Tissue nanotransfection-based endothelial PLCγ2-targeted epigenetic gene editing rescues perfusion and diabetic ischemic wound healing

Abstract

Diabetic wounds are complicated by underlying peripheral vasculopathy. Reliance on vascular endothelial growth factor (VEGF) therapy to improve perfusion makes logical sense, yet clinical study outcomes on rescuing diabetic wound vascularization have yielded disappointing results. Our previous work has identified that low endothelial phospholipase Cγ2 (PLCγ2) expression hinders the therapeutic effect of VEGF on the diabetic ischemic limb. In this work, guided by single-cell RNA sequencing of human wound edge, we test the efficacy of gene-targeted therapeutic demethylation intending to improve VEGF-mediated neovascularization. PLCγ2 expression was diminished in all five identified diabetic wound-edge endothelial subclusters encompassing arterial, venous, and capillary cells. Such low expression was associated with hypermethylated PLCγ2 promoter. PLCγ2 promoter was also hypermethylated at murine diabetic ischemic wound edge. To specifically demethylate endothelial PLCγ2 promoter during VEGF therapy, a CRISPR-dCas9-based demethylation cocktail was delivered to the ischemic wound edge using tissue nanotransfection (TNT) technology. Demethylation-based upregulation of PLCγ2 during VEGF therapy improved wound tissue blood flow with an increased abundance of von Willebrand factor (vWF)⁺/PLCγ2⁺ vascular tissue elements by activating p44/p42-mitogen-activated protein kinase (MAPK) → hypoxia-inducible factor [HIF]-1α pathway. Taken together, TNT-based delivery of plasmids to demethylate the PLCγ2 gene promoter activity led to significant improvements in VEGF therapy for cutaneous diabetic wounds, resulting in better perfusion and accelerated wound closure.

Keywords: CRISPR; DNA methylation; PLCγ2; VEGF therapy; angiogenesis; diabetes; epigenetics; single-cell RNA sequencing; wound.

Graphical abstract

Figure 1

The correlation between low PLCγ2 levels and promoter methylation is observed in the diabetic wound-edge endothelial cell subsets (A) UMAP plot shows single-cell transcriptomes of 66,709 human chronic wound-edge cells (44,979 T2DM [D], n = 5; 21,730 non-diabetic [ND], n = 5). Cellular heterogeneity with 12(0–11) distinct clusters (with %) of cells identified (based on specific markers, right). Each dot represents a cell. (B) Endothelial cells (0, 1, 2, 3, 4) were then sub-setted in five different subclusters. (C) Dot plot represents the classification of endothelial subclusters based on endothelial cell subtype-specific markers. (D) Bar plot showing that diabetic wound-edge has comparatively low RGCC^hiSPARC^hi capillary endothelial cells (subcluster 0) and high SELE^hiACKR1^hi venous endothelial cells (subcluster 2) as compared to ND wound-edge. Student’s t test. (E) Violin plot showing that PLCγ2 expression was significantly down in all five subclusters of endothelial cells despite differences in origin (arterial, venous, or capillary). p < 0.05 (Wilcoxon rank-sum test). (F) Representative immunohistochemical images and (G) analyses of vWF⁺/PLCγ2⁺ colocalization in human diabetic and ND wound-edge. n = 6. Scale bar, 50 μm. White arrows show major blood vessels. ∗p < 0.05 (Student’s t test). (H) Selection of CDH5⁺ tissue elements (green) and collection before and after the laser-capture microdissection (LCM). n = 7 ∗p < 0.05 (Student’s t test) (I) Schematic of human PLCγ2 promoter analyzed through bisulfite sequencing in LCM-captured CDH5⁺ tissue elements. Increased methylation of PLCγ2 promoter was found in diabetic wound edge LCM-captured CDH5⁺ elements (45%) compared to ND wound edge (11%). n = 12 clones from four subjects. ∗p < 0.05 (Student’s t test). T2DM, type 2 diabetes mellitus.

Figure 2

The methylation status of the Plcγ2 promoter in the skin of acute diabetic mice (A) UCSC Genome Browser excerpt showing presence of a CpG island containing 62 CpGs in the promoter of murine Plcγ2 gene indicating that its expression can be governed by DNA methylation marks. (B) Schematic diagram showing STZ-induced diabetic mice. (C) Blood glucose level in normal C57BL/6 and STZ-induced diabetic C57BL/6 at day (d)15 used in this study (n = 5).(D) Methylation status of a region of Plcγ2 promoter (mm10_chr8:117,498,149-117,498,925) analyzed through bisulfite sequencing (methylated CpG, black; unmethylated CpG, white) (n = 10 clones from four animals). (E) Distribution of methylated and unmethylated CpGs in Plcγ2 promoter (ND, top; STZ-induced diabetic, bottom). (F) Expression of Plcγ2 mRNA was analyzed by RT-qPCR. Actb was used as internal control (n = 6 and 5). (G) Representative IHC analysis of PLCγ2 (green) in normal murine skin (n = 7 and 8). (H) Intensity analysis of the images. Scale bar, 100 μm; ∗p < 0.05, Student’s t test. STZ, streptozotocin.

Figure 3

Ischemic conditions prompt DNA hypermethylation in the promoter region, leading to the subsequent downregulation of Plcγ2 in murine skin (A) Schematic diagram showing creation of ischemic monopedicle flap on the dorsum of C57BL/6 mice to study the effect of ischemia on Plcγ2 promoter methylation. (B) Monopedicle dorsal skin flap experiment showing samples collected from different flap regions. The extent of ischemia was categorically characterized by dividing the flap into three parts: proximal (green box), intermediate (blue box), and distal (red box) from cephalad attachment. Scale bar, 5 mm. (C) Methylation status of a region of Plcγ2 promoter (mm10_chr8:117498097-117498981) analyzed through bisulfite sequencing (methylated CpG, black; unmethylated CpG, white) (n = 22, 20, and 19 clones from eight subjects). (D) Distribution of methylated and unmethylated CpGs in Plcγ2 promoter proximal (top), intermediate (middle), and distal (bottom). (E) Representative IHC analysis of PLCγ2 (red) in the monopedicle flap containing all three abovementioned regions. (F) Intensity analysis of the IHC images. Scale bar, 100 μm; n = 4 ∗p < 0.05 intermediate vs. proximal; §p < 0.05 distal vs. proximal (one-way ANOVA, followed by Tukey’s HSD post hoc test).

Figure 4

Endothelial-targeted demethylation of Plcγ2 improves VEGF therapy on diabetic wound-edge vascularization (A) Vector components used for targeted demethylation of Plcγ2 promoter in STZ-induced diabetic mice using CRISPR-dCas9 approach. Endothelial cells targeting was achieved by CDH5 promoter-driven gRNAs. (B) Schematic diagram showing TNT-mediated delivery of VEGF ORFs or/and endothelial demethylation cocktail in bipedicle ischemic wound model of STZ-induced diabetic mice. (C) Blood glucose level in STZ-induced diabetic mice used in this study. n = 17, ∗p < 0.05 (Student’s t test). (D) Respective PeriMed laser speckle-assisted perfusion images (scale bar, 5 mm) and (E) their analysis data presented as arbitrary perfusion units. TNT procedure was done with scramble-gRNA, VEGF only, demethylation cocktail only, or VEGF plus demethylation cocktail CDH5 promoter-driven (endothelial) Plcγ2 gRNA in presence of dCas9-TET1CD cocktail. Perfusion was calculated based on the ratio of the wound area vs. distant normal skin. n = 6–11, ∗p < 0.05 gRNA PLCγ2 vs. scramble-gRNA; ^§p < 0.05 gRNA PLCγ2 + VEGF vs. scramble-gRNA; ^#p < 0.05 gRNA PLCγ2 + VEGF vs. VEGF (two-way ANOVA, followed by Tukey’s HSD post hoc test). (F) Blood flow and pulse pressure at the wound edge were measured using the pulse wave Doppler feature of Vevo 2100. The pulse wave profiles at d10 are shown (left). The increases in blood flow and pulsation as indicated by increase in peaks heights are shown for the respective groups compared to the control. The percentage of VTI and the calculated pulse pressure are shown as bar graphs (right). n = 6, ∗p < 0.05 (one-way ANOVA, followed by Tukey’s HSD post hoc test). VTI, velocity time integral; scr gRNA, scramble-gRNA; VEGF, VEGF ORFs; gRNA PLCγ2, gRNAs for Plcγ2 promoter; STZ, streptozotocin. (G) Wound closure was monitored at different days after wounding in abovementioned murine bipedicle ischemic wounds (D) subjected to TNT by digital planimetry (scale bar, 5mm) (H). n = 5–11, ∗p < 0.05 gRNA PLCγ2 vs. scramble-gRNA; ^§p < 0.05 gRNA PLCγ2+VEGF vs. scramble-gRNA; ^#p < 0.05 gRNA PLCγ2+VEGF vs. VEGF (two-way ANOVA, followed by Tukey HSD post hoc test).

Figure 5

Endothelial cell-specific demethylation of Plcγ2 promoter during VEGF therapy increases angiogenesis pathway in diabetic ischemic wounds (A–D) Immunohistochemical analysis of different vascular markers in the ischemic d10 wound-edge tissue from STZ-induced diabetic mice. The wound edges were treated with demethylation cocktail at d1 post surgery in absence or presence of Plcγ2 gRNAs. Following colocalization, analyses were performed: (A) vWF⁺/PLCγ2⁺ and (B) their colocalization coefficient. (C) vWF⁺/p-PLCγ2⁺ and (D) their colocalization coefficient. Scale bar, 100 μm; n = 4. ∗p < 0.05, one-way ANOVA, followed by Tukey’s HSD post hoc test. scr gRNA, scramble-gRNA; PLCγ2 gRNA, gRNAs for Plcγ2 promoter; STZ, streptozotocin; vWF, von Willebrand factor.

Figure 6

Demethylation of the Plcγ2 promoter during VEGF therapy promotes ischemic tissue angiogenesis via activating MAPK44/42 (A–D) Immunohistochemical analysis of different vascular markers in the ischemic d10 wound-edge tissue from STZ-induced diabetic mice. The wound edges were treated with demethylation cocktail at d1 post surgery in absence or presence of Plcγ2 gRNAs. Following colocalization, analyses were performed: (A) vWF⁺/VEGFA⁺ and (B) their colocalization coefficient. (C) vWF⁺/p-44/42MAPK⁺ and (D) their colocalization coefficient. Scale bar, 100 μm; n = 4. ∗p < 0.05 one-way ANOVA, followed by Tukey’s HSD post hoc test. Scramble-gRNA; PLCγ2 gRNA, gRNAs for Plcγ2 promoter; STZ, streptozotocin; vWF, von Willebrand factor.

Figure 7

PLCγ2 overexpression in endothelial cells increases VEGF expression via activation of MAPK44/42/HIF-1α signaling (A) ELISA analyses of the expression of PLCγ2 and phospho-PLCγ2 (Tyr 753) in the whole-cell lysate. n = 4–6, ∗p < 0.05 (Student’s t test). (B) Expression levels of VEGFA were quantified in media obtained from empty vector and PLCγ2 overexpression vector transfected HMECs. n = 3, ∗p < 0.05 (Student’s t test). (C) Western blot analyses of MAPK 44/42 and (D) phospho-MAPK44/42 expression in the whole-cell lysate obtained from empty vector and PLCγ2 overexpression vector transfected HMECs. β-actin was used as the loading control. The fold change of the expression between the groups is shown below. n = 4, ∗p < 0.05 (Student’s t test). (E) Immunohistochemical analyses were performed for nucleus/HIF-1α and (F) their colocalization coefficient. n = 5. Scale bar, 100 μm; ∗p < 0.05 (Student’s t test). ORFs, open reading frames.