Innate Sex Bias of Staphylococcus aureus Skin Infection Is Driven by α-Hemolysin

This article is featured in In This Issue, p.373

Sex steroid hormones, such as estrogen and testosterone, can differentially modulate host immune responses (1–3), yet the impact of sex and the associated steroid hormones on mechanisms of disease susceptibility has historically been understudied (4, 5). This is especially true for skin-related research (6). Skin is the body’s largest organ and serves as the first line of defense against exogenous insult (7, 8). Although women have a greater incidence of dermatological disorders (6), they also demonstrate accelerated wound healing compared with men (9–12). Yet despite these differences, very few studies have directly addressed sex-dependent effects on skin disease (6).

Between 2000 and 2012, the incidence of skin and soft tissue infections (SSTIs) in the United States is reported to have increased 40% with treatment expenditures increasing from $4.4 billion to $13.8 billion (in 2012 dollars) (13). Although males are often considered to be at increased risk of infection, innate sex bias in infection susceptibility and severity varies by pathogen and site of infection (14, 15). The dominant cause of SSTIs is Staphylococcus aureus (16–18), and in the United States more than half of these infections are caused by methicillin-resistant S. aureus (MRSA) strains (16, 17). Invasive S. aureus SSTI is mediated in large part by α-hemolysin (Hla), a secreted pore-forming toxin that is a major virulence factor causing dermonecrosis and inflammation (19–22). However, despite the prevalence of S. aureus in SSTIs and the contribution of Hla to pathogenesis, sex bias in innate susceptibility to S. aureus SSTI or Hla has not been reported.

Given the significant impact of SSTIs on human health, identifying and understanding sex-dependent differences in susceptibility to S. aureus skin infections could lead to novel treatment options. A recent literature review revealed that, compared with females, males are more often carriers of S. aureus, which may predispose them to infection (23), and males are at increased risk of S. aureus bacteremia (SAB) (14, 23). Given the superior wound healing capacity of females compared with males (9, 10, 12), the protection afforded by estrogen in other models of Gram-positive infection (24, 25), and the contribution of Hla to pathogenesis (19–22), we investigated whether female sex provides innate resistance to S. aureus SSTI in an Hla-dependent manner.

Here, we used a mouse model of S. aureus SSTI (20) to demonstrate that female sex, via the sex hormone estrogen, provides innate resistance to S. aureus SSTI—supporting a physiological component behind the epidemiological data showing a >2-fold higher incidence of S. aureus SSTI in male versus female patients. Specifically, female mice had an estrogen-dependent reduction in dermonecrosis and local inflammatory cytokine levels compared with males, followed days later by significant reductions in bacterial burden. Importantly, this bias was largely driven by a sex-specific response to Hla. Inflammatory cytokine production at the site of S. aureus infection increased in females infected with an isogenic hla deletion strain (LACΔhla) compared with infection with wild-type S. aureus (LAC), consistent with previous studies using female mice (26). However, when comparing the sexes, the opposite was observed in infected male mice, with either no change or with a trend toward reduced inflammatory cytokine levels in the absence of Hla. These differences were conserved following injection of Hla alone, demonstrating independence from bacterial burden. In addition, in ex vivo studies, neutrophils from female mice demonstrated increased S. aureus bactericidal capacity compared with neutrophils from male mice. Given that neutrophils are essential for clearing S. aureus infections (27), this work suggests that sex-specific differences in both skin innate responsiveness to Hla and neutrophil bactericidal capacity play important roles in limiting S. aureus SSTI in females.

Incidence of S. aureus SSTI in male versus female patients was evaluated by searching the Clinical and Translational Science Center Health Facts database at the University of New Mexico (UNM) Health Sciences Center (HSC). The database comprises de-identified electronic health records from more than 600 participating Cerner hospitals (UL1TR001449 National Institutes of Health Clinical and Translational Science Award at the UNM). We queried the database for patients of reproductive age (postpubescent to premenopausal; 18–45 y) receiving treatment for uncomplicated SSTIs between 2011 and 2014. Based on previous studies (28, 29), SSTIs were defined by the International Classification of Diseases, Ninth Revision, (ICD-9) (30) diagnosis codes 680 (carbuncle and furuncle) and 681–682 × (cellulitis and abscess). SSTI codes were matched with same encounter ICD-9 codes for methicillin-sensitive S. aureus (MSSA) (41.11) or MRSA (41.12). Population statistics are based on the number of 18–45-year-old males or females included in the database during the same time frame.

S. aureus USA300 LAC and the LAC hla deletion mutant (LACΔhla) were provided by Dr. F. DeLeo (Rocky Mountain Laboratories, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Hamilton, MT) and Dr. J. Bubeck-Wardenburg (University of Chicago, Chicago, IL), respectively. Bacteria were cultured at 37°C in trypticase soy broth (TSB) to early exponential phase as described previously (31). CFUs were determined by plating serial dilutions on sheep blood agar (BD Biosciences, Franklin Lakes, NJ), and stocks maintained at −80°C in TSB with 10% glycerol.

Animal work protocols were approved by the Institutional Animal Care and Use Committee of the UNM HSC and carried out in the American Association for Accreditation of Laboratory Animal Care’s accredited UNM HSC Animal Research Facility. All animal work was performed in strict accordance with recommendations in the Guide for the Care and Use of Laboratory Animals (32) and the U.S. Animal Welfare Act (33). C57BL/6J and BALB/c mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were acclimated for a minimum of 7 d prior to use. Ovariectomized (OVX) C57BL/6J mice and sham surgery controls were purchased from The Jackson Laboratory. These mice were allowed to recover over several weeks (≥10 as previously described (25, 34)) prior to exogenous administration of vehicle or 17β-estradiol (E2).

The mouse model of dermonecrosis was performed according to previously published protocol (20), with no attempt to target a specific phase of the estrous cycle. Twenty-four hours before infection, 8–12-wk-old male and female mice were anesthetized by isoflurane inhalation, and Nair (Church and Dwight, Ewing, NJ) was used to remove hair from the right flank. On the day of infection, early exponential phase S. aureus at the indicated inoculum was delivered in 50 μl of USP-grade saline (Braun, Irvine, CA) by s.c. injection into the right flank. Mice were weighed prior to infection and daily thereafter. Injection sites were photographed daily and area of dermonecrosis determined by ImageJ analysis (35). At the time of sacrifice, mice were euthanized by CO₂ asphyxiation and skin surrounding the injection site (2.25 cm²) excised and homogenized prior to serial dilution and plating on sheep blood agar for CFU determination. Homogenate was centrifuged at 12,500 × g for 10 min and supernatant (clarified homogenate) stored at −80°C for later cytokine analysis. Stock solutions of E2 (Sigma-Aldrich, St. Louis, MO) were prepared in ethanol then diluted 10-fold in 0.9% NaCl, 0.1% BSA, 0.01% Tween 20 for injection. Where indicated, mice received i.p. injections of 10 μg/kg E2 or vehicle control on day −2 through day +2 relative to infection. Recombinant Hla for dermonecrosis studies was prepared as previously described (36), and 1 μg of purified Hla in 50 μl USP-grade saline was injected s.c. into the right flank as indicated.

Cytokines and myeloperoxidase (MPO) were measured in clarified supernatant from homogenized skin prepared as described above. Cytokine levels were determined using a custom designed multiplex assay (Millipore, Billerica, MA), and MPO levels measured using the ELISA Mouse MPO DuoSet kit (R&D Systems, Minneapolis, MN), each according to the manufacturer’s directions. For cytokine levels below the limit of detection, one-half of the lowest standard concentration was used.

For histologic examination, skin samples were collected on day 3 postinfection or post-Hla injection and a total of six lesions were evaluated for each sex and treatment group. Briefly, a 1-cm² sample of skin containing a lesion or an inoculation site was flattened on thin cardboard, fixed 24–48 h in neutral buffered formalin, then stored in 70% ethanol. The skin sample was bisected through the center of the lesion or site of inoculation and processed routinely for embedding in paraffin, sectioning at 5-μm thickness, and staining with H&E. H&E-stained slides were scanned by an Aperio CS2 scanner (Leica Biosystems, Buffalo Grove, IL) and morphometry was performed using the HALO image analysis platform (Indica Labs, Albuquerque, NM) to determine the length of necrotic epidermis and the area of dermal necrosis in each lesion. Because the necrotic material in abscesses adhered poorly to slides and was often lost during tissue processing, quantification of abscess size is not included.

Infection site tissue (2.25 cm²) was collected in RNALater (Qiagen, Valencia, CA). RNA was isolated and purified using Qiazol and RNeasy Kits (Qiagen), respectively, according to the manufacturer’s directions. High-capacity cDNA reverse transcription kits with RNAse inhibitor, random hexamer primers (Applied Biosystems, Foster City, CA), and a PTC-200 Peltier thermocycler (Bio-Rad, Hercules, CA) were used to generate cDNA. Quantitative PCR was performed on a ViiA 7 Real-Time PCR system (Applied Biosystems) using TaqMan Gene Expression master mix (Applied Biosystems). Gene expression was quantified using QuantStudio software (Applied Biosystems) relative to hprt using Prime Time Predesigned quantitative PCR assays for nlrp3, asc, casp1, or il-1b (Integrated DNA Technologies, Coralville, IA).

Frozen clarified abscess homogenates were thawed rapidly at 37°C and protein concentration measured by A₂₈₀ (Nanodrop 1000 Spectrophotometer; Thermo Fisher Scientific, Wilmington, DE). Equal amounts of total protein were separated by SDS-PAGE on a 4–12% Bis-Tris Bolt gel in MES-SDS running buffer (Life Technologies, Grand Island, NY), prior to transfer to nitrocellulose membranes. Membranes were blocked with 3% nonfat milk in TBS (20 mM Tris, pH 7.5, and 150 mM NaCl) for 1.5 h at 22°C, then probed overnight at 4°C with rabbit anti-mouse anti-pro–caspase-1 + p10 + p12 Ab (Abcam, Cambridge, MA) or rabbit anti-mouse IL-1β (Abcam) in TBS plus 1% nonfat milk. After washing with TBS with 0.1% Tween 20, membranes were developed using SuperSignal Maximum Sensitivity Substrate (Thermo Fisher Scientific, Waltham, MA) following incubation with goat anti-rabbit IgG poly-HRP–conjugated secondary Ab (Thermo Fisher Scientific). Imaging was performed using a Protein Simple FluorChem R imaging system (ProteinSimple, Santa Clara, CA) and band intensity determined using Image Studio Lite software (LI-COR Biosciences, Lincoln, NE) relative to total protein loaded based on SYPRO Ruby staining (Lonza, Allendale, NJ).

The indicated S. aureus isolates were prepared for ex vivo killing assays as follows: 1 d prior to the assay, 2 × 10⁷ CFU/ml of S. aureus was cultured (37°C, 220 rpm) in TSB along with 50 nM autoinducing peptide 1 (Biopeptide, San Diego, CA) to stimulate quorum sensing. After 5 h, bacteria were washed twice with phenol red–free HBSS (Life Technologies, Thermo Fisher Scientific) by centrifugation at 1800 × g at 4°C for 4 min. Bacteria were briefly vortexed and sonicated, suspended in HBSS to 6 × 10⁶ CFU/ml, and stored on ice at 4°C until use. On the day of the assay, bacteria were opsonized in 10% autologous mouse serum for 20 min at 37°C with gentle rotation (10 rpm). Bacteria were then plated for CFU determination.

On the day of the assay, bone marrow was collected by flushing mouse femurs and tibias with ice cold HBSS containing 2 mM EDTA using a 26-gauge needle. RBCs in bone marrow were lysed in 0.2% sodium chloride, after which neutrophils were isolated by density gradient centrifugation using Histopaque 1119 and 1077 (Sigma-Aldrich). Isolated neutrophils were suspended in assay buffer (phenol red–free HBSS with calcium/magnesium [Lonza, Walkersville, MD], 20 mM HEPES, 1% charcoal stripped FBS [JR Scientific, Woodland, CA]) and primed with 100 ng/ml mouse TNF-α (BioLegend, San Diego, CA) for 30 min at 37°C with 5% CO₂.

To allow for phagocytosis, TNF-α–primed neutrophils and opsonized S. aureus were combined at a multiplicity of infection (MOI) of 1 and incubated for 15 min at 37°C with gentle end-over-end rotation. The mixture was then centrifuged at 200 × g (7 min) to remove nonphagocytosed bacteria, before resuspending the pellet in room temperature assay buffer. A sample was sonicated in PBS containing 1% Triton X-100 before serial dilution and plating on sheep blood agar to determine the number of bacteria phagocytosed (t0 CFUs). Percent phagocytosed bacteria was determined relative to the number of opsonized bacteria. The remaining sample was incubated for 1 h at 37°C with gentle end-over-end rotation, before serial dilution and plating for CFU determination (t60 CFUs). Percent killing was calculated as follows based on CFUs at t0 and t60: ((t0–t60) / t0) × 100.

S. aureus USA300 LAC at 1 × 10⁵ CFU/ml was grown overnight in TSB at 37°C with shaking in the presence of vehicle control or the indicated concentrations of E2. Growth was measured by OD at 600 nm (OD₆₀₀) recorded at 15 min increments using a Tecan Systems (San Jose, CA) Infinite M200 plate reader.

Statistical analyses were performed using Prism 7 software (GraphPad Software, La Jolla, CA). Variance was determined by F test and data analyzed as appropriate by unpaired t test or Mann–Whitney U test for nonparametric data. Necrosis data were analyzed by unpaired t test with Welch correction for unequal variance. Multiple comparison analyses were performed by ANOVA with post hoc tests as indicated.

To gain insight into potential sex bias in the incidence of S. aureus SSTI in humans, we queried the UNM HSC’s Clinical and Translational Science Center Health Facts database of electronic health records for patients of reproductive age (postpubescent to premenopausal; 18–45 y) receiving treatment for MSSA or MRSA SSTI between 2011 and 2014. We found a significantly increased incidence of infection in males compared with females (Table I), resulting in an odds ratio of >2.3 for male infection. Given that these findings likely result from both behavioral and biological factors (1), we focused on the biological contribution of sex to S. aureus SSTI.

Using an established mouse model of S. aureus SSTI (20), we compared outcomes in male and female mice on days 1 and 3 after s.c. infection with the well-characterized, highly virulent, MRSA isolate USA300 LAC (37). Although differences in dermonecrosis did not reach significance on day 1 postinfection, by day 3, near the peak of pathogenesis (38), female C57BL/6J mice showed significantly reduced dermonecrosis (>2.5-fold reduction) compared with males (Fig. 1A, 1B). This effect was not restricted to C57BL/6J mice as infected female BALB/c mice also displayed reduced day 3 dermonecrosis versus males (>1.8-fold reduction) (Supplemental Fig. 1A, 1B). Histologically, the lesions in C57/BL6J males and females had a similar appearance, with a clearly delineated central region of epidermal necrosis overlying an area of dermal necrosis which extended into the panniculus carnosus. An intense inflammatory cell infiltrate, consisting predominantly of neutrophils, occupied the dermis adjacent to the central area of epidermal and dermal necrosis. An abscess consisting of degenerate neutrophils filled the subcutis at the base of the lesion (Fig. 1C). Although similar in character, lesions in male and female mice varied significantly in extent, with males having greater lengths of epidermal necrosis (9.57 ± 0.98 versus 6.35 ± 0.75 mm, p = 0.0277) and greater areas of dermal necrosis (1.49±0.57 versus 0.74 ± 0.36 mm², p = 0.0238) than females.

Differences in dermonecrosis between the sexes appeared to be largely independent of bacterial burden at the site of infection, which differed only slightly, but significantly, on day 1 between C57BL/6J males and females (8.9 and 8.6 log CFUs, respectively) (Fig. 1D) and on day 3 between male and female BALB/c mice (8.5 versus 8.2 log CFUs, respectively) (Supplemental Fig. 1C). Similarly, weight loss, a measure of overall morbidity, was reduced on day 1, but not day 3, postinfection in female C57BL/6J mice versus male controls (Fig. 1E, BALB/c shown in Supplemental Fig. 1D). We next asked if there were differences between the sexes in the cytokine response at the site of infection. Although inflammatory cytokines can drive bacterial clearance (38, 39), inflammation at the site of S. aureus SSTI can also contribute to disease severity without improving bacterial clearance (40). Here, however, as might be expected from reduced dermonecrosis, on both days 1 and 3 postinfection female C57BL/6J mice displayed reduced levels of the proinflammatory cytokines IL-1β (2.6-fold, 2.9-fold, respectively), TNF-α (2.0-fold, 2.9 fold), IL-6 (6.9-fold, 2.5-fold), and CXCL1 (3.1-fold, 3.1-fold) at the site of infection compared with males (Fig. 1F). However, IL-10 levels did not significantly differ between the sexes at these time points, suggesting that the differential regulation is cytokine specific. A similar trend was observed for BALB/c mice (Supplemental Fig. 1E). Together, these results demonstrate an innate sex bias in the severity of S. aureus SSTI, with female sex providing a protective benefit against pathogenesis and inflammation.

Sex-specific responses to S. aureus–secreted virulence factors such as Hla could account for the pronounced differences in dermonecrosis and inflammatory cytokine production in infected mice. To test this, we s.c. infected male and female C57BL/6J mice with wild-type S. aureus (LAC) or an isogenic hla deletion mutant (LACΔhla) (41). Due to the trend toward increased bacterial clearance in LAC-infected female versus male mice on day 3 postinfection (Fig. 1D, Supplemental Fig. 1C), we postulated there would be a significant difference in bacterial clearance later postinfection. We compared outcomes over the course of a 7-d infection (see Supplemental Fig. 2 for data from days 1 to 3 postinfection with LACΔhla). By day 7 postinfection, female mice infected with LAC again showed significantly reduced dermonecrosis, but also significantly increased bacterial clearance, compared with infected males (Fig. 2A–C). During LACΔhla infection, male mice showed small areas of dermonecrosis at the site of infection on day 7 (Fig. 2A, 2B), suggesting that males may be more susceptible than females to S. aureus virulence factors other than Hla. Importantly, in contrast to LAC-infected mice, there was no significant difference in day 7 bacterial clearance between male and female C57BL/6J mice infected with LACΔhla (Fig. 2C). These findings suggest that sex-specific differences in innate immune control of S. aureus infection are dependent on the ability of the pathogen to express Hla.

In a previous study of S. aureus skin infection using female BALB/c mice, Hla was shown to suppress local inflammatory cytokine production and neutrophil influx (26). However, by comparing both sexes, our findings suggest that the local inflammatory response during S. aureus SSTI is not only dependent on Hla, but on the sex of the host as well. To test this, we compared inflammatory cytokine levels in the skin of male and female C57BL/6J mice on day 7 postinfection with LAC or LACΔhla. As expected given the reduced dermonecrosis and bacterial burden compared with males (Fig. 2A–C), female mice infected with LAC had lower local levels of IL-1β, TNF-α, IL-6, and CXCL1 (Fig. 2D). Interestingly, unlike earlier time points (Fig. 1F), day 7 postinfection females also had reduced expression of the anti-inflammatory cytokine IL-10, suggesting that the sex differences in cytokine production are both time and cytokine specific. In contrast to LAC-infected mice, however, there was no difference in cytokine levels between male and female mice infected with LACΔhla. Consistent with findings in BALB/c females (26), local IL-1β, TNF-α, and CXCL1 levels were increased in S. aureus infected C57BL/6J female mice in the absence of Hla (Fig. 2D). However, this was not the case with male mice, which showed a trend toward cytokine reduction in the absence of Hla, although this did not reach statistical significance. Statistical analysis (two-way ANOVA) of these findings revealed significant interactions between host sex and Hla for each cytokine (IL-1β, p = 0.0235; TNF-α, p = 0.0132; IL-6, p = 0.0391; CXCL1, p = 0.0088; and IL-10, p = 0.0258). These findings, together with sex- and Hla-dependent differences in pathogenesis and bacterial clearance, demonstrate that interactions between host sex and S. aureus expression of Hla mediate innate sex bias in murine models of S. aureus SSTI.

Hla pore formation in host cell membranes leads to NLRP3 inflammasome activation, resulting in caspase-1 cleavage and activation, and cleavage and secretion of the mature inflammatory cytokine IL-1β (42, 43). Notably, estrogen-mediated suppression of inflammasome activation in another disease model has been reported (44). To gain insight into the lower levels of IL-1β in the skin of LAC-infected female versus male mice, we measured local transcription of inflammasome component genes nlrp3 and asc, as well as casp1 and il-1β, at different time points postinfection. Transcription of nlrp3 was significantly reduced at the site of infection on days 1 and 3, as was il-1β on day 3 postinfection, in females relative to males (Fig. 3A, 3B, Supplemental Fig. 3). In contrast, there was no difference in transcript levels of asc and casp1 between the sexes at either time point (Supplemental Fig. 3C). However, Western blot analyses showed that although there was no difference in infection site levels of pro–caspase-1 or pro–IL-1β between LAC-infected males and females (Fig. 3C–F), both active caspase-1 and mature IL-1β levels (45) were significantly reduced in the females. Together, differences in local transcription of nlrp3 and il-1β, reduced cleavage of caspase-1 and IL-1β, and reduced IL-1β secretion in the skin of LAC-infected female versus male mice suggest the potential for sex-dependent regulation of NLRP3 inflammasome activation in response to S. aureus infection.

The well-characterized role of Hla in NLRP3 inflammasome activation, together with potential differences in regulation of inflammasome activation, suggested that females would be innately resistant to Hla-mediated pathogenesis, independent of the presence of S. aureus. To address this, we s.c. injected male and female mice with purified Hla. Compared to males, female mice displayed a >4-fold reduction in dermonecrosis on day 3 postinjection (Fig. 4A, 4B). Histologically, lesions consisted of clearly defined regions of epidermal and dermal necrosis, with a mild granulocytic infiltrate in the dermis; no abscessation was seen (Fig. 4C). These lesions were observed in all male mice treated with Hla, but in only a single female mouse. Thus, there were marked differences between the two sexes in average length of epidermal necrosis (14.14 ± 1.81 versus 0.38 ± 0.92 mm, p < 0.0001) and average area of dermal necrosis (2.83 ± 0.83 versus 0.04 ± 0.11 mm², p = 0.0004). Compared to males, female mice also had significantly reduced local inflammatory cytokine levels on both days 1 and 3 postinjection (Fig. 4D), whereas IL-10 levels were below the limit of detection (data not shown). Overall, our findings demonstrate sex-specific innate resistance to Hla-mediated pathogenesis.

Estrogen is known to support wound healing as well as modulate trafficking and function of innate immune cells important for bacterial clearance (9, 10, 12, 46–50), suggesting that ovariectomy would reduce protection during S. aureus SSTI. To test this, female mice were OVX or given a sham operation (SO), and allowed to recover for ∼10 wk (25, 34) prior to s.c. infection with S. aureus. Consistent with a role for estrogen in innate resistance to S. aureus SSTI, OVX mice displayed increased dermonecrosis (1.7-fold) compared with SO mice, as well as slightly increased local bacterial burden (7.5–8.0 log CFUs) (Fig. 5A, 5B). Importantly, short-term treatment of OVX mice with exogenous estrogen in the form of E2 restored protection against S. aureus dermonecrosis and reduced weight loss (Fig. 5C–E), although there was no significant difference in bacterial burden with short-term E2 administration (Fig. 5F). Furthermore, as expected based on differences in dermonecrosis, local inflammatory cytokine levels were increased in OVX mice and reduced with E2 treatment (Fig. 5G). In contrast to OVX mice, E2 treatment of SO mice did not further improve disease outcome (Fig. 5C–F), suggesting a threshold for E2-mediated protection in females. Furthermore, short-term treatment of male mice with E2 did not limit the severity of S. aureus SSTI (Supplemental Fig. 4A–C), possibly due to the suppressive effects of testosterone on the immune response (51, 52). Therefore, together with the absence of direct effects of E2 on S. aureus growth (Supplemental Fig. 4D), these findings demonstrate that estrogen is an important regulator of the female innate response to S. aureus pathogenesis.

Neutrophils are critical for clearance of S. aureus (27), and both sex- and hormone-specific differences in neutrophil trafficking and responses to exogenous stimulation have been reported (50, 53–57). Excessive neutrophil accumulation, however, can contribute to inflammation resulting in pathogenic effects (36, 58, 59). Given the differences in day 7 bacterial burden between male and female mice infected with LAC (Fig. 2C), but not with LACΔhla, we first asked whether this was associated with differences in neutrophil presence at the site of infection. To address this, we measured MPO levels at the site of infection, which is widely used as a marker for phagocyte, frequently neutrophil, presence (38). We found that MPO levels were reduced in the skin of LAC-infected females versus males on both days 3 and 7 postinfection (Fig. 6A). However, the opposite was observed in LACΔhla-infected mice, with MPO significantly increased (day 7) in females compared with males. Although these differences could result from varying rates of phagocyte influx or clearance between the sexes, they suggest that increased bacterial clearance in LAC-infected female mice is not dependent on a sustained, elevated phagocyte presence at the site of infection.

Given the importance of neutrophils for clearance of S. aureus (27), increased S. aureus clearance despite reduced MPO levels in LAC-infected female versus male mice could result from differences in neutrophil bactericidal capacity. To determine whether neutrophils from female mice are more efficient at clearing S. aureus compared with those from male mice, we isolated bone marrow neutrophils and assessed phagocytosis and killing of S. aureus ex vivo. Consistent with our in vivo data, female neutrophils showed significantly increased phagocytosis (1.5-fold) and killing (percent phagocytosed bacteria killed) of LAC compared with males (Fig. 6B), resulting in a 2.2-fold reduction in CFUs. Similar to findings in male versus female mice, OVX mice also had higher local MPO levels on day 3 postinfection compared with SO controls, and neutrophils collected from OVX females had reduced ex vivo bactericidal efficacy compared with those from SO mice (Fig. 6C, 6D). Neutrophils from females also showed increased phagocytosis of LACΔhla (1.5-fold) and total CFUs killed (1.4-fold) compared with males, although the percent of phagocytosed bacteria killed did not vary (Fig. 6E). However, neutrophils from both sexes were better able to kill S. aureus ex vivo in the absence of Hla (Fig. 6B, 6D). This suggests that other factors, such as increased inflammatory cytokine production in LACΔhla- versus LAC-infected females, may have contributed to equivalent bacterial clearance in male and female mice infected with LACΔhla (Fig. 2C, Supplemental Fig. 2E). Alternatively, expression of Hla and similarly regulated virulence factors following phagocytosis of S. aureus by human neutrophils has been shown to contribute to cell lysis (60). It is unclear whether murine neutrophil function is impacted by these factors in vivo and whether female neutrophils are uniquely resistant to such effects compared with male neutrophils. Regardless, these results indicate that sex-dependent differences in neutrophil bactericidal capacity may also contribute to the innate sex bias in S. aureus SSTI.

Despite epidemiological evidence of sex differences in infectious diseases and the ability of sex steroids to differentially regulate immune cells (1–3, 14), the biological impact of sex and sex hormones on mechanisms of infection susceptibility and pathogenesis have been largely understudied. Here we used a murine infection model to show that female sex, via the sex hormone estrogen, provides innate resistance to SSTIs caused by the important human pathogen, S. aureus. Remarkably, this resistance is mediated by significant interactions between host sex and the S. aureus–secreted virulence factor Hla. We found suppressed inflammation in the skin of female versus male mice infected with an Hla-producing isolate of S. aureus (LAC)—suppression which was lost when infected with an isogenic hla deletion mutant (LACΔhla). This differential response was conserved following injection with Hla alone, demonstrating a direct, sex-specific response to Hla, independent of bacterial burden. Together, these findings suggest that sex-specific differences in skin innate responsiveness to Hla play an important role in limiting S. aureus SSTI in females. In this era of growing antibiotic resistance, a detailed understanding of the molecular mechanisms driving this sex bias may reveal novel therapeutic approaches to promote host innate defense against infection.

Epidemiological studies have shown pathogen-specific sex bias in susceptibility and severity of a variety of infectious diseases (14, 15, 61–63). However, such reports outnumber research studies aimed at defining the physiological mechanisms driving these differences in infection susceptibility. Regarding S. aureus, males, as well as females using hormonal contraception, show increased S. aureus nasal colonization rates compared with females not on hormonal contraceptives (64). Given that colonization is associated with increased risk of infection, perhaps it is not surprising that males have an increased incidence of SAB versus females (23, 65–68). However, females have increased risk of mortality from SAB, with reports on MRSA and community-acquired SAB showing higher 7- and 30-d mortality rates, respectively, for female versus male patients (23, 69, 70). Consistent with this epidemiological data showing human females are more likely to die of SAB, female mice show reduced survival following i.v. S. aureus infection compared with males (71), and administration of exogenous estrogen to females further increased mortality compared with vehicle-treated controls (72). In contrast, male rabbits were more susceptible than females in a lethal s.c. implant infection model using a toxic shock syndrome toxin (TSST-1) producing S. aureus isolate (73). In this model, neutering or estrogen administration reduced lethality in male rabbits (73, 74), whereas spaying increased infection lethality in females (73), pointing to the powerful contributions of sex steroid hormones on infection susceptibility and disease progression. Similarly, compared with placebo-treated OVX controls, E2 administration to OVX mice protected against S. aureus arthritis and bone loss without affecting bacterial burden (75). More recently, Corriden et al. (76) reported on a mouse model of systemic S. aureus infection in which tamoxifen, a selective estrogen receptor agonist/antagonist, increased survival and enhanced bacterial clearance. This outcome was attributed to tamoxifen’s effects on neutrophil migration and function, including enhanced formation of neutrophil extracellular traps. Interestingly, the ability of tamoxifen to enhance neutrophil function resulted from an increase in intracellular ceramide production and was independent of estrogen receptors. Here, we demonstrate sex bias in S. aureus SSTI and show that sex-specific and estrogen-dependent differences in skin innate responsiveness to Hla drive this bias. It will be important in future studies to investigate sex-specific regulation of the inflammatory response to Hla ex vivo in murine resident skin cells, as well as validating our findings using human neutrophils and skin from both sexes. To better characterize the role of E2 in defense against S. aureus pathogenesis, it will also be revealing to perform more extensive studies using SO and OVX mice, together with short-term E2 administration, at additional time points postinfection using LAC versus LACΔhla or injection with Hla. Also, given the sex difference in ex vivo killing of S. aureus by murine neutrophils and the findings of Corriden et al. (76), it will be important to explore sex-specific and estrogen-dependent differences in murine and human neutrophil chemotaxis and neutrophil extracellular trap formation. Meanwhile, the work reported here, focused on SSTI as the most common form of S. aureus infection (77), expands our understanding of host-pathogen interactions in S. aureus–mediated disease. Furthermore, the male sex bias in susceptibility to S. aureus SSTI or SAB, together with the female bias toward mortality with SAB, may suggest that sex bias varies not only by pathogen, but also by the type and severity of infection.

As the first line of defense against exogenous insult, skin plays a crucial role in the host immune response (7, 8) and sex steroids regulate many aspects of skin physiology. These hormones, specifically androgens such as testosterone in males and estrogens in females, contribute to skin architecture, including differences in epidermal and dermal thickness, as well as to immune system function (78–80). With respect to SSTIs, differences in both the resolution of infection and subsequent tissue restoration or wound healing likely contribute to disease severity. Interestingly, innate sex differences in the skin are perhaps most apparent in the accelerated wound healing ability of females versus males (81). Inhibition of cutaneous wound healing and enhanced inflammation in males has been attributed to testosterone and the androgen receptor (82). In contrast, intact female or OVX mice treated with estrogen show accelerated wound healing and reduced inflammation (12). Similarly, in humans, topical estrogen accelerates cutaneous wound healing in elderly male and female patients, a finding associated with reduced neutrophil influx and increased fibronectin levels (83). Here, we show that short-term administration of E2 to OVX mice restores protection against S. aureus dermonecrosis, further demonstrating the protective role that estrogen plays in the skin. Notably, a recent investigation into the female prevalence of autoimmune diseases identified an associated female-biased transcriptomic signature in the skin that is independent of sex steroid levels (84, 85). Whether such a transcription system independently contributes to female protection against infectious diseases of the skin, and whether it can be exploited to enhance host defense against SSTIs, warrants further investigation.

Sex steroids are known to differentially modulate host immune responses, with estrogens being broadly associated with immunoenhancement and androgens with immunosuppression (3, 51, 86, 87). However, a detailed understanding of the cells and signaling pathways contributing to innate sex bias in susceptibility to a range of infectious diseases remains lacking (2, 14, 63). E2, the predominant form of estrogen produced by females, signals by binding the cytosolic/nuclear estrogen receptors ERα and ERβ, as well as the noncanonical membrane bound G-protein coupled estrogen receptor (88–91). Signaling through each receptor can result in rapid, nongenomic effects such as calcium mobilization and cAMP production, as well as slower transcriptional (genomic) effects mediated by ERα and ERβ binding to gene promoters containing estrogen response elements or via downstream G-protein coupled estrogen receptor activation of transcription factors. Estrogen receptor signaling is further complicated by far ranging but varying cell distributions, as well as differences in expression levels and the existence of different isoforms of ERα and ERβ (89–93). Similarly, androgens signal through the androgen receptor, a ligand-inducible transcription factor that regulates gene expression by binding to androgen response elements (94). Given the complexity of hormone signaling in vivo, particularly in the context of an intact immune system, it is unlikely that protection against an invading pathogen can be attributed to a single hormone or receptor. Rather, it is likely that the estrogen-mediated innate protection against S. aureus SSTI in female mice reported here results from a complex interplay among these receptors and their signaling pathways. Therefore, future studies must seek to unravel the molecular mechanisms controlling innate protection in females.

The biological impact of sex and sex hormones on disease susceptibility and treatment has historically been understudied. Fortunately, recent years have witnessed a growing recognition of this gap (4, 5), culminating in the recent National Institutes of Health call that sex be addressed as a biological variable in their funded research (95). In addition to expanding our understanding of host-pathogen interactions in S. aureus-mediated disease, this work further highlights the significant contributions of sex and sex hormones to infectious disease susceptibility and severity.

The authors have no financial conflicts of interest.