Noninvasive application of mesenchymal stem cell spheres derived from hESC accelerates wound healing in a CXCL12-CXCR4 axis-dependent manner

Abstract

Mesenchymal stem cells (MSC) derived from adult tissues effectively promote wound healing. However, MSC quality varies, and the quantity of MSC is limited, as MSC are acquired through donations. Moreover, the survival and functioning of dissociated MSC delivered to an inflammatory lesion are subject to challenges. Methods: Here, spheres (EMSC_Sp) generated from human embryonic stem cell-derived MSC (EMSC) were directly dropped onto excised wounds in mice; the effects of EMSC_Sp were compared to those of dissociated EMSC (EMSC_Diss). Following transplantation, we measured the extent of wound closure, dissected the histological features of the wounds, determined transcriptomic changes in cells isolated from the treated and control wounds, and evaluated the molecular mechanism of the effects of EMSC. Results: The application of EMSC_Sp onto murine dermal wounds substantially increased survival and efficacy of EMSC compared to the topical application of EMSC_Diss. RNA sequencing (RNA-Seq) of cells isolated from the wounds highlighted the involvement of CXCL12-CXCR4 signaling in the effects of EMSC_Sp, which was verified in EMSC via CXCL12 knockdown and in target cells (vascular endothelial cells, epithelial keratinocytes, and macrophages) via CXCR4 inhibition. Finally, we enhanced the biosafety of EMSC_Sp by engineering cells with an inducible suicide gene. Conclusions: Together, these data suggest the topical application of EMSC_Sp as an unlimited, quality-assured, safe, and noninvasive therapy for wound healing and the CXCL12-CXCR4 axis as a key player in this treatment.

Keywords: CXCL12/CXCR4; Human embryonic stem cells; mesenchymal stem cells; spheroids; wound healing.

Figure 1

Effects of EMSC on wound closure. (A) Experimental scheme for the development of an excisional wound splint model in NOD/SCID mice and the transplantation of EMSC_Sp, EMSC_Diss or the vehicle control. (B) Representative images of wounds in mice treated as above at various time points following wound formation. (C-E) The percentage of the wound area in NOD/SCID mice was measured after the transplantation of EMSCs derived from Envy (C), CT3 (D), and H9 (E) hESC lines. n = 5, 8, and 6 biological repeats in C, D, and E, respectively. *P < 0.05 and **P < 0.01 for EMSC_Sp versus EMSC_Diss or the vehicle control per ANOVA followed by Tukey's multiple comparison test. (F) The percentage of the wound area was measured in NOD/SCID mice after transplantation of BM-MSC_Sp, BM-MSC_Diss or the vehicle. n = 5 biological repeats; per ANOVA followed by Tukey's test, *P < 0.05 and **P < 0.01 for BM-MSC_Sp versus BM-MSC_Diss or the vehicle control. (G) The percentage of the wound area was measured in mice after transplantation of EMSC_Sp, spheres formed by HaCaT cells (HaCaT_Sp), and EMSC lysates. n = 4 biological repeats, *P < 0.05 and **P < 0.01 for EMSC_Sp versus HaCaT_Sp or the EMSC lysates per ANOVA analysis. (H) The percentage of the wound area was measured in Balb/c mice after transplantation of EMSC_Sp, EMSC_Diss or the vehicle. n = 6 biological repeats; per ANOVA followed by Tukey's test, *P < 0.05 and ** P < 0.01 for EMSC_Sp versus EMSC_Diss.

Figure 2

Viability and engraftment of EMSC following their transplantation to wounds. (A) Representative bright-field images of wounds with transplanted iRFP-expressing EMSC_Sp (upper row) or EMSC_Diss (bottom row) were taken on day 0, followed by fluorescent images acquired using the In-Vivo Xtreme System on various days following wound formation. The colored bar indicates the fluorescence intensity. The fold change in the fluorescent intensity is shown in the graph on the right. (B) (a) Immunostaining for GFP⁺ cells in wounds with transplanted Envy hESC-derived EMSC_Sp and EMSC_Diss 14 days after wound formation. Boxed areas are amplified and shown on the right. (b) The percentage of GFP⁺ cells (determined via DAPI staining of the nucleus) per wound section. (C) Flow cytometry to detect GFP⁺ cells in wounds with transplanted Envy hESC-derived EMSC_Sp and EMSC_Diss 7 and 14 days after wound formation. Values represent the percentage of GFP⁺ cells from 3 biological repeats. (D) Cultured cells were isolated following the digestion of wounds with transplanted EMSC_Sp (Envy) 7 days after wound formation. Bright-field (left) and fluorescent (middle) images of the cultured cells were taken and merged (right). Scale bar = 200 μm. (E) Detection of human TK1 gDNA in mice 14 days after EMSC transplantation. One x 10⁶ EMSC_Sp and EMSC_Diss cells were transplanted to the skin wound. After 14 days, wounded skin was dissected, and total gDNA was extracted for PCR to detect human TK1. n = 3 biological repeats. * P < 0.05 for EMSC_Sp versus EMSC_Diss per student t-test.

Figure 3

Histological analyses of wounds with transplanted EMSC_Sp. (A) Masson trichrome staining of normal skin and wounds 14 days after treatment with EMSC_Sp or vehicle control. Boxed areas are amplified and shown on the bottom row. Arrows in the upper row indicate the junction between the wound and surrounding skin. (B) Re-epithelialization of wounds 14 days after treatment with EMSC_Sp, EMSC_Diss or vehicle. Tissues isolated from the wounds were sectioned and immunostained (a) with anti-pancytokeratin antibody. White dashed lines indicate the detected epidermal keratinocytes, and white arrows indicate skin appendages. The epithelial tongue length (b) and the number of skin appendages per section (c) are shown in the bar charts. n = 5 biological repeats; *P < 0.05 for EMSC_Sp versus vehicle and ^#P < 0.05 for EMSC_Sp versus EMSC_Diss per ANOVA followed by Tukey's test. (C) Detection of vascular endothelial cells in wounds 14 days after treatment as described above. Tissues isolated from the wounds were sectioned and immunostained with antibodies against CD31 as a marker for vascular endothelial cells. The results are quantified in a bar graph and displayed as capillary density. n = 5 biological repeats; **P < 0.01 for EMSC_Sp versus EMSC_Diss or vehicle per Kruskal-Wallis test. (D) Differentiation of transplanted EMSC_Sp into blood vessel endothelial cells. Wounds 14 days after EMSC_Sp (Envy) transplantation were isolated as described above and immunostained for both GFP (green) and CD31 (red), a marker for blood vessel endothelial cells. The white boxed area indicates double-positive cells.

Figure 4

Transcriptomic analysis of host cells in EMSC_Sp-treated wounds. (A) RNA sequencing of EMSC_Sp and EMSC_Diss before transplantation. Enrichment analysis of human DEGs in EMSC_Sp versus EMSC_Diss. The bar length represents the -log10 P value. The red and blue bars represent terms enriched for up- and downregulated genes, respectively. (B) Detection of GFP mRNA in the wounds of NOD/SCID mice treated with GFP⁺ Envy EMSC_Sp or EMSC_Diss. GFP transcript levels were detected via RT-PCR. GFP expression in H9 hESC-derived EMSC was used as a negative control. A GAPDH transcript with a conserved sequence between mice and humans was used as a loading control. (C) Heatmap displaying the transcriptomic changes following the treatment of wounds with EMSC_Sp (Wound+Sp) and vehicle (Wound) after various time points (3, 7, and 14 days). Expression levels of up- and downregulated genes that differed by more than 1.5-fold are highlighted in red and blue, respectively, as indicated in the color key above. (D) Venn diagram showing the intersections among up- and downregulated genes between the Wound+Sp with Wound samples at the three time points, as described above. (E) Boxplots showing changes in the expression of genes associated with eight wound healing-related pathways in the Wound (yellow) and Wound+Sp (green) samples at the three time points. Each plot stands for a pathway. (F) Heatmaps displaying significantly enriched DEGs categorized based on gene ontology. The expression level of up- and downregulated genes that differed by more than 1.5-fold changes are shown in red and blue, respectively, as indicated in the color key below.

Figure 5

Role of the CXCL12/Cxcr4 axis in EMSC_Sp-promoted wound healing. (A-B) Pearson's correlation coefficient between the expression of CXCL12 in EMSC_Sp and the expression of Cxcl12-targeted genes in mouse cells based on RNA-seq data from wounds treated with EMSC_Sp and vehicle 0, 3, 7, and 14 days after wound formation (A). The X-axis represents the time, and the Y-axis represents the z-score values of the coefficient. The correlations between Cxcl12-targeted genes with CXCL12 expression and their corresponding P values are listed in (B). (C) CXCL12 expression was decreased in EMSC stably transduced with lentivirus expressing shCXCL12 compared to that in EMSC transduced with lentivirus expressing shLacZ. The knockdown efficiency was determined by qPCR with 3 biological repeats. (D) CXCL12 secretion by EMSC transduced with shNC (control) or shCXCL12 was measured via ELISA. *P < 0.05 per the Mann-Whitney U test. (E-F) The migration of HUVECs towards EMSC_Sp expressing shCXCL12 was reduced compared to HUVEC migration towards EMSC_Sp expressing shNC in a Transwell assay. HUVECs were seeded on the top of the membrane insert, and no cells (vehicle, left), EMSC_Sp expressing shNC (shNC_Sp, right) or shCXCL12 (shCXCL12_Sp, middle) were seeded on the bottom of the Transwell insert. HUVECs that migrated across the membrane were stained with crystal violet and photographed under a microscope. The numbers of migrated HUVECs in the three groups are displayed in a bar chart and represent 3 biological repeats. *P < 0.05 for shNC_Sp versus shCXCL12_Sp or vehicle; and ^#P < 0.05 for shCXCL12_Sp versus vehicle per ANOVA followed by Tukey's test. (G) Transwell assay to detect the EMSC_Sp-oriented migration of HUVECs pretreated with AMD3100 (a CXCR4 antagonist) at various concentrations for 30 min. EMSC_Sp were seeded on the bottom of the Transwell insert to induce HUVEC migration. After incubation for 24 h, the number of HUVECs that migrated across the membrane was counted. The results were determined in three independent experiments. *P < 0.05 and **P < 0.01 versus 0 μM AMD3100 per ANOVA followed by Tukey's test. (H) Transwell assay to detect the EMSC_Sp-oriented migration of HaCaT cells pretreated with AMD3100, with the number of cells that crossed the membrane counted as described above. The results were determined in three independent experiments. **P < 0.01 versus 5, 10 or 40 μM AMD3100 per ANOVA followed by Tukey's test. (I) Transwell assay to detect the EMSC_Sp-oriented migration of RAW264.7 macrophages. The number of cells that crossed the membrane was counted as described above. The results were determined in three independent experiments. *P < 0.05 for shNC versus shCXCL12 or vehicle and ^#P < 0.05 for shCXCL12 versus vehicle per ANOVA followed by Tukey's test. (J-M) Measurement of the wound area (J), capillary density (number of CD31⁺ cells among the total cells per wound section) (K), epithelial tongue length (L), and macrophage (MAC2⁺) percentage (M) in mice after transplantation with shNC_Sp, shCXCL12_Sp or vehicle control. *P < 0.05 and **P < 0.01 for shNC_Sp versus shCXCL12_Sp or vehicle per one-way ANOVA followed by Tukey's or Dunn's posttest.

Figure 6

Evaluation of the biosafety of EMSC_Sp transplantation and a strategy to establish safety-enhanced EMSC. (A) Teratoma formation assay. Fifty-six days after wound formation, no tumors in the wound area transplanted with EMSC_Sp were observed based on photographs of the exterior (a) and interior (b) of the healed skin. In contrast, teratomas in the hind leg s.c. injected with hESC were observed in photographs of the exterior (c) and interior (d) of the injection site. Representative tissues from the three germ layers in sections of the teratomas were observed (e). (B) Quantitation of EMSC retained in mouse organs at various times following the transplantation with EMSC_Sp (Envy) onto wounds. The wounded skin, heart, liver, lung, and kidney were harvested at these time points and processed to isolate gDNA. The amount of GFP (ng) per μg of gDNA was used to determine the number of retained EMSC in each sample. (C) Apoptosis of iC9-expressing EMSC induced by AP20187 in vitro. As shown in the diagram above, after exposure to AP20187, the cells were harvested at the indicated times and processed for flow cytometry via AO/PI staining to determine the ratios of early and late apoptotic cells as well as dead cells. The results of three independent experiments are displayed as the mean ± SEM in a bar chart. (D) AP20187-induced elimination of iC9-EMSC from mice with iC9-EMSC_Sp transplanted onto the skin wound. As shown in the diagram above, AP20187 was i.p. injected into the mice at 10 and 11 days after wound formation, and the wound skin was dissected at day 14 and subjected to immunostaining for the HA Tag fused with the iC9 gene transduced into EMSC. The samples were counterstained with DAPI to detect cell nuclei. HA⁺ cells were detected in samples from only mice without AP20187 treatment but not in those treated with the AP20187. The results were determined in four independent experiments. H-E staining in the bottom panel shows the extent of the re-epithelialization of iC9-EMSC_Sp-transplanted wounds following treatment with the vehicle or CID.

Figure 7

Schematic to summarize the findings of this study.