Topical Prostaglandin E Analog Restores Defective Dendritic Cell–Mediated Th17 Host Defense Against Methicillin-Resistant Staphylococcus Aureus in the Skin of Diabetic Mice

Diabetes is a clinical syndrome associated with deficiency in insulin secretion or action. As a consequence, diabetes causes hyperglycemia, which has been associated with undermined host defense and increased susceptibility to localized and systemic infections (1). Skin and soft tissues are sites of prevalent infections primarily caused by Staphylococcus aureus in people with diabetes (2,3). Methicillin-resistant S. aureus (MRSA) is a frequent causative pathogen of complicated skin and soft tissue infections in patients with diabetes (2) that can be difficult to treat and often require hospitalization and surgery (4).

Innate immune recognition of S. aureus promotes proinflammatory signals that lead to neutrophil recruitment and abscess formation, a hallmark of S. aureus infection (5). Extensive neutrophil apoptosis is also a hallmark of S. aureus infections (5). An additional complication is that neutrophils from patients with diabetes have defective antimicrobial effector functions (6,7).

Clearance of apoptotic cells (efferocytosis) has long been associated with resolution of inflammation (8). However, efferocytosis of infected apoptotic cells (IACs) has been proposed to be an innate antimicrobial mechanism that results in pathogen destruction (9) and comprises a critical link to adaptive immune responses (10). Considering the increased neutrophil recruitment and apoptosis at sites of MRSA infection, we speculated that the uptake of infected apoptotic neutrophils by dendritic cells (DCs) dictate the type of host defense produced against S. aureus in skin infections.

In the skin, DCs are prompted to respond to intracellular and extracellular modifications that induce a rapid change in their function and phenotype accompanied by their migration into secondary lymphoid tissue (11). During this process, DCs require maturation, a process associated with upregulation of class II MHC and costimulatory molecules (CD80, CD86, CD40) (12) and consequent migration to lymph nodes in a CC-chemokine receptor 7 (CCR7)–dependent response to its ligands CCL19 and CCL21 (13). In addition, the lipid mediator prostaglandin E₂ (PGE₂) has been shown to be a key factor for DC maturation and migration through CCR7 (14). Moreover, PGE₂ is required to initiate skin immunity by increasing maturation and migration of Langerhans cells, a subtype of skin-resident DCs (14).

PGE₂ synthesis involves activation of phospholipase A2 family members that mobilize arachidonic acid from cellular membranes. Cyclooxygenase (COX) 1 and 2 then convert arachidonic acid into prostaglandin H₂ (PGH₂), which is further metabolized by specific synthases to generate specific prostanoids. Microsomal PGE synthase 1 is induced during the inflammatory response that acts mainly downstream of COX-2 to convert PGH₂ to PGE₂. The levels of PGE₂ are regulated by the local balance between the COX-2–driven synthesis and 15-hydroxyprostaglandin dehydrogenase–mediated degradation of PGE₂ or the actions of prostaglandin transporters (15). PGE₂ acts through four G-protein–coupled receptors, designated EP1–EP4, which display different tissue distributions and deliver distinct intracellular signals (16). Moreover, efferocytosis has been shown to trigger PGE₂ production (17), although the role of PGE₂ on DC functions during efferocytosis remains to be determined.

Although efferocytosis leads to secretion of anti-inflammatory mediators that promote expansion of regulatory Foxp3 T cells, efferocytosis of Escherichia coli–infected neutrophils by DCs induces interleukin-6 (IL-6), IL-23, and transforming growth factor-β (TGF-β) production, which triggers T-helper 17 (Th17) differentiation (10,18).

Th17 cells are fundamental players in host defense against S. aureus infection (19) by activating keratinocytes and resulting in the induction of neutrophil chemoattractants and antimicrobial peptides (5). Patients with defective Th17 responses exhibit increased susceptibility to S. aureus skin infections (5); however, whether patients with diabetes also exhibit defective Th17 cells during S. aureus skin infection is unknown. Reduced frequencies of IL-17–producing memory CD4⁺ T cells associated with elevated glucose and increased glycated hemoglobin A_1c have been observed in response to Streptococcus pneumoniae in patients with diabetes (20).

We found that localized deficient PGE₂ production impairs DC migration and maturation, which impairs the generation of Th17 cells and drives poor skin host defense against infection. The data show that topical PGE₂ treatment restores immune responsiveness and increases microbial clearance in diabetic mice.

Wild-type C57BL/6 mice (8–12 weeks old) were obtained from The Jackson Laboratory. Mice were maintained according to National Institutes of Health guidelines for the use of experimental animals in the Laboratory Animal Resource Center at Indiana University School of Medicine (Indianapolis, IN). Experiments were performed under a protocol approved by the Indiana University School of Medicine Animal Care and Use Committee.

Diabetes was induced by five daily intraperitoneal injections of low-dose (40 mg/kg) streptozotocin as previously described (21). Mice with blood glucose >300 mg/dL were considered diabetic. Experiments were performed using mice at 30 days after diabetes onset.

Control and diabetic mice were anesthetized (ketamine/xylazine), their backs were shaved, and MRSA USA300 LAC strain (5 × 10⁶ colony-forming units [cfu]) in 50 μL PBS was injected subcutaneously. Lesion and abscess size were monitored daily and determined by affected area calculated by the standard equation for area [area = (π/2) × length × width] (22). Mice were topically treated at the site of infection twice a day for 7 days with 0.03% misoprostol (prepared by emulsifying 150 μg misoprostol in 0.5 g petroleum gel or vehicle control).

Control and diabetic mice were injected subcutaneously with 5 × 10⁶ cfu of MRSA plus 50 μg of CellTrace FarRed DDAO-SE (Invitrogen) as suggested by the manufacturer. Brachial lymph nodes were harvested 36 or 48 h postinfection. Migrating DCs were identified as FarRed⁺CD11c⁺ cells detected by flow cytometry.

Punch biopsy specimens (8 mm) from noninfected or infected skin were harvested at different time points and used for determining bacterial counts, cytokine production, RNA extraction, and cell isolation (23). For bacterial counts, skin biopsy specimens were collected at day 7 postinfection, processed, and homogenized in tryptic soy broth media, and serial dilutions were plated on tryptic soy broth agar. Colonies were counted after incubation overnight at 37°C.

DMSO-differentiated neutrophil-like cells were incubated with green fluorescent protein-MRSA (multiplicity of infection 50) for 2 h. We observed by FACS that 90% of neutrophil-like cells were infected after 2 h of MRSA exposure (data not shown). After phagocytosis, cells were washed, and apoptosis of infected and noninfected cells was induced by using an ultraviolet irradiation crosslinker (5 mJ) and confirmed by annexin-V and 7-aminoactinomycin staining (Supplementary Fig. 2).

Bone marrow cells from control or diabetic mice were differentiated into DCs as previously described (24) in either 5 or 25 mmol/L glucose. After differentiation, bone marrow–derived DCs (BMDCs) were cocultured with IACs (1:3) for 18 h in DMEM containing 5 or 25 mmol/L glucose. Supernatants were collected, and inflammatory mediators were detected by ELISA or enzyme immunosorbent assay (EIA); maturation of BMDCs were analyzed by FACS.

After 18 h of coculture with MRSA-IACs, BMDCs were isolated using CD11c microbeads (Miltenyi Biotec), and cells (1 × 10⁶) were suspended in serum-free DMEM with 5 or 25 mmol/L glucose and plated into the upper chamber of a 24-well Corning Costar Transwell plate (5 μmol/L). To study CCR7-dependent directional migration, CCL19 (100 ng/mL) and CCL21 (100 ng/mL) were placed into the lower chamber in free-serum medium with 5 or 25 mmol/L glucose. Cells that had migrated into the lower chamber 8 h later were photographed, harvested, and counted by FACS.

Skin biopsy specimens were digested with collagenase and processed to obtain a single-cell suspension. For lymphocyte staining, skin and lymph node cells were stimulated for 5 h with 0.1 μg/mL phorbol myristic acid (Sigma), 0.5 μg/mL ionomycin calcium salt, and 10 μg/mL brefeldin A. In all circumstances, before staining with antibodies, cells were treated with anti-FcγR antibodies to prevent nonspecific antibody binding. For lymphocyte staining, cells were fixed and permeabilized using Cytofix/Cytoperm (BD Biosciences) and were stained with the antibodies indicated in the figure legends. For DC staining, cells were fixed in 1% paraformaldehyde and stained with fluorescent-labeled antibodies for 20 min.

Cytokines, including IL-6, IL-1β, IL-17A, IL-10, and TGF-β (eBioscience or R&D Systems); production of PGE₂; PGE metabolites; and 6-keto prostaglandin F1α (Cayman Chemicals) were detected by ELISA or EIA in skin biopsy homogenates or supernatants from cultured BMDCs. In all circumstances, the sensitivity limit of the assays were ∼4–5 pg/mL.

Total RNA was isolated from skin biopsy specimens using lysis buffer (Buffer RLT; QIAGEN). cDNA and real-time PCR were performed as previously published (21) using primers indicated in the figure legends on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad). Primers were purchased from Integrated DNA Technologies. Relative expression was calculated as previously described (21).

Results are shown as mean ± SEM and were analyzed with GraphPad Prism 5.0 software (GraphPad Software, San Diego, CA). For comparisons between two experimental groups, Student t test was used, and for comparisons among three or more experimental groups, one-way ANOVA followed by Bonferroni multiple comparison test was used. P < 0.05 was considered significant.

Initially, we aimed to characterize the events involved in enhanced susceptibility to MRSA skin infection during diabetes. Mice that were diabetic for at least 30 days were more susceptible to MRSA skin infection than nondiabetic mice as seen by increased lesion size and bacterial load measured at day 7 after infection (Fig. 1A–C and data not shown). We observed that MRSA skin infection in diabetic mice led to an increased number of neutrophils in the infected skin by 18 h postinfection, an event that lasted for at least 7 days (Supplementary Fig. 1). After ingestion and killing of bacteria, neutrophils become apoptotic and are eliminated by macrophages and DCs (9). When clearance is not effective, infected apoptotic neutrophils become necrotic, releasing danger-associated molecular patterns that elicit an inflammatory response. Initially, we determined the numbers of dead cells in the infected skin of diabetic and nondiabetic mice. We observed similar increased numbers of apoptotic cells among skin cells isolated 18 h or 7 days after MRSA skin infection in both control and diabetic mice (Supplementary Fig. 2A–C) and increased apoptosis in a human neutrophil-like cell line cultured with MRSA for 18 h (Supplementary Fig. 2C–E).

Bacterial infections associated with significant induction of apoptosis in host tissues preferentially induce Th17 immunity (5). Thus, we first determined the profile of the T-cell subsets in the skin of diabetic mice 7 days after infection. Diabetic mice showed decreased percentages of Th17, but no significant differences were observed in the percentage of regulatory T (Treg) or Th1 cells compared with infected nondiabetic mice (Fig. 1D–G).

Next, we assessed the production of PGE₂, an inflammatory mediator involved in Th17 differentiation, as well as IL-17A and IL-10 in the skin of control and diabetic mice at different time points after infection. Diabetic mice showed significantly reduced PGE₂ levels at day 2 and decreased levels of IL-17A and IL-10 in the skin at day 7 postinfection compared with infected nondiabetic controls (Fig. 1H–J). We then determined whether decreased PGE₂ levels in DCs cultured in low or high glucose or infected skin of diabetic mice correlated with decreased mRNA expression of enzymes involved in PGE₂ synthesis, such as Cox-1 and -2 and Ptges (which encodes microsomal PGE synthase 1). The data show that Ptges, but not Cox-1 and -2, expression is decreased in vivo and in vitro. Indeed, another important regulatory step involved in decreased PGE₂ levels in diabetic mice could be increased PGE₂ metabolism by 15-hydroxyprostaglandin dehydrogenase or deficiency in PGE₂ transporter. We then studied the presence of PGE metabolites in the infected skin of diabetic and nondiabetic mice. We did not detect expression in the skin of all groups tested, and MRSA skin infection increased the production of PGE metabolites in a similar manner in both diabetic and nondiabetic mice 48 h after infection (Supplementary Fig. 5A–C and data not shown). Together, these findings show that decreased PGE₂ foreshadows the production of IL-17A and IL-10, suggesting a temporal relationship between PGE₂ and the generation of Th17 in the infected skin of diabetic mice.

We next studied whether impaired DC migration to skin-draining lymph nodes is associated with the decrease in Th17 subsets during MRSA skin infection in diabetic mice. We tracked DC migration in vivo during MRSA skin infection and determined the numbers and percentages of migrating DCs (FarRed⁺CD11c⁺) and Th17 cells in the skin-draining lymph nodes. Diabetic mice exhibited decreased percentages and numbers of migrating DCs and Th17 cells in the lymph nodes compared with infected nondiabetic mice (Fig. 2A–F).

Langerhans cells have been described as the main cutaneous DC subset capable of inducing a Th17 response (25). Although migrating DCs were decreased in the lymph nodes, we observed an increased percentage of Langerin-positive (Langerin+) DCs in the infected skin of diabetic mice (Fig. 3A–C), suggesting that these cells are retained in the skin and cannot properly migrate to the lymph nodes. Although no differences were observed in numbers or expression levels of maturation markers in Langerin+ DCs (data not shown) in the naive skin of control and diabetic mice, MRSA infection induced MHC II expression in about the same number of Langerin+ DCs in diabetic and control skin (Fig. 3D), but Langerin+ DCs in diabetic skin had lower levels (median fluorescence intensity [MFI]) than control skin (Fig. 3E). MRSA infection also induced a similar frequency of Langerin+ DCs expressing CD86 (Fig. 3F) and at a similar MFI in diabetic and control mice (Fig. 3G). Of note, after MRSA infection, CCR7 was expressed by fewer Langerin+ DCs with reduced MFI in diabetic skin than in control skin (Fig. 3H–K), indicating that Langerhans cells from infected diabetic mice fail to express the machinery necessary to efficiently activate T cells in the draining lymph node.

We investigated whether a high-glucose environment influences maturation, migration, and generation of cytokines by DCs involved in Th17 commitment after engulfment of MRSA-IACs. Initially, BMDCs from nondiabetic and diabetic mice were differentiated in the presence of 5 or 25 mmol/L glucose. We did not detect differences in spontaneous changes in the phenotype or production of cytokines by immature BMDCs from diabetic and nondiabetic mice under resting conditions (Supplementary Fig. 3A–E). However, after recognition of MRSA-IACs under high-glucose conditions, BMDCs showed decreased expression of the maturation markers MHC II, CD86, and CCR7 (Fig. 4A, C, and D) and a decreased ability to migrate through CCR7 toward a gradient of CCL19/CCL21 compared with cells under normal glucose conditions (Fig. 5B and C).

In 5 or 25 mmol/L glucose, BMDCs showed similar capacities to recognize MRSA-IACs, and no changes in efferocytosis index (Supplementary Fig. 3F and G) or cytokine production (Supplementary Fig. 3H–J) were observed. Moreover, BMDCs from control or diabetic mice cultured in either 5 or 25 mmol/L glucose after recognition of MRSA-IACs produced increased amounts of Th17-driving cytokines IL-6, TGF-β, and IL-1β compared with the control BMDCs before efferocytosis (Supplementary Fig. 3K–N). On the other hand, we observed decreased PGE₂ levels in BMDCs cultured in high glucose and challenged with MRSA-IACs (Fig. 5A). Efferocytosis of MRSA-IACs in BMDCs enhanced Cox-2 but not Cox-1 mRNA expression in cells cultured in 5 or 25 mmol/L glucose (Supplementary Fig. 4A and B). However, we observed that efferocytosis of MRSA-IACs in 25 mmol/L glucose led to decreased Ptges and Ptger4 expression, and no changes were seen in Ptger1, 2, and 3 expression by BMDCs after efferocytosis in 5 mmol/L glucose (Supplementary Fig. 4C, E, and F), suggesting that a high-glucose milieu decreases PGE₂ production and signaling in DCs.

Next, we sought to determine whether decreased PGE₂ levels in high glucose is responsible for impaired maturation and migration of BMDCs. To test this hypothesis, we added PGE₂ in BMDC cocultures in high glucose and observed that exogenous PGE₂ restored BMDC migration (Fig. 5B and C) and CD86 and CCR7 (Fig. 4C, D, G, and H) but not CD80 (Fig. 4B) expression. Together, these data show that deficient PGE₂ production accounts for inadequate DC maturation and chemotaxis during efferocytosis of IACs, which influences later events involved in MRSA skin infection in diabetes.

Considering that exogenous PGE₂ in vitro improved BMDC migration in diabetic mice, we investigated whether topical treatment with PGE₂ in vivo improved DC migration to skin-draining lymph nodes as well as Th17 differentiation and host defense after MRSA skin infection in diabetic mice. For these studies, we used misoprostol, which is a PGE₁ analog with agonist activity for EP2, EP3, and EP4 receptors (26) used therapeutically in humans (27).

At 2 days postinfection, we detected increased percentages of migrating DCs (FarRed⁺CD11c⁺) in lymph nodes from misoprostol-treated diabetic mice compared with untreated diabetic mice (Fig. 6A and C). We also observed that misoprostol treatment in diabetic mice increased the Th17 population in skin-draining lymph nodes (Fig. 6B and D) but did not significantly alter Treg and Th1 cells (Fig. 6E and F).

Given that misoprostol improved DC migration and increased differentiation of Th17 in the lymph nodes, we speculated that misoprostol topical treatment could restore the skin host defense in diabetic mice. Diabetic mice treated with misoprostol had smaller lesion and abscess size (Fig. 7A) and lower bacterial load (Fig. 7B) than vehicle-treated diabetic mice. Indeed, the treatment of diabetic mice with misoprostol promoted an increased percentage (Fig. 7C and D) and number (Fig. 7E) of Th17 cells in the skin compared with vehicle-treated diabetic mice but not Treg and Th1 percentages (Fig. 7F and H) or numbers (Fig. 7G and I) compared with vehicle-treated diabetic and control mice.

Next, we determined whether increased skin host defense in misoprostol-treated diabetic mice was accompanied by changes in production of antimicrobial peptides, which are key mediators involved in controlling MRSA skin infection (28). We found that misoprostol-treated diabetic mice expressed significantly increased mbd2 and mbd3 in the skin compared with vehicle-treated diabetic and nondiabetic mice (Fig. 7K and L) and expressed increased, albeit not significantly, mbd1, 4, 5, and Cramp (Fig. 7J and M–O). Together, these data suggest that misoprostol treatment improves skin host defense in diabetic mice by increasing Th17 generation and expression of antimicrobial peptides.

Skin and soft tissue infections are the most important chronic complications of diabetes and may lead to hospitalization, amputation, osteomyelitis, and death (29). The higher incidence of infections in patients with diabetes is related to hyperglycemia, which is known to cause immune dysfunction (30). In the current study, we provide evidence that reduced PGE₂ in the skin of diabetic mice compromises DC maturation and migration to lymph nodes, which inhibits the development of Th17 cells and host defense. These effects correlate with inefficient bacterial clearance and progressive skin infection in diabetic mice (Fig. 8). These findings suggest new therapeutic opportunities for treating skin infections in patients with diabetes.

After being activated by infectious or inflammatory stimuli, DCs upregulate the expression of maturation marks as they migrate to lymph nodes in a CCR7-dependent manner (12). Migrating mature DCs initiate adaptive immune responses through presentation of peripherally acquired antigens to T cells. PGE₂ signaling in Langerhans cells promotes skin immune responses by supporting maturation and migration to skin-draining lymph nodes (14). Of note, initiation of skin immune responses and DC homing is severely impaired in mice lacking the PGE₂ receptor EP4 (14), highlighting the importance of PGE₂ in DC-mediated skin immunity.

The influence of hyperglycemia on inhibition of PGE₂ by endothelial cells and defective maturation and function of DCs in diabetes has been reported (31,32). As well, during wound healing, diabetic mouse skin exhibits a decreased capacity to produce PGE₂ (15). Moreover, diminished production of COX products of arachidonic acid metabolism have been described as an abnormality of polymorphonuclear leukocytes in diabetes (33). In the current study, we determined whether such effects were responsible for decreased skin DC actions and inefficient host defense in diabetes and found that at a late time point, diabetic mice maintained an open lesion, higher bacterial load, and diminished Th17 population in the infected skin compared with nondiabetic mice.

The finding of increased apoptosis during MRSA skin infection directed us to investigate whether efferocytosis by DCs was involved in PGE₂ production and Th17 differentiation during infection. Efferocytosis induces PGE₂ production (17) and is a condition for restricting Mycobacterium tuberculosis replication (9). In addition, in the absence of microbial signals, efferocytosis induces the differentiation of Treg cells, whereas the phagocytosis of E. coli–infected apoptotic neutrophils by DCs induces Th17 cell differentiation (10). Thus, apoptosis during bacterial infection has been shown to create a proper cytokine microenvironment promoted by efferocytosis that triggers Th17 (34). Here, we extend these findings to show that efferocytosis of MRSA-infected skin cells in a hyperglycemic environment impairs PGE₂ production, which affects maturation and migration of DCs, and consequently prevents Th17 development.

However, after efferocytosis of MRSA-IACs, we observed a specific reduction of PGE₂ levels in the supernatant of BMDCs cultured in high glucose, which correlated with decreased Ptges expression but not COX enzymes. We did not detect any differences in the production of prostacyclin (another prostanoid that exerts immunoregulatory properties) or PGE₂ metabolites in the skin of diabetic mice infected with MRSA. These data contrast with previous reports that showed that hyperglycemia enhances prostaglandin degradation by increasing the expression and activity of PGE₂ transporter, leading to diminished PGE₂ levels and angiogenic signaling in diabetic skin (15). The reasons for such discrepancies are unknown, but it could be due to differences in animal models, cells involved in the inflammatory response during angiogenesis, and infection. In addition to reduced PGE₂ production, decreased expression of EP4 by BMDCs after efferocytosis under hyperglycemic conditions has been observed. This effect may be related to the capacity of PGE₂ to enhance its own receptor expression because PGE₂ increases the levels of both Pgter4 mRNA and protein expression in monocytoid cell lines (35) and bone marrow stromal cells, pointing to a mechanism of autoamplification of PGE₂ action (36).

The current results show that a high-glucose environment causes similar maturation and migration defects in BMDC from control and diabetic mice, suggesting that high glucose is the sole driver of the PGE₂/CCR7/Th17 axis during MRSA infection in diabetes. The role of PGE₂ in CCR7 expression has been controversial. Recently, PGE₂ has been shown to induce CCR7 receptor oligomerization rather than expression, resulting in an efficient signaling pathway that enhances migration (13). However, together with other soluble mediators such as tumor necrosis factor-α, IL-1β, and IL-6, PGE₂ has been shown to increase CCR7 expression (37). The current data show that exogenous addition of PGE₂ to medium containing high glucose restored BMDC migration to CCL19/CCL21 chemoattractants and induced CCR7 and CD86 expression, which prompted us to investigate the in vivo effects of PGE₂ treatment on host defense against MRSA skin infection in diabetes.

Because of their roles in angiogenesis and wound healing, PGE₂ analogs have been used to treat patients with diabetic neuropathy (38). We found that diabetic mice receiving topical misoprostol treatment had decreased lesions and bacterial counts as a consequence of improved DC migration to skin-draining lymph nodes and enhanced Th17 immunity.

Th17-derived cytokines, such as IL-17A and IL-22, induce secretion of antimicrobial peptides by neutrophils and keratinocytes (39). We found that topical misoprostol treatment in diabetic mice increased expression of mbd2 and mbd3 antimicrobial peptides, which correlates with increased skin Th17 cells after infection. Although we did not explore the mechanism by which misoprostol promotes higher antimicrobial peptide expression or determine the cell source, keratinocytes have shown a stronger dependence on IL-17 for expression of antimicrobial peptides than other cell types (39). On the other hand, misoprostol might directly upregulate antimicrobial peptide expression on keratinocytes without a particular dependence of IL-17. Therefore, further investigations are needed to fully elucidate the role of PGE₂ in the antimicrobial peptide machinery. Furthermore, COX-2 activity is important for optimal production of human β-defensin 1 and 2 by human keratinocytes in vitro (40). Therefore, we suggest that improving Th17 responses in diabetic mice promotes skin host defense against MRSA through induction of antimicrobial peptide expression. Indeed, in considering the known effects of hyperglycemia in impairing macrophage and neutrophil phagocytosis and killing, which are associated with resistance to antibiotics, endogenous antimicrobial peptides are attractive sources of more sustainable antimicrobial agents.

The current data show that PGE₂ equilibrium is a key player in skin host defense. Reduced PGE₂ in diabetic skin impairs DC activation and migration to skin-draining lymph nodes, compromising the development of adaptive immunity. Delayed priming of Th1 responses against M. tuberculosis in pulmonary infections has been previously demonstrated in diabetic mice, although no association with decreased PGE₂ levels or DC migration were observed (41). In addition, PGE₂ has been shown to direct IL-23 secretion by DCs and to be crucial for Th17 development (42).

PGE₂ may have both positive and negative immune regulatory roles, showing opposite effects depending on its concentration, microenvironment, cell type, and receptor subtype (16,43,44). PGE₂ affects mainly macrophage phagocytosis and killing of bacteria and fungus (45). On the other hand, the effects of PGE₂ on macrophage functions in the skin microenvironment are still unclear. Indeed, neutrophils and macrophages in patients with diabetes have already poor phagocytosis and killing capacities, which lead to defective bacterial clearance compared with cells of control subjects (46,47). Thus, enhancing the endogenous antimicrobial machinery could be an attractive strategy to control bacterial growth when effector functions of innate cells are impaired. Thus, the use of a PGE analog may have some limitations.

In summary, this study provides evidence that PGE₂ production was compromised during recognition of MRSA-IACs by DCs in a hyperglycemic environment. Moreover, topical treatment with misoprostol improved the host defense against MRSA by enhancing DC-mediated Th17 responses and expression of antimicrobial peptides. On the basis of these results, targeting pathways to restore PGE₂ balance in diabetic skin may be a strategy to improve innate and adaptive immune responses in patients with diabetes and skin infections. Identifying and delineating levels and mechanisms by which PGE₂ improves the skin host defense in patients with diabetes may provide an avenue toward development of a therapeutic target.

Acknowledgments. The authors thank Ana Paula Moreira Serezani (Department of Medicine, Vanderbilt University School of Medicine) for suggestions.

Funding. This work was supported by the National Heart, Lung, and Blood Institute (National Institutes of Health grants HL-103777, R01-HL-124159-01, T32-AI-060519, T32-DK-007519), an American Lung Association senior training fellowship, CAPES, and FAPESP (14/17374-3, 12/23580-0, 11/17611-7).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. N.N.D. contributed to the study concept and design, experiments, data analysis, and writing of the manuscript. S.L.B. contributed to the study concept and design and experiments. A.P. contributed to the experiments and data analysis. N.L.G.-B. contributed to the study concept and design. S.W. and Y.M.S. contributed to the experiments. A.I.M. contributed to the study concept and design and data analysis. C.H.S. contributed to the study concept and design, data analysis, and writing of the manuscript. C.H.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.