Cxcr6-Based Mesenchymal Stem Cell Gene Therapy Potentiates Skin Regeneration in Murine Diabetic Wounds

Abstract

Mesenchymal stem cell (MSC) therapies for wound healing are often compromised due to low recruitment and engraftment of transplanted cells, as well as delayed differentiation into cell lineages for skin regeneration. An increased expression of chemokine ligand CXCL16 in wound bed and its cognate receptor, CXCR6, on murine bone-marrow-derived MSCs suggested a putative therapeutic relevance of exogenous MSC transplantation therapy. Induction of the CXCL16-CXCR6 axis led to activation of focal adhesion kinase (FAK), Src, and extracellular signal-regulated kinases 1/2 (ERK1/2)-mediated matrix metalloproteinases (MMP)-2 promoter regulation and expression, the migratory signaling pathways in MSC. CXCL16 induction also increased the transdifferentiation of MSCs into endothelial-like cells and keratinocytes. Intravenous transplantation of allogenic stable MSCs with Cxcr6 gene therapy potentiated skin tissue regeneration by increasing recruitment and engraftment as well as neovascularization and re-epithelialization at the wound site in excisional splinting wounds of type I and II diabetic mice. This study suggests that activation of the CXCL16-CXCR6 axis in bioengineered MSCs with Cxcr6 overexpression provides a promising therapeutic approach for the treatment of diabetic wounds.

Hyperglycemia in diabetes generates non-healing wounds impeding autologous MSCs recruitment and delayed differentiation at the wound site. Das et al. showed that intravenous transplantation of MSCs with CXCR6 gene therapy led to enhanced recruitment, engraftment, re-epithelialization, and neovascularization in regenerated skin at the wound site in diabetic mice, suggesting CXCL16-CXCR6 axis as a potent therapeutic target for treatment of non-healing diabetic wounds.

Introduction

Mesenchymal stem cells (MSCs) are multipotent adult stem cells with the capability to potentiate repair and regenerate injured tissue and modulate the immune response.¹ Recent preclinical studies have shown limited success of MSC transplantation on chronic wounds and other skin disorders.² The impact of MSC paracrine signaling and differentiation into cell lineages that participate in wound tissue regeneration is presently limited due to a very insignificant number of MSCs that actually migrate and engraft at the wound site. Thus, to maximize the therapeutic benefit of MSC transplantation, efficacious strategies are warranted to optimize MSC recruitment at the wound sites.

Chemokines are closely involved in the orchestration of wound healing.^3,4 They are the key mediators of inflammation, neo-vascularization, re-epithelialization, and repair processes during wound healing.5, 6, 7 MSCs express a wide repertoire of chemokine receptors that get activated in the presence of a chemotactic signal and modulate their physiological responses such as migration from their niche to the injury site.⁸ Endogenous bone-marrow-derived MSCs are known to migrate at the cutaneous wound sites.⁹ Thus, it was hypothesized that migration of exogenous bone-marrow-derived MSCs is regulated by the interaction between specific chemokine receptors and secreted chemokines at the site of injury that can act as potent regulators of MSC proliferation, migration, and differentiation.¹⁰ The present study was designed to elucidate the molecular signaling mechanism of the CXCL16-CXCR6 axis on MSC physiology, as well as during transplantation of MSCs onto normal and diabetic wound for skin regeneration. Cxcr6-based gene therapy in MSCs led to a significant increase in migration, higher engraftment, enhanced re-epithelialization, and neo-angiogenesis at the wound site resulting in skin tissue regeneration.

Results

Cognate Ligand-Receptor Expression at the Wound Site and MSCs

To identify predominant chemokines in the mice skin wound microenvironment, we analyzed temporal expression of an array of chemokines at the wound bed (post-wounding/post-surgery [PS] days 0–10), as well as chemokine receptor expression on MSCs using quantitative PCR. A marked increase in mRNA expression of chemokine ligand of CCL family—CCL7, CCL8, CCL24—was observed at the mice wound tissues with highest expression level at day 7 (Figure S1A) along with a significant increase in expression of CXCL2 (intermittently at days 1 and 7) and CXCL16 at day 0 through day 10 post-surgery (Figure S1B). Also, a significant high expression of their cognate receptors CCR3 and CCR10 (Figure S1C), as well as CXCR2 and CXCR6, was observed in our previously well-characterized mouse bone-marrow-derived bonafide MSCs¹¹ as compared with heterogeneous bone marrow stem cells (BMSCs)¹² (Figure S1D). Although earlier reports in literature focused the role of SDF-1/CXCR4 axis on cell migration, in our experimental set up we observed CXCL16 expression to be consistently increased at the wound site thereby suggesting the effect of CXCL16/CXCR6 axis on MSC physiology. Interestingly, recent literature also suggests that activation of CXCL16/CXCR6 axis induces migration, invasion, and recruitment of endometrial stromal cells during endometriosis,¹³ neutrophil to cerebrospinal fluid in pneumococcal meningitis,¹⁴ lymphocytes and monocytes into decidua during first trimester of human pregnancy,¹⁵ leukocytes in atherosclerosis,¹⁶ and rheumatoid arthritis.¹⁷

CXCL16 Induces MSC Migration via Activation of FAK Signaling but Not Proliferation

An increasing concentration of CXCL16 (1–100 ng/mL) did not alter bromodeoxyuridine (BrdU) incorporation in MSCs (Figure S2A) or induce colony-forming efficiency of MSCs (Figure S2B) indicating no mitogenic role of CXCL16 on MSCs. Interestingly, a concentration-dependent increase in MSC migration by CXCL16 (Figure S2C) was perturbed by CXCL16-neutralizing antibody (CXCL16-NAb; Figure S2D), suggesting a plausible role of CXCL16-CXCR6 axis in MSC migration. MSCs treated with CXCL16 (10 ng/mL) induced focal adhesion kinase (FAK), Src, and extracellular signal-regulated kinases (ERK1/2) phosphorylation that was abrogated by CXCL16-NAb (Figure 1A; quantification, Figure 1B), suggesting an activation of downstream effector genes involved in migration. Thus, it was hypothesized that CXCL16 upon binding to CXCR6 on MSCs activates FAK-Src-ERK1/2 that lead to translocation of ERK1/2 from cytosol to nucleus where it acts as a transcriptional co-activator of matrix metalloproteinases (MMPs), enzymes that degrade the extracellular matrix for cell migration (Figure 1C). Among the various classes of these enzymes, gelatinases (MMP2 and MMP9) and collagenases (MMP13) are reported to be involved in progenitor cell migration.18, 19, 20, 21 CXCL16-treatment to MSCs depicted an approximately 4-fold increase in MMP2 expression (Figure 1D) and activity (Figure 1E) that was mitigated in presence of CXCL16-NAb. These observations support our hypothesis in part of CXCL16-mediated MSC migration occurs via induction of FAK-Src-ERK1/2-MMP2 signaling.

Figure 1.

CXCL16 Induces MMP2-Mediated MSCs Migration

(A) CXCL16 treatment in MSCs led to phosphorylation of FAK, Src, and ERK1/2 that was perturbed by CXCL16-NAb. (B) Densitometric analysis of activated signaling mediators in MSCs treated with CXCL16 in the absence or presence of CXCL16 neutralizing antibody. (C) Schematic representation of the signaling pathway. (D) Expression analysis of MMPs as downstream effectors of CXCL16 signaling in MSCs depicted a significant increase in MMP2 expression. (E) CXCL16 treatment led to an increase in MMP2 activity by using gelatin zymography. (n = 3–6 replicates/experiment repeated thrice; p < 0.05 as compared with *control; ^#CXCL16-treated group; ^$CXCL16-treated Cxcr6 overexpression).

CXCR6 Overexpression Induces MMP2-Mediated MSC Migration

MSCs transfected with either pCMV (control) or pCMV-Cxcr6-FLAG expression vectors were subjected to G418 selection, and overexpression was confirmed at RNA and protein level (Figure S3A). Cxcr6 overexpression led to a marked increase in FAK-Src-ERK1/2 phosphorylation. The specificity of the signal transduction pathway was confirmed using pharmacological inhibitors against FAK, Src, ERK1/2, and MMP (Figures 2A and 2B). Cxcr6 overexpression further led to a significant increase in MMP2 expression as compared with the control group (Figure 2C). CXCL16-induced significant increase in MMP2 expression was abrogated in the presence of signaling pathway inhibitors FAK14, PP2, PD98059, and GM6001 in both control and Cxcr6 overexpressing MSCs (Figure 2C) indicating an involvement of FAK-Src-ERK1/2 signaling in MMP2 expression and upregulation. Next, Cxcr6 overexpression led to a significant increase in MSC migration, which was further augmented in the presence of CXCL16 as compared with control MSCs that were reverted in the presence of pathway inhibitors (Figure 2D).

Figure 2.

CXCR6 Overexpression Potentiated MMP2-Mediated MSCs Migration

Cxcr6 overexpression in MSCs led to (A) sequential increase in phosphorylation of FAK, Src, and ERK1/2 by immunoblot analysis, (B) densitometric analysis of activated signaling mediators in MSCs treated with CXCL16 in absence or presence of pathway-specific inhibitors, (C) increased mRNA expression of MMP2, and (D) increased migratory potential that was mitigated in presence of CXCL16-NAb or pharmacological inhibitors of FAK, Src, and ERK1/2. (n = 3–6 replicates/experiment repeated thrice; p < 0.05 as compared with *control; ^#CXCL16-treated group; ^$/ˆˆCXCL16-treated Cxcr6 overexpression).

CXCR6 Overexpression Potentiates Nuclear Translocation of Activated ERK1/2-Mediated MMP2 Promoter Expression

To further dissect the transcriptional regulation of MMP2 by activated FAK-Src-ERK1/2 signaling, we examined intracellular localization of p-ERK1/2 in MSCs. Cxcr6 overexpression led to a marked increase in nuclear translocation of p-ERK1/2 (Figure 3A). CXCL16 treatment markedly increased the nuclear translocation of p-ERK1/2 that was perturbed by CXCL16-NAb, as well as pathway inhibitors of FAK-Src-ERK1/2 but not MMP in both control and Cxcr6 overexpressing MSCs (Figure 3A). Literature suggests that an increased phosphorylation of Sp1 by activated ERK1/2 in the nucleus leads to increased binding and activity of Sp1 at cis-acting regulatory elements on its target promoter.22, 23, 24 This led us to hypothesize that p-ERK1/2 in the nucleus interacts with Sp1 to promote the transcriptional activity of MMP2 gene. In silico analysis of MMP-2 promoter revealed a presence of three putative Sp1 binding sites of which two (BS2, −61 to −66 bp and BS3, −151 to −156 bp) are located within 500 bp upstream to transcription start site (TSS, +1) and one (BS1, +145 to +150 bp) is located between TSS and translational start site (Figure 3B). We, therefore, cloned −272 to +206 bp region (wild-type [WT]) of MMP2 promoter and 5′-deletion (ΔBS3 and ΔBS3+2) MMP2 luciferase reporter constructs as described in the Materials and Methods to identify the specific role of each Sp1 cis-acting elements on MMP-2 transcription (Figure 3C). Cxcr6 overexpression depicted a significant increase in luciferase activity of the WT MMP2 reporter construct containing all three Sp1 binding sites (Figure 3C). Sequential deletion of the distal Sp1 binding sites, ΔBS3 and ΔBS3+2 on the MMP2 promoter, significantly reduced its promoter activity in the presence of Cxcr6 overexpression (Figure 3C). The deletion of ΔBS1 depicted promoter activity higher than control but lower than the WT-MMP2 promoter construct, thereby suggesting the importance of all the three cis-acting elements for its promoter activity. Thus, we further evaluated the effect of CXCL16 and PD98059 on WT-MMP2 promoter activity with gain or loss of Cxcr6 function in MSCs. Cxcr6 expression in MSCs was silenced by transfecting piLenti-Cxcr6 siRNA vectors (ABM, Canada) that was confirmed using RT-PCR and immunofluorescence analysis (Figures S3B and S3C). CXCL16 could significantly induce WT-MMP2 promoter activity in MSCs co-transfected with either control, piLenti-Scr siRNA, or pCMV-Cxcr6 overexpression vector but not in MSCs transfected with piLenti-Cxcr6 siRNA (Figure 3D). Interestingly, PD98059 treatment to MSCs significantly shunted WT-MMP2 promoter activity in the absence/presence of CXCL16 of all the groups (Figure 3D) indicating the critical role of ERK1/2 on the CXCL16-CXCR6 downstream signaling. Finally, the transcriptional modulation of MMP-2 by activated ERK1/2 via Sp1 cis-acting elements was confirmed using chromatin immunoprecipitation (ChIP) assay. The observation of ChIP analysis depicted a higher binding of p-ERK1/2 with the Sp1 cis-acting elements at BS2 and BS3 but not BS1 that was further induced by CXCL16 (Figure 3E). These mechanistic studies clearly suggested that CXCL16-induced CXCR6 mediated activation of the FAK-Src-ERK1/2-MMP2 signaling pathway promotes MSC migration that can be of therapeutic relevance for allogenic MSC transplantation during wound tissue regeneration.

Figure 3.

CXCL16-CXCR6 Signaling Activation Led to Nuclear Translocation of p-ERK1/2 and Transcriptional Regulation of MMP2 Promoter Activity

(A) Confocal immunostaining depicting CXCL16-induced nuclear translocation of p-ERK1/2 in MSCs transfected with pCMV or pCMV-CXCR6-Flag-tagged vector that was inhibited in presence of CXCL16-NAb or pharmacological inhibitors of FAK, Src, and ERK1/2, but not by MMP pan inhibitor. (B) In silico analysis revealed the presence of three putative binding sites of Sp1 on MMP2 promoter (+1, transcription start site and ▲, translation start site). (C) Promoter reporter assay revealing that all three Sp1 binding sites are essential for promoter activity of MMP2. (D) CXCL16-induced promoter luciferase activity was further increased in MSCs overexpressing CXCR6 and inhibited in MSCs with CXCR6 gene silencing, as well as ERK1/2 inhibitor, PD98509. (E) ChIP assay suggesting higher binding of p-ERK1/2 at Sp1 binding sites—BS2 and BS3, but not BS1, that was further induced by CXCL16. (n = 3 replicates/experiment repeated thrice; p < 0.05 as compared with *control, and ^#CXCR6 overexpressing MSC group).

Transplantation of Bioengineered MSCs for CXCR6 Gene Therapy Led to Enhanced Engraftment at the Wound Site

To evaluate the in vivo effects of Cxcr6 overexpression on MSCs, we generated stable MSCs with Cxcr6 gene overexpression (MSC-pLJM1-Cxcr6-EGFP; Figure S4A) and confirmed by GFP tag expression using immunofluorescence (Figure S4B), immunoblot analysis (Figure S4C), and flow cytometry analysis (Figure S4D). The therapeutic relevance of MSCs overexpressing Cxcr6 was evaluated by intravenous (i.v., tail vein) transplantation into a murine model of excisional splinting wound as described earlier.^12,25 Mice were randomly assigned in different groups for transplantation of 2 × 10⁶ stable MSCs either expressing pLJM1-EGFP (MSCs-control) or pLJM1-Cxcr6-EGFP (MSCs-Cxcr6) or piLenti-Cxcr6 shRNA-GFP (MSCs-Cxcr6-KD) lentiviral vectors. Next, to assess the ability of CXCL16 in recruiting MSCs at the wound site, we injected CXCL16 intradermally (i.d., 0.5 μg/100 μL per wound) into the periphery of wounded skin. A temporal study was performed on PS days 1, 3, and 7 to investigate the effect of Cxcr6 overexpression or silencing on MSCs in its recruitment or engraftment at the wound site, as well as other tissues using genomic PCR to analyze EGFP⁺-MSCs that infiltrated in various tissues post-transplantation. A standard curve was plotted using pLJM1-EGFP as described in the methods (Figure 4A, inset). A significantly higher number of EGFP⁺-MSCs were observed in blood (Figure 4A), heart (Figure 4B), lung (Figure 4C), and liver (Figure 4D) tissue on PS day 1 that decreased through PS days 3 and 7. Our observations are in alignment with the prior reports that suggested that MSCs post-systemic administration in a wounded animal get entrapped in the lungs but within 1–3 days post-administration they were also observed to be at the wound site.²⁶ The number of EGFP⁺-MSCs in the wound tissue was observed to be significantly higher in mice transplanted with stable MSCs-Cxcr6 both in absence/presence of CXCL16 ligand i.d. administration as compared with groups transplanted with MSCs-control (Figure 4E). EGFP⁺-MSCs were observed to be lacking in groups transplanted with stable MSCs-Cxcr6-KD in the absence/presence of CXCL16 ligand (Figure 4E). Interestingly, an enrichment of stable EGFP⁺-MSCs-Cxcr6, from blood to wound site, was observed over the healing period at PS day 7 (Figure 4F). Further, histopathological analysis of the regenerated wound tissue in mouse transplanted with stable MSCs-Cxcr6 in presence of CXCL16 ligand (i.d.) administration at the wound site depicted an increased number of proliferating cells (Figures 5A and 5B) and collagen staining (Figures 5C and 5D) at the wound bed as compared with control. These observations are in line with the literature, wherein MSCs have been reported to enhance collagen synthesis via directly secreting collagen type III²⁷ or indirectly under the influence of hypoxia at the wounded tissues by secreting growth factors and cytokines that in turn stimulate the migration of resident fibroblasts and activate to synthesize collagen, which further accelerate wound repair process.²⁸ On the contrary, mice transplanted with stable MSCs-Cxcr6-KD did not depict proliferative cells at the wound site both in the absence or presence of CXCL16 ligand administration (Figure 5A). These data suggested that the CXCL16/CXCR6 axis in MSCs plays a crucial role in cell migration and improved engraftment of MSCs at the wound site and promote wound tissue regeneration.

Figure 4.

Temporal Increase in Engraftment of CXCR6 Overexpressed MSCs at Wound Site

(Inset) Graph representing pLJM1-EGFP plasmid standard curve indicating the number of EGFP transgene copies in 500 ng of DNA, which was calculated from their Ct (cycle threshold) values using the linear equation. Genomic PCR for EGFP-positive cells was performed to analyze the temporal infiltration of transplanted cells in (A) blood, (B) heart, (C) lung, (D) liver, and (E) wound tissues. (F) Significant increase in engraftment of EGFP-positive cells from blood to the wound site over the healing period analyzed. (n = 2 replicates per wound with 2 wounds/mice, N = 5 mice/group; p < 0.05 as compared with *pLJM1-EGFP control; ^#pLJM1-Cxcr6-EGFP).

Figure 5.

CXCL16-CXCR6 Axis Promotes Re-epithelialization and Collagen Deposition at the Wound Site

Representative images of (A) hematoxylin & eosin (H&E) stained regenerated wound skin at PS days 1, 3, and 7, (dotted lines depicting the wound margin; WB indicates the location of wounded/regenerated wound bed). (B) Graph depicting increased number of H&E-stained cells, (C) Sirius red at PS day 7, and (D) graph depicting increased percent area of collagen deposition in regenerated wounds suggesting high re-epithelialization and collagen deposition in groups transplanted with stable MSCs-Cxcr6. CXCL16 (0.5 μg/100 μL, i.d.) at each wound periphery further increased the wound tissue regeneration as compared with other groups. (E, epithelial cells; V, vascular cells; H, hair follicles; and G, sebaceous glands). (n = 3 replicates/wound, N = 5 mice/group; p < 0.05 as compared with ^∗wound control; ^$pLJM1-CXCR6-EGFP Tx).

Transplanted MSCs with CXCR6 Gene Therapy Are Viable at the Wound Site

To evaluate the viability of the transplanted MSCs, we performed co-staining with Ki67 (proliferation marker) and GFP. At PS day 1, the number of EGFP⁺-MSCs was observed to be markedly higher in mice groups transplanted with stable MSCs-Cxcr6 along with CXCL16 ligand administration as compared with control groups (Figures 6A and 6B). No EGFP⁺-MSC-Cxcr6-KD with/without CXCL16 ligand administration were detected in transplanted groups (Figure 6A). We observed an increased number of GFP–Ki67 co-localized cells at PS day 7 (Figure 6A), which is concomitant with existing literature, where in a normal wound-healing process there is an increased epidermal thickness with a concomitant increase in Ki67-positive proliferating keratinocytes in both basal and suprabasal layers.²⁹ Taken together, these data thereby suggest the plausible therapeutic role of Cxcr6 engineered MSCs mediated wound tissue repair/regeneration.

Figure 6.

Engrafted CXCR6 Overexpressed MSCs Are Viable at the Wound Site

(A) Confocal co-immunostaining of GFP and Ki67, a proliferation marker depicted an increased recruitment of transplanted MSCs-Cxcr6 were viable at the wound site. MSC-Cxcr6-KD did not get recruited at the wound site suggesting the migratory role of CXCR6. (B) Significant increase in colocalization of GFP with Ki67 in MSC-pLJM1-CXCR6-GFP as compared with MSC-pLJM1-GFP depicted higher engraftment and viability of transplanted MSC. (n = 3 replicates/wound, N = 5 mice/group; p < 0.05 as compared with ^∗pLJM1-EGFP-MSC Tx; ^$pLJM1-CXCR6-EGFP Tx).

CXCL16 Modulates Transdifferentiation of MSCs into Endothelial and Keratinocyte Lineages

MSCs are known to be transdifferentiated into lineages other than their tri-lineage—adipocytes, chondrocytes, and osteoblasts.³⁰ Thus, we further evaluated the effect of CXCL16 on MSC transdifferentiation into endothelial and keratinocyte cells (in vitro). MSCs transdifferentiated into endothelial cells depicted a marked increase in uptake of DiIAcLDL (Figure 7A) that was further enhanced in the presence of CXCL16 and inhibited in the presence of CXCL16-NAb (Figure 7B). Further, endothelial lineage of transdifferentiated MSCs was confirmed by the marked increase in endothelial cell-specific protein expression—CD31 and Tie-2 by CXCL16 treatment that was perturbed in the presence of CXCL16-NAb (Figure 7C). MSCs did not depict the expression of CD31 or Tie-2, indicating their retention of mesenchymal properties (Figure 7C). Similarly, MSCs transdifferentiated into keratinocytes depicted an increased expression of Pan-cytokeratin (PanCK) keratinocyte marker in the presence of CXCL16 that was mitigated by CXCL16-NAb (Figure 7D). These observations suggest a therapeutic role of CXCL16/CXCR6 axis on enhancing vascularity and epithelialization for an efficacious wound tissue repair.

Figure 7.

CXCL16 Modulates MSCs Transdifferentiation Potential toward Endothelial Cells and Keratinocytes

(A) Endothelial transdifferentiation of MSCs in vitro in EGM-2 medium showing higher DiIAcLDL uptake when treated with CXCL16 as compared with control that was perturbed in presence of CXCL16 neutralizing antibody. (B) Quantification of DiIAcLDL uptake in cells. (C) CXCL16-enhanced induction of endothelial cell transdifferentiation was confirmed by confocal immunostaining of endothelial markers expression, CD31, and Tie-2. (D) Similarly, PanCK keratinocyte marker expression also confirmed MSC transdifferentiation into keratinocytes, suggesting enhanced plasticity of MSCs in the presence of CXCL16 as compared with control. (n = 6 replicates/experiment repeated thrice; p < 0.01 as compared with **α-MEM control, ^$EGM control, and ^##CXCL16 treated group).

Transplantation of MSCs with CXCR6 Gene Therapy Enhances Neovascularization and Re-epithelialization during Wound Tissue Regeneration

Neo-vascularization and re-epithelialization are crucial steps for tissue regeneration that was evaluated using immunofluorescence co-staining of CD31/GFP and PanCK/GFP, respectively in the regenerating tissue sections. We observed an increase in GFP⁺-MSCs at PS days 1 and 3 in groups transplanted with stable MSCs-Cxcr6 along with/without CXCL16 ligand administration (i.d.) as compared with MSCs-control while co-localization of GFP/CD31 (Figure S5A) and PanCK/GFP (Figure S5B) was observed at PS day 7. There was an absence of GFP⁺-MSCs at the wound site in groups transplanted with stable MSCs-Cxcr6-KD all throughout the post-surgery period evaluated (Figures S5A and S5B). Higher magnification images clearly depicted an increased expression of CD31 (Figures S6A and S6B) and PanCK (Figures S7A and S7B) by the Cxcr6 overexpressing GFP⁺-MSCs with comparatively higher in CXCL16 injected (i.d.) group. These findings clearly indicate a conceivable role of CXCL16/CXCR6 axis on wound tissue healing by potentiating neo-vascularization and re-epithelialization via transdifferentiation of the transplanted MSCs into an endothelial and keratinocyte lineage but does not rule out a paracrine regulation by the transplanted MSCs during early healing period to enhance wound tissue regeneration. Thus, suggesting its plausible therapeutic role in regeneration of skin at non-healing diabetic wounds.

Therapeutic Transplantation of MSCs with CXCR6 Gene Therapy Regenerates Skin Tissue in Type I and II Diabetic Wounds

Diabetic patients with foot ulcers often result in chronic non-healing wounds that remain a challenge to the medical fraternity.³¹ To further determine the therapeutic efficacy of MSCs with Cxcr6 gene therapy in healing of diabetic chronic wounds, we generated type I diabetes in C57BL/6J mice using streptozotocin (i.p., 70 mg/Kg body weight) (Figure 8A), as well as used db/db mice, a known model of type II diabetes. High blood glucose levels that were observed in db/db mice and evaluated during pre- and post-dosing suggested the generation of type I diabetes in C57BL/6J mice (Figure 8B). An excisional splinting wound generated in the diabetic mice group was followed by therapeutic transplantation of stable MSCs-Cxcr6 or MSCs-control. Transplantation of stable MSCs-Cxcr6 at PS day 0 in db/db diabetic mice led to an accelerated wound healing over the period as compared with diabetic mice transplanted with MSCs-control (Figure S8A), as well as un-transplanted diabetic controls (data not shown). Similar observations were also reported in type I diabetes group (data not shown). This led us to evaluate the CXCL16 gene expression in diabetic mice wound tissues. Interestingly, a 10- and 16-fold higher expression of CXCL16 was observed in type I and II diabetic wound, respectively, as compared with non-diabetic wounds (Figure S8B). This observation correlates well with the existing literature suggesting an increased level of CXCL16 expression in hyperglycemic condition and serum CXCL16 level have been considered to be an indicator of diabetic nephropathy.³² Furthermore, the rate of wound closure evaluated on a regular interval till PS day 14 (type I, Figure 8C) or 18 (type II, Figure 8D) clearly indicated a significant increase in percent wound closure in diabetic mice transplanted with stable MSCs-Cxcr6 as compared with non-transplanted or MSCs-control. Although an increasing trend of wound closure was observed in MSCs-control transplanted diabetic mice as compared with non-transplanted groups, the therapeutic efficacy was evident in the MSCs with Cxcr6 gene therapy (Figures 8C and 8D). Next, histopathological analysis revealed an increased re-epithelialization and collagen deposition as evident from the H&E and Sirius red staining, respectively, within the granulation tissue in the type I (Figures S9A and S9B) and type II (Figures 8E and S9C) diabetic mice transplanted with stable MSCs-Cxcr6 as compared with MSC-control groups. An enhanced recruitment of Cxcr6 bioengineered-MSCs was evident from co-immunostaining with GFP and CXCR6 that revealed a markedly higher engraftment of these MSCs at the wound bed as compared with MSCs-control group in type I (Figures S10A and S10B) and type II (Figures 8F and S11A) diabetic mice. Often diabetic wounds fail to heal due to lack of neo-vascularization and re-epithelialization. To evaluate the therapeutic efficacy of MSCs with Cxcr6 gene therapy in neo-vascularization and re-epithelialization at the diabetic wound bed, analysis of CD31 and PanCK, respectively, revealed a markedly higher co-localization of GFP/CD31in type I (Figures S10C and S10D) and type II (Figures 8G and S11B) and GFP/PanCK in type I (Figures S10E and S10F) and type II (Figures 8H and S11C) diabetic mice transplanted with MSCs-Cxcr6 as compared with control-MSCs indicating increased transdifferentiation of MSCs toward vascular and epithelial lineages. Our study clearly demonstrated that the therapeutic efficacy of MSCs was enhanced by the Cxcr6 gene therapy that led to increased engraftment at the wound site, rate of wound closure, enhanced re-epithelialization, and neo-vascularization-mediated accelerated wound repair, as well as skin regeneration of diabetic chronic wounds.

Figure 8.

Transplantation of MSCs with CXCR6 Gene Therapy Efficiently Regenerated Skin in Type I and II Diabetic Mice

(A) Flow chart representing type I diabetes model generation. (B) High fasting blood glucose levels (>150 mg/dL) was observed in both type I and II (db/db) diabetic mice. Higher percent wound closure area in diabetic mice transplanted with MSCs-Cxcr6 (C) type I Stz-induced C57BL/6J and (D) type II db/db mice. (E) Higher H&E and Sirius red staining in db/db mice transplanted with MSCs-Cxcr6 depicting an organized layer of dermis with the presence of glands and hair follicles in the regenerated wounds as compared with control MSCs transplanted groups (E, epidermis; D, dermis; H, Hair follicles; G, sebaceous glands; Ad, adipose layer). Increased co-immunostaining of (F) GFP/CXCR6, (G) GFP/CD31, and (H) GFP/PanCK suggesting more recruitment, engraftment, neo-vascularization, and epithelialization in db/db mice transplanted with MSCs-Cxcr6 as compared with MSC-control. (n = 3 replicates/wound, N = 4–5 mice/group; p < 0.05 as compared with *pLJM1-EGFP control; ^#diabetic control).

Discussion

Adult stem cell transplantation including MSC displays great promise as a therapy for wound healing, although effective engraftment and survival of the transplanted stem cells within the wound area remains a foremost limitation. MSCs have demonstrated to regenerate wounds by immune modulation, paracrine interactions, extracellular matrix remodeling, and regeneration of skin.^2,7,33 However, engraftment of MSCs into wound tissue is controlled by numerous processes that include cell recruitment, migration, and adhesion.^3,4 Studies in the past have shown that MSCs possess the capability to migrate into the tissue from the circulation in response to signals that are induced by injury.⁸ Although the mechanisms by which MSCs transmigrate through vascular endothelial cell layer into the tissues are not yet fully understood, a plausible mechanism can be extended to chemokine-chemokine receptor interaction. Several reports suggest a differential expression levels of chemokines and its receptors on MSCs-derived from murine bone marrow⁹ and Wharton’s Jelly or human bone marrow,³⁴ as well as other tissue sources. Interestingly, our observations are partially in accordance with Abumaree et al.,³⁵ who reported an extensive range of chemokine receptors’ mRNA expression of members including CXC, XC, CX3C, and CC families. Although there exist several studies in literature depicting the CXCL12-CXCR4 axis as a migratory signaling pathway in the stem, as well as inflammatory cells, Wynn et al.³⁶ demonstrated, similar to our findings, a relatively low level of CXCR4 expression on MSCs. In addition, less than 1% of MSCs isolated from bone marrow have CXCR4 expression on the cell surface membrane. Inokuma et al.³⁷ revealed that CTACK/CCL27 majorly regulates the recruitment of keratinocyte precursor cells from bone marrow to skin wounds. CCL27 was observed to be upregulated upon skin wound injury, which in turn recruited bone-marrow-derived CD34⁺ cells to the wound site suggesting the activation of particular migratory axis for specific cell type.³⁷ But limitation of the above study was that only approximately 20% of CD34⁺ cells had expressed CCR10, cognate receptor of CCL27. Also, Sasaki et al.³⁸ reported the expression of CCR7, a chemokine receptor for CXCL21, on MSCs and administration of CXCL21 to the wound site enhanced the migration of GFP-labeled MSCs to the wound site that accelerated the wound healing process. In contrast, we observed a significantly higher fold expression of CXCR6 on murine MSCs, administration of CXCL16 at the wound periphery did not improve the MSC recruitment at the wound bed. To circumvent these limitations, we employed a gene therapy approach to obtain 2-fold benefit by transplanting Cxcr6-overexpressing MSCs toward CXCL16 at wound site—(1) for increasing directional migration and (2) for inducing endothelialization and epithelialization. These indispensable phenomena are observed to be abrogated in various pathophysiological conditions like diabetes, where persisting hyperglycemia or hyper-inflammation inhibits the migration and differentiation of different cells at wound site resulting in impaired and delayed wound healing. However, fewer studies have been performed to augment the existing therapies for chronic diabetic wounds based on chemokine modulation for stem cell-mediated skin tissue regeneration, to date. Unlike, therapeutic administration of CCL2 at wound bed for macrophage recruitment,³⁹ a high level of CXCL16 expression in diabetic wound site, as hyperglycemic condition has been reported to increase CXCL16 expression in both pre-clinical and clinical diabetic nephropathy,⁴⁰ exempted the chemokine administration for recruitment of stable MSCs with CXCR6 gene therapy at diabetic wounds in our study. This not only enhanced the migration toward the wound area but also led to define their fate toward endothelial and epithelial lineage.

With the increasing prevalence of diabetes globally, comorbidities are overriding patients’ quality of life, as well as life expectancy. Hyperglycemic condition in diabetic patients results in the generation of non-healing wounds impeding autologous MSC recruitment at the wound site, along with delayed differentiation capacity of diabetic MSCs. In conclusion, our study identified transplantation of stable MSCs with Cxcr6 gene therapy led to enhanced recruitment and engraftment at the wound site in both non-diabetic and type I (Streptozotocin-induced) and type II (db/db) diabetic mice pre-clinical models with enhanced re-epithelialization, neo-vascularization, and collagen deposition in regenerated skin suggesting CXCL16-CXCR6 axis as a potent therapeutic target for treatment of non-healing diabetic wounds that can potentiate skin tissue regeneration.

Materials and Methods

In Vitro Cell Culture and Functional Studies

C57BL/6 mouse bone-marrow-derived MSCs (Cyagen, USA) were treated with increasing concentrations of CXCL16 and DNA synthesis was measured using a BrdU cell proliferation kit (Roche, USA) as described previously.²⁵ Absorbance was measured at 370 nm with a reference wavelength at 490 nm using a spectrophotometer.²⁵ For colony formation assay, MSCs were plated in complete Mesencult Medium (StemCell Technologies, USA) as described earlier¹¹ and treated with CXCL16 (10 ng/mL) and/or its neutralizing antibody. The number of colonies displaying five or more cells was scored under an inverted microscope and scored for colony-forming units (CFUs) according to standard criteria. Cell migration due to chemotactic signaling was evaluated using a modified Boyden chamber assay (Neuro Probes, Gaithersburg, MD) as described earlier.⁴¹ Briefly, the bottom wells of the chamber contained CXCL16, 10 ng/mL, and 2 × 10⁴ serum-starved cells were plated on the upper chamber of each well and incubated at 37°C in absence/presence of neutralizing antibody or pharmacological signaling pathway inhibitors. Cellular migration was determined by counting the number of Hema3 stained cells on the lower surface of membranes, and data were represented as the number of cells migrated upon CXCL16-induced chemotaxis.⁴² To evaluate the effect of CXCL16 on transdifferentiation of MSCs into endothelial and keratinocytes, we plated MSCs in endothelial⁴³ and keratinocyte⁴⁴ differentiation medium in absence/presence of CXCL16 ligand and/or CXCL16 neutralizing antibody. Cells were stained with endothelial and keratinocyte-specific markers and images were captured with an Olympus FV10i confocal microscope.

Cellular Protein Analysis

Immunofluorescent staining was performed as reported earlier. The cells were incubated with anti-Ki67 or anti-p-ERK1/2 rabbit antibody overnight and nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) contained in the mounting medium (Amresco, USA). Confocal images were obtained with an Olympus FV10i confocal microscope (Olympus, USA) using the 40× oil immersion lens. NIH ImageJ software was used to quantitate nuclear localization.²⁵ Separately, cellular protein was extracted from cells treated with different treatment regimes as described earlier and subjected to 10% SDS–PAGE and transferred to a polyvinylidene difluoride membrane. Membrane was incubated with primary antibodies of phosphorylated and pan signaling markers: ERK1/2, Akt, Src, FAK, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) at 4°C overnight and signals were detected using chemiluminescence enhanced chemiluminescence (ECL) detection system (Merck Millipore, USA) as described earlier.²⁵ Quantification of protein bands was performed using NIH ImageJ software.²⁵ Conditioned media from cells treated with CXCL16 (10 ng/mL) in absence or presence of CXCL16 neutralizing antibody was mixed with sample loading buffer and subjected to 8% SDS-PAGE copolymerized with gelatin type B (2 mg/mL) for zymography analysis as described earlier.³¹ Gelatinolytic activities were detected as transparent bands.⁴²

Gene Overexpression and Silencing Studies

MSCs were transfected with pCMV-Cxcr6-FLAG (Origene, MR205339, USA) or pCMV empty vector expression plasmids using lipofectamine (Invitrogen, USA) according to the manufacturer’s instruction followed by culturing the cells in G418 containing medium.⁴⁵ Separately, CXCR6 gene silencing was accomplished by transfection of lentiviral constructs, piLenti-Cxcr6 siRNA-GFP (ABM, i036744, CA) or piLenti-Scr siRNA-GFP control (ABM, LV015-G, CA) as described in the manufacturer’s protocol.⁴⁶ The overexpression or silencing of CXCR6 was analyzed by using quantitative or semiquantitative RT-PCR analysis⁴⁷ using gene-specific primers, immunoblot analysis using anti-FLAG antibody,²⁵ and fluorescence-activated cell sorting (FACS) analysis using phycoerythrin (PE)-conjugated CXCR6 antibody.²⁵

Transcriptional Regulation Studies

In silico analysis of the MMP2 promoter region between −2 Kb to +100 bp revealed the presence of three binding sites of SP1 transcription factor. Promoter constructs were designed to include all three binding sites (WT-MMP2) or deletion constructs containing either two binding sites (ΔBS1 MMP2 or ΔBS3 MMP2) or one binding site (ΔBS3+2 MMP2). These constructs were cloned in pMCS-Green Renilla vector (Thermo Scientific, USA) using restriction enzymes (XhoI and KpnI) and T4 DNA ligase. MSCs were co-transfected with the pMCS-Green Renilla vector containing the desired promoter region along with the internal control vector pCMV-Red Firefly Luc vector and either pCMV-Cxcr6-FLAG overexpression/pCMV (control) or piLenti-Cxcr6-siRNA-GFP silencing/piLenti-Scr siRNA-GFP vectors as described earlier.^42,48 Transfected MSCs were treated with either CXCL16 or ERK1/2 inhibitor-PD98059 and incubated at 37°C for 72 h. Luminescence was measured according to the manufacturer’s protocol. To evaluate the binding efficiency of p-ERK1/2 on to the Sp1 cis-acting elements, we performed ChIP assay as described earlier.^11,25 MSCs transfected with either pCMV-empty or pCMV-Cxcr6 overexpression vectors were treated with/without CXCL16 followed by cross-linking with 1% formaldehyde at 37°C for 10 min and sonicated using Bioruptor (Diagenode, USA) for 20 cycles. Immunoprecipitation of the chromatin was performed overnight at 4°C with anti-p-ERK1/2 antibody as described earlier.^11,25 Precipitates were washed, DNA was eluted and used as a template for PCR amplification with primers specific for region of Sp1 binding sites on the MMP2 promoter sequence (BS1, BS2, and BS3) was quantitatively measured.²⁵

Stable MSCs Generation with Gain or Loss of CXCR6 Function

Cxcr6 gene was sub-cloned from pCMV-Cxcr6-FLAG into pLJM1-EGFP vector (Addgene, plasmid 19319, USA) using standard protocol with NheI and AgeI restriction sites at the 5′- and 3′- end, respectively.⁴⁹ Sub-cloning of Cxcr6 in the pLJM1 vector was confirmed by gene sequencing techniques. pLJM1 empty or pLJM1-Cxcr6-EGFP vector, as well as piLenti-Scr shRNA-GFP or piLenti-Cxcr6-shRNA-GFP vectors, were amplified and separately co-transfected into HEK293T cells along with packaging vectors VSVG and PAX2 plasmids, and the supernatant containing viral particles was collected filtered and used for transducing allogenic MSCs.⁴⁹ Stable MSCs with either CXCR6 overexpression or silencing were generated by puromycin selection as described elsewhere.⁵⁰

Animal Experiments

An excisional splinting wound model was generated in 8- to 10-week-old C57BL/6J mice of either sex as described earlier.^25,31 In this model, to avoid wound closure due to contraction of skin, excisional wounds on the dorsum of the mice were splinted using cyanoacrylate-based adhesive and surgical sutures to restrict the movement of the panniculus carnosus. Prior studies by Park et al.⁵¹ confirmed no significant difference in wound numbers (one versus two), side (left versus right and cranial versus caudal), or size of wounds per splint on the rate of wound healing. Temporal expression of a panel of chemokines was evaluated at the wound site by performing qRT-PCR analysis (post-surgery day 0 to day 10) and compared with un-wounded healthy tissue (dorsal side) of mouse as control. All the target genes expression was normalized to eukaryotic 18S rRNA mRNA and shown as the -fold change.¹¹ Transplantation of stable allogeneic MSCs either overexpressing or silencing Cxcr6 (MSC-pLJM1-Cxcr6-EGFP/MSC-pLJM1-EGFP (control) or MSC-piLenti-Cxcr6 shRNA-GFP/MSC-piLenti-Scr shRNA-GFP) was performed by injecting i.v. (tail vein) 2 × 10⁶ MSCs/mouse³⁸ along with in absence or presence of CXCL16 (0.5 μg/100 μL) administered i.d. at the wound periphery.³⁷ Animal experimentation protocols were approved by the institutional animal ethics committee (Approval No. IICT/06/2015 and IICT/36/2015). The wound healing depicting wound area closure at different time-points post-wound generation (days 1, 3, and 7) in these groups was quantified as described previously.²⁵ Mice were euthanized on PS days 1, 3, or 7, skin samples including the wound and surrounding skin were harvested by using a 10-mm biopsy punch.²⁵ Regenerated skin wound tissue samples from various treatment groups were harvested using a 10-mm biopsy punch from the center of the wound and fixed in 4% paraformaldehyde. Flat horizontal cross-sections of frozen wound skin tissue were then prepared by cutting from epidermis side toward the dermis using cryostat blade, mounted on slides followed by staining with H&E and Sirius red.⁵² Homing and engraftment of the transplanted MSCs were performed by extracting DNA from different tissue samples including heart, liver, lung, blood, and wound tissue samples for GFP transgene expressing positive cells using quantitative PCR analysis as described previously.⁵³ This robust assay with a detection limit of one EGFP+ cell per 100 ng DNA was reported to be sensitive enough for in vivo biolocalization of transplanted cells.⁵⁴ A pLJM1-EGFP plasmid standard curve was constructed and the data was represented as the number of EGFP transgene copies in 500 ng DNA that was calculated from their respective Ct (cycle threshold) using the linear equation from the respective plasmid standard curve. Separately, frozen wound tissue sections from the various groups were stained with specific antibodies against GFP, Ki67, CXCR6, CD31, and PanCK (Cell Signaling Technologies, USA).²⁵

Allogenic MSC Transplantation in Type I and Type II Diabetic Mice

Diabetes was induced by administrating streptozotocin (Stz, 70 mg/kg) intraperitoneally for 5 consecutive days.⁵⁵ The fasting glucose levels were measured once in a week by taking blood from the retro-orbital area using glucose strips and AccuChek glucometer device. After 2 weeks of diabetes induction, mice were grouped and subjected for wound generation surgery, followed by allogeneic transplantation of stable MSCs overexpressing Cxcr6 and wound closure analysis as described earlier. Similarly, 10- to 12-week-old transgenic db/db type II diabetic mice were procured and monitored for their high glucose levels followed by generation of full-thickness excisional splinted wounds. Type I and type II diabetic mice at PS days 14 and 18, respectively, were euthanized and regenerated wound tissues were collected and fixed in paraformaldehyde. The samples were processed for histological and immunostaining analysis as mentioned earlier.^25,56 Animal experimentation protocols for diabetic models were approved by the institutional animal ethics committee (Approval No. IICT/IAEC/48/2018) and institutional biosafety committee (Approval No. IICT/IBSC/03/2018).

Statistical Analysis

Data from experiments were performed at least thrice, which was statistically analyzed by taking mean ± standard deviation (SD) or standard error of the mean (SEM). To determine the difference between treatment group and their respective controls, we used GraphPad Prism version 6.05 to evaluate statistical significance using one-way or two-way ANOVA followed by appropriate analysis such as post hoc or Student’s paired t test. Blots and photomicrographs represent experiments reproduced at least thrice with similar results.

Author Contributions

A.D conceived the idea and designed the study; N.R.D. and K.K. performed the experiments; N.R.D., K.K., and A.D. analyzed the data and wrote the paper.

Conflicts of Interest

The authors declare no competing interests.