Mitochondrial Bioenergetics of Functional Wound Closure is Dependent on Macrophage-Keratinocyte Exosomal Crosstalk

Abstract

Tissue nanotransfection (TNT)-based fluorescent labeling of cell-specific exosomes has shown that exosomes play a central role in physiological keratinocyte-macrophage (mϕ) crosstalk at the wound-site. Here, we report that during the early phase of wound reepithelialization, macrophage-derived exosomes (Exo_mϕ), enriched with the outer mitochondrial membrane protein TOMM70, are localized in leading-edge keratinocytes. TOMM70 is a 70 kDa adaptor protein anchored in the mitochondrial outer membrane and plays a critical role in maintaining mitochondrial function and quality. TOMM70 selectively recognizes cytosolic chaperones by its tetratricopeptide repeat (TPR) domain and facilitates the import of preproteins lacking a positively charged mitochondrial targeted sequence. Exosomal packaging of TOMM70 in mϕ was independent of mitochondrial fission. TOMM70-enriched Exo_mϕ compensated for the hypoxia-induced depletion of epidermal TOMM70, thereby rescuing mitochondrial metabolism in leading-edge keratinocytes. Thus, macrophage-derived TOMM70 is responsible for the glycolytic ATP supply to power keratinocyte migration. Blockade of exosomal uptake from keratinocytes impaired wound closure with the persistence of proinflammatory mϕ in the wound microenvironment, pointing toward a bidirectional crosstalk between these two cell types. The significance of such bidirectional crosstalk was established by the observation that in patients with nonhealing diabetic foot ulcers, TOMM70 is deficient in keratinocytes of wound-edge tissues.

Keywords: TOMM70; functional wound closure; keratinocyte migration; macrophage-derived exosomes; macrophage–keratinocyte crosstalk; tissue nanotransfection; “don’t eat me” plasmid.

Figure 1

Application of a deep learning algorithm for tissue segmentation in WE tissue postimaging mass cytometry identified keratinocyte-mϕ crosstalk as the predominant exosome-mediated paracrine signaling. (A) Schematic diagram showing delivery of plasmid cocktails in WE tissue of C57BL/6 mice at different time points. (B) Schematic diagram showing the staining of murine WE tissue with metal-conjugated antibodies for Imaging Mass Cytometry. Localization of Exo_κ and Exo_mϕ in dermis d3 post-TNT. Scale, 100 and 10 μm. (C) Deep learning algorithm-based marker antibodies were used for tissue segmentation. Based on the segmentation, the skin and WE tissue were divided into six segments: background (BG), hair follicles (HF), epidermis (E), dermis (D), hypodermis (HD), and adipose tissue (A). A deep learning classifier based on DNA1 and DNA2 channels with expansion of the cell body from the nuclei was used for cell segmentation. The white circle indicates the entire cell, and the red circle demonstrates nuclear compartments. (D) t-SNE plot with a binary color bar showing an abundance of F4/80 cell distributions at d0, d3, and d7 post wounding. (E) Abundance of Exo_κ in F4/80⁺ cells was plotted on a tSNE plot with a continuous color bar. (F) Heat map showing the distribution of Exo_κ in different cell types at d0, d3, and d7 post wounding. (G) t-SNE plot with a binary color bar showing an abundance of e-cadherin⁺ cell distributions at d0, d3, and d7 post wounding. (H) Abundance of Exo_mϕ in e-cadherin⁺ cells was plotted on a t-SNE plot with a continuous color bar. (I) Heat map showing the distribution of Exo_mϕ in different cell types at d0, d3, and d7 post wounding. (J) Representative high-resolution confocal microscopic images demonstrating the presence of Exo_mϕ (red) in keratinocytes (green) in murine d5 WE tissue. Scale, 100, 20, and 10 μm.

Figure 2

Macrophage-derived exosomes isolated from WE tissue showed an abundance of mitochondrial proteins. (A) Schematic diagram showing isolation of Exo_mϕ from murine WE tissue. Field emission scanning electron microscopy (FESEM) images demonstrate the presence of Exo_mϕ on RFP magnetic beads. Scale, 5 μm. (B) Particle size distribution of Exo_mϕ from murine WE tissue (n = 10). (C) Representative transmission electron microscopy (TEM) and FESEM images of Exo_mϕ. Scale, 100 nm. (D) Schematic diagram of endosomal pathway showing early, intermediate, and late endosomal markers. (E) Bead flow cytometric analysis of murine Exo_mϕ conjugated with supermagnetic Dynabeads functionalized with the RFP antibody showing binding of RAB5A_PE, RAB7A_FITC, and RAB11A_AF647 antibodies. The histogram demonstrates the shift in FITC and AF647 fluorescence after binding with the antibodies. The mean percentage of beads with Exo_mϕ was mentioned over the marker bar. (F) Antibody array of Exo_mϕ isolated from murine skin and WE tissue. ALIX, TSG101, CD63, CD81, FLOT1, exosomal marker; EpCAM = epithelial cell adhesion molecule; ANXA5, Annexin 5. *Positive control for HRP detection of exosomes derived from the human serum. (G) Exo_mϕ were isolated and quantified using NTA from murine skin and d5 WE tissue after nanotransfection with the Lyz2 plasmid cocktail (n = 10). (H) Zeta potentials of Exo_mϕ isolated from murine skin and d5 WE tissue at physiological pH (pH 7.4). Each gray dot corresponds to one technical replicate, and the blue and red dots correspond to biological replicates (n = 5). (I) Fourier Transform Infrared (FTIR) spectra of Exo_mϕ isolated from murine skin and d5 WE tissue. (J) Two-dimensional (2D) score plot constructed from principal component analysis of FTIR spectra of Exo_mϕ isolated from murine skin and d5 WE tissue (n = 3). (K) Bead flow cytometric analysis of murine Exo_mϕ conjugated with supermagnetic Dynabeads functionalized with the RFP antibody showing binding of TOMM70_AF647, Prohibitin 1_AF647, and VDAC1_AF647 antibodies. The histogram demonstrates the shift in fluorescence after binding with the antibodies. The mean percentage of beads with Exo_mϕ was mentioned over the marker bar. Data are representative of three independent experiments. Data in panels (D) and (E) are shown as mean ± standard error of the mean (SEM) and are analyzed by two-tailed unpaired Student’s t-test.

Figure 3

TOMM70 is actively packaged in Exo_mϕ that augments keratinocyte migration. (A) Representative immunocytochemistry of RAB5A (red) and TOMM70 (green) in murine proinflammatory mϕ. TOMM70 signal was used for mitochondria morphology analysis. MiNA tool on the ImageJ interface was used to show the skeletonized images of the mitochondrial network. Colocalization of TOMM70 with RAB5A was shown as white dots. The red fluorescence of RAB5A was superimposed with the mitochondrial network (green) for the quantification of mean mitochondrial branch length. Scale, 10 μm. (B) Quantification of the mean mitochondrial branch length from areas with and without RAB5A colocalization (n = 8). (C) Schematic diagram demonstrating the experimental setup of Spatial Frequency Domain Imaging (SFDI) of the murine dorsal excisional wound. (D) Digital and SFDI photographs of murine dorsum 6 h post wounding. Scale, 1.5 mm (E). (E) Quantification of cutaneous tissue O₂ saturation at WE (<2 mm) and skin (>2 mm) analyzed from SFD images (n = 8). (F) Immunohistochemistry of TOMM70 with DAPI counterstaining in d3 murine WE tissue. White dashed lines indicate the dermal–epidermal junction. Scale, 100 μm. (G) Schematic experimental design for respirometric study and cell migration assay. (H) Representative phase-contrast microscopic images of murine keratinocytes at 0 h and 80 h showing migration following treatment with Exo_mϕ isolated from murine skin (d0) and d5 WE tissue. Scale 200 μm. (I) Keratinocyte migrations were quantified and expressed as percentage closure. For each image, the distance of the migrating front at the top, middle, and bottom was measured (gray dots) and the mean value was plotted (blue, red) (n = 5). (J) Baseline OCR in murine keratinocytes after the addition of Exo_mϕ isolated from murine skin (d0, blue dots) and d5 WE tissue (red dots) was plotted graphically. (K, L) ATP production and glycolysis in murine keratinocytes as calculated from OCR and ECAR (n = 10). (M) Energy phenotype of keratinocytes analysis as calculated from OCR and ECAR (n = 10). Data in panels (B, E, I, K, and L) are shown as mean ± SEM and are analyzed by two-tailed unpaired Student’s t-test.

Figure 4

Bidirectional crosstalk between WE mϕ and resident keratinocytes is critical for functional wound closure. (A) Design of Krt14-promoter-driven tetraspanins plasmid-connected via the IRES element with “don’t eat me”-CD47 sequence with “in-frame” GFP and RFP reporter. (B) Representative phase-contrast confocal images showing coexpression of RFP (red) and GFP (green) in murine keratinocytes. Scale, 20 μm. (C) Flow cytometric analysis of murine Exo_κ on pan CD magnetic beads showing dual positivity of GFP and RFP on “don’t eat me” Exo_{κ-GFP-CD47-RFP}. (D) Schematic experimental design. (E) Digital photographs of the excisional stented punch wound (6 mm) at d0 and d12 post wounding in C57BL/6 mice treated with either TNT_κ-GFP or TNT_{κ-GFP-CD47-RFP}. (F) Quantification of the wound area by digital planimetry following TNT_κ-GFP or TNT_{κ-GFP-CD47-RFP}. Scale, 2 mm (n = 20). (G) TEWL in C57BL/6 mice at d12 post wounding following TNT_κ-GFP or TNT_{κ-GFP-CD47-RFP}. The gray dots represent normal skin (n = 16). (H) Representative H&E staining of wounds of C57BL/6 mice at d12 post wounding following either TNT_κ-GFP or TNT_{κ-GFP-CD47-RFP}. Scale, 1000 μm. The blue vertical arrowhead represents complete reepithelialization. The red arrowhead represents the WE. (I) Morphometric analysis showing an epithelial gap, wound length, and granulation tissue area in C57BL/6 mice at d12 post wounding following either TNT_κ-GFP or TNT_{κ-GFP-CD47-RFP} (n = 6). (J) Representative coimmunofluorescence staining of F4/80 (red) with iNOS (green) and DAPI counterstaining in the WE granulation tissue at d12 post wounding in C57BL/6 mice following either TNT_κ-GFP or TNT_{κ-GFP-CD47-RFP}. (K, L) Quantification of F4/80 and iNOS intensity in WE tissue at d12 post wounding in C57BL/6 mice following either TNT_κ-GFP or TNT_{κ-GFP-CD47-RFP}. Each corresponds to one quantified ROI, except the blue and red dots, which correspond to the mean of each mouse. At least 5 ROI per mouse (n = 4 and 5). (M) Heat map showing the expression of proinflammatory markers COX2, MCP-1, and MIP-1α in WE granulation tissue at d12 post wounding in C57BL/6 mice following either TNT_κ-GFP or TNT_{κ-GFP-CD47-RFP} (see Figure S5h). Data in panels (C, F, G, I, K, and L) are shown as mean ± SEM and are analyzed by two-tailed unpaired Student’s t test.