The following mouse strains were used in this report: Ets1^-/- (Ets1 KO) (2), IL17RA^-/- (IL17RA KO) (32), and Ets1^-/-IL17RA^-/- (DKO mice) as well as wild-type (WT) littermate controls that were obtained by breeding heterozygotes of the above strains. IL17RA KO mice used in this study were obtained from Amgen, Inc (31).. Since the loss of Ets1 results in perinatal lethality on an inbred C57BL/6 background (33), all mice were maintained on a mixed C57BL/6 × 129Sv genetic background. Blood was obtained under anesthesia using either retro-orbital bleed or cardiac puncture. Animals used for experiments were euthanized with CO₂ induction followed by cervical dislocation. Mice were housed under specific pathogen-free (SPF) conditions for the duration of the studies. All studies were approved by the University at Buffalo Institutional Animal Care and Use Committee (IACUC).

Spleen and lymph nodes were isolated from WT, Ets1 KO, IL17RA KO, and DKO mice, and single-cell suspensions were prepared. Cells were incubated with Ghost Dye Violet 510 (Tonbo Biosciences, San Diego, CA, USA) to stain dead cells. Cells were subsequently stained with surface antibodies for B- and T-cell marker proteins and by intracellular staining for key transcription factors. Fluorescent signals were collected with an LSRII or Fortessa flow cytometers. Antibodies for flow cytometry were obtained from BD Biosciences, BioLegend, eBioscience, Miltenyi Biotec, or R&D Systems. The following antibodies were used in flow analysis of B-cell subsets: B220 (clone RA3-6B2), CD21 (clone 7G6 or REA800), CD23 (clone B3B4), CD73 (clone eBioTY/11.8), CD80 (clone 16-10A1), CD138 (clone 281-2), FAS (clone Jo2), IgG1 (clone RMG1-1), PD-L2 (clone TY25), and peanut agglutinin (PNA, biotinylated from Vector Labs, Burlingame, CA, USA). The following antibodies were used in flow analysis of T-cell subsets: CD4 (clone GK1.5), CD8b (clone eBioH35-17.2), PD1 (clone 29F.1A12), and CXCR5 (clone L138D7). These antibodies were used for intracellular staining of key transcription factors regulating T helper cell subset differentiation: Tbet (clone 4B10), Foxp3 (clone FJK-16s), RORγT (clone B2D), GATA3 (clone L50-823), and BCL6 (clone 7D1). For isolation of skin-associated lymphocytes, depilated whole skin was harvested and floated on EDTA-free trypsin (Thermo Fisher) overnight at 4°C. The epidermis and dermis were then separated. The epidermal sheets were digested further in trypsin supplemented with 0.01% DNAse I (Millipore Sigma). The dermis was digested with 0.25 mg/mL Liberase (Millipore Sigma) and 1 mg/mL DNAse I (Millipore Sigma). The resulting cell suspensions were enriched for lymphocytes using Lymphoprep (Stemcell Technologies). Following enrichment, cells from the dermis and epidermis were combined and cultured overnight in the presence of 10 U/mL IL-2 (BioLegend). The following antibodies were used in flow analysis of skin-associated γδ T-cell populations: CD4 (clone GK1.5), CD8b (clone eBioH35-17.2), TCRγδ (clone GL3), Vγ5 (clone 536), and CCR6 (clone 29-2L17).

Spleen and lymph nodes from mice were harvested, mechanically disrupted, and passed through a 70-μm filter. Naive CD4⁺ T cells from the sample were purified by magnetic bead separation using the CD4⁺CD62L⁺ isolation kit (Miltenyi Biotec). Cells were plated in 96-well plates coated with anti-CD3 antibody (5 μg/mL, BD, clone 145-2C11) at a density of 3 × 10⁶ cells/mL. Cells were plated in Tfh-conditioning media (RPMI complete media with 10% fetal bovine serum), 10 μg/mL anti-IL4 (BD Biosciences, clone 11B11), 10 μg/mL anti-IFNγ (BD Biosciences, clone XMG1.2), 10 μg/mL anti-TGFβ (R&D Systems, clone 1D11), 30 ng/mL IL6 (Shenandoah Biotechnology), and 50 ng/mL IL21 (Shenandoah Biotechnology) in the presence of soluble anti-CD28 (2 μg/mL, BD, clone 37.51). Polarized cells were harvested for Tfh phenotyping by flow cytometric analysis 4 days after plating.

For surface phenotype and transcription factor analysis, harvested cells were stained with Live/Dead Fixable Aqua dead cell stain (Thermo Scientific Fisher) to exclude dead cells. Cells were subsequently stained for surface markers (CD4, TCRβ, PD1, CXCR5) and then fixed and permeabilized using the Intracellular Fixation and Permeabilization Buffer Set (eBioscience). Fixed cells were stained with antibody to BLC6 (clone 7D1) or isotype control.

ELISA to detect total serum IgM and IgG and autoantibodies was performed as previously described (2). Serum IgE levels were detected using the BioLegend Mouse IgE ELISA Max Deluxe Set. Maxisorp 96 ELISA well plates were coated overnight with 10 μg/mL of antigens, and ELISA was performed as previously described (2).

Single-cell suspensions were prepared from spleen and lymph nodes and plated on ELISpot plates to detect IgM- and IgG-secreting cells as previously described (2, 15). Spots were counted with an automated counter, and the number of antibody-secreting cells per million total cells plated was calculated.

Hep2 hepatoma cells on glass slides were incubated with 1:40 dilutions of mouse serum and subsequently with fluorescein isothiocyanate (FITC)-conjugated anti-mouse IgG. Staining was performed, and images were captured with a HELmed Integrated Optical System (HELIOS, AESKU Diagnostics).

Skin samples harvested from DKO and control mice were processed in TRIzol (Thermo Fisher), and bulk RNA was then purified. Purified RNA was then analyzed by QuBit/Quant-IT and Fragment Analyzer (Agilent) for quality control. cDNA library preparation was completed using TrueSeq RNA sample preparation kit (Illumina) followed by 50-bp single-end sequencing on an Illumina HiSeq 2500. Analysis of the raw reads using FASTQC v0.11.9 application was used as a quality control metric following sequencing.

The raw reads were then mapped to the GRCm38 (the mm10 reference mouse genome) using TopHat2 (34). The aligned reads were then quantified using featureCounts v1.5.3 to generate a raw count matrix. The raw count matrix was then processed in R to generate transcripts per million normalized expression values as previously described in Wagner et al. (35). Differential gene expression analysis was conducted comparing knockout samples to WT controls using DEseq2 v1.24.0 with genes being identified as statistically significantly differentially expressed with a log2fold change ≥1 and an false discovery rate (FDR) value of ≤0.1. RNA-sequencing data are available under GEO DataSets accession number GSE237696.

Upregulated and downregulated differentially expressed genes (DEGs) identified by the above analysis were then subjected to GO term enrichment analysis. DEGs were input into the online tool g;Profiler (36) to assess for gene ontology (GO) terms significantly enriched among the upregulated and downregulated genes.

To determine if DKO mice have spontaneous colonization of the skin with staphyloccal bacteria, we swabbed the ventral side of the neck and upper thorax and plated swabbed bacteria on mannitol-salt-agar (MSA) plates that specifically promote growth of staphylococcal species while suppressing the growth of other bacteria. Plates were incubated overnight at 37°C, and the colonies were subsequently counted. Genomic sequencing was used to confirm the species of staphylococcal bacteria. To determine if DKO mice are susceptible to exogenous S. aureus infection, a methicillin-resistant Staphylococcus aureus (MRSA) USA300 strain NRS-384 carrying a bioluminescent marker (the lux operon from Photorhabdus luminescens) was obtained from Dr. Roger Plaut at the Food and Drug Administration (FDA) (37). Bacteria were grown in tryptic soy agar to log phase, harvested, washed with PBS, and then resuspended at 200 million cells/mL. Mice to be infected were anesthetized followed by shaving the back skin and making three small topical cuts in the skin. Approximately 10 µL of resuspended bioluminescent S. aureus was introduced into the cuts. Mice were given buprenorphine to control pain. Infected mice were imaged on days 1, 3, 7, 14, and 21 post-infection using an In vivo imaging system (IVIS) imager that detects the bioluminescent signal. Luminescence values were normalized to the signal at day 1 to control for differences in cut size and depth between animals.

Ears were harvested from mice and then treated with Nair to remove hair. The two leaflets of the ear were separated from the underlying cartilage. The ear leaflets were then floated on a 3.6% ammonium thiocyanate solution dermis side down to dissociate the dermis from the epidermis. The epidermal sheet was then peeled off and fixed using 4% formaldehyde. Normal goat serum supplemented with Triton X-100 was used to block and permeabilize the tissue followed by overnight staining using FITC anti-TCRγδ (clone GL3). The epidermal sheets were mounted on a slide using VECTASHIELD Antifade Mounting Medium containing 4’,6’-diamidino-2-phenylindole (DAPI), and the entire ear field was imaged using a Leica TCS SP8 confocal microscope system.

WT and IL17RA KO animals were injected with 100 μg of Ultra-LEAF Purified Anti-mouse T cell receptor (TCR) Vγ5 antibody (clone 536) or Ultra-LEAF Purified Syrian Hamster IgG Isotype control antibody (clone SHG-1). For validation of depletion, epidermal sheets were harvested from ear leaflets as previously described (Jameson et al., 2004). Epidermal sheets were stained with FITC-anti-mouse TCRγδ (clone GL3) and TOPRO-3 Iodide (Thermo Fisher). For determination of DETC contribution to skin S. aureus infection immune responses, WT and IL17RA KO animals were injected with antibody as above and infected 2 days post-injection as described above. The burden of S. aureus was measured using an IVIS imager on days 1, 2, 3, 4, 5, 7, 10, 14, 21, 24, 28, 32, and 35 days post-infection with signal normalized to day 1 in order to control for cut size and depth variation.

One-way ANOVAs or nonparametric Kruskal–Wallis tests were performed followed by Tukey’s or Dunn’s multiple comparison test, respectively, except for the autoantibody ELISAs and IVIS imaging, where 2-way ANOVAs were used followed by a Bonferroni posttest. In all cases, error bars are standard error of the mean (SEM), and p < 0.05 was considered significant.

In order to test the role of IL-17 in the autoimmune phenotype of Ets1-deficient mice, we crossed Ets1 KO mice with mice lacking the IL17RA to generate DKO mice. We analyzed DKO and control WT and single knockout mice at 3–6 months of age. While we anticipated that loss of IL-17 signaling might result in reduced immune cell activation and autoimmune phenotypes, we instead found that DKO mice had a stronger phenotype than Ets1 KO mice. This was shown by greatly enlarged peripheral lymph nodes ( Figures 1A, B ) and more modestly enlarged spleens in the DKO mice as compared to controls ( Figure 1A , Table 1 ). ELISpot assays showed that there was a dramatic increase in antibody-secreting cells in DKO mice compared to Ets1 KO mice, especially IgG-secreting plasma cells in the lymph node ( Figure 1C ). Levels of serum IgM and IgE are higher in Ets1 KO than WT controls, whereas total serum IgG titers are not significantly increased (16, 17). Serum from DKO mice showed a similar increase in IgM as that found in Ets1 KO serum but had a much stronger increase in serum IgG and IgE ( Figure 2A ). On average, the serum IgE levels were >400-fold higher in DKO as compared to WT mice and ~5-fold higher in DKO as compared to Ets1 KO.

DKO mice had higher titers of autoantibodies against dsDNA and Smith antigen (Sm) than Ets1 KO mice ( Figure 2B ). Hep2 cell staining demonstrated that Ets1 KO and DKO mice produce anti-nuclear and anti-cytoplasmic IgG autoantibodies ( Figure 2C ). Unexpectedly, we also found that IL17RA KO mice produce autoantibodies as well. The intensity of Hep2 staining was in general stronger using serum from DKO mice than Ets1 KO or IL17RA KO, indicating increased levels of autoantibodies in DKO mice ( Figure 2D ).

To further explore the phenotype of DKO mice, we analyzed B- and T-cell differentiation status using flow cytometry. DKO mice have increased total B-cell numbers in the spleen and lymph nodes compared to controls ( Figure 3A , Table 1 ). Similarly, DKO mice showed a striking increase in the percentages and numbers of plasma cells and IgG1 class-switched B cells compared to Ets1 KO mice ( Figures 3B, C ; Supplementary Figures S1, S2 ). Like Ets1 KO mice, DKO showed a strong reduction of marginal zone B cells in the spleen ( Figure 3D ). We found that there was an increase in the numbers of germinal center B cells (B220⁺PNA⁺Fas⁺) in the lymph nodes and an increase in the numbers of B cells with a memory phenotype (B220⁺CD80⁺PDL2⁺) in both the spleen and the lymph nodes of DKO mice as compared to controls ( Figures 3E, F ; Supplementary Figures S3, S4 ). These results indicate that DKO mice have a greatly expanded germinal center response and an overall high level of B-cell activation.

In the T-cell compartment, we found that while the overall percentage of CD4⁺ T cells in the spleen was similar in all genotypes ( Figure 4A ), the percentage of CD4⁺ T cells in the lymph nodes of DKO mice was reduced ( Figure 4B , Table 1 ). However, because both IL17RA KO and DKO lymph nodes are larger than those of control mice, the total number of CD4⁺ T cells was in fact higher in IL17RA KO and DKO mice than in WT or Ets1 KO controls ( Figure 4C , Table 1 ). Many CD4⁺ T cells in the lymph nodes of IL17RA KO and especially DKO mice expressed CD44, a marker of an activated or memory phenotype ( Figure 4D ). Furthermore, there was an increased number of cells with a Tfh phenotype (CD4⁺ PD1-hi CXCR5-hi BCL6⁺) in Ets1 KO and DKO spleen and lymph nodes ( Figure 4E , Supplementary Figure S5 ). To further examine the propensity of DKO T cells to become Tfh, we isolated naive CD4⁺ T cells from the spleens and lymph nodes of DKO and control mice and stimulated them in vitro under conditions that promote Tfh differentiation. CD4⁺ T cells from Ets1 KO and DKO mice showed increased differentiation to PD1⁺ CXCR5⁺ ICOS⁺ BCL6⁺ Tfh cells ( Figure 4F , Supplementary Figure S6 ).

We also examined the expression of Foxp3 in CD4⁺ T cells to assess the numbers of Treg cells. A previous report had indicated that there was a 3–4-fold reduction in the percentages of Foxp3⁺ Tregs in the spleens of Ets1 KO mice (17). In our studies, we found that the number of Foxp3⁺ Tregs was similar in the spleens and lymph nodes of WT, Ets1 KO, and IL17RA KO mice and was elevated in DKO mice ( Supplementary Figures S7A, B ). However, similar to a previous study (17), we did find that the intensity of Foxp3 staining was lower in CD4⁺ T cells from mice lacking Ets1, especially in lymph node T cells ( Supplementary Figure S7C ). We also analyzed staining of Tbet, GATA3, and Rorγt in spleen and lymph node cells to determine if there was spontaneous differentiation of CD4⁺ T cells to Th1, Th2, or Th17 fates. The total number of CD4⁺ T cells that stained with GATA3 and RORγt antibodies was elevated in DKO mice ( Supplementary Figures S7C, D ).

By 6 months of age, most DKO mice began to develop skin dermatitis and lesions, often on the ventral surface of the neck ( Figure 5A ). Such lesions were not found on Ets1 KO mice in our colony and were found at a lower rate and lesser severity in IL17RA KO mice. RNA-sequencing experiments using RNA isolated from the skin of DKO and control mice showed upregulation of many pathways associated with an inflammatory immune response, including the response to S. aureus infection ( Supplementary Figure S8A ). Indeed, many cytokines and chemokines involved in the skin immune response were highly overexpressed in the skin of DKO mice as compared to control mice ( Supplementary Figure S8B ).

Mice lacking IL17RA are known to have a defect in clearing S. aureus from the skin (27). Given this observation and the results of RNA-sequencing, we sought to determine whether staphylococcal bacteria could be recovered from the skin of DKO. Swabbing the ventral neck of mice resulted in the recovery of high levels of staphylococcal bacteria from the skins of 100% of DKO mice tested (whether or not they showed visible lesions), while staph was cultured from only two of six IL17RA KO mice tested using this technique. The level of staph bacteria recovered from IL17RA KO mice was also lower than that found in DKO mice ( Figure 5B ). Genomic sequencing showed that recovered bacteria were Staphylococcus xylosus ( Supplementary Figure S8C ), a staph species prominent on mouse skin (38–40) and a frequent cause of skin infections in mice (41–45).

To determine whether DKO mice have an impaired immune response to staphylococcal bacteria, we experimentally infected cuts in the skin of DKO and control mice with a bioluminescent strain of S. aureus (37). S. aureus and S. xylosus are closely related staphylococcal bacteria, and both are common colonizers of the skin. Because bioluminescent S. xylosus is not available, examining DKO responses to S. aureus was used to determine if there is a generalized impairment of anti-staphylococcal immunity in these mice. We monitored the bioluminescent signal at days 1, 3, 7, 14, and 21 after infection. In this assay, Ets1 KO mice did not show any increased susceptibility and cleared the infection with kinetics similar to WT mice ( Figure 5D ). As previously reported, mice lacking IL17RA showed a susceptibility to S. aureus with initially delayed clearance ( Figure 5C ) but were able to fully clear the infection by day 21 post-infection. On the other hand, DKO mice initially started to clear the infection and on day 3 actually showed a signal that was lower than that of IL17RA KO mice and comparable to that of WT and Ets1 KO controls. However, by day 7, the infectious signal in DKO mice increased dramatically and remained elevated until day 21 and beyond ( Figure 5C ). Additionally, although the initial infections were made in the skin of the upper back, in some mice, the infection in DKO mice spread to the face/neck area ( Figure 5D ). We considered whether the susceptibility to skin infection by staphylococcal bacteria might be due to autoimmune-mediated skin damage, given the strongly enhanced autoimmunity in DKO mice. To test whether autoimmune skin damage promotes susceptibility to staph infection, we tested autoimmune MRL/lpr as compared to control non-autoimmune MRL/MpJ mice in the bioluminescent staph infection model. MRL/lpr mice are known to have autoimmune-mediated skin damage, with lesions similar to human cutaneous lupus (46, 47). Despite this, we found that MRL/lpr mice could clear exogenous skin infections with S. aureus with the same kinetics as control MRL/MpJ mice ( Figure 5E ). Thus, the presence of autoimmune skin damage per se does not lead to a defect in staph clearance.

Mice completely lacking γδ T cells have been shown to have a defect in the clearance of staph infections from the skin, while mice lacking αβ T cells were able to clear (27). We found that DKO mice and Ets1 KO mice lack γδ TCR⁺ cells in the epidermis ( Figure 6A ), while IL17RA KO and WT mice have these cells. These results were verified using flow cytometry to identify Vγ5⁺ and Vγ5^- γδ T cells in the skin. As shown in Figure 6B , DKO and Ets1 KO mice lack Vγ5⁺ DETC cells in the skin, while IL17RA KO and WT mice have such cells. On the other hand, all genotypes of mice had Vγ5^- γδ T cells in the skin, and these tended to be slightly increased in the skin of DKO and IL17RA KO mice.

The specific role of DETCs in responding to skin staph infection is unknown. To determine if DETCs might be required for clearance of staph infection, we depleted DETCs using anti-Vγ5 antibody ( Figure 6C ). Since DETCs are the only γδ T cells that express the Vγ5 receptor, the depletion is specific to DETCs and does not affect other γδ T-cell subsets. DETCs were depleted from both WT and IL17RA KO mice, and the mice were subsequently infected with bioluminescent staph and the infection followed with IVIS imaging. As shown in Figure 6D , IL17RA KO mice that had DETCs depleted initially showed normal kinetics of clearance. However, by day 20 post-infection, the bioluminescent signal in depleted IL17RA KO mice began to increase. This was not seen in IL17RA KO mice injected with isotype control Ab or in WT mice injected with either isotype control or Vγ5 antibody. Similar to what we saw in a subset of DKO mice, we found that the infection in some IL17RA KO depleted for DETCs shifted from the initial site of infection on the upper back to the face region (not shown). These results indicate that DETCs cooperate with pathways induced by IL-17 signaling to mediate complete clearance of staph infections from the skin.