at SciVerse ScienceDirect

Biomaterials 34 (2013) 112e120

Contents lists available

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

The effect of heme oxygenase-1 complexed with collagen on MSC performance inthe treatment of diabetic ischemic ulcer

Chunli Hou a, Lei Shen b, Qiang Huang c,d, Jianhong Mi a, Yangxiao Wu a, Mingcan Yang a, Wen Zeng a,Li Li a, Wen Chen a, Chuhong Zhu a,*

aDepartment of Anatomy, Key Lab of Biomechanics, Third Military Medical University, Chongqing 400038, ChinabDepartment of Anatomy, Qiqihar Medical College, Heilongjiang 161006, ChinacDepartment of Orthopaedics, Southwest Hospital, Third Military Medical University, Chongqing 400038, ChinadDepartment of Orthopaedics, Traditional Chinese Medicine Hospital, Shaping Ba District, Chongqing 400038, China

a r t i c l e i n f o

Article history:Received 30 July 2012Accepted 12 September 2012Available online 8 October 2012

Keywords:Diabetic ischemic ulcersBone marrow mesenchymal stem cellsHeme oxygenase-1CollagenWound healing

* Corresponding author.E-mail address: zhuch99@yahoo.com (C. Zhu).

0142-9612/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.biomaterials.2012.09.022

a b s t r a c t

Diabetic ischemic ulcer is an intractable diabetic complication. Bone marrow mesenchymal stem cells(BMSCs) have great potential in variety of tissue repair. In fact, poor cell viability and tolerance limit theirability for tissue repair. In addition, it is difficult for stem cells to home and locate to the lesion. In thisstudy, we explore whether over-expression of heme oxygenase-1 (HO-1) in BMSCs complexed withcollagen play an important role in treatment of diabetic ischemic ulcers. In vitro, over-expression of HO-1promoted the proliferation and paracrine activity of BMSCs and the conditioned medium of BMSCsaccelerated HUVECs migration and proliferation. These processes were closely related to Akt signalingpathway and were not dependent on Erk signaling pathway. In vivo, in order to make BMSCs directly acton the wound, we choose a solid collagen as a carrier, BMSCs were planted into it, ischemic wounds ofdiabetic mice were covered with the complex of BMSCs and collagen. The results indicate that thecomplex of HO-1-overexpressing BMSCs and collagen biomaterials can significantly promote angio-genesis and wound healing. These preclinical findings open new perspectives for the treatment of dia-betic foot ulcers.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Diabetic ischemic ulcer is an intractable diabetic complication.There has been no good treatment method to date, and amputationis often the end result, creating a significant economic pressure andburden for both patients and society. In recent years, stem celltransplantation has been considered as a new therapeutic strategyfor diabetic foot ulcers.

Bone marrow mesenchymal stem cells (BMSCs) have been usedin a variety of tissue engineering strategies because of their multi-differentiation potential, immunomodulatory and tissue repairproperties [1]. Wu and his colleagues [2] reported that theimplantation of BMSCs can significantly promotewound healing byincreasing the formation of epithelial cytoplasm and local angio-genesis in wounds. In addition, artificial skin including BMSCs orinjecting BMSCs around the wound can also accelerate woundhealing in diabetic rats [3,4]. Although BMSCs have great thera-peutic potential for damaged tissue, poor cell viability and

All rights reserved.

tolerance limit their ability for tissue repair. Studies have confirmedthat diabetes can damage BMSCs, decrease their proliferation andsecretion as well as their anti-apoptosis and differentiationpotential [5e8]. Exogenous BMSCs can survive for only a short timein subjects and ulcers relapse quickly, how to enhance stem cellviability in diabetic ischemic ulcers micro-environment is a keyproblem.

Oxidative stress induced by high glucose is believed as thecentral pathogenesis of the chronic complications of diabetes, andit is also an obstacle for stem cell therapy. The blockage of oxidativestress is very beneficial for the prevention and treatment of chronicdiabetic complications. It is well known that heme oxygenase-1(HO-1) is an important antagonist of oxidative stress. HO-1 playsan important role in regulating many pathophysiological processes,especially in protecting tissues and organs, and it has beena research focus in many fields, including research on organtransplantation, ischemia-reperfusion injury, and cardiovasculardiseases [9e11]. One study found that the expression of HO-1 wassignificantly decreased in diabetics in comparison with non-diabetics. Tang et al. [12] reported that HO-1 can prevent graftcell death in the ischemic heart. The up-regulation of HO-1 can

C. Hou et al. / Biomaterials 34 (2013) 112e120 113

reduce oxidation products and protect endothelial cells againstdamage and loss, thus reducing the vascular complications ofdiabetes.

At present, cell homing and positioning to the lesion is stilla difficult problem in stem cell therapy. In order to make BMSCsdirectly act on the wound, we choose a solid collagen as a carrier.BMSCs were planted into the solid collagen biomaterials, thewounds were coveredwith the complex of BMSCs and collagen.Wespeculate that the new complex plays an important role in thetreatment of diabetic ischemic ulcers. In further, we hope to explorea non-invasive, convenient and more effective method to treatischemic diabetic ulcers of the lower limb.

2. Materials and methods

2.1. Cell culture and cell group

We purchased human umbilical vein endothelial cells (HUVECs) and GFP-transfected human bone marrow mesenchymal stem cells (BMSCs) from CyagenBiosciences (Guangzhou, China). The cells were routinely cultured according to theinstructions of the supplier. Cells were used between passage 4 and passage 6 in thefollowing experiments. Cobalt protoporphyrin (Copp) is known as an inducer of HO-1, it can specifically promote HO-1 expression in a variety of cells. On the contrary,Sn-protoporphyrin (Snpp) is an inhibitor of HO-1, it can inhibit HO-1 expression inmany kinds of cells. In this experiment, we up-regulated or inhibited HO-1expression by stimulating BMSCs with Copp (Santa, 20 mM) or Snpp (Sigma,10 mM). The cells were divided into five groups; one group (control group) wascultured in a high-glucose medium (30 mM glucose), and the other four groups werecultured in a high-glucose medium with different stimulating factors, includingCopp, or Snpp, or Copp and Akt inhibitor (Copp þ Akt�), or Copp and Erk inhibitor(Copp þ Erk�). These media were used for the treatments described below.

2.2. Expression of HO-1 in BMSCs

In the experiment, we extracted cytoplasm protein by protein extraction kit(Beyotime Institute of Biotechnology, China) and determined the role of Copp orSnpp on HO-1 expression in BMSCs through the ELISA assay (R&D Systems)according to the manufacturer’s instructions.

2.3. The effect of HO-1 on the proliferation and paracrine of BMSCs

2.3.1. MTT assayBMSCs were harvested with 0.25% trypsin-EDTA (Hyclone Laboratories) when

the closure rate was about 80% and were transferred to a 96-well plate at 105

cells/well. The BMSCs were cultured in different medium (Copp, or Snpp, orCopp þ Akt�, or Copp þ Erk�) for 3 days. MTT (3-[4, 5-dimethylthiazol-2-yl] 2, 5-diphenyltetrazolium bromide) was dissolved in PBS at 5 mg/ml and filter steril-ized. An aliquot (20 ml) of the MTT solutionwas added to each well incubation of theBMSCs continued at 37 ℃, 5% CO2 for 4 h. The medium was removed, and 150 ml ofDMSO was added to each well. After 15 min of orbital shaking, the plate was placedin an incubator for 5 min to eliminate air bubbles. The absorbance at awavelength of492 nm was measured using a plate reader.

2.3.2. ELISA assayThe conditioned medium was collected after BMSCs were cultured for 3 days.

Vascular endothelial growth factor (VEGF) in the conditioned culture supernatantswere determined using ELISA assay (R&D Systems) according to the manufacturer’sinstructions.

2.4. The function of the conditioned medium of BMSCs to proliferation and migrationof HUVECs

2.4.1. MTT assayBMSCs were harvested with 0.25% trypsin-EDTA when the confluence rate was

approximately 80% and were transferred to a 96-well plate at 105 cells/well. Themedium consisted of RPMI 1640 with the different treatments (conditionedmediumof BMSCs/1640 ¼ 1/5); the ensuing steps were the same as described above (2.3.1MTT assay).

2.4.2. Cell scratch experimentHUVECs were plated in 6-well plates at 5 � 105 cells/well. When cells covered

the bottom of the well, a scratch was made with a sterilized P1000 pipette tip (awidth of 0.4e0.5 mm), and the HUVECs continued to be cultured in the differentmedia (conditioned medium of BMSCs/1640 ¼ 1/5) after cleaning the stripped cellswith PBS. After 24 h, the HUVECs were fixed in 4% paraformaldehyde and were thenstained with 0.5% Crystal violet (Sigma). The scratches were analyzed with an

OLYMPUS BX50 microscope, and the scratch area was measured using IPP software(Media Cybernetics).

2.4.3. Transwell chamber experimentHUVECs were harvested with 0.25% trypsin-EDTA when the confluence rate

achieved 75%, and 104 cells were loaded into the upper chamber of a 24-welltranswell migration insert (pore size of 5 mm). The lower chamber containedbasal medium with the different conditioned media of the BMSCs (conditionedmedium of BMSCs/1640 ¼ 1/5). After 6 h, the cells on the upper side of themembrane were wiped away, and cells on the lower side of the membrane werefixed with 4% paraformaldehyde and stained with 0.5% crystal violet. The cellmorphology was immediately determined by light microscopy, and the number ofcells migrating into the lower chamber was determined.

2.5. Growth of BMSCs in biomaterials

The hybrid electrospinning nano-fiber mesh was made by polylactic acid 80%,silk protein 10% and collagen 10% using a previously described procedure [13].BMSCs were seeded in the biomaterials at 8 � 103 cells/material; cells were dividedinto three groups. The first group was added high concentration glucose (30 mM) ascontrol group; the second group was added high concentration glucose (30 mM) andCopp (20 mM), and the third groupwas added high concentration glucose (30 mM) andSnpp (10 mM). Three days after, BMSCs were fixed with 4% paraformaldehyde anddetermined by Leica TCS SP5 laser scanning confocal microscope.

2.6. Animal procedures

Six- to eight-week-old male C57BL/6J mice were streptozotocin-induced withdiabetes (Sigma, i.p., 40 mg/kg). Four weeks after the diabetes induction, the micewere depilated on their hindlimb area and underwent bilateral hindlimb ischemiaby ligature and electro-coagulation under anesthesia. Full-thickness wounds weresimultaneously created in the dorsal thigh skin of both legs using a sterile 5-mm-wide biopsy punch. To evaluated the healing potential of HO-1-overexpressingBMSCs in the model of diabetic ischemic ulcer. The wounds were covered withcollagen alone or collagen containing 2�104 BMSCs or 2�104 HO-1-overexpressingBMSCs or 2 � 104 HO-1-lowexpressing BMSCs. For a detailed experimental proce-dure, please see references [14e17]. All the procedures complied with the standardsstated in the Guide for the Care and Use of Laboratory Animals (Third MilitaryMedical University) and were approved by the ethics committee of Third MilitaryMedical University.

2.7. The rate of wound closure and histology

Photographs were taken on day 0, day 7 and day 14, the wound area wasanalyzed using IPP software 6.0 (Media Cybernetics). On day 14 postoperative, theanimals were sacrifice using an overdose of anesthetic (0.3% pentobarbital, i.p.), andthe wounds and surrounding skin were removed. The tissues were fixed in 4%paraformaldehyde, and the survival and distribution of the BMSCs were analyzed byfluorescence microscopy (BMSCs with green fluorescence). The vascular endothelialcells were labeled by von Willebrand Factor (vWF, 1:200, over night; BOSTER Corp.,Wuhan, China). Capillary profiles were recognized by fluorescence microscope, thecapillary density was determined in a blind manner [14,18].

2.8. Statistical analysis

The experimental data were analyzed by SPSS 18.0, t-test and variance analyseswere performed and the results were expressed as the mean � SEM. Each experi-ment was repeated at least three times. All p values were two-tailed, and p value<0.05 was considered statistically significant.

3. Results

3.1. Expression of in HO-1 BMSCs

In an Enzyme-Linked Immunoassay (Fig.1), high concentrationglucose promoted HO-1 expression of BMSCs in the early stage, itindicates that HO-1 plays an important role in cellular oxidativestress reaction. Copp significantly improved HO-1 expression inBMSCs, and the difference was significant compared with highglucose group (#p < 0.01). On the contrary, Snpp significantlyinhibited HO-1 expression in BMSCs. So in the following experi-ments, we achieved the up-regulation or inhibition of HO-1expression in BMSCs throuth Copp or Snpp respectively.

Fig. 1. Copp and Snpp regulated HO-1 expression in BMSCs. HO-1 expression wasdetermined by ELISA when BMSCs were cultured for 72 h. *p < 0.01 (n ¼ 8) versus thecontrol group (BMSCs cultured in normal medium); #p < 0.01 (n ¼ 8) versusCopp þ high glucose group; $p < 0.01 (n ¼ 8) versus Snpp þ high glucose group.Valuesare mean � SEM.

C. Hou et al. / Biomaterials 34 (2013) 112e120114

3.2. The function of HO-1 to proliferation and paracrine activity ofBMSCs

To explore the effect of HO-1 on BMSCs growth in high-glucosemedium, we used MTT assay to examine BMSCs proliferation.When we added Copp to the high-glucose medium, the prolifera-tion rate of BMSCs was significantly higher than the control group(*p < 0.01), the OD value (492 nm) of Copp group was 132% of thatin control group, which suggests that HO-1 can promote BMSCsproliferation. On the contrary, the proliferation rate of BMSCs wassignificantly lower than the control group ($p < 0.01) when theexpression of HO-1 was inhibited by Snpp, the OD value was 63% ofthat in control group. When we added Copp and Akt inhibitortogether to the high-glucose medium, the proliferation rate ofBMSCs significantly decreased versus Copp group (#p < 0.01), theOD value was as the same as control group. However, there waslittle difference between Copp þ Erk inhibitor group and Coppgroup. These results demonstrate that the over-expression of HO-1can promote BMSCs proliferation. In contrast, when the expressionof HO-1 is inhibited, the proliferation rate of BMSCs declines

Fig. 2. HO-1 promoted BMSCs proliferation and VEGF secretion. (a) MTT assay showing thethe control group; #p < 0.01 (n ¼ 18) versus the Copp þ Akt inhibitor group. $p < 0.01(n ¼conditioned medium was determined by ELISA. *p < 0.01 (n ¼ 18) versus the control groupcontrol group. Values are mean � SEM.

significantly. In addition, the process relies on the Akt signalingpathway and the Erk signaling pathway is not involved (Fig. 2a).

VEGF is a critical cytokines for angiogenesis. In this experi-ment, we used an ELISA to detect the VEGF level in the condi-tioned medium of BMSCs. We found that the amount of VEGF inCopp group was 152% compared with control group. On thecontrary, which were significantly lower than control group whenthe expression of HO-1 was inhibited by Snpp ($p < 0.01), theamount was 57% compared with control group. In addition, theamount of VEGF in Copp þ Akt inhibitor group was 50%compared with Copp group, the difference was significant(#p < 0.01). But there was no significant difference betweenCopp þ Erk inhibitor group and Copp group. These resultsdemonstrate that over-expression of HO-1 can promote BMSCs tosecrete VEGF. In contrast, when the expression of HO-1 isinhibited, the secretion of BMSCs declines significantly. In addi-tion, the process relies on the Akt signaling pathway and the Erksignaling pathway is not involved (Fig. 2b).

3.3. The role of conditioned medium of BMSCs on proliferation andmigration of HUVECs

The proliferation and migration of endothelial cells are veryimportant for vascular repair and regeneration. Through the aboveexperiments, we conclude that HO-1 can promote proliferation ofBMSCs and secretion VEGF. So we speculate that the conditionedmedium of BMSCs plays a role in the proliferation and migration ofendothelial cells. In order to confirm our speculation, the prolifer-ation of HUVECs was detected through MTT assay and scratchexperiments (Figs. 3aec). In addition, the migration of HUVECs wasverified through a transwell chamber experiment (Fig. 3d and e). InMTT assay, the OD value of Copp group was 165% of the controlvalue and was 175% of the value of Snpp group, it indicates that themedium of BMSCs þ Copp group can improve HUVECs prolifera-tion. In addition, the OD value of Copp þ Akt inhibitor group was64% of the value of Copp group, but there was no significantdifference between Copp þ Erk� group and Copp group (Fig. 3a).The result of cell scratch experiment was consistent withMTTassayresult (Fig. 3b and c). In transwell chamber experiment, theconditioned medium of Copp group promoted significantlyHUVECs migration, and the migration rate was 140% comparedwith control group. The migration rate of Copp þ Akt inhibitorgroup descended sharply compared with Copp group (#p < 0.01),but there was no significant difference between Copp þ Erkinhibitor group and Copp group. These observations suggest that

OD value (492 nm) of BMSCs cultured for 72 h in each group. *p < 0.01 (n ¼ 18) versus18) versus the control group. Values are mean � SEM. (b) The VEGF concentration in; #p < 0.01 (n ¼ 18) versus Copp þ Akt inhibitor group. $p < 0.01 (n ¼ 18) versus the

Fig. 3. The conditioned medium of BMSCs induced HUVECs proliferation and migration. (a) MTT assay showing the OD value (492 nm) of HUVECs cultured for 72 h in each group.*p < 0.01 (n ¼ 18) versus the control group; #p < 0.01 (n ¼ 18) versus the Copp þ Akt inhibitor group. Values are mean � SEM. (b) The scratch closure rate in HUVECs cultured for24 h in each group. *p < 0.01 (n ¼ 18) versus the control group; #p < 0.01 (n ¼ 18) versus the Copp þ Akt inhibitor group. Valuesare mean � SEM. (c) Photographs of scratches inHUVECs cultured for 24 h in each group, scale bar ¼ 250 um. (d) HUVECs were placed in the upper chamber of a 24-well transwell migration insert. The medium in the lowerchamber contained BMSC-conditioned medium. After 24 h, HUVECs migrated from the upper chamber to the lower chamber in each group, scale bar ¼ 100 um. (e) The effect ofBMSC-conditioned medium on HUVEC migration. *p < 0.01 (n ¼ 10) versus control group; #p < 0.01 (n ¼ 10) versus the Copp þ Akt inhibitor group. Valuesare mean � SEM.

C. Hou et al. / Biomaterials 34 (2013) 112e120 115

the conditioned medium of HO-1-overexpressing BMSCs canpromote the proliferation and migration of HUVECs, and theseprocesses were dependent on the Akt signaling pathway, whereasthe Erk signaling pathway was not involved.

3.4. The effect of HO-1 on the proliferation of BMSCs in biomaterials

To make the cells remain in the ulcer surface, we need a suitablecarrier dressing. In this experiment, we chose a solid collagen. We

C. Hou et al. / Biomaterials 34 (2013) 112e120116

observed that the number of BMSCs activated by HO-1 significantlyincreased as described above (Fig. 4); it indicates that the solidcollagen is suitable for BMSCs growth and HO-1 can promoteBMSCs to grow better on collagen biomaterials.

3.5. The contribution of the new complex of stem cell andbiomaterials to wound healing

To study the impact of the new complex of stem cells andcollagen on wound healing, we produced wounds and limbischemia in diabetic mice. The wound was covered with collagenalone or collagen containing 2 � 104 BMSCs or 2 � 104 HO-1-overexpressing BMSCs or 2 � 104 HO-1-lowexpressing BMSCs.We found that collagen with HO-1-overexpressing BMSCs was

Fig. 4. HO-1 promoted BMSCs proliferation in collagen biomaterials. (aec) BMSCs culturedglucose (a) or high glucose þ Copp (b) or high glucose þ Snpp (c), scale bar ¼ 50 mm (def)scale bar ¼ 50 mm. (g) The cell count of BMSCs cultured for 72 h hours in the biological matewith the Snpp group. The values are the mean � SEM.

more effective for wound healing (n ¼ 6 per group) than the othertreatments either on day 7 or day 14 (*p < 0.01, #p < 0.01). Thehealing rate of control group (collagen alone) was slowest speed onday 7, but there was little difference among the three groups(control, BMSCs and BMSCs þ Snpp) on day 14. It indicates that thecomplex of collagen and HO-1-overexpressing BMSCs play animportant role in treating intractable diabetes ischemic ulcers(Fig. 5).

3.6. Effect of the new complex of stem cell and biomaterials onblood vessel restoration and angiogenesis

Vascular repair and angiogenesis play an important role inwound healing process. In vivo, experiment results indicate that

for 72 h in collagen biomaterials and collagen biomaterials supplemented with highSynthesized maps composed of collagen biomaterials with the control, Copp and Snpp,rial. *p < 0.01 (n ¼ 12) compared with the control group; #p < 0.01 (n ¼ 12) compared

Fig. 5. Effect of the new complex of stem cells and collagen on wound healing. (a) The collagen biomaterials for tissue engineering. (b) A 5-mm-wide full-thickness wounds. (c) Thediabetic foot wounds were covered by collagen biomaterials. (d) The rate of wound healing on day 7 and day 14. Data are expressed as mean � SEM, n ¼ 6 per group. BMSCs þ Coppcan accelerate wound closure than any other group. *p< 0.01 BMSCs þ Copp group versus BMSCs group on day 7; #p< 0.01 BMSCs þ Copp group versus BMSCs group on day 14; (e)Typical photographs of wound healing for the above groups.

C. Hou et al. / Biomaterials 34 (2013) 112e120 117

BMSCs þ Copp group can significantly accelerate the wound heal-ing, so we speculate that HO-1-overexpressing BMSCs can promotewound vascular repair and regeneration. To determine the rela-tionship between exogenous BMSCs and the number of vascular inwound, wounds surrounding skin were analyzed on day 14 post-operative when animal sacrificed by an overdose of anesthetic.Exogenous BMSCs were traced by green fluorescence. Capillaryprofiles were recognized by immuno-histochemical staining usingrabbit von Willebrand Factor (vWF, 1:200, over night; BOSTERCorp., Wuhan, China) followed by Rhodamine-conjugated goatanti-rabbit IgG (1:100; BOSTER Corp., Wuhan, China). Results showthat there is some green fluorescence (survival BMSCs) inBMSCs þ Copp group, and part of those BMSCs have differentiatedinto endothelial cells identified by vWF (Fig. 6jel, shown by thearrows). which indicates that HO-1 can improve BMSCs survival inwound microenvironment, and BMSCs can differentiate intoendothelial cells. We can not see any transplanted BMSCs (greenfluorescence) in BMSCs group or BMSCs þ Snpp group on day 14(Fig. 6g and o). These results show that HO-1 can protect BMSCsfrom injury induced by high glucose, ischemia and hypoxia inwound microenvironment. HO-1-overexpressing BMSCs canaccelerate the repair and regeneration of the local wound capillarythrough differentiation and paracrine mechanisms.

In order to evaluate the function of BMSCs to vascular repaira-tion and angiogenesis, the capillary density was determined in

a blind manner [14,18]. Results showed that the capillary density inBMSCs þ Copp group was 33 n/mm2, the number was significantlyhigher than any other group. It indicates that the HO-1-overexpressing BMSCs can promote wound healing by improvingthe restoration of blood vessels and angiogenesis (Fig. 6q).

4. Discussion

Chronic wounds in diabetes are mainly due to the lack ofangiogenesis [19]. How to initiate and effectively promote angio-genesis is one efficacious way. Preliminary evidence supports thepotential of BMSCs for the healing of skin ulcers [2,20,21]. BMSCscan differentiate and release pro-angiogenic factors to accelerateangiogenesis and promote wound healing; thus, BMSCs havepowerful potential in the treatment of diabetic ischemic foot ulcers.As previously mentioned, the local micro-environment in diabeticischemic ulcers is detrimental to the survival of BMSCs, and exog-enous BMSCs can survive for only a short time [2]. Therefore, howto improve the tolerance and vitality of BMSCs is a key concern.

HO-1 is one of the most important endogenous antioxidantfactors. The over-expression of HO-1 can reduce sugar-mediatedcell growth inhibition and apoptosis of human micro-vascularendothelial cells. In addition, the over-expression of HO-1 canincrease BMSCs survival in ischemic and infracted myocardium[22,23]. Relevant studies have also shown that the expression and

Fig. 6. The distribution of BMSCs and capillary density in wound on day 14. (aed) Typical photographs for control group (collagen alone). (eeh) Typical photographs for BMSCsgroup (collagen containing BMSCs). We can not find any survival BMSCs (green fluorescence) from the photograph; (iel) Typical photographs for BMSCs þ Copp group (collagencontaining BMSCs þ Copp). We can find that there are many survival BMSCs in the wound (green fluorescence) and BMSCs merged with vascular wall (shown by the arrows), whichindicates that HO-1 can improve BMSCs survival in wound microenvironment, and BMSCs can differentiate into endothelial cells. (mep) Typical photographs for BMSCs þ Snppgroup (collagen containing BMSCs þ Snpp). We can not find any survival BMSCs (green fluorescence) in the wound. Scale bar ¼ 50 um (a, e, i, m); Scale bar ¼ 25 um (bed, feh, gel,nep). (q) Capillary density in wounds was analyzed in a blind manner. *p < 0.01 BMSCs group versus control group; #p < 0.05 BMSCs þ Copp group versus BMSCs group; $p < 0.01BMSCs þ Copp group versus BMSCs þ Snpp group. Data are expressed as mean � SEM, n ¼ 12 per group.

C. Hou et al. / Biomaterials 34 (2013) 112e120118

activity levels of HO-1 in STZ-induced diabetic rats are significantlydecreased compared with non-diabetic control rats [24]. In vitro,our experiment results show that high concentration glucose canpromote HO-1 expression in BMSCs, it indicates that HO-1 as animportant antioxidant factor can significantly increase in the earlystage when cells were stimulated, but the increase can not resistthe damage caused by high glucose. On the basis of the abovestudies, we achieve over-expression of HO-1 in BMSCs throughCopp. In addition, collagen is a kind of safe and absorbablebiomaterials, so we choose a solid collagen as a carrier of stem cells.

In vivo, we covered the wound with collagen containing differentgroups of BMSCs. Results show that the complex of collagen andHO-1-overexpressing BMSCs is the most effective for woundclosure. On day 14 after cell transplantation, fluorescent tracerresults show that there are some survival BMSCs around thewoundonly in BMSCs þ Copp group, and part of them have differentiatedinto vascular endothelial cells. Furthermore, capillary density assayresults show that HO-1-overexpressing BMSCs can promote capil-lary repair and regeneration, accelerate wound healing. Accordingto some reports, BMSCs can affect other nearby cells though the

C. Hou et al. / Biomaterials 34 (2013) 112e120 119

release of a large number of soluble growth factors and cytokines,and this process is affected by the growth conditions and thevitality of the BMSCs [25e27]. In vitro, our results also confirm thatHO-1-overexpressing BMSCs have strong proliferation and para-crine ability than other treatments. These findings indicate that theover-expression of HO-1 can accelerate angiogenesis throughpromoting BMSCs proliferation and differentiation. Although HO-1-overexpressing BMSCs play an important role in promotingwound healing, further work should be performed to clarifywhether the cells can maintain their efficacy over longer experi-mental periods.

VEGF is the most important pro-angiogenesis factor: it hasa strong ability to promote angiogenesis, and it can also stimulatecell proliferation, delay senility, inhibit apoptosis and promote cellsurvival [28e30]. Our experimental results indicate that the over-expression of HO-1 can significantly promote BMSCs to secreteVEGF and the conditioned medium of HO-1-overexpressing BMSCscan significantly promote migration and proliferation of HUVECs,so we believe that HO-1-overexpressing BMSCs can promotewound healing through not only BMSCs themselves proliferationand differentiation but also promoting surrounding endothelialcells migration and proliferation, then accelerate angiogenesis andrestore blood supply in the wound. In addition, previous evidenceindicates that hypoxia is vital to the differentiation of MSCs intoendothelial cells [31]. Under hypoxic conditions, MSCs can secretemore VEGF by autocrine process, which induces MSC differentia-tion. In this experiment, themicroenvironment of diabetic ischemiculcers is hypoxic and ischemic, and the viability of HO-1-overexpressing BMSCs is very good. Therefore, we can concludethat the over-expression of HO-1 can protect BMSCs from injuryinduced by diabetic ischemic wound microenvironment, HO-1-overexpressing BMSCs can promote angiogenesis and woundhealing through the following two approaches: one is to increaseproliferation and differentiation of BMSCs through autocrineactivity, the other is to recruit more endothelial cells to the woundthrough paracine activity.

Akt is a cytosolic protein kinase (or alternatively protein kinaseB, PKB). It is closely related to VEGF-mediated angiogenesis and cellprotective effects [32]. Akt signaling pathway plays an importantrole in the survival and apoptosis of cells. Many growth factors andsurvival factors can activate Akt signaling pathway. In addition,extracellular signal-regulated kinase (Erk) is amember of theMAPKfamily. It is the core of the signal network involved in regulating cellgrowth, development and division. As these two signaling path-ways are very important for angiogenesis and life activity of cells,we wondered whether they were involved in the mechanism ofHO-1 action. Our results show that the HO-1 can promote prolif-eration and paracrine of BMSCs, the process is mainly dependent onthe Akt signaling pathway and not dependent on the Erk signalingpathway.

5. Conclusion

The over-expression of HO-1 can enhance stem cell viability andparacrine activity in diabetic ischemic ulcers micro-environment,the new complex of HO-1-overexpressing BMSCs and collagenbiomaterials is effective in stimulating wound cicatrization, and theprocess is associated with Akt signaling pathway and not depen-dent on the Erk signaling pathway. These preclinical findings opennew perspectives for the treatment of diabetic foot ulcers.

Acknowledgments

This work was supported by National Science Foundation ofChina (No:31100707), the Joint Research Fund for Overseas Chinese

Young Scholars (No:31028008), Natural Science Foundation Projectof CQ CSTC (cstc2011jjjq0016, 2010BB5176) and The MilitaryTwelfth Five-year Key Project (BWS11C056).

References

[1] Sasaki M, Abe R, Fujita Y, Ando S, Inokuma D, Shimizu H. Mesenchymal stemcells are recruited into wounded skin and contribute to wound repair bytransdifferentiation into multiple skin cell type. J Immunol 2008;180:2581e7.

[2] Wu Y, Chen L, Scott PG, Tredget EE. Mesenchymal stem cells enhance woundhealing through differentiation and angiogenesis. Stem Cells 2007;25:2648e59.

[3] Tark KC, Hong JW, Kim YS, Hahn SB, Lee WJ, Lew DH. Effects of human cordblood mesenchymal stem cells on cutaneous wound healing in leprdb mice.Ann Plas Surg 2010;65:565e72.

[4] Inoue H, Murakami T, Ajiki T, Hara M, Hoshino Y, Kobayashi E. Bioimagingassessment and effect of skin wound healing using bone-marrow-derivedmesenchymal stromal cells with the artificial dermis in diabetic rats.J Biomed Opt 2008;13:064036.

[5] Jin P, Zhang X, Wu Y, Li L, Yin Q, Zheng L, et al. Streptozotocin-induced diabeticrat-derived bone marrow mesenchymal stem cells have impaired abilities inproliferation, paracrine, antiapoptosis, and myogenic differentiation. Trans-plant Proc 2010;42:2745e52.

[6] Cramer C, Freisinger E, Jones RK, Slakey DP, Dupin CL, Newsome ER, et al.Persistent high glucose concentrations alter the regenerative potential ofmesenchymal stem cells. Stem Cells Dev 2010;19:1875e84.

[7] Stolzing A, Sellers D, Llewelyn O, Scutt A. Diabetes induced changes in ratmesenchymal stem cells. Cells Tissues Organs 2010;191:453e65.

[8] Khan M, Akhtar S, Mohsin S, Khan SN, Riazuddin S. Growth factor pre-conditioning increases the function of diabetes-impaired mesenchymal stemcells. Stem Cells Dev 2011;20:67e75.

[9] Tsubokawa T, Yagi K, Nakanishi C, Zuka M, Nohara A, Ino H, et al. Impact ofanti-apoptotic and anti-oxidative effects of bone marrow mesenchymal stemcells with transient overexpression of heme oxygenase-1 on myocardialischemia. Am J Physiol-Heart C 2010;298:H1320e9.

[10] Zeng B, Ren XF, Lin GS, Zhu CG, Chen HL, Yin JC, et al. Paracrine action of HO-1-modified mesenchymal stem cells mediates cardiac protection and functionalimprovement. Cell Biol Int 2008;32:1256e64.

[11] Li M, Kim DH, Tsenovoy PL, Peterson SJ, Rezzani R, Rodella LF, et al. Treatmentof obese diabetic mice with a heme oxygenase inducer reduces visceral andsubcutaneous adiposity, increases adiponectin levels, and improves insulinsensitivity and glucose tolerance. Diabetes 2008;57:1526e35.

[12] TangYL, TangY, ZhangYC, Qian KP, Shen LP, PhillipsMI.Mesenchymal stem cellsprotected with a hypoxia-regulated heme oxygenase-1 vector: a potentialstrategy to prevent graft cell death in the ischemic heart (abstract). Circulation2004;110:3e3.

[13] Zeng W, Yuan W, Li L, Mi JH, Xu SC, Wen C, et al. The promotion of endothelialprogenitor cells recruitment by nerve growth factors in tissue-engineeredblood vessels. Biomaterials 2010;31:1636e45.

[14] Barcelos LS, Duplaa C, Krankel N, Graiani G, Invernici G, Katare R, et al. HumanCD133þ progenitor cells promote the healing of diabetic ischemic ulcers byparacrine stimulation of angiogenesis and activation of Wnt signaling. Circ Res2009;104:1095e102.

[15] Emanueli C, Graiani G, Salis MB, Gadau S, Desortes E, Madeddu P. Prophylacticgene therapy with human tissue kallikrein ameliorates limb ischemiarecovery in type 1 diabetic mice. Diabetes 2004;53:1096e103.

[16] Himeno T, Kamiya H, Naruse K, Harada N, Ozaki N, Seino Y, et al. Beneficialeffects of exendin-4 on experimental polyneuropathy in diabetic mice. Dia-betes 2011;60:2397e406.

[17] Costanza E, Maria BS, Alessandra P, Tiziana S, Anna FM, Alessandra S, et al.Prevention of diabetics-induced microangiopathy by human tissue kallikreingene transfer. Circulation 2002;106:993e9.

[18] Madeddu P, Emanueli C, Pelosi E, Salis MB, Cerio AM, Bonanno G, et al.Transplantation of low dose CD34þKDRþ cells promotes vascular andmuscular regeneration in ischemic limbs. Faseb J 2004;18:1737e9.

[19] Marrotte EJ, Chen DD, Hakim JS, Chen AF. Manganese superoxide dismutaseexpression in endothelial progenitor cells accelerates wound healing in dia-betic mice. J Clin Invest 2010;120:4207e19.

[20] Falanga V, Iwamoto S, Chartier M, Yufit T, Butmarc J, Kouttab N, et al. Autol-ogous bone marrow-derived cultured mesenchymal stem cells delivered ina fibrin spray accelerate healing in murine and human cutaneous wounds.Tissue Eng 2007;13:1299e312.

[21] Urban VS, Kiss J, Kovacs J, Gocza E, Vas V, Monostori E, et al. Mesenchymalstem cells cooperate with bone marrow cells in therapy of diabetes. Stem Cells2008;26:244e53.

[22] Zeng B, Lin GS, Ren XF, Zhang Y, Chen HL. Over-expression of HO-1 onmesenchymal stem cells promotes angiogenesis and improves myocardialfunction in infarcted myocardium. J Biomed Sci 2010;17:80e7.

[23] Tang YL, Tang Y, Zhang YC, Qian KP, Shen LP, Phillips MI. Improved graftmesenchymal stem cell survival in ischemic heart with a hypoxia-regulatedheme oxygenase-1 vector. J Am Coll Cardiol 2005;46:1339e50.

[24] Hu CM, Lin HH, Chiang MT, Chang PF, Chau LY. Systemic expression of hemeoxygenase-1 ameliorates type 1 diabetes in NOD mice. Diabetes 2007;56:1240e7.

C. Hou et al. / Biomaterials 34 (2013) 112e120120

[25] Kinnaird T, Stabile E, Burnett MS, Lee CW, Barr S, Fuchs S, et al. Marrow-derived stromal cells express genes encoding a broad spectrum of arteriogeniccytokines and promote in vitro and in vivo arteriogenesis through paracrinemechanisms. Circ Res 2004;94:678e85.

[26] Steingen C, Brenig F, Baumgartner L, Schmidt J, Schmidt A, Bloch W. Charac-terization of key mechanisms in transmigration and invasion of mesenchymalstem cells. J Mol Cell Cardiol 2008;44:1072e84.

[27] Xu MF, Uemura R, Dai Y, Wang YG, Pasha Z, Ashraf M. In vitro and in vivoeffects of bone marrow stem cells on cardiac structure and function. J Mol CellCardiol 2007;42:441e8.

[28] Roskoski R. VEGF receptor protein-tyrosine kinases: structure and regulation.Biochem Bioph Res Co 2008;375:287e91.

[29] Wang P, Xie ZH, Guo YJ, Zhao CP, Jiang H, Song Y, et al. VEGF-inducedangiogenesis ameliorates the memory impairment in APP transgenic mousemodel of Alzheimer’s disease. Biochem Bioph Res Co 2011;411:620e6.

[30] Yan D,Wang X, Li D, LiuW, Li M, Qu Z, et al. Macrophages overexpressing VEGFtarget to infarcted myocardium and improve neovascularization and cardiacfunction. Int J Cardiol 2011. http://dx.doi.org/10.1016/j.ijcard.07.026.

[31] Li QA, Yao D, Ma JH, Zhu JA, Xu XH, Ren YL, et al. Transplantation of MSCs incombination with netrin-1 improves neoangiogenesis in a rat model of hindlimb ischemia. J Surg Res 2011;166:162e9.

[32] Ackah E, Yu J, Zoellner S, Iwakiri Y, Skurk C, Shibata R, et al. Akt1/proteinkinase B alpha is critical for ischemic and VEGF-mediated angiogenesis. J ClinInvest 2005;115:2119e27.