Commensal Bacteria Lipoteichoic Acid Increases Skin Mast Cell Antimicrobial Activity against Vaccinia Viruses

Mast cells (MCs) are derived from a common progenitor of CD34 bone marrow cells. Released into the circulation in an undifferentiated state, MCs arrive at the periphery of the epithelia and position themselves within close proximity to blood vessels. This location allows them to come into contact with potential pathogens that initially cross the skin barrier (1, 2). MC activation subsequently leads to the expression of receptors and production of mediators that influence the activity of neutrophils, T cells, and dendritic cells (3).

A recent study from our group demonstrates the following: MC presence protects mice from vaccinia virus (VV) skin infection; MC degranulation is required for protecting mice from VV; neutralizing Ab to the L1 fusion entry protein of VV inhibits degranulation apparently by preventing S1PR2 activation by viral membrane lipids; and antimicrobial peptide (AMP) release from MC granules is necessary to inactivate VV infectivity (4, 5).

MCs express certain TLRs, which may contribute to their ability to sense and respond to invading pathogens (6, 7). TLRs are pattern recognition receptors for conserved molecular patterns of pathogenic microorganisms (8, 9). Some of the microbial components that activate TLRs have been characterized. TLR1 and TLR6 can interact with TLR2 and are activated by bacterial lipoproteins (10).

TLR2 ligands are very well-known activators of cathelicidin AMP in MCs (11) and are also the most represented ligands on the skin surface (12).

The present study aims at demonstrating that MC antiviral function can be modulated through TLR2 ligands. In our study, we demonstrate that the administration of lipoteichoic acid (LTA, a TLR2 ligand) increases MC capacity to kill VV. We also show how bacterial byproducts present at the epidermal surface maintain MC production of AMPs and enhance MC innate immune function through TLR2 activation. Our findings suggest that control of AMPs as well as MC expression of TLR2 is critical for MC antipathogenic activity. This study provides novel insight into the interaction between the skin microbiota and MCs, and suggests previously unexplored therapeutic avenues to fight skin infections.

MC-deficient mice (C57BL/6-Kit^Wsh−/−) were a donation from Dr. P. Besmer’s laboratory (Developmental Biology Program, Memorial-Sloan Kettering Cancer Center, Cornell University, NY). The animals were bred at our facility. The Veteran Affairs and Institutional Animal Care and Use Committee approved all animal experiments. These mice have been extensively studied since they were generated (13, 14). Kit^Wsh−/− mice bearing the W-sash (W^sh) inversion mutation have MC deficiency, but lack anemia and sterility. Adult Kit^Wsh−/− mice had a profound deficiency in MCs in all tissues examined, but normal levels of major classes of other differentiated lymphoid cells. These mice may develop in adulthood myeloid and megakaryocyte dysplasia in the spleen (15, 16). In our case, 20–30% mice exhibit splenomegaly. Hematopoietic abnormalities extend to the bone marrow and are reflected by neutrophilia and thrombocytosis. Kit^Wsh−/− mice can accept transplanted genetically compatible bone marrow-derived cultured MCs with normal c-kit gene expression. The reconstitution of MCs can be done intradermally by adoptive transfer of these cells via i.p., intradermal, or i.v. injection, without the development of other donor-derived hematopoietic cells (4, 17). The levels of lymphoid cells, including TCR γδ, are normal in adult Kit^Wsh−/− mice, and these animals do not exhibit a high incidence of spontaneous pathology affecting the skin, stomach, or duodenum (18, 19). Cnlp^−/− C57BL/6 mice were generated in Dr. R. Gallo’s laboratory (University of California San Diego as previously described (20). Cnlp^−/− has been backcrossed in C57BL/6 and used with C57BL/6 background for the last 5 y. Tlr2^−/− were acquired and bred in our facility. Sex-matched wild-type C57BL/6 littermate mice were used as wild-type controls throughout the study. Cathelicidin-deficient (Cnlp^−/−) mice were obtained from R. Gallo (University of California San Diego).

Primary MCs were generated by extracting bone marrow cells from the femurs of 5- to 8-wk-old mice and culturing cells in RPMI 1640 medium (Invitrogen) supplemented with 10% inactivated FBS (Thermo Fisher Scientific), 25 mM HEPES (pH 7.4), 4 mM l-glutamine, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 50 μM 2-ME, 100 IU/ml penicillin, and 100 μg/ml streptomycin. Murine rIL-3 (1 ng/ml; R&D Systems) and recombinant murine stem cell factor (20 ng/ml; R&D Systems), both shown to support the in vitro growth and differentiation of the MC precursor, were also included. After 5 wk, MCs were consistently generated, as confirmed by the expression of CD117 and FcεRI and by staining with toluidine blue. MC cultures were derived from Cnlp^−/−, Tlr2^−/− C57BL/6 mice, and their wild-type littermate mice.

For degranulation experiments, compound 48/80 was used. Compound 48/80 is used to promote MC degranulation in an IgE-independent way (21).

For TLR2 ligand experiments, cells were pretreated with either PBS or LTA (Sigma-Aldrich) at a concentration of 10 μg/ml, according to previous reports (22–24).

Cell viability was quantified by trypan blue staining (25).

The purified VV (strain Western Reserve) and recombinant VV vB5R-GFP are a gift of B. Moss (National Institutes of Health) and has been previously described (26). BS-C-1 monkey kidney cells were grown to confluence in DMEM (Invitrogen) supplemented with 10% FBS for propagation of VV. Plaque assays were performed on BS-C-1 monkey kidney cells, as previously described (27), with supernatant of tissue homogenate or cell culture media.

University of California San Diego Institutional Animal Care and Use Committee System subcommittee on animal studies approved procedures. The backs of sex-matched adult littermates were shaved, and hair was removed by chemical depilation (Nair; Church & Dwight). Mice were inoculated with 10⁶ PFU VV by scarification with 15 pricks into the skin. The volume used to administer 10 × 6 PFU VV by scarification is 10 μl/mouse. Lesions were recorded 2 d later, and sizes were measured using ImageJ 1.42q. After 2 d, skin lesions and spleens were collected and homogenized with a Mini-Beadbeater (Biospec Products) for assessing cytokine concentrations by ELISA, VV titer by plaque assay, and virus gene expression by quantitative real-time PCR. For TLR2 ligand experiments, mice were pretreated with either PBS or LTA (Sigma-Aldrich) intradermal injection of 100 μl at 0.5 mg/ml 24 h before virus inoculation. For cathelicidin experiments, MC-deficient Kit^Wsh−/− mice were reconstituted with bone marrow-derived MCs from Cnlp^−/−, Tlr2^−/− C57BL/6 mice, and their wild-type littermate mice, and 1 × 10⁷ MCs were injected intradermally to eight locations on the shaved backs of recipient mice. For cromolyn experiments, wild-type C57BL/6 littermate mice were injected i.p. with 10 mg/kg body weight of cromolyn sodium salt (Sigma-Aldrich) once every day for 4 d before experiments, and VV was inoculated on the fourth day after cromolyn injection; then cromolyn was still injected i.p. once every day until mice were sacrificed (28).

We used TRIzol reagent (Invitrogen) to isolate total RNA. We used 1 μg total RNA for cDNA synthesis by the iSCRIPT cDNA Synthesis Kit (Bio-Rad), according to the manufacturer’s instructions. We performed real-time RT-PCR in an ABI 7300 Real-Time PCR system (Applied Biosystems). We purchased the primers and probes used for real-time RT-PCR from Applied Biosystems. We did RNA analysis by using the TaqMan Master Mix reagents kit (Applied Biosystems). We used the comparative ΔΔC_T method to determine the quantification of gene expression. We normalized the target gene expression in the test samples to the endogenous reference GAPDH level and reported them as the fold difference relative to GAPDH gene expression in untreated baseline control (29). We performed all the assays in triplicate and repeated the experiments more than two times.

VV early gene expression was evaluated by using quantitative real-time PCR normalized to housekeeping gene GAPDH. Briefly, MCs were cultured in 48-well plates at a concentration of 10⁶ cells/well. VV (10⁶ PFU) was added to corresponding wells and incubated for 1 h in a total volume of 100 μl, then 500 μl MC media was added for additional 24 h. RNA was isolated from cultured cells or homogenized tissues using TRIzol (Invitrogen), according to the manufacturer’s instructions. After reverse transcription with the iScript cDNA synthesis kit (Bio-Rad Laboratories), real-time PCR was performed by using an ABI 7300 Real-Time PCR system (Applied Biosystems). The following primer sequences were used to assay for the vaccinia gene transcripts: forward, 5′-GCCAATGAGGGTTCGAGTTC-3′ and reverse, 5′-AACAACATCCCGTCGTTCATC-3′. This region of the genome encodes a subunit of a DNA-directed RNA polymerase expressed within 2 h of viral entry (30). The TaqMan probe, 6FAM-5′-ATTGAATTCTCTTCCCGCGGATGCTG-3′, was purchased from Applied Biosystems. Samples were amplified in MicroAmp Optical 96-well reaction plates (Applied Biosystems) in a 25-μl vol containing 2× TaqMan Master Mix (Applied Biosystems), 200 nM forward primer, 200 nM reverse primer, 80 nM probe, and template RNA. Thermal cycling conditions were 50°C for 2 min and 95°C for 10 min for the first cycle. Subsequently, samples were amplified for 50 cycles at 95°C for 15 s and 60°C for 1 min.

MCs were stained with 1 μg/ml anti-TLR2 Ab (eBioscience), according to the manufacturer’s instructions. Cells were analyzed with the Guava EasyCyte 8HT two-laser, six-color microcapillary-based benchtop flow cytometer (Millipore) for assessing TLR2 expression.

Human skin sections were stained with 1 μg/ml anti-LTA Ab (Abcam), according to the manufacturer’s instructions. Slides were mounted in ProLong Anti-Fade reagent with DAPI (Molecular Probes). We imaged the sections using the Bx51 research microscope (Olympus) and X-Cite 120 fluorescence illumination systems (EXFO Photonic Solutions). Some mouse skin and human skin sections were stained with toluidine blue to identify MCs.

Supernatants from in vitro MC culture were collected 24 h after VV infection (multiple of infections [MOI] = 1), and 100 μl was applied to a nitrocellulose membrane (Whatman) using S&S Minifold I Dot-Blot System (Schleicher & Schuell). The membrane was blocked in Odyssey Blocking Buffer (LI-COR) and then incubated with a rabbit polyclonal anti-mCamp Ab (0.25 μg/ml) (Abs against the CAMP peptide were prepared by Quality Controlled Biochemicals, Hopkinton, MA) in Odyssey Blocking Buffer for 1 h. Incubation with the secondary Ab, IRDye 680 goat anti-rabbit secondary Ab (LI-COR) diluted 1:15,000 in Odyssey Blocking Buffer, was performed for 1 h. Immunoreactivity was visualized using an Odyssey Infrared Imaging System (LI-COR), according to the manufacturer's instructions. All incubations were performed at room temperature.

All data are presented as mean ± SD. At least three independent experiments were performed to assess the reproducibility of the different experiments. The two-tailed t test and one-way or two-way ANOVA with Bonferroni’s posttest of GraphPad Prism Version 4 were used to determine significance between two groups or multiple groups. For all statistical tests, p values <0.05 were considered statistically significant (*p < 0.05, **p < 0.01, ***p < 0.001).

Previous studies demonstrated that the presence of bacteria recruits MCs to the infected area (31, 32). We wanted to also prove that bacterial byproducts, present at the epidermal surface, are a relevant in vivo mechanism to enhance MC innate immune function, not only when an infection is present, but also under physiological conditions. We used an anti-LTA Ab specific for Staphylococcus epidermidis LTA (LTA-SE) to stain normal human skin for presence of this ligand (Fig. 1A–F). As shown in Fig. 1A–F, LTA from S. epidermidis is abundant in the epidermis and follicles and arrives in close contact with MCs, and therefore, positioned to stimulate MCs in close contact with the skin barrier. Secondly, we stained the same section of skin with toluidine blue (Fig. 1C, 1F) to identify the presence of MCs. To further investigate whether the presence of S. epidermidis products could modify MC recruitment, we injected the skin of mice with LTA and after 24 h we determined the number of MCs present in the dermis. LTA recruits MCs to the skin surface, as demonstrated in Fig. 1G. Injection of LTA into the skin did not induce clinical inflammation of the skin (data not shown), confirming what has been previously described (33). However, a small but significant increase in TNF-α (1-fold increase) was noted (Fig. 1H).

After we proved that the presence of LTA from the surface induces MC recruitment at the skin interface, we proceeded to demonstrate that the observed MC presence was functionally positive. First, we studied the direct effect of LTA on MC antimicrobial function.

TLR signaling pathways in MCs is a strong activator of their innate immune system (6), especially of the AMP cathelicidin (11, 34). It is known that TLR2 and TLR4 ligands activate cathelicidin production in MCs (11, 34). Therefore, we looked at the expression of bone marrow-derived receptor-MC TLR2 after in vitro maturation and found that TLR2 is highly expressed on mature MCs (Fig. 2A). When stimulated with LTA, MCs begin to express Cnlp mRNA, the gene for cathelicidin, at 2 h, with a peak at 4 h (Fig. 2B). We recently published that MC cathelicidin is essential for MC capacity to kill VV (35), and in a different publication (4) we verified that LTA directly induces Cnlp gene expression in MCs. Therefore, we hypothesized that LTA will increase MC resistance to viral injury.

LTA is known to induce higher cathelicidin expression in MCs compared with LPS as measured by FACS analysis (11), and LTA from S. epidermidis is abundant in the epidermis and follicles and arrives in close contact with MCs, as we showed in Fig. 1. Therefore, we chose LTA to stimulate TLR2 in MCs to test the hypothesis that MCs stimulated by TLR ligands and bacterial byproducts increase their load of AMPs and in turn their antiviral capacity. We challenged mouse MCs with LTA for 24 h and subsequently inoculated with VV at a MOI of 1. After an additional 24 h, the DNA-directed RNA polymerase expression (30) of VV was measured. Fig. 2C shows that LTA-pretreated wild-type MCs significantly increased antiviral activity through suppressing the viral growth rate when compared with non-LTA–pretreated wild-type MCs. To verify that the observed suppression of viral growth was mediated by cathelicidin, we analyzed the expression of cathelicidin in the wild-type cells after LTA application with and without VV infection (Fig. 2D). The increased resistance to the viral infection after LTA pretreatment was mirrored by an increase in cathelicidin expression (Fig. 2D). Cells pretreated with LTA and then treated with VV demonstrated an even stronger elevation of cathelicidin expression, suggesting that LTA pretreatment provides a better innate response to VV. Therefore, MCs preconditioned through TLR2 are more efficient at combating VV.

We demonstrated that LTA both induces cathelicidin expression in MCs and enhances the antiviral activity of these cells in a TLR2-dependent manner. Next, we will demonstrate more conclusively that the LTA-induced antiviral activity is primarily due to the induction of cathelicidin. MCs deficient in cathelicidin (Cnlp^−/− MC) were differentiated in vitro from bone marrow of Cnlp^−/− mice. The cells were stimulated with LTA for 24 h and inoculated with VV. As shown in Fig. 2E–H, wild-type cells pretreated with LTA showed a decreased expression of VV mRNA (Fig. 2E), a decreased percentage of cell death (Fig. 2F), and a lower number of VV virions measured by GFP-VV by FACS (Fig. 2G, 2H) when compared with cathelicidin-deficient (Cnlp^−/− MC) LTA-pretreated cells. We also measured mouse β-defensin 3, which has been implicated in viral response (36), and we found no expression of mouse β-defensin 3 in MCs even after stimulation with VV or LTA (data not shown).

Experiments were performed in vivo to confirm the importance of TLR2 ligand LTA in potentiating the antiviral host response (Fig. 3A–D). We treated the skin of wild-type mice and MC-deficient mice (Kit^{W-sash−/−}) with LTA or PBS for 24 h. Following this treatment, the skin was excised and evaluated for cathelicidin expression. The LTA-pretreated wild-type skin showed a higher level of Cnlp expression (Fig. 3A) compared with untreated wild type, whereas Kit^{W-sash−/−} skin did not show any change in Cnlp expression after LTA treatment (Fig. 3B). These data demonstrate that the presence of a TLR2 ligand increases cathelicidin expression in the skin and that this expression is linked to the presence of MCs. To confirm the relevance of this observation during virus infection, we inoculated 10⁶ PFU VV to the lower backs of Kit^{W-sash−/−} mice and their wild-type littermates (C57BL/6) 24 h after pretreatment of skin with LTA to allow time for the protein to be synthesized (37, 38). The wild-type mice pretreated with LTA showed smaller lesions compared with those of the nonpretreated littermate (Fig. 3C, 3D), and the Kit^{W-sash−/−} mice showed no change in lesion size (Fig. 3C, 3D). However, at an observation time of 72 h, besides exhibiting MCs, Kit^{W-sash−/−} mice and wild-type mice showed a similar cell infiltrate in the presence of VV (Fig. 3E, 3F), and the Kit^{W-sash−/−} pretreated mice improved more than the Kit^{W-sash−/−} nonpretreated mice (data not shown).

To further evaluate the importance of TLR2 signaling for the production of MC cathelicidin, we studied Tlr2^−/− MCs in vitro. Tlr2^−/− MCs failed to increase Cnlp following LTA stimulation (Fig. 4A). As expected based on the prior observations discussed above, Tlr2^−/− MCs infected by VV were more prone to infection, as measured as cell death, when compared with wild-type cells (Fig. 4B) even if their capacity to degranulate was equivalent to the wild-type cells (data not shown). When MC degranulation was blocked by cromolyn, wild-type cell susceptibility to infection went up (Fig. 4B). Upon analysis of Cnlp expression, we verified that Tlr2^−/− cells were still able to express Cnlp upon viral challenge, but to a significantly lower extent than the wild type (Fig. 4C).

To test the role of TLR2 in vivo, we reconstituted the Kit^{W-sash−/−} mice with Tlr2^−/− MCs and challenged the mice with VV. As measured by surface analysis, the size of the lesions developed by the Kit^{W-sash−/−} reconstituted mice was equivalent to the Kit^{W-sash−/−} not reconstituted (Fig. 4D–G); however, when we measured the level of VV expression (72 h from the VV inoculation), the mice with Tlr2^−/− reconstituted cells had a higher VV expression and a lower cathelicidin expression than the Kit^{W-sash−/−} not reconstituted mice (Fig. 4H, 4I). This confirms that, despite a superficial spread of the virus, the deep infection was contained by the Cnlp production induced by the virus in MCs and other inflammatory cells recalled by MCs at the site of the infection.

Skin is naturally bathed in a mixture of commensal bacteria. Gram-positive bacteria are the most represented commensal flora on the skin, including Staphylococcus spp., Micrococcus spp., and various Gram-positive rods (39). The most frequently cultured bacterial species on the skin is S. epidermidis (40). Many of these organisms express, as do all Gram-positive bacteria, LTA, a TLR2 ligand. Therefore, we conducted an investigation to demonstrate that superficial LTA is relevant in vivo, not only when an infection is present, but also under physiological conditions. We used an anti-LTA Ab specific for LTA-SE to stain normal human skin for a byproduct of S. epidermidis. The LTA-SE is present in the upper dermis and, therefore, is positioned to stimulate MCs in close contact with the skin barrier. This experiment demonstrates that the influence of bacterial byproducts on MCs observed in vitro has the potential to be a powerful mechanism in vivo. Our investigation started from this observation and went beyond; we proved that LTA stimulation of the skin increases MC recruitment. It was already known that MCs could respond to LTA when they are preconditioned with TNF-α (4, 24, 41). However, in our study, the increase in TNF-α was minimal without any sign of inflammation.

Various studies of the mucosa have revealed the significance of microflora to the host’s innate immune system (33, 40, 42–44). Other observations include studies on how indigenous microbiota enable expansion and maintenance of CD8 memory T cells in the lung and how gut microflora influence inflammatory bowel disease (45). In the intestines, microbiota protect the host by educating the immune system and preventing pathogenic infections. This microflora benefits the host through enhancing systemic immunity in specific organs, such as the lung. In the mouth, >500 species of bacteria protect the mucosa from infections by preventing colonization by pathogenic yeasts and other bacteria. Our data support the hypothesis that bacteria on the skin surface influences MC responses to pathogens, including viruses. The epithelial cells are situated in an opportune position, not only to act as a physical barrier, but also to serve as first responders against invading pathogens (46). A recent publication demonstrated that TLR2 ligands from the skin microbiome have an essential role in driving the skin innate response on keratinocytes (33) and in retinal muller glia (47). In addition, the TLR2 pathway has been implicated in MC cytokine modulation (48). A recent paper also reported that the LAD2 MC line increased their capacity to uptake bacteria after stimulation with TLR2 ligands (49). However, a direct correlation between increased AMP production and MC viral resistance has not been reported before. Kurt-Jones et al. (50) reported that TLR2 is a target for HSV; in addition, Zhu et al. (51) reported a lack of proinflammatory cytokines after challenging Tlr2^−/− mice with VV. We previously demonstrated that MCs play a key role in defending the host from viral infection and that LTA increases MC AMP expression. Therefore, we hypothesized that the skin microflora, through TLR2 signaling, is critical for viral defense.

In this study, we demonstrate that LTA, through TLR2 activation, stimulates MC responses to VV by inducing cathelicidin. LTA-preconditioned wild-type MCs demonstrate a greater antiviral response as compared with the response in pretreated Tlr2^−/− MCs, and this response was abolished when cromolyne blocks granule release, confirming that AMP release is important. To further confirm that LTA improved the cell capacity to kill VV through cathelicidin AMP, we used MCs deficient in cathelicidin; these cells did not improve their killing capacity on VV after stimulation with LTA.

Parallel to this in vitro observation, our in vivo LTA-pretreated mice presented an improved response against VV infection. This response was mostly visible in the first 24-h window when MCs are already present in the skin and neutrophils are not present yet. After 3 d, the LTA-pretreated Kit^{W-sash−/−} mice improved over the Kit^{W-sash−/−} nonpretreated mice. This change is attributable to the fact that Kit^{W-sash−/−} mice are MC deficient, but not cathelicidin deficient. After 3 d, the infiltrate from other cells compensated for the relative cathelicidin deficiency due to MC absence. The infiltrate includes macrophages, neutrophils, followed by granulocytes and T lymphocytes (52, 53). The major contributor to cathelicidin will be represented by neutrophils (54).

LTA-induced cathelicidin production enables MCs to efficiently kill viruses. Consequently, we predicted that MCs lacking TLR2 could not have been preactivated by LTA. Our prediction was confirmed by our second set of experiments.

We demonstrated that TLR2 is required to maintain MC cathelicidin levels. Therefore, when a virus is recognized, the AMP expression is further induced. As shown by our Tlr2^−/− in vitro experiments, Tlr2^−/− MCs can still produce cathelicidin on viral contact; however, without the possibility of being preconditioned through TLR2 ligands, cathelicidin is not upregulated, and therefore, MC-preloaded granules are not potent enough to adequately fight the infection in the skin. Our Tlr2^−/− MC reconstitution experiment demonstrates that during infection, mice reconstituted with Tlr2^−/− MCs had a superficial widespread involvement of infection in which the contact with the microbiome is expected to play the most important effect. Current literature supports the concept that MCs play a relevant role in fighting infections. Our findings suggest that the expression and control of AMPs and TLRs on MCs are key to this transformation. Our study demonstrates that the presence of bacterial byproducts at the skin surface participates in the regulation of MC antiviral activity through TLR activation.

LTA, released by saprophytic bacteria, has the potential to be a potent mechanism to keep MCs vigilant at the port of entry of infections, and to increase their capacity to fight viruses that use skin as an entry port.

We thank Dr. Besmer and the Developmental Biology Program, Memorial Sloan-Kettering Cancer Center at Cornell University for donation of the Kit^Wsh−/− mice. We acknowledge Dr. Bernard Moss from the National Institutes of Health for the gift of the vB5R-GFP VV and Dr. Gallo from University of California San Diego for the gift of the Cnlp^−/− mice.

The authors have no financial conflicts of interest.