of various stem-cell markers. We identified the stromal cell cultures with two characteristic stem-cell markers, Sca-1 and CD133. Expression of these two markers was significantly reduced in the BMSC KO cultures (5.5 ± 1.5 and 2.9 ± 1% percentage of positive cells, respectively, compared with 40.3 ± 0.8 and 9 ± 1.2%, respectively, in WT BMSCs; Figure 6A and E), suggesting that the presence of miR-106b∼25 is important for the expansion of progenitor cells in vitro.

Figure 6

(A) Expression of stem-cell marker Sca-1and CD133 in bone marrow-derived stromal cell cultures obtained from miR-106b∼25 wild-type or knockout (B and C) CD133-positive and (D and E) Sca-1 positive cells.

Numbers of Sca-1-, C-kit-, or Flk-1-positive cells in bone marrow of miR-106b∼25 knockout mice

In view of the marked decrease in the expression of stem-cell markers in the miR-106b∼25 KO BMSC cultures, we were interested in assaying their expression in freshly isolated bone marrow cells. Bone marrow obtained from miR-106b∼25 WT and KO mice were stained for Sca-1, C-kit, and Flk-1 and analysed by flow cytometry. The numbers of Flk-1+ and Sca-1+Flk-1+ progenitors were slightly decreased in the KO mice compared with the WT (2.8 ± 0.6% percentage of positive cells and 1.48 ± 0.45% compared with 4.3 ± 0.8 and 2.18 ± 0.4%, respectively), but the difference between them was not significant (P > 0.1; Figure 7A and B).

Figure 7

Expression of Sca-1, Flk-1, and C-kit in freshly isolated bone marrow cells. Representative images of (A and B) miR-106b∼25 wild-type(n = 5) and knockout mice (n = 5), and from (C and D) miR-106b∼25 wild-type (n = 4) and knockout mice (n = 4) 7 days after the induction of hindlimb ischaemia. (E) The number of Sca-1 positive cells in ischaemic muscles (n = 4) obtained from Mir-106b∼25 wild-type and knockout mice.

To examine whether hypoxia differentially influences the expansion of bone marrow-derived stem cells in miR-106b∼25 WT and KO mice, we assayed the expression of the stem-cell markers Sca-1, C-kit, and Flk-1 at 7 days after induction of hindlimb ischaemia. The number of Sca-1-positive cells (as a percentage of the total number of bone marrow cells) was significantly increased in the miR-106b∼25 KO mice compared with the WT (8.4 ± 0.4 vs. 5.7 ± 0.16%; P = 0.0009 Figure7C and D), whereas the numbers of C-kit-positive and Flk-1 positive cells in the WT and KO mice were similar. The number of Sca-1 positive cells in ischaemic muscles was reduced in miR-106b∼25 KO mice (Figure 7E), but this difference was not statistically significant (P = 0.07).

Overexpression of individual members of the miR-106b∼25 cluster increases viability, proliferation, and migration of endothelial cells

To investigate the possible role of the miR-106b∼25 cluster in murine endothelial cells, we infected the H5V cell-line with retroviruses encoding miR-106b, miR-93, miR-25, or HTR-control virus. Retroviral infection of miR-93 increased H5V proliferation (Figure 8A), and infection with miR-25 increased cell migration towards VEGF (Figure 8B). Cell viability after H2O2 treatment was increased after transduction with each of the three plasmids (Figure 8C). Also, miR-93 and miR-25 reduced apoptosis of 293HEK cells upon transient transfection (Figure 8E). Interestingly, the expression of miR-106b∼25 in H5V endothelial cells was up-regulated after 24 h of 1% O2 hypoxic treatment, but not after 6 h (Figure 8J). However, overexpression of miR-93, miR-25, or miR-106b had no effect on the Matrigel tube formation capacity of H5V cells in vitro (Figure 8F and I).

Figure 8

Overexpression of individual members of the miR-106b∼25 cluster increases viability, proliferation, and migration of H5V endothelial cells. MiR-25, miR-106b, or miR-93 was overexpressed in H5V endothelial cells by retroviral infection. HTR was used as a control virus. (A) Proliferation rate of endothelial cells infected with miR-25, miR-106b, miR-93 or HTR viruses. Proliferation was determined by the XTT Cell Proliferation Assay Kit 48 h after infection. Results are expressed as fold of HTR. (B) H5V migration towards 10 ng/mL of VEGF. (C and D) Cell viability was assayed by XTT after treatment of the cells with 800 µM H2O2 or 1% hypoxia for 24 h. (E) Flow cytometry images of 293HEKcells transfected with miR-25, miR-106b, miR-93, or HTR. (F–I) Matrigel tube formation was assayed after plating H5V cells on Matrigel-coated dishes 48 h after viral infection of the cells with miR-25, miR-106b, miR-93, or HTR. (J) RNA was extracted from H5V endothelial cells after 6 or 24 h of hypoxia (1% O2). Relative expression was determined by qRT–PCR and normalized to the expression of U6. Relative expression is expressed as fold of H5V normoxia (*P < 0.05).

Deletion of the miR-106b∼25 cluster reduces aortic ring capillary sprouting

To further study the role of the miR-106b∼25 in endothelial cell function, we employed the aortic ring angiogenesis assay. The number of capillaries grown from aortas of miR-106b∼25 KO mice was reduced when compared with WT aortic rings. The same difference was observed when aortic rings were subjected to hypoxia (1% O2 for 10 h) (Figure 9A–E).

Figure 9

Decrease in capillary sprouting of miR-106b∼25 knockout aortic rings embedded in matrigel and cultured for 6 days, either under normoxia (A and B) or hypoxia (C and D) (1% O2 for 10 h). n = 10 per group. (E) The average number of capillaries per group.

Discussion

MicroRNAs have been shown to orchestrate many biological processes pertaining to the formation of new blood vessels, including proliferation, migration, and differentiation of endothelial cells. MiR-93 overexpression in U68 glioblastoma cells was previously shown to promote tumour angiogenesis and cell survival. Here, we report the novel finding that the miR-106b∼25 cluster is a key participant in the recovery of ischaemic muscles in mice, and describe the central role of this cluster in the regulation of angiogenic properties of BMSCs.

The entire 106b∼25 microRNA cluster is up-regulated in ischaemic skeletal muscle as evident from the results of real-time PCR and in situ hybridization. Such up-regulation strongly supports the contention that miR-106b∼25 is involved in the regeneration attempt, partly because it is apparent 7 and 21 days after the ischaemic injury but not in the initial stages of ischaemia. Up-regulation of the miR-106b∼25 miR cluster was also observed here in vitro, in murine endothelial cells subjected to hypoxia. Furthermore, the up-regulation of miR-93 was more robust than that of the other two mature miRNAs of the cluster, suggesting differential roles of the individual members of the 106b∼25 cluster and raising intriguing questions concerning the post-transcriptional regulation of mature miRNAs derived from a common precursor.

Using a miR-106b∼25 KO mouse model, we were able to show that post-ischaemic vascularization in these mice is impaired, as indicated by the decreased blood flow seen on LDPI and by reduced capillary density. Injection of naked DNA plasmid has been shown to be a feasible approach for gene transfer in skeletal muscle.22 By re-expression of miR-106b∼25 via local injection of a plasmid encoding this cluster, we were able to partially reverse the mutant phenotype. One possible reason for our failure to reverse it fully might be that miR-106b∼25 expression is important also in other cellular components outside the ischaemic limb. It is well established that circulating and bone marrow-derived progenitors6,7 home and migrate to the site of an ischaemic injury, where they contribute significantly to the regeneration process both by activating resident endothelial cells and by becoming incorporated into newly forming vessels. It is therefore not surprising that this process was not subject to influence by local gene delivery to the ischaemic muscle. Further support for this hypothesis came from our observation that local injection of the plasmid did not improve the blood flow in the WT mice, which already express high levels of the miR-106b∼25 cluster after ischaemia. Naked plasmid injection, however, is considered a relatively inefficient method of gene delivery, and therefore we cannot rule out the possibility that full reversal of the mutant phenotype could have been achieved by more efficient gene delivery methods.

We also isolated and cultured BMSCs to investigate the role of the miR-106b∼25 cluster in functional properties of relevance for the neovascularization process. We found that BMSCs devoid of the miR-106b∼25 cluster were more susceptible to H2O2-induced apoptosis than the WT cells. Also, the migratory response to hypoxia-induced cytokines such as VEGF and SDF1 is reportedly of great importance to the regeneration process mediated by peripheral progenitor cells.23 We found that the presence of miR-106b∼25 is important for cell migration towards VEGF. Our in vitro Matrigel assay showed that the tube formation capacity of BMSCs devoid of the miR-106b∼25 cluster was reduced when the cells were plated either alone or with murine endothelial cells. However, conditioned media obtained from BMSCs of either WT or KO mice did not differentially influence tube formation in the murine endothelial cells. This suggests that the miR-106b∼25 cluster is an intrinsically important contributor to the vasculogenic properties of bone marrow cells. Expression of this cluster might also play a role in the activation of endothelial cells by BMSCs, but this process appears to be contact dependent. However, when the BMSCs were subjected to hypoxic treatment, differential cytokine expression was detected in conditioned media from the WT and the KO BMSCs. These findings suggest that the Mir-106b∼25 KO BMSCs have reduced capacity to promote tissue repair by secretion of angiogenic cytokines under hypoxia. It was previously shown that the miR-106b∼25 cluster target the tumour suppressor PTEN,21 which is a negative regulator of the PI3K/AKT pathway and angiogenesis.24 Here, we show that transfection of the miR-106b∼25 cluster reduces PTEN protein levels and increases AKT phosphorylation. It was previously shown that PTEN deficiency24 or overexpression of AKT25 promotes secretion of pro-angiogenic cytokines of BMSCs. Therefore, it is tempting to speculate that the miR-106b∼25 cluster enhances the angiogenic response of BMSCs by targeting PTEN. Indeed, we confirmed the role of PTEN by showing reversal of the diminished tube formation and migratory capacity of miR-106b∼25 BMSCs upon PTEN inhibition.

Lastly, we examined whether the presence of miR-106b∼25 was important for the expression of various stem-cell markers. Relative to the WT, we found a marked reduction in Sca-1-positive and CD133-positive cells in our miR-106b∼25 KO cell cultures. This might be attributable to increased apoptosis that is restricted mainly to the Sca-1-positive subpopulation in the KO mouse cell cultures, rather than a differential proliferation rate, as the latter was found to be similar in the WT and the KO cells.

The numbers of Sca-1, C-kit, and Flk-1 positive cells were also assessed in vivo, with or without the induction of hindlimb ischaemia. In the absence of ischaemia, the numbers of Sca-1 positive cells were similar in the WT and the KO miR-106b∼25 mice. At 7 days after induction of ischaemia; however, the numbers of Sca-1-positive cells in the bone marrow of the KO mice were increased. This might be explained by the weaker regenerative capacity of ischaemic muscles in miR-106b∼25 KO mice than in the WT, as was evident from the reduced blood flow seen on LDPI 7 days after ischaemia induction. Thus, the bone marrow of these mice might be overstimulated because of the persistence of the ischaemic state.

The miR-106b∼25 cluster may also play an important role in endothelial cell function: all three mature microRNAs, and especially miR-93, conferred protection against H2O2-induced damage. Also, miR-93 increased the proliferation rate of endothelial cells and miR-25 increased the migratory response to VEGF. Moreover, capillary growth from miR-106b∼25 KO aortic rings was significantly diminished. Altogether, these results suggest that the reduced blood flow in 106b∼25 KO mice is a consequence of attenuated post-ischaemic neovascularization, which can be explained by impaired functioning of bone marrow-derived and endothelial cells.

In conclusion, the present study shows for the first time, the importance of the miR-106b∼25 cluster for post-ischaemic neovascularization in a murine model of hindlimb ischaemia. In addition to demonstrating the essential role of this cluster in the angiogenic response of BMSCs and endothelial cells, these data improve our understanding of the regeneration mechanisms in ischaemic processes, and consequently may have therapeutic implications.

Funding

This work was partially funded by the Dr Herman Schauder Memorial Fund of the Sackler Faculty of Medicine, Tel Aviv University, Israel.

Conflict of interest: This work was performed in partial fulfilment of the requirements for a PhD degree by Jonathan Semo, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel.

Acknowledgements

MDH1 constructs were a generous gift from Dr Anne Brunet (Stanford University, Stanford, CA, USA). The miR-106b∼25 KO mice were kindly provided by Dr Tyler Jacks (Massachusetts Institute of Technology, Cambridge, MA, USA). The pMSCV-miR plasmids were kindly provided by Dr Reuven Agami (The Netherlands Cancer Institute, Amsterdam).

References

View Abstract