Full experimental procedures have been presented previously (Mahmoud et al., 2016; Mey et al., 2018; Chen et al., 2022). Briefly, two groups of participants were recruited from the Chicago, IL, metropolitan area, a type 2 diabetes mellitus (T2DM) group (n = 8) and a lean healthy control (LHC) group (n = 10). Following an approximate 12 h overnight fast, blood samples were acquired via antecubital vein and subdermal needle biopsies of the vastus lateralis were obtained under local anesthesia (Lidocaine HCl 2%) using a 5 mm Bergstrom cannula with suction. All studies were approved by the Institutional Review Board of the University of Illinois at Chicago and performed in accordance with the Declaration of Helsinki. Written informed consent was obtained from all research participants during the initial screening visit.

All animal procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health) and as approved by the Institutional Animal Care and Use Committee. WT (C57BL/6J), Lepr ^db+/db+ (C57BLKS/J leptin receptor mutant strain, or db/db) exhibit an obesity-induced type 2 diabetes phenotype with total body weight of over 40 g, insulin resistance, and blood sugar level over 500 mg/dL by 9 weeks of age, which are nearly double that of WT littermate controls; Chen et al., 2022), and eNOS ^−/− mice were purchased from Jackson Laboratory (Bar Harbor, ME). Four 5 mm full thickness excisional wounds were made on the dorsal skin of mice (male, 9 weeks old) using a standard skin biopsy punch (Acuderm Inc. Fort Lauderdale, FL) under ketamine (100 mg/kg) and xylazine (5 mg/kg) anesthesia. 30 μl of IgG or neutralizing anti-CCL28 antibody (Ab) at indicated concentrations were applied topically to wounds in F-127 Pluronic gel (Sigma, St. Louis, MO; 25% gel in saline) immediately after wounding (Mirza et al., 2009). Wound size was determined as previously described (Zhao et al., 2016; Chen et al., 2020) and wound tissues were collected at indicated times for biochemical or histological analysis.

F-127 Pluronic gel, DTT, BSA, and RIPA buffer were from Sigma (St. Louis, MO). n-Octyl-β-d-glucopyranoside (ODG) was from RPI Corp (Mt Prospect, IL). Mouse neutralizing anti-CCL28 antibody and mouse ELISA kits were purchased from R&D Systems (Minneapolis, MN). Mouse anti-CCR10 and mouse anti-eNOS antibodies were from Santa Cruz Biotechnology (Dallas, TX). 4′,6-Diamidino-2-phenylindole (DAPI), fluorescently labeled secondary antibodies, Trizol and SYBR Green PCR mix were from ThermoFisher Scientific (Waltham, MA). Mouse anti-eNOS and mouse anti-actin antibodies were from BD Biosciences (San Diego, CA). Rabbit monoclonal anti-CD31 and Griess Reagent kit were from Cell Signaling Technology (Danvers, MA). Nitrocellulose membrane was from Bio-Rad Laboratories (Hercules, CA). SuperSignal West Femto Kit and Restore Western Stripping buffer were from Pierce (Rockford, IL). Skin punch biopsy tool was from Acuderm Inc. (Fort Lauderdale, FL). Cell strainer was from VWR Scientific (Franklin, MA). Control and CCL28 adenoviral expression vectors (Adv-Ctl, Adv-CCL28) were generated by Dr. Jody L. Martin (UIC Vector Core and Department of Pharmacology, University of California, Davis).

After shaving WT C57BL/6 mice (9 weeks old), 100 µl Adv-CCL28 or Adv-Ctl (1 × 10⁹ particles/ml) was subcutaneously injected into 4 labeled areas of dorsal skin per mouse. Five mm wounds were produced at each labeled area 21 days after Adv injection. Wound sizes were measured daily as described previously (Zhao et al., 2016; Chen et al., 2020), and then wounds were collected on day 7 and subsequently prepared for further analysis.

Human plasma, mouse skin wounds, or blood serum were collected, lysed, and prepared for ELISA measurements of mouse CCL28, IL-6, TNF-α, IL-1β and VEGF according to manufacturer’s instructions. For subcutaneous adipose tissue, the tissue was collected and minced. EBM-2 (without FBS) was added at a ratio of 3:1 (v/w) and the tissue was incubated for 24 h at 37°C. The conditioned medium was collected and filtered through a 70 μm cell strainer for ELISA measurements.

Total cellular RNA from mouse skin wounds and human tissue (collection of LHC and T2DM patient samples are described in Methods 2.1) was extracted using TRIzol. mRNA level used to assess relative gene expression was examined by real-time RT-PCR, employing a SYBR Green PCR mix as described previously (Chen et al., 2022). All primers were purchased from Integrated DNA Technologies IDT (Coralville, IA). Relative gene expression was determined by the ΔΔC _t method based on GAPDH levels as described previously (Chen et al., 2013).

Dorsal skin or wounds were collected at indicated times and lysed in RIPA buffer for Western blotting. For Co-IP experiments, samples were lysed in 2% ODG (Tris buffer) as described previously (Chen et al., 2012; Chen et al., 2018).

For H&E staining, slides with 5 μm tissue sections were baked at 60°C for 30 min and stained with an Autostainer XL (Leica Microsystems, Wetzlar, Germany) using an established protocol (Chen et al., 2020; Chen et al., 2022). Immunohistochemistry of CD31 was performed in skin wound sections of db/db mice (6 mice per group) as described previously (Chen et al., 2022). Briefly, for wound tissue analysis of CD31 immunostaining, formalin-fixed paraffin-embedded sections were probed with a 1:50 dilution of anti-CD31 Ab overnight at 4°C, and then 1 h at room temperature with 1:500 dilution of a secondary antibody conjugated with Alexa-555. Images were obtained with a Zeiss LSM 880 confocal microscope. Vascular density, determined from 6 high-power fields per section and represented as a percentage of vascular tissue area occupied by CD31 relative to the total area of the field of view, was analyzed using Zeiss Zen and ImageJ software as described (Chen et al., 2022).

All data were analyzed using the unpaired Student’s t-test or one-way ANOVA, followed by post hoc Tukey test. Unpaired Student’s t-test was used to compare two groups. If more than two groups were analyzed, one-way ANOVA followed by Tukey test was used. All data are expressed as mean ± SEM. Statistical significance was considered at p < .05.

Recently we reported that levels of CCL28 and CCR10 were significantly higher than CCL27 and CCR3 in the dorsal skin of diabetic mice. Thus, the CCL28/CCR10 chemokine signaling axis was chosen for further study (Chen et al., 2022). We observed an elevated level of CCL28 in plasma (Figure 1A) and subdermal biopsies from the vastus lateralis (Figure 1B) in patients with T2DM, compared to lean healthy control (LHC) donors. Similar to human samples, the level of CCL28 was significantly higher in plasma (Figure 1C) and dorsal skin (Figure 1D) in obese and diabetic db/db mice, as compared to wild type (WT) C57BL/6 mice. Previously, we reported a reduction in eNOS expression (Mahmoud et al., 2016; Chen et al., 2018) and excessive levels of CCL28/CCR10 in both patients with T2DM and diabetic mice (Chen et al., 2022). We next sought to establish a correlation between plasma and adipose tissue CCL28 levels in T2DM, as compared to WT mice. We observed an elevated level of CCL28 (by 50%; Figure 1E), IL-6 (one-fold; Figure 1F), TNF-α (2.5-fold; Figure 1G), and IL-6 (1.7-fold; Figure 1H), while VEGF level was reduced by about 50% (Figure 1I), as determined by ELISA of conditioned medium from db/db subcutaneous adipose tissue.

To investigate whether overexpression of CCL28 per se, which we found to be increased in plasma, dorsal skin and adipose tissue of db/db mice (Figures 1C–E), results in elevation of CCR10 and reduction of eNOS expression, adenovirus expressing CCL28 (Adv-CCL28) or an empty control vector (Adv-Ctl), was injected subcutaneously in the dorsal skin of WT mice. After 3 weeks, the skin was collected for ELISA and Western blot analysis. ELISA measurements indicated enhanced expression of CCL28 (Figure 2A), IL-1β (Figure 2B), IL-6 (Figure 2C), and TNF-α (Figure 2D) following infection with Adv-CCL28, compared with Adv-Ctl. Interestingly, following overexpression of CCL28, Western blotting (Figure 2E) showed significantly increased CCR10 expression (Figures 2E,F) but reduced eNOS level (Figures 2E,G) in the dorsal skin. Taken together, these results suggest that overexpression of CCL28 in dorsal skin results in elevation of CCR10 leading to downregulation of eNOS, along with augmented release of pro-inflammatory cytokines known to contribute to delayed wound healing in diabetic subjects.

Next, we investigated whether blockade of excessive CCL28 with a neutralizing antibody affects skin wound healing in diabetic mice. In a pilot dose-finding study, .5 μg/mL anti-CCL28 Ab was topically applied to dorsal skin wounds on both WT and db/db mice (Supplementay Figure S1A). We observed reduced wound sizes on day 3 (by 36% in WT mice; by 12% in db/db mice) and day 7 (by 40% in WT mice; by 12% in db/db mice) following anti-CCL28 Ab treatment, respectively, as compared with IgG control Ab (Supplemental Figure S1B), suggesting that reducing CCL28 level in the dorsal skin may improve wound healing.

To further investigate the effect of anti-CCL28 Ab on diabetic wound healing, different doses (0–4 μg) of anti-CCL28 Ab were topically administrated to dorsal skin wounds on db/db mice (Figure 3). Low doses of .5 and 1 μg anti-CCL28 Ab significantly reduced wound sizes by day 8, whereas higher doses, such as 2 and 4 μg delayed wound healing as shown on day 10 (Figure 3A). Specifically, 1 μg anti-CCL28 Ab was identified as the optimum dose for promoting wound healing (Figures 3A,B), resulting in a nearly completely closed wound by day 10 compared with the same concentration of control IgG (Figure 3C).

After treatment with different doses of anti-CCL28 Ab, skin wounds from db/db mice were collected on day 10 and prepared for further examination. eNOS expression by Western blotting indicated there was a dose-dependent effect of CCL28 Ab (Figure 4A), with maximum eNOS expression observed also at the dose of 1 μg anti-CCL28 Ab per wound. In the wound tissue lysates, levels of CCL28 (Figure 4B) and IL-6 (Figure 4C) determined by ELISA reached a minimum at the dose of 1 μg anti-CCL28 Ab indicating a reduction in inflammation by neutralization of excessive CCL28. In addition, the real-time RT-PCR level of anti-inflammatory cytokine IL-4 (Figure 4D) showed a pattern opposite to that of proinflammatory IL-6 (Figure 4C), but also reached maximum level following 1 μg anti-CCL28 Ab treatment. These data suggest that neutralization of excessive CCL28 reversed the inflammatory environment and facilitated wound healing in diabetic mice. In addition, mRNA level of eNOS (Figure 4E) paralleled that of eNOS protein (Figure 4A), with maximum levels achieved with 1 μg anti-CCL28 Ab. Interestingly, CCR10 level was also maximally reduced at the same 1 μg dose of anti-CCL28 Ab. Taken together, these data suggest that depletion of excessive CCL28 by Ab neutralization led to enhanced eNOS expression and reduced CCR10 level, along with a switch from a proinflammatory to anti-inflammatory environment that collectively are thought to have contributed to improved wound healing in diabetic animals.

With the optimum dose of 1 μg anti-CCL28, day 10 eNOS expression level determined by Western blot analysis was about 2.5-fold higher than that observed in control IgG treated db/db mouse wounds, while CCR10 level decreased by about 20% (Figure 5A). Previously, we reported that CCR10 binding to eNOS increases following addition of its ligand CCL28 (Chen et al., 2020; Chen et al., 2022) to human ECs. Consistent with this observation, inhibition of CCL28 signaling in diabetic skin wounds with a neutralizing anti-CCL28 Ab reduced the interaction between CCR10 and eNOS in mouse wounds, as assessed by Co-IP, as compared to IgG control treated wounds (Figure 5B). Interestingly, in the anti-CCL28 Ab treated wounds, we observed elevated levels of both NO (nitrite; Figure 5C) and VEGF (Figure 5D). Furthermore, anti-CCL28 treatment resulted in a thicker epidermis, as well as dermis and hypodermis, in day 10 skin wounds as observed in H&E stained wound sections (Figure 5E). Microvessel density, indicated by CD31 immunohistochemistry, was also significantly enhanced in anti-CCL28 Ab treated wounds as compared with IgG control (Figure 5F). These results suggest neutralization of CCL28 reduces CCR10-eNOS interaction, thereby increasing eNOS/NO level, which may have further lead to an increase in VEGF production as well as angiogenesis that collectively facilitated diabetic wound healing.

eNOS-deficient mice exhibit impaired wound healing and angiogenesis defects (Lee et al., 1999) and thus we assessed levels of CCL28 and proinflammatory cytokines IL-1β, IL-6 and TNF-α by ELISA of wound lysates. As shown in Supplementary Figure S2A-D, CCL28 and cytokines were significantly elevated in the dorsal skin of eNOS ^−/− mice. In addition, CCR10 mRNA level was significantly increased (Supplementary Figures S2E). Following treatment with 1 μg/mL anti-CCL28 Ab, skin wound size on eNOS ^−/− mice was significantly reduced starting on day 3, and by day 10 was 46% smaller than those in the IgG control group (Supplementary Figures S2F-S2G). These results indicate anti-CCL28 has both eNOS dependent and independent effects on wound healing. Based on these findings, neutralization of excessive CCL28 appears to have at least two effects: 1) reduced expression of pro-inflammatory cytokines and CCR10, and 2) blockade of CCR10 activation and its interaction with eNOS in endothelial cells, resulting in upregulation of eNOS/NO level and VEGF production, both of which are thought to facilitate wound healing in diabetic mice.

Based on the above findings, upregulation of chemokine CCL28 in the db/db diabetic mouse promotes inflammation and sustained CCR10 receptor activation and binding to eNOS in endothelial cells. Excessive activation of CCR10 on endothelial cells inhibits eNOS activity and downregulates eNOS expression and NO production leading to a further increase in inflammation, decrease in angiogenesis, and delay in wound healing. We propose that neutralization of excessive CCL28 with 1 μg anti-CCL28 Ab reduces and perhaps normalizes CCL28 signaling in endothelial and non-endothelial CCR10 expressing cells resulting in reduced inflammation and eNOS downregulation, thereby improving wound healing in diabetic mice by NO dependent and NO independent anti-inflammatory and proangiogenic mechanisms (Figure 6).