Cutaneous wound healing in aged, high fat diet-induced obese female or male C57BL/6 mice

Body composition

Over the course of the 8 week feeding experiment, an increase in body weight was stimulated by HFD in both males and females (Figure 2A, 2B; Supplementary Table 1). The most evident differences in body weight gain was detected between females (Figure 2A) and males (Figure 2B; Supplementary Table 2). The stable, linear increase in body weight was substantially greater for males relative to females, regardless of age.

The difference in body weight gain between HFD and LFD reached statistical significance within 2 weeks of diet (p<0.05) for young males and 1 week for old (p<0.05), whereas for females this was achieved after 4 weeks of diet for young (p<0.05) and 3 weeks for old (p<0.05). The different response between female and male mice to HFD indicates a clear role of sex to dietary regiment (Supplementary Tables 1–3).

Weekly body composition analyzes performed by nuclear magnetic resonance (NMR) confirmed sex differences in body weight gain (Figure 2C–2F; Supplementary Table 4–9). The increase in body weight in males and females fed HFD was achieved primarily through fat mass accumulation (Figure 2C, 2D). The statistically significant differences in body fat mass gain between old HFD vs LFD male mice were apparent after 1 week of diet (p<0.001) and continued to increase until the end of dietary program (p<0.001), whereas for old females it was delayed until the 3^rd week (p<0.05, Figure 2C, 2D; Supplementary Table 4–6). No differences in body fat content was observed in mice fed LFD.

The lean mass content did not change substantially during the 8 weeks of feeding study in the young or old females (Figure 2E). Interestingly, for males fed HFD, the lean mass content increased (Figure 2F) indicating that both fat and lean mass contributed to overall body weight gains (Figure 2B; Supplementary Table 7–9). The glucose tolerance test (GTT) performed at the end of the 8^th week showed impaired glucose tolerance for HFD mice which was most severe for males (Figure 2G).

Histological analysis of the skin at the end of 8^th week of the HFD or LFD program.

Mammalian skin is comprised of three structural layers: epidermis, dermis and subcutaneous white adipose tissue (sWAT), which in rodents is separated from dermis by a thin layer of skeletal muscle known as the panniculus carnosus (Figure 3A) [22]. Recent studies have further defined the heterogeneity of the dermis [22–24], pointing out its structural division into papillary and reticular dermis, and dermal white adipose tissue (dWAT) [22, 23] (Figure 3A) identified as the layer of adipocytes within the reticular dermis of the skin [25].

To analyze the impact of age and diet on the skin of young and old mice after LFD or HFD, we performed histological measurements of the thickness of the epidermis, the dermis, and the dWAT (Figure 3A). The thickness of epidermis demonstrated no significant differences (Figure 3B–3D) among analyzed groups. The measurement of the total dermis thickness (comprising the fibroblast-rich dermis + dWAT) showed increases in mice fed HFD regardless of age (Figure 3B, 3D, Supplementary Table 10). While measurements of the fibroblast-rich dermis alone decreased in old animals, regardless of diet (Figure 3B, 3D; Supplementary Table 13), the increase in dWAT thickness, as a result of HFD in old mice, compensated for the loss of fibroblast rich dermis (Figure 3B, 3D; Supplementary Tables 11, 12). As a consequence, dermal layer was significantly thicker in HFD relative to LFD old mice and comparable to that observed in young, HFD animals (Figure 3B, 3D).

Age, sex and obesogenic environment differentially modulate skin wound healing parameters

Skin wound healing occurs in overlapping but distinct stages characterized by inflammation, new tissue formation and remodelling [2]. The inflammatory phase is marked by recruitment to the wound of monocytes which subsequently differentiate into macrophages and together with resident cells (keratinocytes at the wound edge, fibroblasts and endothelial cells) produce MCP-1 (monocytes chemoattractant protein 1), a strongly chemotactic cytokine [26] which further mediates the recruitment of monocytes, mast cells and lymphocytes to the site of injury [27]. To assess the inflammatory phase of skin wound healing in different mice cohorts we measured levels of MCP-1 (Figure 4A), the expression of CD68 mRNA (Figure 4B) and the immunohistochemical localization of macrophage markers: Mac-2 (galectin-3) (Figure 4C) and CD68 (Figure 4D) in skin tissues. The peak of MCP-1 levels was detected at day 3 after wounding (Figure 4A; Supplementary Table 14). Skin injury in old mice evoked the highest levels of MCP-1 that was not affected by diet (LFD vs HFD) (Figure 4A; Supplementary Table 15). CD68 mRNA analysis revealed the highest levels of expression at post-wounded day 3 which gradually decreased as the skin wound healing process progressed to day 21, falling to levels comparable to those expressed by uninjured skin (Figure 4B, Supplementary Tables 16–19). Interestingly, in contrast to MCP-1, the highest levels of CD68 mRNA expression were detected for young mice, regardless sex (Figure 4B, Supplementary Table 19). The effect of age and diet on CD68 mRNA expression was apparent throughout the skin wound healing process (days 3-21) (Supplementary Table 19). Immunohistological localization of Mac-2 positive cells was identified in the epidermal layer, consistent with reports associating Mac-2 epidermal expression with differentiation and maturation of suprabasal keratinocytes [28] in the wound area, in granulation tissues and in the wound margin (Figure 4C). The accumulation of Mac-2 positive cells was observed in skin sections collected at day 3 (Figure 4C) and, to a lesser extent, day 7 (data not shown) after injury. The most robust infiltration of Mac-2 was observed in old HFD mice (Figure 4C). CD68 immunoreactivity was exclusively detected in the dermal part of the skin after wounding (Figure 4D). Positivity for CD68 was observed among cells spread throughout the post wounded dermis, blood vessels and some adipose tissue (Figure 4D). The strongest CD68 reaction product deposition at post wounded day 7 was observed in dWAT tissues localized at wound margin of old HFD mice (Figure 4D).

New tissue formation involves re-epithelialization, angiogenesis, collagen synthesis and ECM production [5]. The aim of the re-epithelialization process is to cover the wound with the new epithelium. Re-epithelialization involves keratinocyte migration, proliferation and differentiation that is essential for successful cutaneous healing [29, 30]. The analysis of post wounding skin sections stained with: haematoxylin and eosin (H&E) or immunostained for the presence of keratin 16 (Figure 4E), a marker of keratinocytes induction, showed the rate of re-epithelialization measured by percentage of area covered by newly formed epithelium in the wound. The fastest rate of wound re-epithelialization was detected for young, LFD mice: 63.4 % (±11.5) at day 3 and 100% (±19.9) at day 7 (Figure 4F). In the remaining groups of mice, the re-epithelialization rate was significantly slower. At day 3^rd after injury, young, HFD mice displayed 24.8% (±10.9), old LFD mice 25.9% (±14.0) and old HFD mice 35% (±12.2) of wound coverage. The combined effect of age and diet on re-epithelialization process was detected at day 3^rd after wounding (Supplementary Table 20–22). At the day 7, the percentage of re-epithelialization were 86.2% (±14) for young HFD, 70.5% (±17.2) old LFD and 78.4% (±12.2) old HFD mice. By day 14 wounds of all studied cohorts were fully re-epithelialized (Figure 4E, 4F).

The skin wound healing process is tightly coordinated by Mmps enzymes [31]. Mmps function as regulators of each phase of skin wound healing contributing to the inflammatory and remodelling stages through their participation in the degradation of structural extracellular matrix (ECM) components, activation of other Mmps, release of growth factors from the cell membrane or ECM, and the shedding of cell adhesion molecules [32]. The analysis of Mmp-9 (Figure 5A; Supplementary Tables 23–26) and Timp-1 (Figure 5C; Supplementary Tables 27–30) revealed a surge in mRNA expression in post-injured skin tissues at day 3 for both males and females followed by their gradual decrease to reach the initial baseline (day 0) levels at day 14. The highest levels of Mmp-9 mRNA expression at post wounded day 3 was observed for old HFD mice, indicating the impact of two parameters: age and diet (Figure 5A; Supplementary Tables 23–26). The surge in Mmp-9 expression at day 3 accompanied a burst of Timp-1 mRNA where Timp-1 expression was higher for old HFD females relative to males (Figure 5C; Supplementary Tables 27–30). The combined effect of: diet, age and sex was detected at day 3^rd after wounding for Timp-1 mRNA surge (Figure 5C; Supplementary Tables 27–30). The analysis of Mmp-9 protein expression at day 3 by Western blot confirmed that age and HFD diet stimulated its expression (Figure 5B).

Likewise, Mmp-13 is a metalloproteinase which regulates skin wound healing through participation in the growth of granulation tissue, organization of myofibroblasts, and the formation of large blood vessels [33]. Similar to Mmp-9, peak Mmp-13 mRNA expression for both females and males was observed in samples collected at day 3 and day 7 after injury (Figure 5D). In contrast to Mmp-9 (compare Figure 5A), the levels of Mmp-13 expression at post-injured day 3 and 7 was the highest in young HFD females (Figure 5D; Supplementary Tables 31–34). The two variables: age and sex at day 3^rd and three variables: age, sex and diet at day 7 act in combination to increase the expression of Mmp-13 mRNA (Figure 5D; Supplementary Tables 31–34). At post-wounded days 14 and day 21 the expression of Mmp-13 decreased and approached the baseline levels observed in non-injured skin samples (Figure 5D).

Other factors that affect the skin wound healing process are transforming growth factors beta 1 and 3 (Tgfβ-1 and Tgfβ-3). Tgfβ-1, so called pro-scarring, is one of the first cytokines to elicit inflammatory cell recruitment. The skin wound healing process evokes Tgfβ-1 mRNA (Figure 6A) and protein (Figure 6B) expression at post-wounded days 3 and 7. The highest expression of Tgfβ-1 was detected in post-wounded skin from old female mice regardless of diet (Figure 6A). The upregulated Tgfβ-1 gene expression at day 3 (p<0.001), 7 (p<0.01) and 14 (p<0.05) is substantially greater in old HFD females than relative to old HFD males (Figure 6A; Supplementary Tables 35–38). The expression levels of Tgfβ-3, a cytokine which plays an important role in wound repair and is presumed to be a potential mediator of scar reduction and skin healing improvement [34], showed the increase for both females and males at post-wounding day 7 (Figure 6C, 6D; Supplementary Tables 39–42). For old HFD females Tgfβ-3 mRNA expression remained elevated at day 14 after injury (Figure 6C) indicating the interaction among: sex (female) and diet (HFD) and post-wounding day (day 14) (Supplementary Table 42).

The restoration of post-injured skin tissues requires collagen biosynthesis and turnover that affect the post-wounded skin quality and outcome. To evaluate collagen content during skin wound healing, we analyzed collagen 1 and 3 mRNA expression, total collagen content through hydroxyproline assays, and analysis of post-wounded area on Masson’s trichrome stained histological sections (Figure 7). The highest levels of collagen 1 and 3 mRNA in uninjured and post-injured skin tissue samples were detected in young male mice (Figure 7A, 7B, Supplementary Tables 43–50) whereas the expression in aged males was significantly lower and comparable to young and old females (Figure 7A, 7B; Supplementary Tables 43–50). Evidently, age (young) and sex (males) but not diet have the combined effect on the collagen1 and 3 gene expression in the skin (Figure 7A–7B; Supplementary Table 46–50). Further analysis of collagen was performed through hydroxyproline content measurement (Figure 7C) and the histological analysis of Masson’s Trichrome stained skin sections (Figure 7D, 7E) at post-wounded day 14 and 21. The combined effect of age and diet for hydroxyproline content was detected at day 21 after wounding (Supplementary Table 51). The highest percentage of fibrosis area analyzed in histological sections was quantified for young and low fat diet mice at both day 14 and 21 (Figure 7D, 7E; Supplementary Tables 52, 53).

Animals

The experimental animal procedures performed in these studies were approved by the Ethics Committee of University of Warmia and Mazury (Olsztyn, Poland), No. 22/2015.

The C57BL/6J (B6) mice for the study were generated through colonies established at the animal facility in Institute of Animal Reproduction and Food Research of Polish Academy of Sciences. B6 mice were originally purchased from The Jackson Laboratory (Bar Harbor, ME USA). At the beginning of the experiment, adult (2-3 months), and old (16-18 months), male and female (young; n=24 per group and old; n=16-24 per group) B6 mice were assigned into separated groups fed for a period of 8 weeks either a low fat diet (LFD, 13 kcal% fat; PicoLab Rodent Diet 20 5053) or high fat diet (HFD 59 kcal% fat; TestDiet AIN-76A; LabDiet) (scheme of the experiment Fig. 1). At the end of every week of a diet (LFD or HFD), body mass was measured and body composition was analyzed. Measurements of lean and fat mass were acquired by nuclear magnetic resonance (NMR) using the Minispec LF90II (Bruker Optics). After 8 weeks of diet, glucose tolerance test (GTT) was performed, as described by Anunciado-Koza et al. [65]. Briefly, after fasting for 4 h from 8:00 to 12:00, mice received an intraperitoneal injection of 20% glucose solution (2 g/kg of body weight; Sigma-Aldrich by Merck). Blood glucose levels were determined using a blood glucose monitor (Accu-chek, Roche).

Wounding and collecting skin tissue samples

The day before wounding 5 months old (young) and 21 months old (old) female and male B6 mice, fed LFD or HFD were anesthetized with isoflurane and shaved in the dorsal area. On the following day (day 0), mice were anesthetized with isoflurane and four full-thickness excisional wounds were created on the back of mice using a 4 mm biopsy punch (Miltex). Excised skin samples (4 mm in diameter) were collected, immediately frozen in liquid nitrogen and stored at -80°C until analysis (day 0, uninjured control). B6 mice were sacrificed at days 3, 7, 14 and 21 after wounding. Post-injured skin areas were collected with 8 mm diameter biopsy punches. Tissue samples (n=6-8 per time point) for RNA or protein isolation were immediately frozen in liquid nitrogen and stored in -80°C till further use. Samples for histology were fixed in formalin (n=3-6 per time point).

Masson trichrome staining, histological measurement of the thickness of layers of the skin and analysis of fibrosis

Formalin-fixed sections of uninjured and post-injured skin tissues were deparaffinized by immersion in xylene and then rehydrated. The samples were stained with Trichrome Stain Kit (Modified Masson's) (ScyTek Laboratories, Inc.) For the histological analysis of uninjured skin and measurements of fibrosis in post-injured skin tissues, Image J image analysis software was used (National Institutes of Health (NIH) Image).

Analysis of re-epithelialization

The percentage of re-epithelialization was calculated as previously described by Low et al. [26] and Noguchi et al. [66]. Briefly, the width of wounds and the distance of migrating epithelium were measured on post-injured skin sections (days 3, 7, 14 after injury) stained with hematoxylin and eosin (H&E). The percentage of re-epithelialization was calculated according to the formula: [length of the extending epidermal tongues]/[length of the wound] x100% (number of mice n = 3-5 per group, number of measurements n = 3-10 per group) Measurements were performed using ImageJ software (NIH Image).

Monocyte chemotactic protein 1 (MCP-1) detection assay

The ELISA assay for quantitative determination of MCP-1 concentrations in tissue homogenates was performed according to the manufacturer’s protocol (Cusabio). Collected skin tissue samples from day 0 (controls), and post-wounded day 3 and day 7 were weighed before further preparation. The samples were homogenized in PBS using 100 mg tissue per 1ml of PBS. The average weight of skin tissue samples varied between 30 and 80 mg therefore an appropriately adjusted volume (between 0.3 to 0.8 ml) of PBS was used for each tissue sample. Tissue homogenates were stored overnight at -20°C. After two freeze and thaw cycles, the homogenates were centrifuged (5 minutes/5000 x g, 2-8°C) and the supernatant was collected, aliquoted and stored at -80°C until performing the assay.

Immunohistochemistry

Formalin-fixed sections of post-injured (day 3 and 7) skin tissues were processed, embedded in paraffin and sectioned. Slides were stained with Masson’s trichrome (Trichrome Stain Kit, Abcam) for collagen presence using standard protocol. Immunohistochemical staining for the presence of Mac-2 was performed with anti- LGALS3 (1:2500, mouse monoclonal, LSBio), CD68 (1:200, rabbit polyclonal, Abcam) and Cytokeratin 16 (1:300, rabbit polyclonal, LsBio). Antibody binding was detected with the ABC complex (Vectastain ABC kit, Vector Laboratories, Inc.). In control sections primary antibodies were substituted with non-specific-immunoglobulin G (IgG). Peroxidase activity was revealed using 3.3′-diaminobenzidine (Sigma-Aldrich by Merck) as a substrate. Slides were counterstained with hematoxylin (Sigma-Aldrich by Merck). Sections were visualized using Olympus microscope (BX43), photographed with Olympus digital camera (XC50) and analyzed with Olympus CellSens Software.

RNA isolation and quantitative RT-PCR

Total RNA was extracted from skin samples using TRI Reagent (Sigma-Aldrich by Merck) according to the manufacturer’s instructions. Quantity and quality of RNA was verified on NanoDrop 1000 (Thermo Fisher Scientific) and through analysis of agarose gel after electrophoresis. cDNA was synthesized from 500 ng of total RNA using High-Capacity cDNA Reverse Transcription Kit with RNase Inhibitor (Applied Biosystems by Thermo Fisher Scientific). To measure the levels of Collagen 1, Collagen 3, Hprt-1, Mmp-9, Mmp-13, Tgfβ-1, Tgfβ-3 and Timp-1, mRNA expression, Single Tube TaqMan® Gene Expression Assays (Life Technologies Thermo Fisher Scientific) were used. Amplification was performed using 7900HT Fast Real-Time PCR System under conditions: initial denaturation for 10 min at 95°C, followed by 40 cycles of 15 sec at 95°C and 1 min at 60°C. Each run included standard curve based on aliquots of pooled skin RNA. All samples were analyzed in duplicates. Hprt-1 was chosen as the most stable housekeeping gene during cutaneous wound healing after analysis described previously [67]. mRNA expression levels were normalized to the reference gene Hprt-1 and multiplied by 10.

Protein isolation and western blot analysis

Frozen post-injured skin samples collected at day 0, 3 and 7 were homogenized in liquid nitrogen with mortar and pestle in 500 μl of RIPA buffer containing protease inhibitor cocktail (Sigma-Aldrich by Merck), Phosphatase inhibitor cocktail (Sigma-Aldrich by Merck) and Phenylmethanesulfonyl fluoride (PMSF, Sigma-Aldrich by Merck) further sonicated (3 × 20 s, 20 kHz) with Vibro-Cell VCX 130 PB ultrasound sonicator (Sonics). Protein concentration was measured by the infrared (IR)-based protein quantitation method using a Direct Detect® Infrared Spectrometer (Merck) as previously described [68]. Thirty micrograms of proteins were separated on 12% SDS-polyacrylamide gels and transferred onto polyvinylidene difluoride membranes (Merck Millipore). The membranes were incubated separately with antibodies: anti- Mmp-9 (1:1000, rabbit polyclonal, Merck Millipore), anti-Tgfβ-1 (1:500, rabbit polyclonal, Biorbyt), anti-Tgfβ-3 (1:200, rabbit polyclonal, Abcam) and anti-β-actin (1:1000 mouse monoclonal, Abcam), followed by fluorescent secondary antibodies IRDye 800 (1:5000 goat anti-rabbit, Rockland) and Cy5.5 (1:10000, goat anti-mouse, Rockland). Bands were visualized using the ChemiDoc™ Touch Imaging System (Bio-Rad) and analyzed using Image Lab Software (Bio-Rad) according to the manufacturer’s protocol.

Measurement of hydroxyproline

Collagen assay was prepared as described before [69]. Briefly, 8 mm diameter frozen skin punches, were homogenized in 2 ml of phosphate buffered saline (Sigma-Aldrich by Merck) and stored in 4 °C overnight. The next day, 0.5 ml aliquots were hydrolyzed with 0.25 ml of 6N HCL for 4.5 h at 120 °C. To construct a standard curve hydroxyproline concentrations from 0 to 20 μg/ml were used (Sigma-Aldrich by Merck). 20 μl of each sample and standard curve point was added to a 96 well plate and incubated for 20 min at room temperature with 50 μl of chloramine T solution (282 mg chloramine T, 2 ml n-propanol, 2 ml distilled water and 16 ml citrate acetate buffer [5% citric acid, 7.24% sodium acetate, 3.4% sodium hydroxide and 1.2% glacial acetic acid]). Next, 50 μl of Ehrlich's solution (2.5 g 4-(dimethyloamino) benzaldehyde, 9.3 ml n-propanol and 3.9 ml 70% perchloric acid) was added and the plate was incubated for 15 min at 65°C. After cooling down the samples the plate was read at 550 nm on a microplate reader (Multiskan Sky Microplate Spectrophotometer, Thermo Fisher Scientific™).

Statistical analysis

In order to analyze impact of gender, age, diet and timepoint on body parameters linear mixed-effects models were used, in which mice were included as random intercept and mentioned factors and all possible interactions were included as fixed effects. Similarly, mixed-effect models were prepared for skin tissues parameters; however the random effects were insignificant, so linear models were used instead. Based on such models, lsmeans were calculated (least-squared means - means calculated based on model’s coefficients) and compared (for mixed-effects models Kenward-Roger method for degrees of freedom was used; Tukey p-value adjustment used where appropriate). Results are presented in tables and plots (lsmeans with standard error). Moreover, optimal models were obtained using backward stepwise method. All calculations were performed in R (ver. 3.5.3) using packages: lme4, lmerTest, emmeans and tidyverse.