Mesenchymal stem cell-based gene therapy with prostacyclin synthase enhanced neovascularization in hindlimb ischemia

Atherosclerosis 206 (2009) 109–118

Contents lists available at ScienceDirect

Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis

Mesenchymal stem cell-based gene therapy with prostacyclin synthase enhanced neovascularization in hindlimb ischemia Masakazu Ishii a,1 , Yasushi Numaguchi a,b,∗,1 , Kenji Okumura b,c , Ryuji Kubota b , Xiuyang Ma a , Ryuichiro Murakami b , Keiji Naruse d , Toyoaki Murohara b a

Department of Medical Science of Proteases, Nagoya University School of Medicine, Nagoya, Japan Department of Cardiology, Nagoya University Graduate School of Medicine, Nagoya, Japan Department of Cardiovascular Medicine, Nagoya University School of Medicine, 65 Tsurumai, Showa, Nagoya 466-8550, Japan d Department of Cardiovascular Physiology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, 2-5-1 Shikada, Okayama 700-8558, Japan b c

a r t i c l e

i n f o

Article history: Received 27 September 2008 Received in revised form 21 February 2009 Accepted 21 February 2009 Available online 11 March 2009 Keywords: Angiogenesis Cell therapy Hindlimb ischemia Mesenchymal stem cell Prostacyclin

a b s t r a c t Objective: Bone marrow cell therapy contributes to collateral formation through the secretion of angiogenic factors by progenitor cells and muscle cells per se, thereby presenting a novel option for patients with critical limb ischemia. However, some cases are refractory to this therapy due to graft failure. Therefore, we used genetic modification of mesenchymal stem cells (MSCs) to overexpress a vasoregulatory protein, prostacyclin (PGI2 ), to examine whether it could enhance engraftment and neovascularization in hindlimb ischemia. Methods and results: We engineered the overexpression of PGI2 synthase (PGIS) within MSCs, which resulted in higher expression levels of phosphorylated Akt and Bcl-2 than in control. Under hypoxic conditions, the overexpression of PGIS led to upregulated expression of cyclooxigenase-2 and peroxisome proliferator-activated receptor ␦, following a 40% increased rate of proliferation in MSCs. We then produced unilateral hindlimb ischemia in C57BL6/J mice, which were injected either with MSCs transfected with GFP, with MSCs overexpressing PGIS, or with vehicle. Laser Doppler analyses demonstrated that the administration of MSCs effectively recovered blood perfusion, and that the peak blood flow was reached within 7 days of surgery in mice with MSCs overexpressing PGIS, which was earlier than that in mice with MSCs transfected with GFP. This beneficial effect was correlated to enhanced collateral formation and muscle bundle proliferation. Conclusion: Sustained release of PGI2 enhanced the proangiogenic function of MSCs and subsequent muscle cell regrowth in the ischemic tissue suggesting potential therapeutic benefits of cell-based gene therapy for critical limb ischemia. © 2009 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Peripheral artery disease (PAD), mainly caused by atherosclerosis, is characterized by reduced blood flow through narrowed or blocked arteries. Patients in the earlier stages of PAD can lower their risk factors through lifestyle changes, such as smoking cessation, control of diabetes or hypertension, increasing their physical activity, and adopting a low cholesterol diet [1]. However, advanced cases require surgical options, such as angioplasty and bypass surgery, in combination with medication. Based on the positive

∗ Corresponding author at: Department of Medical Science of Proteases, Nagoya University School of Medicine, 65 Tsurumai, Showa, Nagoya 466-8550, Japan. Tel.: +81 52 744 2265; fax: +81 52 744 2265. E-mail address: [email protected] (Y. Numaguchi). 1 Both authors contributed equally to this work. 0021-9150/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2009.02.023

results of preclinical studies [2], intramuscular injection of bone marrow-derived progenitor cells, or cell therapy, has been performed worldwide and many clinical trials have demonstrated that this therapy is safe and efficacious for the “no option” patients with critical limb ischemia [3,4]. It is assumed that the mechanism by which cell therapy effectively improves limb ischemia is due to the secretion of angiogenic factors from the implanted progenitor cells and the activation of muscle cells [5,6]. Therefore, to obtain the maximum benefit of cell therapy, the survival and functioning of implanted progenitor cells at the homing sites are desirable. However, we have experienced some cases that are refractory to this therapy due to graft failure. Prostacyclin (PGI2 ), a clinically proven vasodilator, upregulates the intracellular production of cAMP by binding to its receptor, which is postulated to be responsible for the multiple effects of PGI2 , such as vasodilation, antiplatelet aggregation, and the inhibition of smooth muscle proliferation [7–9]. In addition to

110

M. Ishii et al. / Atherosclerosis 206 (2009) 109–118

these conventional functions of PGI2 , it has recently been reported that activation of the cyclooxigenases (COXs)/PGI2 /peroxisome proliferator-activated receptor ␦ (PPAR␦) pathway is an important mechanism underlying the proangiogenic function of progenitor cells [10,11]. Previous reports have demonstrated that gene therapy for PAD combined with hepatocyte growth factor (HGF) and PGI2 synthase (PGIS) synergistically improved blood perfusion after hindlimb ischemia compared with the administration of control mesenchymal stem cells (MSCs) [12]. MSCs, which are potentially used in cell therapy, are known to secrete many angiogenic growth factors including HGF, vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), and stromal cell-derived factor-1␣ (SDF-1␣) [5,13–15]. Thus, we set out to confirm the efficiency of a cell-based gene therapy that is a combination of cell therapy and PGIS gene overexpression within MSCs. In the present study, to assess the effects of PGIS overexpression on the fate of cells and the angiogenic functions of MSCs, we investigated whether the sustained release of PGI2 could promote cell growth and enhance proangiogenic function, and confirmed the efficacy of cell therapy using MSCs engineered to overexpress PGIS in a murine hindlimb ischemia model. 2. Methods 2.1. Construction of recombinant adenovirus vectors We constructed an adenoviral vector encoding green fluorescent protein (GFP) and PGIS (GenBank accession no. BC061814) under the control of tetracycline-regulatable bidirectional expression (AdGFP–PGIS), as well as adenoviral vectors expressing PGIS or GFP alone (Ad-PGIS and Ad-GFP, respectively) [16]. Briefly, GFP-tagged cDNA containing the full-length rat PGIS gene [17] was inserted into the adenoviral DNA construct provided by BD Biosciences Clontech (Palo Alto, CA, USA). Virus particles were purified, concentrated with a titering kit (BD Biosciences Clontech) and stored at −80 ◦ C before use in experiments [18]. 2.2. Cell culture for MSCs and HUVECs Human MSCs were purchased from Cambrex Bio Science Walkersville, Inc. (PT-2500, Walkersville, MD, USA) and cultured with Dulbecco’s modified eagle medium (DMEM) (BD Biosciences Clontech) containing 10% fetal bovine serum (FBS) and antibiotics (50 ␮g/mL gentamycin, 50 ng/mL amphotericin B) at 37 ◦ C, 5% CO2 and 95% air [19] or for the hypoxia challenge, cells were incubated at 37 ◦ C, 5% O2 , 5% CO2 and 90% N2 for the indicated periods. MSCs at passages 3–5 were used in the experiments. Flow cytometry experiments following cell surface antigens were checked and found to be positive for CD105, 166, 29, and 44 and negative for CD14, 34, and 45. Human umbilical vein endothelial cells (HUVECs) were a kind gift from Dr. Keiji Naruse, Okayama University. The cells were cultured in an endothelial growth medium, HuMedia-EG2 (Kurabo, Osaka, Japan), supplemented with 2% FBS, 10 ng/mL human epithelial growth factor, 5 ng/mL human fibroblast growth factor-B, 1 ␮g/mL hydrocortisone, 10 ␮g/mL heparin, 50 ␮g/mL gentamicin, and 50 ng/mL amphotericin B. 2.3. Adenovirus-mediated gene transfer to MSCs In vitro gene transfer to MSCs was performed by incubation with Ad-GFP or Ad-GFP–PGIS at a multiplicity of infection (MOI) of 100, with 1 ␮g/mL doxycycline. Cells were then cultured for 36–48 h [16,18]. The expressions of GFP and GFP–PGIS were induced by

culturing the cells in the same low-serum medium without doxycycline (see supplementary data). 2.4. Measurements of prostanoids and proangiogenic growth factors To assess the secretion of prostanoids such as PGI2 and thromboxane A2 (TxA2 ) and proangiogenic growth factors from MSCs, 1 × 106 MSCs were plated in serum-free medium on 6-well plates. After 24-h incubation, the medium was used for determination of 6-ket PGF1␣ (a stable metabolite of PGI2 ) and TxB2 (a stable metabolite of TxA2 ) using the respective enzyme immunoassay kits (Cayman Chemical, Ann Arbor, MI) and the levels of VEGF, bFGF, HGF and SDF-1␣ were measured using immunoassay kits (Quantikine, R&D Systems) [14]. 2.5. Cell proliferation, apoptosis, and migration assays Cell proliferation and apoptosis assays were performed as previously described [18] (see supplementary data). Migration assay was performed with transwell (Corning, Acton, MA, USA) 24-well tissue culture plates composed of a polycarbonate membrane with 8-␮m pores. The inner chamber membrane was coated with 50 ␮L of Matrigel solution (50 ␮g/mL, BD BioScience) at 4 ◦ C overnight to avoid polymerization, and then rinsed with DMEM. Cells were then seeded on the inner chamber of the transwell plate at a concentration of 1 × 105 cells/100 ␮L. The inner chamber was placed into the outer chamber filled with 600 ␮L of serum-free DMEM for MSCs or EG2 for HUVECs, which contained growth factors, or the conditioned medium (CM) described above, and incubated for 6 h at 37 ◦ C. Cells that migrated onto the outer surface of the membrane were fixed with cold methanol and 4% paraformaldehyde, and then stained by the May–Giemsa method. The numbers of migrated cells were counted in four to six randomly chosen fields of the duplicated chambers at a magnification of ×200 for each sample. 2.6. Immunoblotting analysis MSCs were cultured and exposed to the adenovirus in a 6-well plate. After 24 h of quiescence, the medium was replaced with 10% FBS/DMEM and cells were incubated for further 30 min. Cells were then lysed in Laemmli sample buffer (Sigma, St. Louis, MO), and the cell lysate was used for immunoblotting analysis, as previously described [16–18]. To assess the direct effect of PGI2 on Akt phosphorylation, MSCs were incubated with 1 or 10 ␮M of the PGI2 analogue (iloprost, Sigma) for 24 h. The following antibodies were used in this study: rabbit antiPPAR␦ and HIF-1␣ (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA), rabbit anti-COX-1 and -2, rabbit anti-PGIS (Cayman chemical), rabbit anti-Akt and anti-phospho Akt (Ser473), (Cell Signaling, Beverly, MA, USA), mouse anti-Bcl-2 and ␤-actin (Sigma) [16,18–20]. 2.7. Administration of MSCs or Ad-GFP–PGIS The investigation conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). C57Bl6/J mice weighing 25–28 g were obtained from SLC (Nagoya, Japan), and cared for and used in accordance with the guidelines of the National Institutes of Health. Murine MSCs were prepared by first flushing out the femurs with 5 mL of DMEM repeatedly (each n = 4), then collecting by centrifugation at 1500 × g for 5 min, and then by harvesting for 7–10 days in DMEM with 10% FBS. To distinguish MSCs from other cell types, such as fibroblasts, a proportion of the MSCs was checked for

M. Ishii et al. / Atherosclerosis 206 (2009) 109–118

several molecular markers and cell surface antigens, as previously described [19]. MSCs were transfected with Ad-GFP or Ad-GFP–PGIS at MOI 100. In the pilot study, the group that received a viral injection of Ad-GFP by 107 plaque forming units (PFU) as a control for Ad-GFP–PGIS injection was also evaluated. However, no significant differences or adverse effects were observed, including hemodynamic parameters, the degree of improvement of the perfusion recovery from ischemia, and edemas. Then, mice, were divided into the following 4 groups: (1) control group, which received PBS as a vehicle; (2) viral injection group, which received intramuscular injections of Ad-GFP–PGIS by 107 PFU; (3) MSCs overexpressing GFP group; and (4) MSCs overexpressing GFP–PGIS group (each n = 8). After ligature of the right femoral artery, 106 cells in 200 ␮L of PBS were injected into the ischemic muscles at 6–8 ischemic sites [21]. Systolic blood pressure and heart rate were measured before and 21 days after surgery by the tail-cuff method (BP98A, Softron, Tokyo,

111

Japan) [16–18]. On postoperative day 21, we evaluated the extent of digit necrosis and digit loss in the ischemic hindlimbs and classified the magnitude as follows: grade 0, no digit ulcers or necrosis; grade 1, ulcer in the tip of the digit; grade 2, necrosis extending to less than half of a digit; grade 3, necrosis extending to half of a digit; and grade 4, auto-amputation of the whole digit (n = 30 digits/group). The kinetics of PGIS protein expression levels and PGI2 production levels in the muscle tissues were measured at the indicated times by Western blotting and ELISA, as described above (n = 3 each). 2.8. Laser Doppler blood flowmetry We measured the ratio of ischemic (left)/normal (right) hindlimb blood flow using laser Doppler blood flowmetry (LDBF; moor LDI, Moor Instrument, Devon, UK), as described previ-

Fig. 1. Effects of PGIS overexpression on MSCs (A). Western blotting was repeated three times and representative images are shown. Significant PGIS expression was observed. Activation of Akt-Bcl-2 pathway (B). PGIS overexpression promoted phosphorylation of Akt and protein expression of Bcl-2 in MSCs. Phospo-Akt (Ser473): phosphorylation site at serine 473 of Akt. Effects of exogenous PGI2 on Akt phosphorylation (C). Iloprost significantly enhanced the phosphorylation levels of Akt only in the PGIS-transfected MSCs in a dose-dependent manner.

112

M. Ishii et al. / Atherosclerosis 206 (2009) 109–118

ously [21,22]. Perfusion blood flow was expressed as the ischemic (left)/normal (right) limb flow ratio (see supplementary data). 2.9. Histological assessment Twenty-one days after surgery, mice were killed and perfusion fixed in 4% paraformaldehyde. The excised adductor muscle tissues were snap-frozen in liquid N2 , sectioned at 4 ␮m, and mounted on glass slides [21]. To evaluate the homing of injected MSCs, GFP-positive cell counts per 1000 nuclei were calculated in MSC overexpressing GFP–PGIS group (n = 5). To evaluate the density of capillary, arteriole, and muscle bundles after cell therapy, five to six sections from each of the control and cell therapy groups were immunostained with rat monoclonal anti-mouse CD31 antibody (Pharmingen; 20 pg/mL) and rabit polyclonal antibody for ␣ smooth muscle actin (␣SMA, Abcam Inc., Cambridge, MA; 25 pg/mL). 2.10. Statistical analysis Data are presented as mean ± SEM values. Statistical analysis of multiple comparisons among the groups used 1-way ANOVA followed by the Bonferroni test. Perfusion recovery among groups was assessed by 2-way repeated measures ANOVA followed by the Bonferroni test. Multiple comparisons in nonparametric analysis were performed by the Kruskal–Wallis test. A probability value of P < 0.05 was considered statistically significant.

3. Results 3.1. PGIS overexpression activates Akt-Bcl-2 signaling within MSCs The transfection efficiency of the adenoviral vector encoding PGIS one day after Ad-GFP–PGIS transfection was 95.7%, quantified by the number of GFP-positive cells per attached cells on the dishes (n = 5). PGIS protein expression confirmed by Western blotting was 7.2-fold higher in Ad-GFP–PGIS-transfected MSCs than in Ad-GFP transfected MSCs (P < 0.001, n = 3 each) (Fig. 1A). We then evaluated the cell survival signaling of the AktBcl-2 pathway within MSCs after PGIS overexpression, which would confirm that the enhancement of this pathway directly leads to an increase in stem cell survival and tissue regeneration [15,23–25]. At basal conditions, PGIS overexpression in MSCs increased the phosphorylation of Akt by 3.2-fold (P < 0.001, n = 3 each) and subsequent Bcl-2 protein expression by 1.8-fold (P < 0.01, n = 3 each, Fig. 1B), compared with Ad-GFP-transfected MSCs. To assess the direct effect of PGI2 on Akt phosphorylation, MSCs were treated with a PGI2 analogue, iloprost. Iloprost significantly enhanced the phosphorylation levels of Akt only in PGIS-transfected MSCs in a dose-dependent manner, while iloprost treatment alone did not significantly elevate the phosphorylation levels of Akt at the observed concentrations (Fig. 1C).

Fig. 2. Effects of PGIS overexpression on the PGI2 /PPAR␦ pathway in MSCs under conditions of hypoxia and serum starvation. Under hypoxic conditions, PGIS overexpression induced COX-2 protein expression as well as Bcl-2 (n = 3) (A and B). Production of prostaglandins (C). Hypoxia-induced PGI2 production determined by a metabolite, 6-ketoPGF1␣, in MSCs. MSCs overexpressing PGIS are greater under conditions of both normoxia and hypoxia. No difference was observed in TxA2 production determined by a metabolite, TxB2, among the groups under both conditions. COX-2 and PPAR␦ were induced under serum starvation within 12 h in MSCs, while COX-1 was not induced (D). In MSCs overexpressing PGIS, the levels of COX-2 and PPAR␦ were higher than those in control at 12 h (both P < 0.05) (E). Data are presented as mean ± SEM.

M. Ishii et al. / Atherosclerosis 206 (2009) 109–118

3.2. PGIS overexpression upregulates COX-2 and enhances PGI2 production in MSCs To assess the effects of PGIS overexpression on the COXs/ PGI2 /PPAR␦ pathway in MSCs, we evaluated protein expression levels of COX-1, COX-2, and PPAR␦ under basal, hypoxic (5% O2 ), and starved (0.2% FBS/DMEM) conditions. At basal conditions, Bcl-2 protein expression was significantly higher in Ad-GFP–PGIStransfected MSCs than in Ad-GFP-transfected MSCs, while the levels of COX-2 and PPAR␦ were similar among groups (Fig. 2A). Under hypoxic conditions, COX-2 protein expression was significantly enhanced by 2.1-fold in Ad-GFP–PGIS-transfected MSCs and this was associated with enhanced expression of Bcl-2 protein (Fig. 2B). At basal conditions, Ad-GFP–PGIS-transfected MSCs consistently produced a 3.8-fold higher amount of PGI2 (determined by measuring 6-keto PGF1␣) compared with that from Ad-GFP-transfected MSCs (see supplement Fig. 2), and the level was increased by 5.7-fold under hypoxia (P < 0.05 vs. normoxia). In contrast, the pro-

113

duction of TXB2 was not altered among the groups before or after hypoxia. Under serum starved conditions over 12 h, protein expression of COX-2 was significantly enhanced in Ad-GFP–PGIS-transfected MSCs than in control or Ad-GFP-transfected MSCs by 3.2-fold (n = 3, P < 0.05), and then gradually decreased reaching basal levels over 48 h (Fig. 2C and D). The levels of COX-1 showed no changes during the observation period. In Ad-GFP–PGIS-transfected MSCs, the level of PPAR␦ was enhanced over 24 h, and at 12 h, the level was significantly higher than that of GFP and control by 2.1-fold (n = 3, P < 0.05). 3.3. Analyses of cell proliferation and apoptosis After gene transfer, PGIS overexpression enhanced cell proliferation in MSCs by as much as 1.4-fold, confirmed by a WST-1 assay 4 days after gene transfer (Fig. 3A). Similarly, PGIS overexpression enhanced cell proliferation in HUVECs as much as 1.7-fold (Fig. 3B).

Fig. 3. Effects of PGIS overexpression on cell proliferation and apoptosis. Proliferation of MSCs and HUVECs after PGIS or GFP overexpression was assessed by WST-1 assay. (A) MSCs overexpressing PGIS had higher levels of proliferation at 4 days after gene transfection than control and GFP groups (P < 0.05 and P < 0.01, respectively). NS: no significance. (B) Similarly, PGIS overexpression within HUVECs resulted in higher levels of proliferation than other groups (P < 0.05). Apoptosis was assessed by a singlestranded DNA detection kit. Although PGIS overexpression did not affect apoptosis rate under normoxia condition (C), the rates of apoptosis under conditions of hypoxia (D) and starvation (E) were significantly reduced (P < 0.05 vs. control and GFP groups). The values are shown as arbitrary units that depict signal intensity measured at OD450 nm.

114

M. Ishii et al. / Atherosclerosis 206 (2009) 109–118

M. Ishii et al. / Atherosclerosis 206 (2009) 109–118

Although PGIS overexpression did not affect apoptosis rate under normoxia condition, the rates of apoptosis under conditions of hypoxia and starvation were significantly reduced (Fig. 3C–E). 3.4. Effects of paracrine factors on migration In the migration assay, we found that paracrine factors secreted from Ad-GFP–PGIS-transfected MSCs enhanced the migration of MSCs, with a 1.5-fold increase observed in the migration of MSCs with the CM from Ad-GFP–PGIS gene-transfected cells compared with those with the CM from Ad-GFP gene-transfected cells (Fig. 4A and B, each n = 3, P < 0.01). Similarly, CM from the Ad-GFP–PGIS gene-transfected cells enhanced the migration of HUVECs compared with CM from Ad-GFP gene-transfected cells (Fig. 4C and D, each n = 3, P < 0.05). Next, we measured proangiogenic growth factors in the culture medium. VEGF, bFGF, HGF and SDF-1␣ were detectable, and no difference was observed in the concentration of each paracrine factor in the mediums of control and Ad-GFP gene-transfected MSCs. (Fig. 4E) However, in the medium of MSCs overexpressing PGIS, the levels of VEGF, bFGF, HGF and SDF-1␣ were significantly higher than those of control (P < 0.01 for VEGF and HGF and P < 0.05 for bFGF and SDF1␣). 3.5. Perfusion recovery after MSC-based PGIS gene therapy in ischemia model Regarding the homing of injected MSCs, the GFP-positive cell count per 1000 nuclei was 2.3 ± 0.3 in MSC overexpressing GFP–PGIS group (see supplementary data). No significant changes in hemodynamic parameters such as blood pressure and heart rate were observed among groups (see supplementary data). Regarding the kinetics of PGIS protein expression and 6-keto PGF1␣ secretion levels, PGIS protein overexpression and PGI2 secretion peaked at 3 days after injection and decreased over 21 days. At 21 days, the baseline levels were reached (see supplementary data). Viral injection with Ad-GFP–PGIS into ischemic muscle tissue did not improve perfusion recovery over 21 days and the time course was similar to that of control mice (Fig. 5A and B ). Notably, these mice had edema at injection sites and lower limbs (Fig. 5C). In groups of control, viral injection with Ad-GFP–PGIS, and injection of MSCs overexpressing GFP, necrosis occurred in foot digits due to hypoperfusion; however, this was not observed in the group treated with MSCs overexpressing PGIS. Perfusion recovery in the groups given MSC injections was better than in control and viral injection groups. The recovery of the MSC–PGIS group was accelerated as early as 7 days after cell therapy and the blood perfusion reached a similar level to that in healthy limbs. The perfusion recovery of the MSC–PGIS group was significantly better than that of the other groups as assessed by 2-way repeated measures ANOVA followed by the Bonferroni test (P < 0.05). The necrotic digit score in the MSC–PGIS group was significantly lower than that in the other groups, as assessed by the Kruskal–Wallis test (P < 0.01) (Fig. 5D). A large number of capillaries and arterioles were detected in the ischemic muscle of the MSC-transplanted groups (Fig. 5E). Quantitative analysis demonstrated that the capillary/muscle fiber

115

ratio was the highest in the MSC–PGIS group, followed by the MSC–GFP group, adenoviral injection group, and control group (2.2 ± 0.2, 1.8 ± 0.2, 1.2 ± 0.2, and 1.2 ± 0.2, respectively; P < 0.05 vs. MSC–GFP group and P < 0.01 vs. adenoviral injection group and control). Similarly, the arteriole density per section was the highest in the MSC–PGIS group (8.2 ± 0.3, 5.3 ± 0.2, 3.9 ± 0.2, and 3.6 ± 0.2, respectively; P < 0.05 vs. MSC–GFP group and P < 0.01 vs. adenoviral injection group and control). This was associated with an increase in muscle bundle density in the MSC–PGIS group, followed by the MSC–GFP group, control group, and adenoviral injection group (316 ± 27, 218 ± 25, 208 ± 24, and 196 ± 27 bundles per section, respectively; P < 0.01 vs. other groups). 4. Discussion In the present study, we examined the efficiency of a new strategy for therapeutic angiogenesis, MSC-based gene therapy coupled with PGIS, for critical limb ischemia. Overexpression of PGIS caused sustained release of PGI2 from MSCs and enhanced the proliferation, migration and survival of MSCs, with upregulation of the Akt-Bcl-2 axis and COX-2 under conditions of serum starvation and hypoxia. In a mouse hindlimb ischemia model, MSC-based gene therapy caused a significantly greater improvement than either MSC transplantation or adenoviral gene transfer. Numerous subsequent clinical trials have proved the safety and efficacy of therapeutic angiogenesis not only for PAD, but also for other cardiovascular diseases, such as myocardial infarction and cardiomyopathy [25]. However, accumulating data show that a certain percentage of patients do not obtain the benefit of therapeutic angiogenesis due to graft failure. To address this issue, we examined the administration of genetically modified cells in a murine ischemia model. Firstly, we adopted MSCs as a vehicle for cell therapy because MSCs are superior in transfection efficiency, durability under hypoxia and malnutrition, and in the secretion potential of growth factors such as VEGF, FGF, HGF, and SDF-1, compared with other candidate cells, such as mononuclear cells or endothelial progenitor cells (EPCs) [13–15,25]. Just as we had expected MSCs to be a reservoir of proangiogenic chemokines and growth factors, MSCs secreted VEGF, HGF, and SDF-1 by 2- to 8fold with PGIS overexpression when compared with control. These beneficial effects may lead to cell survival of MSCs via paracrine mechanisms. There are many candidate genes for cell-based gene therapy and further investigations are required to search for the optimal gene(s) [22–25]. In the present study, based on the results of our previous studies and those of other researchers, we examined PGIS overexpression within MSCs because of the multiple beneficial effects from the sustained release of PGI2 [7–9,16,17]. If we consider the mechanisms by which the administration of genetically modified cells enhances neovascularization and accelerates the reparative process of damaged muscle tissues, the targets for cellbased gene therapy and genes of interest can be categorized into 5 groups according to the steps involved in stem cell-based treatment of ischemic muscle tissues: (1) stem cell homing (e.g., SDF-1, monocyte chemotractant proteins, HGF, insulin growth factor-1 (IGF1), and CXCR4); (2) stem cell migration (e.g., integrins); (3) stem cell engraftment and survival (e.g., Akt, Bcl-2, and SDF-1);

Fig. 4. Effects of PGIS overexpression on migration in MSCs and HUVECs. The conditioned medium from MSCs enhanced MSC migration more than PDGF-BB-induced MSC migration (A). Representative images of migrated MSCs. +PDGF: MSCs stimulated with PDGF-BB; +CM GFP: MSCs stimulated with conditioned medium from MSCs overexpressing GFP; +CM PGIS: MSCs stimulated with conditioned medium from MSCs overexpressing PGIS. The number of migrated MSCs was largest in MSCs stimulated with conditioned medium from MSCs overexpressing PGIS, followed by the GFP group, PDGF group, and control (P < 0.001 vs. PDGF group and P < 0.01 vs. GFP group) (B). The conditioned medium from MSCs also enhanced HUVEC migration more than the EG2-induced HUVEC migration (C) Representative images of migrated HUVECs. +EG2: HUVECs stimulated with EG2. P < 0.05 vs. EG2 group and GFP group (D). Paracrine factors were detected in conditioned medium. (E) In the medium from MSCs overexpressing PGIS, the levels of VEGF, bFGF, HGF and SDF-1␣ were significantly higher than those of control († P < 0.01 for VEGF and HGF and * P < 0.05 for bFGF and SDF-1␣). The columns depict control, the medium from MSCs overexpressing GFP and from MSCs overexpressing PGIS, respectively.

116

M. Ishii et al. / Atherosclerosis 206 (2009) 109–118

Fig. 5. Perfusion recovery after cell-based gene therapy. (A) Representative images of LDBF 7 days after surgery. Blood perfusion of the ischemic hindlimb markedly increased in the MSCs overexpressing PGIS group. (B) Quantitative analysis of hindlimb blood perfusion. Ischemic (left)/normal (right) limb flow ratio (I/N ratio) was significantly higher 3 and 7 days after surgery in the MSCs overexpressing PGIS group. The increase in the I/N ratio was significantly higher than in the control and adenoviral injection group. * P < 0.01 vs. control; † P < 0.05 vs. control and GFP group; # P < 0.01 vs. control and adenoviral injection groups. MSC–PGIS: MSC overexpressing PGIS group; MSC–GFP: MSC overexpressing GFP group; Ad-PGIS: adenoviral injection group. (C) Representative images of hindlimbs. Note that the administration of an adenoviral vector encoding PGIS evoked edema at injection sites. A red circle depicts edema. Mice in groups other than the MSC–PGIS group had necrosis in foot digits. Yellow arrow heads indicate necrosis. (D) The necrotic digit score of the MSC–PGIS group was significantly lower than that of other groups as assessed by the Kruskal–Wallis test (P < 0.01, n = 30 each). (E) Representative images of CD31 and ␣SMA immunostaining in ischemic muscles. The number of capillaries and arterioles markedly increased in the MSC–GFP and MSC–PGIS groups, with higher levels in the MSC–PGIS groups as assessed by one-way ANOVA followed by the Bonferroni test (in each analysis P < 0.01 vs. control and adenoviral injection groups and P < 0.05 vs. MSC–GFP group, n = 5–6 sections each). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

M. Ishii et al. / Atherosclerosis 206 (2009) 109–118

117

Fig. 5. (Continued ).

(4) secretion of paracrine factors (e.g., interleukin-1␤, VEGF, FGFs, HGF, IGF-1, nitric oxide, and PGI2 ); and (5) stem cell fusion or differentiation [25]. Our strategy using sustained release of PGI2 from MSCs would have many advantages in the steps of migration, engraftment, survival and secretion of paracrine factors. PGIS overexpression in MSCs had additive effects, such as enhancement of survival signaling via the Akt-Bcl-2 axis and COX-1 or -2/PGI2 /PPAR␦ pathway compared with MSCs overexpressing GFP. Regarding the PGI2 –PPAR␦ axis, the activation of PPAR␦ by agonists enhances the proliferation of endothelial cells and angiogenesis via VEGF receptors [26]. Similarly, it has been demonstrated that the PGI2 –PPAR␦ axis regulates angiogenic function in EPCs [11]. These novel effects from the activation of PPAR␦ coupled with the conventional pathway through PGI2 receptors may affect endothelial cells

and EPCs around transplanted MSCs to induce proliferation and angiogenesis. Furthermore, our finding that overexpression of PGIS induced HGF secretion from MSCs may suggest that the autocrine mechanism of MSC also contributes considerably to the recovery of ischemia. Regarding combination therapy with HGF and PGIS, Morishita et al. have already confirmed the efficacy of co-transfection of naked DNA plasmids of HGF and PGIS in a hindlimb ischemia model [12]. In the present study, we attempted direct intramuscular injections of an adenoviral vector encoding PGIS. However, injections of the adenoviral vector caused edema at injection sites and had no beneficial effect in the reparative process in a hindlimb ischemia model. Further investigations are required for the clinical application of combined therapy with HGF and PGIS using viral vectors, in terms

118

M. Ishii et al. / Atherosclerosis 206 (2009) 109–118

of optimal delivery techniques and the vehicles for cell-based gene therapy. 5. Conclusion We have demonstrated that the overexpression of PGIS enhanced the proliferation, migration and survival of MSCs via the Akt-Bcl-2 axis, and MSC-based PGIS gene therapy caused significantly greater improvement in hindlimb ischemia than by the administration of MSC alone, or by adenoviral gene transfer with PGIS. This strategy may become a novel option for therapeutic angiogenesis in patients with PAD. Conflict of interest None. Acknowledgments This study was performed mainly at the Institute for Laboratory Animal Research, Nagoya University, Japan. Funding: This work was partially supported by a grant from the Ministry of Education, Science, and Culture of Japan (no. 19590811). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.atherosclerosis.2009.02.023. References [1] Norgren L, Hiatt WR, Dormandy JA, et al., TASC II Working Group. Intersociety consensus for the management of peripheral arterial disease. Int Angiol 2007;26(2):81–157. [2] Shintani S, Murohara T, Ikeda H, et al. Augmentation of postnatal neovascularization with autologous bone marrow transplantation. Circulation 2001;103(6):897–903. [3] Tateishi-Yuyama E, Matsubara H, Murohara T, et al., Therapeutic Angiogenesis using Cell Transplantation (TACT) Study Investigators. Therapeutic angiogenesis for patients with limb ischaemia by autologous transplantation of bone-marrow cells: a pilot study and a randomised controlled trial. Lancet 2002;360(9331):427–35. [4] Miyamoto K, Nishigami K, Nagaya N, et al. Unblinded pilot study of autologous transplantation of bone marrow mononuclear cells in patients with thromboangiitis obliterans. Circulation 2006;114(24):2679–84. [5] Kinnaird T, Stabile E, Burnett MS, et al. Marrow-derived stromal cells express genes encoding a broad spectrum of arteriogenic cytokines and promote in vitro and in vivo arteriogenesis through paracrine mechanisms. Circ Res 2004;94(5):678–85. [6] Tateno K, Minamino T, Toko H, et al. Critical roles of muscle-secreted angiogenic factors in therapeutic neovascularization. Circ Res 2006;98(9):1194–202.

[7] Moncada S, Vane JR. Pharmacology and endogenous roles of prostaglandin endoperoxides, thromboxane A2 and prostacyclin. Pharmacol Rev 1979;30:293–331. [8] Furchgott RF, Zawadzki JV. The obligatory role of endothelial cells in the relaxation of arterial smooth muscle by acetylcholine. Nature 1980;299:373– 6. [9] Lam JYT, Chesebro JH, Badimon L, Fuster V. Exogenous prostacyclin decreases vasoconstriction but not platelet thrombus deposition after arterial injury. J Am Coll Cardiol 1993;21:488–92. [10] Santhanam AV, Smith LA, He T, Nath KA, Katusic ZS. Endothelial progenitor cells stimulate cerebrovascular production of prostacyclin by paracrine activation of cyclooxygenase-2. Circ Res 2007;100(9):1243–5. [11] He T, Lu T, d’Uscio LV, et al. Angiogenic function of prostacyclin biosynthesis in human endothelial progenitor cells. Circ Res 2008;103(1):80–8. [12] Hiraoka K, Koike H, Yamamoto S, et al. Enhanced therapeutic angiogenesis by cotransfection of prostacyclin synthase gene or optimization of intramuscular injection of naked plasmid DNA. Circulation 2003;108(21):2689–96. [13] Kinnaird T, Stabile E, Burnett MS, et al. Local delivery of marrow-derived stromal cells augments collateral perfusion through paracrine mechanisms. Circulation 2004;109(12):1543–9. [14] Iwase T, Nagaya N, Fujii T, et al. Comparison of angiogenic potency between mesenchymal stem cells and mononuclear cells in a rat model of hindlimb ischemia. Cardiovasc Res 2005;66(3):543–51. [15] Gnecchi M, He H, Noiseux N, et al. Evidence supporting paracrine hypothesis for Akt-modified mesenchymal stem cell-mediated cardiac protection and functional improvement. FASEB J 2006;20(6):661–9. [16] Imai H, Numaguchi Y, Ishii M, et al. Prostacyclin synthase gene transfer inhibits neointimal formation by suppressing PPARdelta expression. Atherosclerosis 2007;195:323–32. [17] Numaguchi Y, Naruse K, Harada M, et al. Prostacyclin synthase gene transfer accelerates reendothelialization and inhibits neointimal formation in rat carotid arteries after balloon injury. Arterioscler Thromb Vasc Biol 1999;19:727–33. [18] Yokouchi K, Numaguchi Y, Kubota R, et al. l-Caldesmon regulates proliferation and migration of vascular smooth muscle cells and inhibits neointimal formation after angioplasty. Arterioscler Thromb Vasc Biol 2006;26(10):2231– 7. [19] Ishii M, Koike C, Igarashi A, et al. Molecular markers distinguish bone marrow mesenchymal stem cells from fibroblasts. Biochem Biophys Res Commun 2005;332(1):297–303. [20] Flusberg DA, Numaguchi Y, Ingber DE. Cooperative control of Akt phosphorylation, bcl-2 expression, and apoptosis by cytoskeletal microfilaments and microtubules in capillary endothelial cells. Mol Biol Cell 2001;12(10):3087– 94. [21] Inoue N, Kondo T, Kobayashi K, et al. Therapeutic angiogenesis using novel vascular endothelial growth factor-E/human placental growth factor chimera genes. Arterioscler Thromb Vasc Biol 2007;27(1):99–105. [22] Couffinhal T, Silver M, Zheng LP, et al. Mouse model of angiogenesis. Am J Pathol 1998;152:1667–79. [23] Mangi AA, Noiseux N, Kong D, et al. Mesenchymal stem cells modified with Akt prevent remodeling and restore performance of infarcted hearts. Nat Med 2003;9:1195–201. [24] Li W, Ma N, Ong LL, et al. Bcl-2 engineered MSCs inhibited apoptosis and improved heart function. Stem Cells 2007;25:2118–27. [25] Tongers J, Roncalli JG, Losordo DW. Therapeutic angiogenesis for critical limb ischemia: microvascular therapies coming of age. Circulation 2008;118(1):9–16. [26] Piqueras L, Reynolds AR, Hodivala-Dilke KM, et al. Activation of PPARbeta/delta induces endothelial cell proliferation and angiogenesis. Arterioscler Thromb Vasc Biol 2007;27(1):63–9.