TLR2 Signaling in Skin Nonhematopoietic Cells Induces Early Neutrophil Recruitment in Response to Leishmania major Infection

Abstract

Neutrophils are rapidly recruited to the mammalian skin in response to infection with the cutaneous Leishmania pathogen. The parasites use neutrophils to establish the disease; however, the signals driving early neutrophil recruitment are poorly known. Here, we identified the functional importance of TLR2 signaling in this process. Using bone marrow chimeras and immunohistology, we identified the TLR2-expressing cells involved in this early neutrophil recruitment to be of nonhematopoietic origin. Keratinocytes are damaged and briefly in contact with the parasites during infection. We show that TLR2 triggering by Leishmania major is required for their secretion of neutrophil-attracting chemokines. Furthermore, TLR2 triggering by L. major phosphoglycans is critical for neutrophil recruitment to negatively affect disease development, as shown by better control of lesion size and parasite load in Tlr2^−/− compared with wild-type infected mice. Conversely, restoring early neutrophil presence in Tlr2^−/− mice through injection of wild-type neutrophils or CXCL1 at the onset of infection resulted in delayed disease resolution comparable to that observed in wild-type mice. Taken together, our data show a crucial role for TLR2-expressing nonhematopoietic skin cells in the recruitment of the first wave of neutrophils after L. major infection, a process that delays disease control.

INTRODUCTION

The skin is one of the first barriers and lines of defense against invading pathogens. Keratinocytes are a major constituent of the epidermis, where they are a source of cytokines and growth factors (Pivarcsi et al., 2004). In response to pathogens, keratinocytes play multiple roles in the control of cutaneous diseases, secreting chemokines and cytokines that help shape the local microenvironment and attract neutrophils to the site of infection.

Leishmaniases are vector-borne diseases transmitted by sand flies that cause a spectrum of diseases that manifest as self-healing cutaneous lesions, mucocutaneous lesions, or the more severe visceral form that is fatal if not treated. After the bite of an infected sand fly, both the epidermis and the dermis where the parasite is deposited are damaged. There is an increasing interest in better understanding the role of keratinocytes during the onset of an immune response after infection with Leishmania species (Descatoire et al., 2017; Ehrchen et al., 2010; Eidsmo et al., 2007; Gasim et al., 1998; Scorza et al., 2017). Early neutrophil recruitment is rapidly observed after sand fly infection or needle inoculation of a high dose of Leishmania major in mice. Neutrophils can have either a protective or a deleterious impact on disease evolution, depending on the infecting Leishmania species and the host, as reviewed by Carlsen et al. (2015) and Hurrell et al. (2016).

Healthy human skin keratinocytes express toll-like receptor (TLR) 2 (Kawai et al., 2002; Kollisch et al., 2005; Lebre et al., 2007; Li et al., 2009), a TLR that recognizes several pathogen-associated molecular patterns, including peptidoglycans (Schwandner et al., 1999; Yoshimura et al., 1999), glycosylphosphatidylinositol anchors from protozoan parasites (Campos et al., 2001), and Leishmania lipophosphoglycan (LPG) (Spath et al., 2000). Whether TLR2 signaling in skin cells contributes to early neutrophil recruitment at the onset of L. major infection and the type of cells involved are unknown.

Here, we show that within hours of infection, Leishmania surface glycoconjugates trigger TLR2 signaling in nonhematopoietic cells including keratinocytes, inducing their release of neutrophil chemoattractants. This mechanism contributes to delayed control of disease.

RESULTS

L. major induces CXC chemokine expression in infected skin and primary keratinocytes

KC (CXCL1), MIP-2 (CXCL2), and lipopolysaccharide-induced LIX (CXCL5) chemokines play a major role in neutrophil recruitment. To investigate their induction in the first hours of infection, L. major parasites were injected intradermally into the ear dermis of C57BL/6 mice, and the kinetics of chemokine mRNA levels were analyzed in infected skin during the first hours after infection. Infection increased the mRNA levels of all three chemokine transcripts, but with distinct kinetics (Figure 1a). To assess the contribution of keratinocytes in chemokine secretion, primary keratinocytes were derived from the skin of C57BL/6 neonates and exposed to L. major. Significantly higher levels of KC, MIP-2, and LIX were released in response to L. major (Figure 1b). To investigate whether chemokine secretion resulted from parasite internalization, mCherry-L. major parasites were generated. The frequency of infection was analyzed by flow cytometry 16 hours after incubation with keratinocytes and compared with that observed after infection of macrophages and neutrophils. A clear mCherry⁺ population was detected among macrophages and neutrophils, but only a low frequency of keratinocytes appeared to be infected (Figure 1c). To further analyze the infection of keratinocytes, fluorescent parasites were incubated with primary keratinocytes, and parasite presence was determined by confocal microscopy 16 hours later. No parasites could be detected within the cells, but Leishmania species interacted with keratinocytes through the flagellar tip or the flagellar base or with the posterior pole of the parasite (Figure 1d). These results suggest that surface recognition of L. major by keratinocytes triggers the secretion of MIP-2, KC, and LIX by keratinocytes.

Figure 1. Leishmania major parasites stimulate chemokine secretion by keratinocytes but do not infect them.

(a) Chemokine mRNA levels in ears injected with L. major or phosphate buffered saline. Results are expressed as fold increase ± standard deviation compared with the expression detected in uninfected skin. The data are representative of two experiments. (b) Primary keratinocytes were cocultured with L. major promastigotes at a multiplicity of infection of 10 for 24 hours. Chemokine levels ± standard deviation were measured in supernatants by ELISA. ***P < 0.001 between keratinocytes not exposed and exposed to L. major. (c) mCherry-expressing L. major was incubated with macrophages, neutrophils, or primary keratinocytes. The frequency of infection was analyzed by flow cytometry. A representative plot ± standard deviation is shown, and the frequency is given in the bar graph. (d) CMFDA-stained L. major parasites were incubated with macrophages or keratinocytes for 16 hours. L. major presence was analyzed by fluorescent microscopy. Representative pictures are shown, with a magnification of the indicated area shown for keratinocytes. Scale bar = 5 μm. Lm, Leishmania major.

TLR2 expression is required to recruit neutrophils after L. major infection

TLRs are pathogen receptors at the surface of various body cells, including keratinocytes. Among all TLRs expressed, TLR2 showed the highest mRNA expression level in primary keratinocytes of naïve mice (Figure 2a). To assess the role of TLR2 and the downstream adapter protein MyD88 in L. major-induced rapid neutrophil recruitment in the skin, Tlr2^−/− and MyD88^−/− mice were infected intradermally with L. major, and CXC chemokine levels were assessed in the ear skin. Absence of TLR2 and MyD88 in infected ears resulted in a markedly reduced induction of KC, MIP-2, and LIX mRNA (Figure 2b). Neutrophil accumulation assessed in the infected ear by flow cytometry and histology peaked at 24 hours after infection. A significantly reduced neutrophil number was observed in the ear skin of Tlr2^−/− mice (Figure 2c-e), accounting for most of the decrease observed in the absence of MyD88 (Figure 2c). These data suggest a predominant role for TLR2 signaling in neutrophil recruitment early after infection and a minor role for other TLRs in this process.

Figure 2. Neutrophil attraction to Leishmania major inoculation site is TLR-2 dependent.

(a) Relative TLR mRNA expression ± standard deviation was determined in naïve primary keratinocytes. (b) mRNA levels ± standard deviation of the indicated chemokines were assessed at the indicated time points after infection in infected ears of WT, Myd88^−/−, and Tlr2^−/− mice. Results are expressed as fold increase relative to expression in uninfected ears. (c, d) WT, Tlr2^−/−, and MyD88^−/− mice were infected intradermally with 10⁶ L. major promastigotes. (c) Flow cytometry gating strategy and (d) mean number ± standard deviation of neutrophils recruited were measured at the indicated times by flow cytometry. ***P < 0.001, L. major-infected WT versus Tlr2^−/− or MyD88^−/− mice. n = 4 mice/group . (e) Representative hematoxylin and eosin staining of histological sections of ears from WT and Tlr2^−/− mice at 0, 4, and 24 hours after L. major infection. The data are representative of three independent experiments. Scale bar = 100 μm. h, hour; HPRT, hypoxanthine-guanine phosphoribosyl transferase; TLR, toll-like receptor; WT, wild type.

L. major LPG triggers TLR2 signaling and neutrophil recruitment

LPG is one of the most commonly expressed surface molecules on infectious promastigote parasites that also express other phosphoglycan-containing molecules (Sacks et al., 2000). To check the importance of LPG on the TLR2-dependent induction of CXC chemokine secretion by keratinocytes, wild-type (WT) or Tlr2^−/− primary keratinocytes were isolated and exposed to L. major or LPG. Both induced KC and MIP-2 secretion in WT keratinocytes, a process markedly decreased in Tlr2^−/− keratinocytes (Figure 3a and b). To show the role of LPG in TLR2-induced early neutrophil recruitment in vivo, WT mice were infected with WT L. major, L. major deficient for LPG alone (lpg1^−/−), or L. major deficient for LPG and all phosphoglycan-containing molecules (lpg2^−/−). The complemented add-back (AB) lpg1^−/− or lpg2^−/− parasites were used as controls. Injection of L. major deficient in lpg1^−/− significantly reduced neutrophil recruitment, and adding back lpg1 (lpg1AB) restored the capacity to recruit neutrophils in these parasites (Figure 3c). Injection of lpg2^−/− parasites resulted in even stronger inhibition of neutrophil recruitment to the site of infection, with levels that were comparable to those observed after injection of medium. Adding back lpg2 to the L. major mutant (lpg2AB) restored neutrophil recruitment upon infection (Figure 3d). Very few neutrophils were attracted after injection of WT, AB, and lpg1^−/− or lpg2^−/− L. major parasites in Tlr2^−/− mice (Figure 3e). These data show that LPG and other phosphoglycans present at the surface of L. major promastigotes trigger TLR2, resulting in local secretion of chemo-attractants that recruit most of the neutrophils observed at the site of infection during the first 24 hours after infection.

Figure 3. Triggering of TLR2 by Leishmania major LPGs induce neutrophil recruitment to the site of infection.

(a, b) Primary WT and Tlr2^−/− keratinocytes were exposed to L. major or L. major-purified LPG, and 24 hours later, cell-free supernatant was collected and tested for KC and MIP-2 content by ELISA. Representative mean values ± standard error of the mean from one of four experiments are shown. n ≥ 3. (c) Kinetics of neutrophil recruitment ± standard deviation in the ears of C57BL/6 mice after injection of WT, lpg1^−/− (lpg1¹-knockout), and lpg1 AB L. major and (d) lpg2^−/− (lpg2-knockout) and lpg2 AB L. major. (e) Kinetics of neutrophil recruitment ± standard deviation during the first day after infection in Tlr2^−/− mice with the indicated L. major parasites. The data presented are representative of three independent experiments (n = 3–5 mice/group per time point) with similar results. *P < 0.05, **P < 0.01. AB, add back; LPG, lipophosphoglycan; TLR, toll-like receptor; WT, wild type

TLR2 triggering of nonhematopoietic skin cells induces early neutrophil recruitment

To further investigate the source of TLR2-dependent neutrophil chemoattractants in the skin, we generated bone marrow (BM) chimeras (Figure 4a). Eight weeks later, the chimeras were infected with L. major. Twenty-four hours after infection, the presence of CD11b⁺Ly6G⁺ neutrophils was analyzed by flow cytometry (Figure 4b). Neutrophil recruitment was similar in the ear of WT and irradiated WT mice that received Tlr2^−/− or WT BM (Figure 4c). In contrast, irradiated Tlr2^−/− mice that received WT BM showed markedly reduced neutrophil recruitment 24 hours after L. major infection (Figure 4d). These data show that TLR2 expression on nonhematopoietic or radio-resistant cells is required for neutrophil recruitment at the site of parasite inoculation. In addition, 2 weeks after infection, decreased parasite burden was observed selectively in irradiated Tlr2^−/− that received WT BM (Figure 4e and f), further showing a positive correlation between neutrophil number at the onset of infection and subsequent parasite burden.

Figure 4. TLR2 signaling in nonhematopoietic cells is involved in early neutrophil recruitment.

(a) WT and TLR2 recipient mice were irradiated and reconstituted with bone marrow (BM) from WT or Tlr2^−/− donor mice as depicted. As a control, WT BM was transferred into irradiated WT mice. At 8 weeks later, BM-reconstituted mice were inoculated intradermally in the ear with 10⁶ Leishmania major organisms. (b) Representative gating strategy and (c, d) number ± standard deviation of Ly6G⁺ neutrophils recruited to the site of injection 24 hours after infection, as quantified by flow cytometry analysis. The data are representative of two independent experiments. (e, f) Two weeks after infection, the parasite burden ± standard deviation in the infected ears of the indicated chimeras was analyzed by limiting dilution analysis. The data are representative of two independent experiments including ≥3 mice/group. *P < 0.05, ***P < 0.001. TLR, toll-like receptor; WT, wild type.

To visualize the cells producing early KC and MIP-2, WT mice were infected with L. major, and 4 hours later, ear skin was isolated and subjected to immunofluorescence staining. High levels of KC staining were observed predominantly in the Epcam⁺ keratinocytes of infected WT and Tlr2^−/− mice (Figure 5a and data not shown) but not when the primary antibody was omitted (see Supplementary Figure S1 online). Similar levels of KC protein were found in the epidermis of naïve mice (Figure 5a), suggesting that this chemokine is constitutively expressed and stored in keratinocytes. In contrast, most of the MIP-2 staining was observed in the dermis and not the epidermis. Similar to KC, MIP-2 protein was present in the ear skin of both naïve and infected WT mice (Figure 5b), suggesting that this chemokine is also constitutively expressed and stored in dermal cells; no staining was observed in absence of the primary antibody (see Supplementary Figure S1). Upon L. major infection, dot-like MIP-2⁺ staining appeared in WT dermis, often in close proximity to cells harboring large quantities of intracellular MIP-2 (Figure 5b and c), suggesting that the dots represent chemokines released by the dermal MIP-2 source. These MIP-2 dots were observed only locally at the site of swelling and inflammation in WT dermis (Figure 5c), in close proximity to infiltrating myeloid cells. In Tlr2^−/− mice, a similar number of MIP-2⁺ cells was present in naïve and infected ear skins; however, the infection triggered less dot-like MIP-2⁺ staining, suggesting less release from the major dermal MIP-2 source. This finding correlated with reduced attraction of myeloid cells, composed mainly of neutrophils (Figure 5d, and see Supplementary Figure S1). Quantification of MIP-2⁺ cells and released dots indicated a significant increase in MIP-2⁺ released dots, specifically in inflamed areas of L. major infected ears of WT mice but not of Tlr2^−/− mice (Figure 5e). CD11b⁺ and NIMP-R14⁺ cells and Ly6G⁺ neutrophils (Figure 5c and, and see Supplementary Figure S1) showed only little co-localization with cytoplasmic MIP-2 staining, suggesting that MIP-2 is not being expressed at high levels by monocytes and neutrophils. To further define the MIP-2–expressing cell type and rule out the contribution of potentially radioresistant macrophages in early neutrophil recruitment, the dermis was stained for mannose receptor (CD206) and F4/80, two additional dermal macrophage markers. No co-localization with MIP-2 was observed, ruling out dermal macrophages as major MIP-2 producing cells. Similarly, no co-localization was observed on CD207⁺ Langerhans cells (Figure 5f). Absence of detectable levels of CD45 expression on most MIP-2⁺ cells in the skin of naïve and infected mice confirmed their nonhematopoietic origin. MIP-2⁺ cells were negative for vascular endothelial cadherin staining and not associated with collagen type IV⁺ basement membranes, ruling out lymphatic or endothelial endothelium as well (Figure 5f). Collectively, these data indicate that in the first hours after infection, L. major triggering of TLR2 on nonhematopoietic skin cells, including mostly keratinocytes and dermal stromal cells, leads to KC and MIP-2 release and thereby neutrophil recruitment.

Figure 5. MIP-2– and KC-producing skin cells in the first hours of Leishmania major infection.

Staining for (a) KC and DAPI in naïve and L. major-infected WT and Tlr2^−/− mice 4 hours after infection. (b) Staining for MIP-2 and DAPI in WT naïve and infected mice. Magnifications of the indicated area are shown below. The arrows point to MIP-2⁺ dots presumably showing released chemokine. (c) Noninflamed and inflamed areas of infected WT skin samples stained for MIP-2 and for NIMP-R14^highCD11b⁺ neutrophils. Below is an enlargement of the defined areas. (d) Similar staining as in c performed for naïve and L. major-infected Tlr2^−/− mice. (e) Quantification of MIP-2⁺ cells (arrow head) and dots (small arrow). The number of MIP-2⁺ events/50,000 μm² for WT and Tlr2^−/− ears is shown ± standard deviation. n = 4 ears/group, representative of two independent experiments. (f) Co-localization of MIP-2 was assessed for CD206⁺F4/80⁺ dermal macrophages, CD207/langerin⁺ Langerhans cells, CD45⁺ hematopoietic cells, vascular endothelial cadherin⁺ endothelium, and collagen IV⁺ endothelium-associated basement membranes. Representative images are shown. Scale bar = 100 μm. ***P < 0.001. c, cartilage; ns, not significant; WT, wild type.

Absence of TLR2 signaling reduces disease development after L. major infection

To assess the global impact of TLR2 on the disease, Tlr2^−/− mice were infected intradermally with L. major, and lesion development was measured. Tlr2^−/− mice developed a significantly smaller lesion compared with WT mice, as represented by lesion score. To bypass the TLR2 signaling required to attract neutrophils, 10 μg of KC was injected intradermally in Tlr2^−/− mice at the time of infection, a dose that recruited a similar number of neutrophils compared with that observed after L. major infection. Tlr2^−/− mice injected with KC developed a lesion score similar to that observed in WT mice, whereas injection of KC in WT mice did not change lesion development (Figure 6a). Eighty days after infection, Tlr2^−/− mice infected in the presence of exogenous KC had a parasite load significantly increased compared with that measured in the ear of similarly infected Tlr2^−/− mice (Figure 6b). To further assess if this phenotype was linked to the restored recruitment of neutrophils, Tlr2^−/− or WT mice were injected with 10⁶ inflammatory WT or Tlr2^−/− neutrophils at the time of infection. The injected inflammatory WT and Tlr2^−/− neutrophils did not differ functionally (see Supplementary Figure S2 online). Lesion development in Tlr2^−/− mice that received WT or Tlr2^−/− neutrophils was similar to that observed in L. major-infected WT mice (Figure 6c and d), with a comparable parasite load observed 35 days after infection that was significantly higher than that observed in Tlr2^−/− ears (Figure 6e), confirming the absence of functional deficiency in Tlr2^−/− neutrophils. Conversely, injection of WT or Tlr2^−/− neutrophils in WT mice did not affect lesion size development (Figure 6c and d), and parasite load was similar in WT mice that received WT or Tlr2^−/− neutrophils. Unlike Tlr2^−/− mice, in addition to transferred neutrophils, L. major also massively recruited neutrophils in the infected skin of WT mice, providing a protecting parasite shelter and resulting in a subsequent higher parasite load observed in the WT mice (Figure 6f). Altogether, these data show that L. major infection induces rapid TLR2 signaling in skin nonhematopoietic cells, triggering chemokine-mediated neutrophil recruitment, a process favoring transient survival of parasites within neutrophils and delaying subsequent development of a protective response.

Figure 6. TLR2-dependent neutrophil recruitment affects lesion development and parasite control.

(a) WT and Tlr2^−/− mice were co-injected or not with KC at the time of Leishmania major infection, and lesion score development (± standard error of the mean) was measured (n = 5 mice/group). (b) The parasite load ± standard deviation was determined by quantitative real-time reverse transcriptase–PCR 80 days after infection.*P < 0.05. The data are representative of two experiments. (c) WT or Tlr2^−/− inflammatory neutrophils were injected in the ear dermis of WT or Tlr2^−/− mice together with L. major. Lesion development was monitored with a caliper and compared with that of mice not coinjected with neutrophils. Score size ± standard error of the mean of lesions is shown. n = 5 mice per group. (d) Representative pictures of ear at 35 days after infection are shown, and (e, f) parasite load ± standard deviation was evaluated by quantitative real-time reverse transcriptase–PCR. P < 0.05, **P < 0.01, Results are representative of two experiments. ns, not significant; TLR, toll-like receptor; WT, wild type.

DISCUSSION

During the first day of infection, L. major uses newly recruited neutrophils as a transient shelter, resisting destruction by the otherwise efficient killing machinery of neutrophils (Regli et al., 2017). We show here that KC, MIP-2, and LIX, three major neutrophil-attracting chemokines, are induced locally in the skin in the first hours of infection, and we show that prestored KC and MIP-2 are released within 4 hours of infection. Up-regulation of KC mRNA and CXCL6 were previously shown to be associated with neutrophil recruitment after L. major infection (Muller et al., 2001; Uyttenhove et al., 2011). In addition to chemokines, other factors produced by the parasites (van Zandbergen et al., 2002) or contributed by the sand fly during natural infection, such as egested bacteria, salivary gland products, and a proteophosphoglycan gel, are likely to also participate in neutrophil recruitment (de Moura et al., 2010; Dey et al., 2018; Giraud et al., 2018). Thus, multiple factors may contribute and synergize to promote neutrophil recruitment at the site of infection. We show here that during the first hours after infection, TLR2 signaling induced by parasite phosphoglycans (especially LPG) expressed at the surface of the parasite plays a major role in rapid neutrophil migration to the site of parasite inoculation.

Leishmania species and their LPGs were reported to be recognized in vitro via TLR2 expressed on macrophages, neutrophils, natural killer cells, or dendritic cells (Becker et al., 2003; Charmoy et al., 2007; de Veer et al., 2003; Faria et al., 2014; Huang et al., 2015; Kavoosi et al., 2010; Tolouei et al., 2013). Prostaglandin E2 secretion by macrophages in response to Leishmania donovani was also shown to be TLR2 dependent (Bhattacharjee et al., 2016). However, a role for TLR2 signaling in nonhematopoietic cells had not been reported. Here, we show that in response to L. major, TLR2 is the most induced TLR expressed in keratinocytes and that secretion of neutrophil chemoattractants by keratinocytes is TLR2 dependent. Furthermore, recognition of purified L. major LPG induced the secretion of chemokines by primary keratinocytes in a TLR2-dependent manner. In vivo, recognition by TLR2 of L. major LPG and other phosphoglycan-containing molecules was shown to be crucial in the early recruitment of neutrophils after infection with L. major. In contrast to macrophages and neutrophils, L. major was not internalized by keratinocytes, in line with previous reports (Blank et al., 1993; Mbow et al., 2001; Scorza et al., 2017). However, as shown by confocal microscopy, L. major interacted with keratinocytes at locations where LPG and proteoglycans are enriched (Sadlova et al., 2010). This resulted in the secretion of neutrophil-attracting chemokines in a TLR2-dependent process. BM chimera and immunofluorescence microscopy showed the predominant early role of two main L. major-induced neutrophil attractants, KC and MIP-2, produced by keratinocytes and dermal stromal cells, respectively. KC and MIP-2 chemokines appeared prestored in nonhematopoietic cells of naïve skin, which is in line with former studies (Johansson et al., 2015; Oynebraten et al., 2004), suggesting that in response to TLR2 triggering, rapid release of these prestored chemokines can occur, in addition to de novo synthesis. In the epidermis, keratinocytes are the main nonhematopoietic cell type that can sense microbes. In the dermis, various nonhematopoietic cells exist, including endothelial cells, fibroblasts, and adipocytes (Lai and Gallo, 2008). We showed that TLR2 induces KC release by keratinocytes. Currently, we cannot rule out the possibility that radioresistant Langerhans cells also produce some KC. Very few keratinocytes, and no endothelial cells, expressed MIP-2. The exact source of nonhematopoietic cells releasing MIP-2 remains to be determined. Despite the low or lack of chemokine co-localization with CD45⁺ hematopoietic cells, a few dermal macrophages showed co-staining with MIP-2. Collectively, these data suggest that chemokine secretion by macrophages and Langerhans cells does not seem to be responsible for most neutrophils recruited.

Neutrophils play important roles not only in the cutaneous forms but also in the experimental visceral form of the disease (Dey et al., 2018; McFarlane et al., 2008; Sacramento et al., 2015; Smelt et al., 2000). Decreased neutrophil recruitment in the liver and spleen of Leishmania infantum-infected Tlr2^−/− mice was recently described (Sacramento et al., 2017), further implying the importance of TLR2 signaling in the inflammatory response to infection with Leishmania species inducing visceral diseases. Leishmania species exposure to human primary keratinocytes induced cytokine and chemokine production in vitro; however, L. infantum was a better inducer than L. major (Scorza et al., 2017), suggesting that there may exist differences between the impact of distinct Leishmania species on keratinocytes function, as previously observed for neutrophils (Hurrell et al., 2016).

We show here that L. major-infected Tlr2^−/− mice developed a significantly smaller lesion size and better controlled parasite load than WT mice. These data are in line with a deleterious role for neutrophils early in infection. The presence of neutrophils in the skin during the first day of infection, through their release of cytokines and other granule factors (Tecchio et al., 2014), most likely contributes to shaping the microenvironment at the site of infection with a significant impact on subsequent lesion development, in line with previous reports (Hurrell et al., 2015; Peters et al., 2008).

In contrast, Tlr2^−/− mice infected subcutaneously in the footpad with L. major were reported to become transiently susceptible to infection (Halliday et al., 2016). The differences observed between this report and our study may be explained by the distinct sites used to inoculate the parasites. During the first hours of infection, neutrophils are poorly recruited after subcutaneous infection, whereas they are strongly recruited after dermal parasite inoculation (Ribeiro-Gomes et al., 2014), a site corresponding to parasite delivery by the sand fly. In line with our data, treatment of mice injected subcutaneously in the footpad with TLR2 agonists increased neutrophil recruitment to the site of infection (Huang et al., 2015), and treatment with anti-TLR2 antibodies suppressed the anti-inflammatory response but not parasite load (Komai-Koma et al., 2014).

Altogether, we show here a crucial role for TLR2 signaling in skin nonhematopoietic cells in the early recruitment of neutrophils. This affects the evolution of cutaneous leishmaniasis, showing a pathogenic role for early TLR2 signaling in the skin, opening avenues for modulating disease onset.

MATERIALS AND METHODS

Mice

C57BL/6 mice were from Harlan (Envigo, Huntington, UK). MyD882^−/−Tlr2^−/− mice backcrossed onto a C57BL/6 background (from S. Akira, University of Osaka, Japan) were bred and maintained under pathogen-free conditions (UNIL, Epalinges, Switzerland). The maintenance and care of mice complied with and the studies were approved by the ethical guidelines of the state of Vaud ethics committees.

Parasites

L. major bearing homozygous deletions of LPG1 (lpg1^−/− or δlpg1⁻) or LPG2 (lpg2^−/− or Δlpg2⁻) genes and their respective complemented ABs were generated previously in the LV39 clone 5 background (Spath et al., 2000, 2003). These and L. major LV39 (MRHO/SU/59/P) were maintained as previously described (Tacchini-Cottier et al., 2000).

Keratinocyte culture and chemokine production

Newborn mouse primary keratinocytes were cultivated and grown as previously described (Missero et al., 1996). Confluent monolayers of keratinocytes were exposed to 10⁶/ml parasites in Leydig cell medium. Chemokines were measured by ELISAs (R&D Systems, Minneapolis, MN). For immunofluorescence, L. major was labeled with CMFDA (Molecular Probes, Eugene, OR) fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton, and counterstained with rhodamine phalloidin (Molecular Probes), mounted, and analyzed with an Axio-plan2 fluorescent microscope (Zeiss, Jena, Germany).

Two-step SYBR green reverse transcriptase–PCR

mRNA was purified (Qiagen, Hilden, Germany) and reverse transcribed using SuperScript II RT (Invitrogen Life Technologies, Carlsbad, CA). Quantitative PCR was performed using the primers listed in the Supplementary Materials online.

L. major inoculation, ear explant, and neutrophil recovery

Mice were injected intradermally with 10⁶ stationary phase, wild-type, or mutant L. major or 5 × 10⁵ metacyclic promastigotes in a volume of 10 μl of DMEM. Atthe end of the experiment, mice were killed, and ears were prepared as previously described (Charmoy et al., 2010). Stained cells were analyzed in phosphate buffered saline 2% fetal calf serum with a FACSCan or FACS Calibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ) and analyzed with FlowJo (Ashland, OR) software.

Immunofluorescence and flow cytometry

Ears were prepared for immunohistology as previously described (Fasnacht et al., 2014). Details are provided in the Supplementary Materials. Antibodies against Ly6G, CD4, and CD19 (BioLegend) and against CD8, CD11b and CD45, and AnnexinV (e-Bioscience, San Diego, CA) were used for flow cytometry. L. major-infected keratinocytes were counterstained with rhodamine phalloidin (Molecular Probes) and analyzed with a fluorescent microscope, Axio-plan2, coupled with an ApoTome system (Zeiss). Immunohistologic quantification of MIP-2⁺ cells and dots was performed with ImageJ software (National Institutes of Health, Bethesda, MD). The area of determined regions was calculated, and the number of MIP-2 positive events/area were counted and expressed as events/50,000 μm₂.

Generation of mCherry-fluorescent parasites

L. major LV39 mCherry parasites were generated by transfecting log-stage promastigotes with linearized mCherry expression plasmid using the AMAXA nucleofection system (Lonza, Basel, Switzerland) and selecting clones with 50 μg/ml hygromycin B (Calbiochem Merck, Darmstadt, Germany), as detailed in the Supplementary Materials. In selected experiments, L. major was labeled with 5 μmol/L CMFDA according to the manufacturer’s protocol (Molecular Probes). Reactive oxygen species formation was measured with the dihydrodamine 123 probe (Sigma-Aldrich, St. Louis, MO) or luminol (Carbosynth, Newbury, UK) as described by the manufacturers.

BM chimera

Recipient mice were irradiated (900 rad) and reconstituted with 1 × 10⁷ donor mouse BM cells. Six weeks later, mice were bled, and the presence of CD4⁺CD8⁺CD19⁺ cells was determined by flow cytometry. Chimeric mice were inoculated with L. major in the ear; 24 hours after infection, the recruitment of neutrophils was analyzed by flow cytometry, and 2 weeks after infection, parasite load was determined by limiting dilution analysis (Charmoy et al., 2010).

Transfer of neutrophils

Mice were injected intraperitoneally with 5 × 10⁷ L. major parasites, and neutrophils were collected 4 hours later and purified using magnetic activated cell sorting (MACS)–negative selection (Myltenyi Biotec, Gladbach, Germany). Purity of neutrophils (>96%) was assessed by flow cytometry. Next, 10 μg of KC, a dose resulting in a similar number of neutrophils recruited in the infected ear 24 hours after injection than the number observed after infection with 10⁶ L. major organisms (60,360 ± 4,002 vs. 65,810 ± 8,181 neutrophils/ear), was injected intradermally with or without L. major. In other experiments, 5 × 10⁵ C57BL/6 or Tlr2^−/− neutrophils were co-injected in the ear dermis with 10⁶ stationary phase L. major promastigotes. Development of ear lesions was measured, and a score was given (Schuster et al., 2014). Parasite presence was quantified with quantitative real-time reverse transcriptase–PCR, using Leishmania Kmp11 gene-specific primers or limiting dilution analysis, as previously described (Ives et al., 2011).

Statistical analysis

Results are expressed as the mean ± standard deviation. Statistical differences between groups were analyzed using the t test for unpaired data. A value of P < 0.05 was considered significant.

Abbreviations:

AB

add back

BM

bone marrow

LPG

lipophosphoglycan

TLR

toll-like receptor

WT

wild type