Skin Wound Healing Is Accelerated and Scarless in the Absence of Commensal Microbiota

The commensal microbiota consists of microorganisms that are present on body surfaces covered by epithelial cells exposed to the external environment such as skin and the gastrointestinal tract (1). It has been shown that the impact of the commensal microbiota on health and disease can be wide-ranging from protection against pathogenic microorganisms, absorption of nutrients, and vitamin production to modulating the development and homeostasis of immune system (2–6).

Wound healing is a highly dynamic process that involves a complex sequence of cellular and biochemical events. In adult mammals, after an early inflammatory stage characterized by infiltration of neutrophils and macrophages, the formation of a fibroproliferative tissue rich in immature collagen bundles and newly formed blood vessels takes place and prompt re-epithelization occurs. Finally, the maturation phase involves dermis collagen remodeling and scar budding (7) as the usual outcome of tissue repair.

Wound inflammatory leukocytes may interfere with virtually all phases of tissue repair (8–12). On one hand, neutrophils (11) and inflammatory classically activated macrophages (13) may impair wound healing. On the other hand, macrophages may range from a proinflammatory to an anti-inflammatory/angiogenic/healing phenotype that serves to resolve inflammation and promote healing (14).

Our group has demonstrated that germ-free (GF) animals (which have no commensal microbiota) display anti-inflammatory and reduced hypernociception responses in models of ischemia and reperfusion, systemic LPS, and carrageenan-induced pain (15–17). Conversely, the chronic inflammatory response in murine models of colitis, arthritis, and asthma may be exacerbated or prolonged in GF animals (18). Animals free of commensal microbiota therefore represent a valuable tool for understanding the intrinsic mechanisms involved in tissue repair, especially when considered the importance of a controlled inflammatory response after injury. However, studies evaluating the influence of commensal microbiota on the healing of skin wounds are very scarce.

The first published studies about the influence of commensal microbiota on the healing emerged in the 1960s when histological examination of oral wounds indicated no differences between GF mice and their respective conventional (CV) controls animals regarding inflammatory response and collagen production (19). In contrast, in a comparative study using GF and CV rats, a more intense inflammatory reaction in skin wounds from the latter animals was observed (20). Likewise, incisional wounds of CV mice exhibited higher tensile strength than in GF mice (21). More recently, Lai et al. (22) showed that the resident skin commensal microbiota was able to modulate local inflammatory response after injury.

In this study, we evaluated the influence of commensal microbiota on wound closure rate and tissue repair of excisional skin wounds by using GF mice. Our main findings suggest that in the absence of any contact with microbiota wound healing is accelerated and is characterized by reduced content of neutrophils, increased markers of macrophage polarization to an anti-inflammatory/wound healing phenotype, better angiogenesis stimulation into the wound bed, and reduced scar formation.

All procedures complied with the standards stated in the Guide for the Care and Use of Laboratory Animals and were conducted under conditions approved by the local animal ethics committee, Comissão de Ética no Uso de Animais/Universidade Federal de Minas Gerais (protocol number 122/11). Eight- to 10-wk-old male and female GF Swiss were derived from a GF nucleus (Taconic Farms, Germantown, NY) and maintained in flexible plastic isolators (Standard Safety Equipment, Marietta, OH) using classical gnotobiology techniques (23). Conventional CV Swiss mice were derived from GF matrices and considered conventional only after six generations in the conventional facility. The process of colonizing GF mice with microbiota from CV mice was performed as described previously (15). Briefly, fecal samples from CV mice were homogenized in saline (10%) and administered by oral gavage to GF mice. Three weeks later, these animals were conducted to excisional wound healing model, as described below. To assess whether there was adequate colonization of GF mice, fecal samples were cultured using a thioglycollate test (15). All experimental procedures in GF mice were conducted under aseptic conditions to avoid contamination of animals (16).

Mice were anesthetized and their dorsum shaved. Four excisional wounds were created on the dorsum of mice with the aid of a sterile 5-mm circular punch, removing the entire thickness of the skin. After surgery, mice were caged individually. The area of wounds was measured at the indicated time points with a digital caliper, and the results were expressed as percentage closure relative to original size (1 − [wound area]/[original wound area] × 100) (24). Suggestive signs of local infection were not detected in the wounds area.

The extent of neutrophil accumulation in the wound tissue was measured by assaying myeloperoxidase (MPO) activity, as described previously (25, 26). Briefly, after euthanasia with an overdose of anesthetic, the wound and surrounding skin area were removed and snap frozen in liquid nitrogen. On thawing and processing, the tissue (100 mg tissue/1 ml buffer) was homogenized in 0.02 mol/l NaPO4 buffer (pH 4.7) containing 0.015 mol/l Na-EDTA, centrifuged at 10,000 × g for 10 min, and the pellet was subjected to hypotonic lyses. After a further centrifugation, the pellet was resuspended in 0.05 mol/l NaPO₄ buffer (pH 5.4) containing 0.5% hexadecyltrimethylammonium bromide and rehomogenized. Suspensions were then subjected to three freeze-thaw cycles using liquid nitrogen and centrifuged for 15 min at 10,000 × g; supernatants were used for MPO assay. The assay was performed by measuring the change in OD at 450 nm using 1.6 mM 3,3′-5,5′-tetramethylbenzidine (Sigma-Aldrich, St. Louis, MO) dissolved in DMSO (Merck, Rahway, NJ) and 0.003% H₂O₂ (v/v) dissolved in phosphate buffer (0.05 M Na₃PO₄ and 0.5% hexadecyltrimethylammonium bromide [pH 5.4]). Results were expressed as the relative unit that denotes activity of MPO.

The extent of macrophages accumulation in the wound tissue was measured by assaying N-acetyl-β-d-glucosaminidase (NAG) activity, as described previously (26, 27). Briefly, the tissue (100 mg tissue/1 ml buffer) was homogenized in 0.9% saline containing 0.1% Triton X-100 (v/v) and centrifuged at 4°C for 10 min at 3000 × g; supernatants were used for NAG assay. The assay was performed by measuring the change in OD at 405 nm using p-nitrophenyl-N-acetyl-β-d-glucosaminide (Sigma-Aldrich) and 0.2 M glycine buffer (pH 10.6). Results were expressed as the relative unit that denotes activity of NAG.

Blood (5 μl) was collected from the tail vein of anesthetized mice and diluted in Turk’s solution; total cell counts were performed in a modified Neubauer chamber. In addition, a blood smear was prepared and stained with Panótico Rápido (a Diff Quick–based stain, Laborclin, Brasil). The percentage of monocytes and neutrophils was determined using standard morphological criteria (25).

Animals were euthanized by an overdose of anesthetic. Wounds and surrounding skin were removed and perpendicularly cut into two halves. One-half was immediately frozen for molecular biology studies, whereas the other one was fixed in formalin (4% in isotonic saline) and further processed for histological analyses. To determine the expression of genes associated with inflammatory and alternative macrophage activation, total RNA was obtained from wounds using TRIzol (Invitrogen, Carlsbad, CA), according to manufacturer’s instructions. Total RNA was reverse transcribed with SuperScript III (Invitrogen) as described by the manufacturer. Real-time quantitative PCR was performed on an ABI PRISM Step-One sequence-detection system (Applied Biosystems, Carlsbad, CA) using SYBR Green PCR Master Mix (Applied Biosystems). The relative expression level of genes was determined by the 2^{−ΔΔ cycle threshold (Ct)} method, and data were normalized by 18S ribosome subunit expression levels. All reactions were replicated. Primers were generated for nos2 (forward, 5′-AGC ACT TTG GGT GAC CAC CAG GA-3′; reverse, 5′-AGC TAA GTA TTA GAG CGG CGG CA-3′), IFN-γ (forward, 5′-ACA ATG AAC GCT ACA CAC TGC AT-3′; reverse, 5′-TGG CAG TAA CAG CCA GAA ACA-3′), fizz1 (forward, 5′-ACC TTT CCT GAG ATT CTG CCC-3′; reverse, 5′-CAG TGG TCC AGT CAA CGA GTA AGC-3′), dectin1 (forward, 5′-GGA ATC CTG TGC TTT GTG GTA GTA G-3′; reverse, 5′-GGA AGG CAA GAC TGA GAA AAA CCT C-3′), mrc1 (forward, 5′-TCT TTT ACG AGA AGT TGG GGT CAG-3′; reverse, 5′-ATC ATT CCG TTC ACC AGA GGG-3′), arg1 (forward, 5′-TGA CAT CAA CAC TCC CCT GAC AAC-3′; reverse, 5′-GCC TTT TCT TCC TTC CCA GCA G-3′), and 18S (forward, 5′-CGT TCC ACC AAC TAA GAA CG-3′; reverse, 5′-CTC AAC ACG GGA AAC CTC AC-3′).

Samples were collected as described above. Five-micrometer paraffin-embedded sections were stained with H&E or toluidine blue and examined under a light microscope (Olympus BX43 with a camera Olympus Q-color 5). Sections were also stained with Picro-Sirius for collagen quantification and analyzed under polarized light. Morphometric analyses were performed on digital images using ImageProPlus 7.0 Software.

The extent of epithelialization (distance that the neoepithelium had extended from the margin of the wound as defined by the presence of hair follicles in nonwounded skin; known as epithelial tongue) and scar tissue formation (area comprising from the epidermal–dermal junction down to the panniculus carnosus and from the borders of the wound—outlined using the freeform outline tool in ImageProPlus to produce a pixel-based measurement then converted to square micrometers) was determined on H&E-stained sections. The number of mast cells was determined by counting cells in the entire area of wound tissue on toluidine-stained sections. Organization and maturation of collagen bundles were assessed on Sirius Red–stained sections. All histomorphometric analyses were performed in a blind manner.

Animals were euthanized by an overdose of anesthetic and the wounds and surrounding skin area were removed and snap frozen in liquid nitrogen. On thawing and processing, the tissue was homogenized in extraction solution (100 mg tissue/1 ml), containing 0.4 M NaCl, 0.05% Tween 20, 0.5% BSA, 0.1 mM PMSF, 0.1 mM benzethoniumchloride, 10 mM EDTA, and 20 KI aprotinin, using Ultra-Turrax. The suspension was then spun at 10,000 × g for 10 min at 4°C. The supernatant was used for ELISAs, which were performed using kits from R&D Systems (Minneapolis, MN) for murine IL-10, VEGF, total TGF-β1, TNF-α, and CXCL1/keratinocyte-derived chemokine (KC), according to the manufacturer’s instructions. All samples were assayed in duplicate. The threshold of sensitivity for each cytokine/chemokine was 7.5 pg/ml.

Analyses were performed using the GraphPad Prism 5.3 software. Results are presented as the mean ± SEM. Comparisons between two groups were carried out using Student t test for unpaired data. Three or more group comparisons were carried out using one-way ANOVA, followed by Student-Newman-Keuls multiple comparisons test. Two-way ANOVA was used for graph lines to verify the interaction between the independent variables time and strain and was followed by Bonferroni posttest. A p value < 0.05 was considered significant.

First, to evaluate whether the absence of commensal microbiota had any impact on wound closure after an excisional skin injury, we followed the rate of macroscopic wound closure. We observed that GF mice exhibited a significant increase in percentage of wound closure after day 3 postwounding compared with CV animals (p < 0.001) (Fig. 1A).

These macroscopic findings were confirmed by histological assessment of epithelialization at day 7 postwounding. On H&E-stained paraffin sections, the length of the epithelial tongues (an indicative of the neoepithelium migration extent from the margin of the wound) was significantly increased in GF mice in comparison with CV control mice (p < 0.001) (Fig. 1B). These findings indicate that enhanced epithelialization contributed to the accelerated wound closure of GF mice.

All steps of the healing process are controlled by a wide variety of cytokines. We observed an increase in wound levels of the proinflammatory cytokine TNF-α after injury in both GF and CV groups. However, although in CV mice TNF-α dropped to basal levels at day 1 postinjury, in GF mice it was kept at higher levels at least till day 3 (p < 0.05) (Fig. 2A). Interestingly, levels of the anti-inflammatory cytokine IL-10 were higher in the nonwounded skin of GF mice when compared with CV mice, remaining significantly high until day 3 after wounding (p < 0.05 for all groups) (Fig. 2B). In wounds from CV mice, similar to TNF-α kinetics, levels of IL-10 decreased to basal after its peak at day 1 postinjury. Levels of CXCL1/KC, a chemoattractant for neutrophils, did not show any difference between groups (Fig. 2C). These results suggest that GF mice are able to promptly respond to skin injury by increasing levels of inflammation mediators.

Neutrophils, macrophages, and mast cells are known to rapidly be recruited into the wound area (9). To evaluate accumulation of neutrophils and macrophages into wounds, we used MPO- and NAG-based activity assays, respectively. MPO is the most abundant enzyme in neutrophils, and NAG is expressed in high levels in activated macrophages; both have been shown to be useful and reliable markers for neutrophil and macrophage infiltration, respectively (26, 27). We observed that neutrophilic accumulation was delayed and of lower magnitude into wounds of GF mice when compared with CV controls (Fig. 3A). Although neutrophil accumulation into wounds of CV animals reached the peak at day 1 after injury, this peak was displaced to day 3 in GF animals. At day 7, MPO activity in wound bed had returned to basal skin levels in both groups. As shown in Fig. 3A, at day 1 after injury, neutrophils accumulation into wounds of GF mice was significantly reduced when compared with CV mice (p < 0.001).

In wounds of GF mice, macrophage accumulation reached the peak at day 1, and levels of NAG activity were significantly higher than those in wounds of CV mice (p < 0.01) (Fig. 3B). Moreover, macrophage content in wounds of GF mice remained high until day 3 (p < 0.05), as compared with CV mice, which showed increased levels of macrophages only later, at day 7 postinjury. These data suggest that, contrary to neutrophils, macrophage accumulation in GF mouse wounds happens earlier and at higher levels than in CV mice. Supplemental Fig. 1 illustrates the predominance of neutrophils in wounds of CV mice and of mononuclear cells in wounds of GF mice at day 1 postwounding.

Of note, when we evaluated the number of circulating cells in the peripheral blood (Table I), we observed that nonwounded GF mice have ∼80% less neutrophils in the peripheral blood than CV mice. After wounding, there was a gradual increase in blood neutrophil number in GF mice, although it was still lower than in CV mice. Similarly, we observed that nonwounded GF mice had ∼60% less monocytes in the peripheral blood than CV mice. Nevertheless, 24 h after injury, although the number of blood monocytes dropped to the lowest levels in CV mice, this cell type remained basically unaltered in the blood of GF mice.

To quantify the number of mast cells at wound sites, we used toluidine blue histological staining. Similarly to macrophages, the number of mast cells in GF nonwounded skin was already higher than in CV mouse nonwounded skin. After wounding, the number of mast cells in the wound bed rose significantly and reached the highest levels at day 1 postinjury in both groups. Interestingly, although the number of mast cell decreased significantly after day 1 postinjury in CV mice, mast cells remained high throughout the observational period in GF mice (Fig. 3C).

We then evaluated which subtype of macrophages would be accumulating in the wounds of GF and CV mice. To phenotypically characterize the macrophages infiltrating the wound tissue, we analyzed the expression of mRNA for alternatively activated macrophage-related genes (Dectin-1, Mannose receptor-1, Fizz-1, and Arginase-1) and for classically activated macrophage-related genes (IFN-γ and inducible NO synthase [iNOS]). We observed that alternatively activated healing macrophage-related genes were more highly expressed in wound tissue of GF mice at days 3 and 7 after wounding, as compared with CV mice. In contrast, genes related to classically activated macrophages were less expressed in GF and highly expressed in CV mice at days 3 and 7 postinjury (Fig. 4).

The kinetics of angiogenesis in the wound tissues was analyzed by evaluating capillary density in H&E-stained sections. At day 3 after wounding, the capillary density in the wound of GF mice was significantly increased relative to CV mice (p < 0.001), and it remained high until the end of the experimental time (Fig. 5A, 5C). Levels of VEGF, an important angiogenic growth factor, were measured by ELISA. In line with angiogenesis, levels of VEGF were significantly higher in wounds of GF mice when compared with CV mice (p < 0.05) (Fig. 5B).

The extent of collagen deposition and alignment during dermal healing may determine the severity of scar tissue formation and consequently loss of function compared with the original tissue. We then looked at collagen deposition on the wound area 14 d after tissue injury by staining histological sections with Picro–Sirius red. The levels of type I larger collagen fibers, seen as bright orange to red bundles, were similar in GF and CV mice. However, levels of type III thinner collagen fibers, seen as green fibrils, were significant higher in wounds from GF compared with CV mice (Fig. 6A, 6C, 6D). Levels of total TGF-β1, a profibrogenic cytokine, were significantly higher in wounds from CV mice when compared with GF mice (p < 0.01 at day 3 and p < 0.05 at day 7) (Fig. 6B). Interestingly, scar tissue area was significantly larger in CV mice when compared with GF mice (Fig. 6E–G).

To demonstrate that alterations observed in GF mice are due to the absence of commensal microbiota, GF mice were colonized with feces of CV mice for 21 d. After that, excisional wounds were created on conventionalized (CVZ) mice. Interestingly, we observed that CVZ mice restore the phenotype of CV mice, as seen by the wound closure rate, inflammatory cell accumulation and scar tissue formation (Fig. 7). Wounding cytokines at day 3 after surgery were also unaltered (TNF-α 725 ± 261 versus 975 ± 218, IL-10 602 ± 273 versus 855 ± 188, and CXCL1 2913 ± 314 versus 3327 ± 499, CV versus CVZ). These results suggest that the accelerated wound healing phenotype observed in the skin of GF mice is in fact due to the absence of commensal microbiota.

Surface tissues, such as skin and intestinal tract, are continuously exposed to a number of microorganisms, most of which are harmless or beneficial to the host. Although there is an increasing amount of literature systematically evaluating gut and skin microbial communities diversity in healthy and disease (28) and also evaluating the role of gut microbiota on the development and modulation of immune system (29), the influence of microbiota on skin wound healing is largely unknown. In this study, we disclose the impact of microbiota on epithelization, inflammation, angiogenesis, and scar formation after excisional skin injury by a comparative study in GF and CV mice. To the best of our knowledge, the present work provides the first direct evidence for innate/genetically encoded wound healing mechanisms in adult skin excisional lesions in contrast with the ones that occur in the setting of body interactions with microorganisms.

In adult mammals, wound healing is a highly dynamic process that involves a complex and overlapping sequence of cellular and biochemical events that range from an immediate response to damage of skin cells and invading microbial signals to inflammatory and angiogenic responses and finally wound fibroplasia and scar formation (7). Damage or microbial signals activate pattern recognition receptors such as TLRs in leukocytes that then trigger antimicrobial defense and/or inflammatory signaling cascades (30). Interestingly, Lai et al. (22) showed that the resident skin commensal microbiota is able to modulate local inflammatory responses after injury in an epithelial-derived TLR dependent manner. In the case of GF mice, the only cue to stimulate the healing response just after skin injury are signals from the damage of skin structures once these animals are devoid of microbiota and the whole process occurs in sterile conditions. In other words, the response to skin injury in GF animals is triggered by host-derived damage associated molecular patterns and activation of disturbed resident cells. Activation of inflammatory cells is consequently an integral part of wound healing.

Neutrophils are one of the earliest immune cells recruited to the site of injury. The major function of these cells is to protect the host from infection by combating invading microorganisms and clearing cellular debris. However, activated neutrophils secrete a battery of bioactive substances, such as proteases and reactive oxygen intermediates, which in excess, can lead to tissue damage (8). In fact, Dovi et al. (11) demonstrated that neutropenia induced by an antineutrophil serum accelerated the rate of wound epithelial closure without altering the overall quality of the dermal healing process in mice. In this study, besides accelerated epithelial closure, we found reduced neutrophil content in 24-h wounds of GF mice, although CXCL1 levels were produced to the same extent as in CV mice in response to injury. Therefore, our data are consistent with the idea that reduced neutrophil infiltration into wound site correlates with accelerated wound closure. Mechanistically, our results suggest that the delayed infiltration of neutrophils after injury in GF mice is, at least partially, due to the neutropenia in these animals before injury and not due to reduced local production of neutrophil-related chemoattractants. Indeed, levels of CXCL1 were unaltered in GF mice as compared with their CV controls. However, we cannot exclude the possibility that neutrophils from GF mice were hyporesponsive to chemoattractants (18).

Alongside with the influx of neutrophils, circulating monocytes enter the wound in response to damage or microbes and differentiate into mature macrophages. The key regulatory role of macrophages in driving a successful tissue repair has been a matter of discussion. There is evidence to suggest that the inflammatory response might serve the function of preventing infection in detriment of a fibrotic repair with scar formation. This has been observed in embryos and PU.1 null mice that heal without excessive inflammation and scarring (12, 31, 32). Nevertheless, proper activation of wound macrophages appears central to skin repair as depletion of these cells results in severe disturbance of the healing process, for instance, by impairing angiogenesis and collagen deposition (33–35). In this study, in contrast to reduced neutrophil infiltration, we observed a significant increase in macrophage infiltration into wounds of GF when compared with CV mice, supporting a beneficial role of wound macrophages for skin repair.

One of the hallmarks of macrophages is their ability to become activated in response to exogenous and endogenous “danger” signals with the potential of enhancing inflammation. With the eventual elimination of the insult, macrophages contribute to the resolution of the inflammatory response (36). In fact, macrophages are a diverse and dynamic population of cells that can perform a wide range of critical functions in wounding healing. Activated macrophages make up a spectrum of activation status varying from a classical inflammatory phenotype (M1) to a nonclassical or alternative phenotype (M2), also referred to as “repair macrophages,” that promote wound healing and angiogenesis (14, 37, 38). Interestingly, alternatively activated macrophage-related genes were highly expressed in wound tissue of GF mice, suggesting the predominant presence of this macrophage phenotype in the absence of microbiota. In fact, the most important point when considering the role of monocytes/macrophages in skin wound healing in GF mice is possibly the predominance of cells with the M2 phenotype in the wounds of those animals. These cells are known to contribute to the resolution of the inflammatory process and to stimulate angiogenesis as well as the production of type III collagen by fibroblasts. As a consequence, we observed an increase in the kinetics of the healing process and better quality of dermis remodeling, favoring a regenerative instead of a fibrotic repair process.

Corroborating this idea, we found high levels of IL-10 in wounds of GF animals. IL-10 is a regulatory cytokine with pivotal functions in the control of inflammation and immune-mediated tissue damage. IL-10 may not only decrease the inflammatory response to injury but also create an environment favorable to differentiation of regulatory M2 macrophages and regenerative wound healing (39). In fact, IL-10 may be responsible for the scarless repair observed in fetal skin (40). Globally, this could additionally explain the accelerated wound healing without scarring and the predominant expression of genes related to M2 macrophages phenotype in the wounds of GF mice. Although IL-10 levels were increased and IL-10 can actually decrease inflammation, there were high levels of the proinflammatory cytokine TNF-α and much inflammation as seen by the increase of macrophages and mast cells in the wound bed. These data suggest a controlled inflammatory process in GF animals that favored successful wound healing.

Mast cells are able to release a variety of soluble mediators, but their function is less understood when compared with other inflammatory cells. Although traditionally viewed as effector cells of allergic reaction and parasitic diseases, an important role for mast cells in tissue homeostasis and wound healing is now increasingly recognized (41–43). Of note, in this study, we found a high and sustained infiltration of mast cells into wounds of GF animals during the whole experimental period in contrast to a transient peak of mast cell infiltration 24 h after wounding in CV mice. On one hand, this inflammatory cell type seems to play an important role in the proliferation phase where angiogenesis is essential for provision of oxygen and nutrients to the nascent tissue. Similarly to M2 macrophages, mast cells release angiogenic growth factors, including VEGF and metalloproteinases that prepare surrounding tissue for angiogenesis during skin repair (44). In fact, the rich content of M2 macrophages and mast cells strongly suggests a connection with the high levels of VEGF and high number of capillaries in the wounds of GF animals. On the other hand, mast cells can limit inflammatory skin reaction by producing IL-10 (45) that, in turn, can downregulate mast cell FcεRI IgE receptor expression supporting protection against skin allergy sensitization (46).

In addition to inflammation and angiogenesis, both macrophages and mast cells are also able to regulate fibroplasia at wound sites, for instance, by releasing TGF-β1, an important profibrogenic growth factor (47). Interestingly, GF mice showed lower levels of TGF-β1 at wound sites that directly correlated with scar formation in injured skin of these animals when compared with CV mice. It has been shown that collagen subtype deposition may predict future scar formation. In fact, the fetal skin, the best model for scarless healing, is known to contain a greater proportion of type III collagen in comparison with type I collagen and this differential collagen deposition during fetal skin healing is thought to contribute to scarless wound healing (48). In this study, we observed that GF wounds have a higher proportion of type III collagen as well as a minimum area of scar tissue.

Finally, to prove that alterations observed in GF mice after excisional skin injury were indeed due to the absence of commensal microbiota, we conventionalized GF mice by transferring feces from CV mice into GF mice. The conventionalization process restored the phenotype of CV mice as observed by the similar rate of wound closure, inflammatory cell accumulation, cytokine release, and scar tissue formation. Our results suggest that, in the absence of microbiota, skin wound healing is scarless and basically dependent on a mosaic of intrinsic mechanisms of sensing and reacting to damage plus the activation of a non-microbial-primed “immature” immune system (2, 14, 16). In this sense, a parallel may be drawn with the fetal skin wound healing that occurs in sterile conditions in utero (i.e., in the absence of microbiota). Both GF and fetuses share unique properties, including reduced neutrophil content, high levels of IL-10 and VEGF, low levels of TGF-β1 production, deposition of an extracellular matrix rich in type III collagen, and minimum scar formation after skin injury (49). This observation is suggestive of the recapitulation of aspects of the fetal regenerative phenotype in the postnatal skin of animals that never had any contact with commensal microbiota. From this point of view, one could speculate that the progressive and dynamic contact with microbiota is a key link underlining scarring after wound healing in adults. Perhaps, the acquisition of the ability to deal with infection at the site of wound healing pays the price of greater scarring under usual nonsterile conditions. Further studies are required to better understand potential fibrotic and nonhealing mechanisms of microbiota in skin.

Taken together, our findings suggest that skin wound healing is accelerated and scarless in the absence of commensal microbiota because of a controlled inflammatory process characterized by low accumulation of neutrophils and high levels of alternatively activated macrophages as well as because of increased angiogenesis at wound sites. Molecularly, it was associated to elevated levels of IL-10 and VEGF and low levels of TGF-β1 from the beginning of the healing process. Understanding how commensals regulate the healing process provides not only new directions in the pathophysiology of wounds but also could support strategies to treat wounds by manipulating microbiota.

We thank Marcelo Gomes and Valner Augusto Mussel for animal care and Ilma M. Souza for technical support.

The authors have no financial conflicts of interest.