Syk^fl/fl and Syk^fl/flMrp8-cre⁺ mice, were previously described (Van Ziffle & Lowell 2009), p47phox^−/− mice were a gift from Prof. Romani (University of Perugia) and were previously described (35). Tcrb _-/- mice were a gift from Prof. Constantin (University of Verona). C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). All mice used in this study were on a C57BL/6 background and kept in a specific pathogen-free facility.

For induction of psoriasis-like skin inflammation, mice at 8–12 wk of age received a daily topical dose of 62,5 mg of commercially available IMQ cream (5%) (Aldara Cream™, Meda AB) or control cream (vaseline) on their shaved backs for 6 consecutive days as previously described (28, 38). On the fourth or seventh day, the animals were euthanized. Back skin was isolated, and half was fixed in 10% formaldehyde for histopathology analysis while the other half was finely chopped and stored in RNAlater (Ambion) for quantitative real-time PCR (qRT-PCR) or digested, as described below, to achieve single-cell suspensions for flow cytometry analysis.

Mice were injected intra-peritoneally (i.p.) with 300 μg of rat anti-mouse Ly6G Ab (clone 1A8; BioXcell) or isotype control Rat IgG2a (clone 2A3; BioXCell), dissolved in 300 ul phosphate-buffered saline (PBS) every other day from day 0 to day 6.

Skin tissue (2 cm X 2 cm) was cut from dorsal skin of the mouse. After removing subcutaneous tissue and collagen intensively with forceps, the skin was cut into small pieces and digested with 0,4 mg/ml Liberase TM (Roche Ltd.) and 0,5 mg/ml DNase I (Sigma) in RPMI 1640 medium (Sigma) for 1 hour. Single cell suspension was obtained by shredding with gentle Macs Dissociator (Miltenyi Biotec) and filtering with 70 μm and 40 μm cell strainer in series. Lymph nodes were mechanically dissociated by two frosted microscope slides and passage through a 70 μM cell strainer to yield a single-cell solution. Cells were resuspended in phosphate buffered saline containing 2% (vol/vol) fetal calf serum, 2 mM EDTA and maintained at 4°C. For flow cytometry, 1–2×10⁶ cells were stained. Non-specific binding was blocked by pre-incubation with 0.5 µg anti-CD16/32 (2.4G2, Biolegend) and 100 µg mouse IgG (Sigma). Surface staining was performed with the following anti-mouse Abs: Ly6G(1A8), TCRαβ (H57-597), CD62L (MEL-14), CD11b (M1/70), CD45 (30-F11), I-Ab (MHCII)(AF6-120.1), CD44 (IM7), TCR γ/δ (GL3) from Biolegends; Ly6C (AL-21), CD11c (HL3), CD3 (145-2C11) and GR-1 (RB6-8C5), from BD Biosciences. After final wash, cells were resuspended in staining/wash buffer containing 1 mg/ml propidium iodide (PI; Sigma-Aldrich) for viability staining according to the manufacturer’s instructions. For intracellular cytokine staining, the cells were activated for 4 hours in phorbol 12-myristate 13-acetate (PMA; 50 ng/ml) and ionomycin (750 ng/ml) in the presence of brefeldin A (1 mg/ml). Thereafter, cells were surface-stained, washed, and then fixed and permeabilized using the eBioscience kit as previously described (39). Intracellular staining was performed with anti-mouse IL-17A (TC11-18H10.1; eBioscience) or its relevant isotype control mAbs. Sample fluorescence was measured by a seven-color MACSQuant Analyzer (Miltenyi Biotec), while data analysis was performed by using FlowJo software Version 8.8.6 (Tree Star, Ashland, OR, USA).

Real-time reverse transcription-PCR was performed, as previously described (40), using total RNA isolated from 30 mg of the skin by RNeasy Fibrous Tissue Mini Kit (QIAGEN) and utilizing the following gene-specific primer pairs (all purchased from Invitrogen) ( Supplementary Table 1 ). Data were calculated by Q-Gene software (http://www.gene ) quantification.de/download.html) and expressed as mean normalized expression (MNE) units after RPL32 normalization.

Dorsal skin samples (3 mm) were obtained by a transversal cut of the central skin area, fixed in 10% neutral buffered formalin and embedded in paraffin blocks by using a Tissue-Tek^® Tissue Embedding Console System from Diapath (Bergamo, Italy). The paraffin blocks were cut into 3 µm thick cross-sections and stained with hematoxylin and eosin following the standard procedure (immersion in Mayer’s hematoxylin: 2 minutes; immersion in eosin: 3 min) by using a Leica Microsystem Autostainer XL ST5010 (Milano, Italy). Epidermal thickness was determined by measuring the average interfollicular distance under the microscope in a blinded manner. Pictures were taken using Leica DFC 300FX Digital Color Camera on a Leica DM 6000 B microscope at a 100x magnification. For γδ T cell and neutrophil immunohistochemical staining, 4 µm formalin-fixed, paraffin-embedded tissue sections were double stained after appropriate antigen retrieval with rat anti-mouse RORgt (dilution 1:50, clone AFKJS-9, eBiosciences, San Diego, CA, USA) and Ly6G (dilution 1:400, clone 7/4, Cedarlane, Burlington, On, Canada).The first immune reaction was revealed using rat-on-mouse HRP-Polymer (Biocare Medical, Concord, CA, USA) and developed by diaminobenzidine; the slides were then incubated with anti-Ly6G, revealed using rat-on-mouse AP-Polymer and developed with vector Blue chromogen (Vector Laboratories, Newark, CA, USA). Slides were then counterstained with hematoxylin. Slides were photographed using the DP73 Olympus digital camera mounted on an Olympus BX60 microscope and resized using Adobe Photoshop.

Data were expressed as the mean ± SD and analyzed using GraphPad Prism Version 5 software (GraphPad Software, Inc.). The comparison of variables was performed using two-tailed Student t- test (for comparison between two groups) or a 1-way ANOVA with Bonferroni’s posttest (used for multiple comparisons), Dunnett’s post-test (when multiple comparisons to control group were made). P-values of less than 0.05 were considered significant and symbols indicate significant increases: *^/#, P <0.05; **^/##, P ≤ 0.01; ***^/###, P ≤ 0.001; ^****/####, P ≤ 0.0001. Graphs were elaborated using GraphPad Prism Version 5 software (GraphPad Software, Inc.).

This includes extended methods, one Table and four Figures.

To investigate the specific contribution of neutrophils to the development of IMQ-induced psoriasis, we performed neutrophil depletion by injecting anti-Ly6G (clone 1A8) Ab, or isotype control Ab, in mice treated with IMQ (or vaseline control cream), for 3 or 6 consecutive days as originally described by Vanderfits et al. (28). First, we confirmed that the anti-Ly6G–treatment successfully depleted neutrophils in lymph nodes and the skin of either vaseline or IMQ-treated mice after both 3 and 6 days of treatment ( Supplementary Figures 1A–C ). Interestingly, neutrophil depletion did not significantly affect epidermal thickening, the most utilized and reproducible clinical parameter utilized to quantify disease severity in this model (27), up to 3 days of IMQ treatment ( Figure 1A ). However, differently from what was previously published by others (24, 25), we observed an unexpected significant increase of epidermal thickening in neutrophil-depleted mice, as compared to control mice, upon 6 days of IMQ treatment ( Figures 1A, B ). Consistently, the expression of skin-associated psoriatic genes by qRT-PCR, such as Lipocalin-2 (Lcn2) and S100 calcium binding protein A7/psoriasin (S100A7) was significantly higher in dorsal skin of mice IMQ-treated receiving anti-Ly6G Ab, as compared to control IgG-treated mice ( Figure 2 ). Strikingly, we also observed that, upon IMQ treatment, mice devoid of neutrophils manifested a significantly increased expression of cytokines implicated in the IL-23/T17 axis, including IL-23, IL-17, IL-22, CXCL1 and IL-6, as compared to control IgG-treated mice ( Figure 2 and data not shown). Neutrophil depletion, instead, did not significantly affect the expression of other inflammatory cytokines induced by IMQ treatment, such as IL-1β, IL-36 and IL-1α ( Figure 2 ).

Overall, these data suggest a novel potential role for neutrophils as negative modulators of disease progression and of the IL-23/T17 axis in IMQ-induced psoriasis.

We then performed a careful characterization of the CD45⁺ cells infiltrating the draining lymph nodes and the skin of IMQ-treated mice receiving anti-Ly6G Ab, or control IgG, by flow cytometry, utilizing the gating strategies previously described (38). Interestingly, we found that neutrophil-depleted mice displayed a strongly increased accumulation of γδ T cells in the draining lymph nodes after 6 days of IMQ treatment ( Figure 3A ). Besides the total number, also the number of CD44^highCD62L^low effector γδ T cells ( Figure 3B ) and of IL-17-producing γδ T cells ( Figures 3 C, D ) were significantly increased, indicating that not only the numbers but also the activation state of these cells was profoundly affected by the depletion of neutrophils. No significant differences in the infiltration of αβ T cells ( Figure 3E ), monocytes/macrophages ( Figure 3F ) and DCs ( Figure 3G ) were instead found in the draining lymph nodes of anti-Ly6G–treated mice when compared to controls.

Notably, a strong expansion of dermal γδ TCR^low T cells ( Figure 4A ), but not of monocytes/macrophages, DCs or αβ T cells ( Figures 4B–D ), was also evident in the dorsal skin of anti-Ly6G-treated, as compared to control IgG-treated, mice after 6 days of IMQ treatment. It is worth pointing out that, under our experimental conditions, γδ T cells and neutrophils infiltrated the lymph nodes and the skin of IMQ-treated mice with similar kinetics, and that the infiltration of both cell types appeared much more consistent after 6 rather than 3 days of IMQ-treatment ( Supplementary Figures 1, 2 ). Interestingly, γδ T cells and neutrophils infiltrating the skin dermis of IMQ-treated mice appear to be in close contact ( Supplementary Figure 3 ). Collectively, our findings suggest neutrophils’ potential negative regulatory role toward the infiltration and expansion of γδ T cells in both the lymph nodes and skin of IMQ-treated mice.

Previous findings have highlighted the capacity of neutrophils to both positively and negatively modulate the effector functions of γδ T cells (41–43). Therefore, we tested the immunomodulatory roles of neutrophils on the proliferation and the production of IL-17 by γδ T cells stimulated in vitro with plate-bound anti-CD3 Abs and soluble anti-CD28 Abs in the presence of 100 ng/mL IL-23 and ˜1 ng/mL IL-1β, as previously described (31, 38). As shown in Figures 5A, B , neutrophils inhibited both the proliferation and the production of IL-17, respectively, by activated γδ T cells. Given that the degree of this inhibitory effect was ratio-dependent ( Figures 5A, B ), in all subsequent experiments we used the 5/1 neutrophil/T cell ratio, a condition in which we obtained a strong and reproducible inhibition of γδ T cell functions by neutrophils. In agreement with previous studies (42, 44), we found that the addition of either catalase (a H₂O₂ scavenger) or of diphenyleneiodonium (DPI, a NADPH oxidase inhibitor) almost completely reverted the immunosuppressive functions of mouse neutrophils on γδ T cells ( Figures 5C, D ). Other inhibitors of neutrophil’s effector functions such as pentoxyfilline (PTX, a degranulation inhibitor) or L-arginine [an arginase-1 (ARG1) inhibitor] were effective neither on the proliferation nor on the production of IL-17 in activated γδ T cells ( Figures 5C, D ). In line with these observations, neutrophils isolated from p47^{phox -/-} mice, that lack NOX2 activity, were unable to effectively inhibit γδ T cell proliferation in vitro ( Figure 6A ). Consistently, by performing a flow cytometric measurement of ROS production, we also observed that wild-type (WT) neutrophils, but of not p47^phox-/- neutrophils, produce ROS in the presence of γδ T cells in the culture ( Figure 6B ). Finally, in line with the fact that the inhibitory functions of different immunosuppressive neutrophil populations have been shown to occur through direct cell contact-dependent mechanisms (45–48), we found that the capacity of neutrophils to inhibit γδ T cell proliferation was significantly lower if neutrophils were physically separated from T cells by the use of transwells ( Figure 6C ).

Taken together, data suggest that neutrophils inhibit γδ T cell functions via a cell contact-dependent ROS production.

Spleen tyrosine kinase (Syk), a member of nonreceptor tyrosine kinases, transmits signals in neutrophils from a variety of immunereceptors, including Fcγ receptors (FcγRs) and adhesion molecules, such as β2 integrins and P-Selectin glycoprotein ligand 1 (PSGL-1) (49–51). As a consequence, Syk ^-/- neutrophils display impaired effector functions, including the production of ROS and the release of granule contents, in response to several inflammatory stimuli (50, 51). Syk-based signaling in neutrophils alone was previously shown to be critical for appropriate host defense to Staphylococcus aureus (37) or the development of inflammatory arthritis (36), suggesting the relevance of this signaling pathway in neutrophils during immune responses. Therefore, we decided to utilize mice carrying the specific deletion of Syk in neutrophils [Syk^fl/flMrp8-cre⁺ mice (36, 37),], available in our laboratory, as an experimental model to test whether the specific impairment of this signaling pathway in neutrophils was sufficient to affect their interactions with γδ T cells in IMQ-induced psoriasis. Consistently, Syk ^-/- neutrophils failed to produce ROS and to inhibit the proliferation of γδ T cells in our in vitro experimental conditions ( Figures 7A, B ). These data validated therefore Syk as a crucial signaling molecule involved in the modulation of the neutrophil capability to inhibit γδ T cell proliferation via a contact-dependent ROS production.

We next performed the IMQ-induced psoriasis model in Syk^fl/flMRP8-cre⁺ mice, to evaluate the effect of Syk-deficiency in neutrophils on disease development. Histological section measurement of dorsal skin in Syk^fl/flMRP8-cre⁺ mice failed to show significant variations in epidermal thickness as compared to control mice (consisting of a mix of Syk^+/+MRP8-cre⁺Syk^+/+MRP8-cre^- mice) after 6 days of IMQ-treatment ( Figure 8A ). However, similarly to neutrophil-depleted mice, Syk^fl/fl MRP8-cre⁺ mice manifested an enhanced expression of skin-associated psoriatic genes, such as S100A7 and Lcn2, as well as a specific increase in the expression of cytokines implicated in the IL-23/T17 axis, including IL-23, IL-22, IL-17, CXCL1 and IL-6 after 6 days of IMQ treatment ( Figure 8B and data not shown). Furthermore, also the number of total and activated γδ T cells producing IL-17 ( Figures 9A–C ), was increased in the draining lymph nodes of Syk^fl/fl MRP8-cre⁺ mice, as compared to control mice, after 6 days of IMQ treatment. In a similar fashion, the number of dermal γδ T cells ( Figure 9D ) infiltrating into the skin of IMQ-treated of Syk^fl/fl MRP8-cre⁺ mice was significantly increased as compared to IMQ-treated control mice.

It is noteworthy to remark that, in line with the fact that Syk is not directly involved in controlling neutrophil migration to the inflammatory sites (51), we did not notice any significant difference in the capacity of Syk^-/- neutrophils to infiltrate the lymph nodes and the skin in response to IMQ treatment ( Supplementary Figures 4A, B ).

Overall, data suggest that Syk-dependent signaling pathways controlling neutrophil effector functions, but not neutrophil migration, are required for neutrophil-mediated inhibition of γδ T functions in vitro, and in IMQ-induced psoriasis in vivo.