Microorganisms are increasingly being recognized to have profound effects on the host they invade (Shropshire and Bordenstein, 2016). In many cases, host behaviors are altered upon infection by diverse pathogens, including bacteria (Cameron and Sperandio, 2015). These observed changes in host behavior are often thought to be beneficial to the microorganisms, as they may increase the rate of transmission (Cézilly and Perrot-Minnot, 2005).

Staphylococcus aureus is both a commensal and extremely versatile pathogen and one of the most common causes of bacterial community and healthcare-associated infections in humans (Lowy, 1998). S. aureus primarily causes skin and soft tissue infections (SSTIs), bloodstream infections, and pneumonia. The pathogenicity of S. aureus is determined by the extended repertoire of toxins produced by this bacteria (Kong et al., 2016). In vitro and in vivo studies in animals have identified pore-forming toxins, such as Panton-Valentine leukocidin (PVL), alpha-hemolysin (Hla), phenol-soluble modulin-alpha (PSMα), and delta-hemolysin (Hld), as major virulence factors involved in the pathophysiology of staphylococcal skin infections (Wang et al., 2007; Kobayashi et al., 2011; Lipinska et al., 2011; Syed et al., 2015). These toxins are capable of targeting a wide variety of immune cells during infection, such as human polymorphonuclear leukocytes, monocytes, and macrophages, and can significantly contribute to dampening both innate and adaptive immune response to S. aureus infection (Pozzi et al., 2015).

The expression of PVL, Hla, and PSMs is regulated in different manners by accessory gene regulator (agr), a quorum sensing system in S. aureus. Hld is encoded by RNAIII, a non-coding RNA that orchestrates the expression of most virulence factors (Novick et al., 1993), including PVL and Hla (Vandenesch et al., 1991; Dumitrescu et al., 2011). In contrast, PSMα expression is regulated by AgrA, the response regulator of the two component system that activates the agr system (Lina et al., 1998; Queck et al., 2008).

Skin to skin contact via the hands plays a significant role in the spread of S. aureus to new hosts not only in hospitals, but also in nursing homes and child care settings, among others (Bloomfield and Scott, 1997). Increasing the frequency of contact between one's hand and skin colonized or infected by S. aureus due to itching and scratching behavior may enhance the transmission potential of the pathogen. The induction of scratching behavior in animals during experimental skin infection by S. aureus was described a long time ago (Wagner et al., 1997) but only recently described in humans when skin infections induced by community-acquired methicillin-resistant S. aureus (CA-MRSA) were misidentified by both patients and physicians as spider bites because they were very erythematous, indurated, and itchy, sometimes with a central dermo-necrosis (Suchard, 2011).

The skin is considered a major interface of the body for the host defense, not only as a passive barrier, but also through the immune system. Innate immune cells residing in the skin, such as Langerhans cells, dendritic cells, and dermal mast cells, provide cutaneous immune surveillance (Kupper and Fuhlbrigge, 2014). Mast cells are leukocytes originating from hematopoietic progenitor cells and located at host interfaces with the environment, such as the skin, pulmonary, and digestive mucosa. Mast cell differentiation and maturation are different according to the organs in which they are located. Two major phenotypes of mature mast cells are differentiated by granule content and the receptors expressed: Phenotype T contains mainly tryptase, and phenotype TC contains mainly tryptase and chymase (Galli et al., 2011).

These cells are able to recognize pathogenic agents and trigger the inflammatory process through complex inter-cellular communication mediated by several mediators released by mast cells. Thus, mast cells participate in the first line of defense in innate immunity against pathogens, including bacteria (Abraham and St. John, 2010).

Mast cells express several receptors capable of recognizing pathogens that belong to pattern recognition receptor (PRR) family involved in the recognition of pathogen-associated molecular patterns (PAMPs), such as Toll-like receptors (TLRs) TLR1 to TLR9, Nod-like receptors (NLRs), and C-type lectin receptors (CLRs), including dectin-1 (Urb and Sheppard, 2012; St. John and Abraham, 2013). Pathogen recognition induces several activation pathways in mast cells, resulting in the release of mediators, such as intra-cytoplasmic granules (histamine, protease, tryptase, tumor necrosis factor [TNF]), lipid-derived eicosanoids (leukotrienes and prostaglandins), and cytokines or chemokines, including TNF, IL-4, and IL-6 (Abraham and St. John, 2010; Urb and Sheppard, 2012; St. John and Abraham, 2013). After bacterial invasion, mast cells preferentially release IL-8 and TNF to promote polymorphonuclear neutrophil (PMN) recruitment at the infection site (Urb and Sheppard, 2012). First, TNF, IL-6, and IL-8 promote PMN chemotaxis. Next, TNF and eicosanoid trigger an up-regulation of adhesion molecules on the endothelial cell surface and vascular permeability, allowing PMN adhesion to endothelial cells and diapedesis (Abraham and St. John, 2010). In addition, mast cells can directly kill pathogens via the production of antimicrobial peptides known as cathelicidins (Di Nardo et al., 2003).

Mast cells can recognize bacteria through TLR-2 and TLR-4, which recognize the lipopolysaccharides of Gram-negative bacteria and peptidoglycans of Gram-positive bacteria, and through TLR-5, which recognizes flagellin (Abraham and St. John, 2010). Mast cell mediators are also known to activate neuroreceptors on sensory nerve fibers involved in the induction of pruritus (Ständer et al., 2008).

Only a few studies have examined the interaction between S. aureus and mast cells. S. aureus induces mast cell activation through peptidoglycan-sensing by TLR2 (Feng et al., 2007). Two recent studies also confirmed that S. aureus-exposed mast cells release pro-inflammatory cytokines, such as IL-8, IL-3, IL-13, and TNF-α (Rönnberg et al., 2014; Swindle et al., 2015). Furthermore, S. aureus internalization by mast cells is enhanced by Hla and induces the release of extracellular traps, antimicrobial compounds, and pro-inflammatory cytokines (Abel et al., 2011; Goldmann et al., 2016; Johnzon et al., 2016). Nakamura et al. (2013) also demonstrated in a mouse model that S. aureus Hld induces skin mast cell activation, promoting the establishment of atopic dermatitis, an allergic skin disease. However, no data are yet available on the effect of other staphylococcal toxins involved in skin pathophysiology on human mast cells.

The aim of our study was to investigate whether S. aureus produces virulence factors during skin infection that may increase the rate of transmission by stimulating itching and scratching behavior through mast cell activation. We examined the impact of virulence factors involved in skin infection pathophysiology on skin mast cell activation and verified there in vivo production in human S. aureus skin abscess samples.

Despite progress in our knowledge of S. aureus pathophysiology, virulence factors that favor its transmission remain enigmatic. Hands play a significant role in the spread of S. aureus to new hosts through contact with skin colonized or infected by S. aureus. In the skin, dermal mast cells act in close interaction with nerve fibers and contain substances that act as mediators of itching (Ständer et al., 2008). In this study, we examined whether S. aureus pore-forming toxins that are involved in the pathophysiology of skin infection can induce mast cell activation/lysis. We used HMC-1 cells derived from a patient with leukemia. Even though this cell line has two major deficiencies (cytokine-independence and lack of a functional FcεR), it is widely used and a faithful in vitro model of normal mast cells (Drexler and MacLeod, 2003). We verified that HMC-1 cells produce tryptase, one of the best indicators of mast cell activation (Schwartz et al., 2003; Guhl et al., 2010), and express CD88 similar to human skin mast cells (Füreder et al., 1995; Werfel et al., 1996) (Supplementary Figure 5).

Our results clearly demonstrate that PSMα1, PSMα3, and Hld induce dose-dependent tryptase and LDH release by HMC-1 cells. We suspect that tryptase liberation was mainly the result of mast cell lysis. In contrast, PVL and Hla did not induce human HMC-1 cell degranulation or lysis. This finding was unexpected because König et al. (1995) reported that PVL induces histamine release by human basophilic granulocytes, immune cells that are morphologically and functionally related to mast cells, especially given that the HMC-1 cell line expresses C5aR, the PVL receptor (König et al., 1995; Spaan et al., 2013). Two major hypotheses are possible to explain this discrepancy. Either HMC-1 cells do not express enough CD88 to make the cell sensitive to PVL and/or there is an additional specific co-receptor necessary for PVL to be active that is missing from HMC-1 cells. Supplementary investigations should be performed to test these hypotheses.

We found a similar cytotoxic effect of PSMα1 and PSMα3 on HMC-1 cells as those previously reported with human neutrophils. The lethal concentration 50% was achieved in both mast cells and human neutrophils at 25 μg/mL PSMα1 (Löffler et al., 2010). However, for PSMα3, we observed only 20% lysis of mast cells at the concentration that results in the 80% lysis of neutrophils (i.e., 50 μg/mL) (Löffler et al., 2010). We observed the same level of Hld activity against mast cells as Nakamura et al.; using 10 μg/mL we measured 13.30 ± 3.01% tryptase release, whereas they reported β-hexosaminidase release of ~10% (Nakamura et al., 2013). The only difference was that β-hexosaminidase release by Hld seems to not be related to cell lysis of the mouse MC/9 mast cell line, which is in contrast to what we observed with HMC-1 cells. This may be due to the different origins of the mast cell lines.

In order to determine whether PSMα1, PSMα3, and Hld are expressed in vivo, 23 pus samples recovered from human S. aureus skin abscesses were analyzed. We detected the presence of Hld in 65% of the analyzed clinical specimens and PSMα1 in 17% of the samples, but no PSMα3. The concentration of Hld and PSMα1 detected in pus samples was 50- and 10-fold lower, respectively, than the concentrations produced in vitro and did not correlate. The weakest toxin concentration in vivo may indicate that the toxins were degraded more rapidly in pus than in supernatants, despite similar storage procedures. Only 50% of pus samples contained sufficient Hld levels to induce HMC-1 cell lysis, but it was 100% with the in vitro supernatants. In vitro Hld production was 10-fold higher than mean PSMα1 production and 30-fold higher than PSMα3 production as described previously (Berlon et al., 2015).

Two studies recently demonstrated that PSM and Hld production by MRSA and MSSA is higher among SSTI isolates than either infectious endocarditis or hospital-acquired pneumonia (Berlon et al., 2015; Qi et al., 2016). These observations suggested that the level of PSM and Hld correlates more with the disease than the methicillin-resistant status. As PSM and Hld production is regulated by the agr, the level of PSM and Hld production may reflect agr activity and its requirement for the development of skin infection in humans, as in animal models of S. aureus skin infection (Montgomery et al., 2010; Cheung et al., 2011). Looking at the impact of agr alleles in our sample, strains in agr group 1 produced more Hld, PSMα1, and PSMα3 than those from groups 2 and 3, despite CA-MRSA belonging mainly to agr group 3 (Tristan et al., 2007).

The involvement of PSMα1–3 in SSTI pathogenesis was demonstrated by Wang et al., who showed that S. aureus strains producing PSMα are more virulent, notably through the capacity of PSMα to lyse neutrophils (Wang et al., 2007). Recently, Syed et al. highlighted the involvement of S. aureus PSMα, including Hld, in the induction of skin inflammation through pro-inflammatory cytokine release from keratinocytes (Syed et al., 2015). Skin mast cells are very important immune cells and participate in the fight against pathogens, but they can be deleterious when over-stimulated (Abraham and St. John, 2010; Urb and Sheppard, 2012; St. John and Abraham, 2013). This deleterious effect has been shown in mice, with Hld participating in the development of atopic dermatitis (Nakamura et al., 2013). Therefore, our results suggest that S. aureus produces Hld during acute skin infection, and to a lesser extent PSMα1, which induces mast cell degranulation and/or lysis directly or indirectly through skin inflammation. This human skin mast cell degranulation activates neuroreceptors on sensory nerve fibers involved in the induction of pruritus and explain the itchy feeling during S. aureus skin infections (Zhu et al., 2015). We speculate that the scratching behavior induced by S. aureus though Hld and PSM production promotes S. aureus transmission via hand-to-skin contact in the presence of S. aureus-infected skin lesions, thereby promoting S. aureus epidemicity. As highlighted previously, agr is a key regulatory system for the maintenance of S. aureus virulence, which is considered a key factor for host-to-host transmission (Massey et al., 2006). Moreover, Hld is encoded by RNAIII, the effector of agr (Novick et al., 1993). We hypothesize that the Hld–agr association contributes to the elective transmission of virulent S. aureus strains with a functional agr system. Altogether could participate in the epidemic success of certain S. aureus strains, such as the pandemic CA-MRSA USA300 clone that strongly expresses toxins with the help of the agr system, induces itchy skin infections, and is pandemic in the USA (DeLeo et al., 2010; Cheung et al., 2011; Suchard, 2011).

In conclusion, our results demonstrate that Hld and PSMα1 are produced by S. aureus in vivo and induce human mast cell degranulation and lysis. As mast cell mediators are known to be involved in the induction of pruritus, we hypothesize that S. aureus may promote its transmission via the hands through scratching of infected skin lesions prompted by Hld and PSM production.

This study was conducted according to the principles of the Declaration of Helsinki and its subsequent amendments, the guidelines for Good Clinical Practices (CPMP/ICH/135/95), the French regulations and the guidelines of the ethical committees of the participating hospital, Centre Hospitalier Lyon Sud (Hospices civils de Lyon, Pierre Benite, France), and Groupement Hospitalier Est (Hospices civils de Lyon, Bron, France) (DC-201-1306). Only specimens from human S. aureus skin abscesses were collected as part of the routine management of patients was examined.

EH, VL, BD, MB, AT, GL, and OD designed the experiments. EH, CC, CB, FB, JS, RC, and AP performed the experiments. EH, VL, BD, GL, and OD wrote the manuscript.

This work was supported in part by the LABEX ECOFECT (ANR-11-LABX-0048) of Université de Lyon, within the program “Investissements d'Avenir” (ANR-11-IDEX-0007) operated by the French National Research Agency (ANR).