Blood from anonymized, healthy volunteers (20 ml per donor) was provided under informed consent and collected through the Novartis Tissue Donor Program (TRI0128) in accordance with the Swiss Human Research Act and approval of the responsible ethic committee (Ethikkommission Nordwest- und Zentralschweiz number: 329/13). Anonymized buffy coats from healthy volunteers were collected through the InterRegionale Blutspende of the Swiss Red Cross in Bern, Switzerland in compliance with article 2.2 lit. b of the Swiss Human Research Act.

Fresh skin discards obtained from healthy, female, Caucasian subjects undergoing aesthetic or reconstructive, abdominoplastic surgical resection were supplied by Alphenyx, Marseille, France. Studies were conducted in accordance with the ethical principles originating in the Declaration of Helsinki. Authorizations for the studies (AC-2014-2141 and IE-2014-749 for exportation) were obtained from the French Ministry of Research with the approval of the French Ethical Committee (Comité de Protection des Personnes). All subjects provided written informed consent before participating in the studies.

Cpd A, [(S)-N-(8-((4-(5-(tert-butyl)-1,2,4-oxadiazol-3-yl)-3-methylpiperazin-1-yl)methyl)-7-methylimidazo[1,2-a]pyridin-6-yl)-6-methylnicotinamide] was synthesized as described in Figure S1. Briefly, the 2-methylpyrimidine-5-carboxamide group of (S)-N-(8-((4-(5-(tert-butyl)-1,2,4-oxadiazol-3-yl)-3-methylpiperazin-1-yl)methyl)-7-methylimidazo[1,2-a]pyridin-6-yl)-2-methylpyrimidine-5-carboxamide [prepared as described in Hintermann et al. (34)] was cleaved with sodium hydroxide. The intermediate amine was subsequently converted into Cpd A by reaction with 6-methylnicotinic acid using T3P as the amide-coupling reagent.

The published RORγt inhibitor SR2211 (42) and the pan-JAK inhibitor CP690550 (43) were purchased from Tocris.

Compound-mediated disruption of the interaction between the human RORγt-LBD (264-518, produced at Novartis) with a biotinylated RIP140 derived co-activator peptide (biotinyl-NH-Ahx-NSHQKVTLLQLLLGHKNEEN-CONH₂, Thermo Scientific) was quantified by TR-FRET as previously described (34, 44). Briefly, Cpd A was incubated with a mixture consisting of 6 nM His₆-tagged RORγt-LBD, 90 nM biotinylated RIP140 co-activator peptide, 0.45 nM Cy5-labeled streptavidin (GE Healthcare), 1.5 nM Europium-labeled anti-His₆ antibody (Perkin Elmer). After 1 h of incubation, the emissions were measured at 615 and 665 nm. Concentration-response curves were generated and IC₅₀ values were calculated using the GraphPad Prism software package.

HUT78 T-cells stably expressing full-length eGFP-RORγt or eGFP-empty vector were stimulated and Cpd A-induced inhibition of IL-17A production was measured as described before (45). Briefly, HUT78 cells stably expressing RORγt (5 × 10⁴/well) were stimulated with PMA (8 ng/ml, Sigma) plus CD3 mAb (2.5 μg/ml, clone OKT-3, BioXcell) and incubated in X-vivo 15 medium containing 10% FCS with diluted Cpd A. After 48 h of incubation at 37°C, the supernatants were collected and IL-17A or IL-2 were quantified by ELISA according to the manufacturer‘s specifications (eBioscience or Biolegend).

HUT78 T-cells stably expressing full-length eGFP-RORγt or eGFP-empty vector were generated as described before (45). Cells were transfected with plasmids containing 4 × ROREs inserted into the pGL4.17 luciferase Firefly reporter plasmid (kindly provided by Dr. Juergen Reinhardt, Novartis). As a control construct, a mutated or a non-functional form of 4 × RORE (Mut-RORE), was transfected in eGFP-RORγt HUT78 T-cells. The plasmids contained the following sequences where the bold nucleotides represent the putative RORγt binding site or the mutated form of the RORE:

Wild-type:

GGTAAGTAGGTCATGGTAAGTAGGTCATGGTAAGTAGGTCATGGTAAGTAGGTCATCGTGAC

CCATTCATCCAGTACCATTCATCCAGTACCATTCATCCAGTACCATTCATCCAGTAGCACTG

Mutated:

GGTAAGTACCTCATGGTAAGTACCTCATGGTAAGTACCTCATGGTAAGTACCTCATCGTGAC

CCATTCATGGAGTACCATTCATGGGAGTACCATTCATGGAGTACCATTCATGGAGTAGCACTG

Two days after transfection, cells were incubated in X-vivo 15 medium containing 10% FCS and 700 μg/ml Geneticin (Corning) and stable clones expressing the 4 × ROREs or mutated-ROREs were obtained by limiting dilution. The luciferase expression of the individual cell clones was measured using Luciferase detection buffer (Perkin Elmer) and the RORC mRNA expression of the selected clones were quantified by qRT-PCR.

Stable eGFP-RORγt HUT78 cells expressing either the 4 × ROREs or mutated-ROREs, eGFP-empty vector HUT78 cells expressing 4 × ROREs (5 × 10⁴/well) were stimulated with PMA (8 ng/ml, Sigma) plus CD3 mAb (2.5 μg/ml, clone OKT-3, Bioxell) and incubated in X-vivo 15 medium containing 10% FCS with diluted Cpd A. After 48 h of incubation at 37°C, the supernatants were removed and 40 μl of Luciferase detection buffer (Steady lite plus Reporter Gene assay system, Perkin Elmer) was added to the cells and further processed according to the manufacturer‘s instructions. The total luminescence signal of Firefly expression was measured by EnVision Multilabel Plate Reader (Perkin Elmer).

Selectivity against other ROR family members was assessed in a GAL4 reporter gene assay as previously described (44). Briefly, Jurkat cells stably expressing the pGL4.35 plasmid (Promega) containing nine GAL4 binding sites inducing the transcription of the luciferase reporter gene were transiently transfected with the human RORα, RORβ or RORγt-LBD/GAL4-DBD constructs. After 24 h post-transfection, cells were incubated with Cpd A for an additional 24 h before luciferase activity was determined using an EnVision Multilabel Plate Reader (Perkin Elmer).

Total, naive and memory CD4⁺ T cells were isolated from Peripheral blood mononuclear cells (PBMCs) from healthy volunteers by density gradient centrifugation using Ficoll Paque (GE). Total, naive or memory CD4⁺ T-cells were prepared by immunomagnetic separation using the respective CD4⁺ T-cell enrichment kits resulting in untouched highly purified subsets (EasySep™, Stem Cell). Purity was assessed by flow cytometry and was between 95 and 98% for the total CD4⁺ T-cells. The purity of naive (CD45RA⁺CD45RO⁻) and memory (CD45RA⁻CD45RO⁺) CD4⁺T-cells ranged between 95 and 98%, respectively. For T-cell stimulation experiments, cells were treated as previously published (44). Briefly, T-cells were stimulated with plate-bound monoclonal antibodies against CD3 (1 μg/ml, clone OKT-3, BioXcell) and CD28 (1 μg/ml, clone CD28.2, BioLegend). For Th17 polarization assays, IL-6 (20 ng/ml), TGF-β1 (5 ng/ml), IL-1β (10 ng/ml), and IL-23 (10 ng/ml) were added. For Th1 polarization assays, IL-12 (10 ng/ml) and anti-IL-4 antibody (5 μg/ml), for Th2 polarizations, IL-4 (10 ng/ml) and anti-IFN-γ antibody (5 μg/ml) were added. For Th0 polarization assays, cells were only stimulated with CD3 and CD28 antibodies. Supernatants were collected after 3 or 7 days of incubation for total or naive/memory CD4⁺ T-cell polarizations, respectively. Cytokine concentrations were quantified by ELISA according to the manufacturer's instructions (eBioscience or BioLegend).

PBMCs from freshly drawn heparinized blood originating from healthy volunteers were isolated by density gradient centrifugation using Ficoll Paque (GE). PBMCs (1 × 10⁵/well) were incubated for 96 h in the presence of DMSO or Cpd A (1 μM) together with anti-CD3 (3 ng/ml, clone OKT-3, BioXcell), anti-CD28 (30 ng/ml, clone CD28.2, BioLegend), TGF-β1 (0.2 ng/ml, Biolegend), IL-2 (15 ng/ml, Biolegend), anti-IFN-γ antibody (5 μg/ml, Biolegend), and anti-IL-4 (5 μg/ml, Biolegend). Cells were prepared for CD4/CD25 surface staining, followed by intracellular FoxP3 staining.

Th17 cells obtained from in vitro cultures after 72 h were incubated for 4 h at 37°C with 5 ng/ml of PMA (Sigma) and 500 ng/ml of ionomycin (Calbiochem) in the presence of 10 μg/ml Brefeldin A (Sigma). The cells were washed, fixed and permeabilized in Cytofix/Cytoperm buffer (BD Biosciences) according to the manufacturer‘s instructions prior to intracellular staining with PE labeled IL-17A antibody (eBioscience) and APC-labeled IFN-γ antibody (eBioscience).

For FoxP3 detection within the CD4⁺ CD25 high T-cells, surface staining was performed for 20 min with FITC labeled CD4 antibody and anti-CD25 BV421 antibody (BD Bioscience). After surface staining, the cells were washed, fixed and permeabilized according to the manufacturer's instructions (eBioscience) prior to intracellular staining with APC labeled FoxP3 antibody (eBioscience). Data were acquired on a LSR Fortessa (BD Bioscience) and analyzed using FlowJo software (Tree Star). FoxP3 expression was determined in gated CD4⁺ CD25 high T-cells that were treated with DMSO or with Cpd A.

To assess whether Cpd A affected STAT3 activity we induced STAT3 phosphorylation in Ba/F3 cells stably expressing the human IL-21 receptor (kindly provided by Dr. Max Warncke, Novartis). Cells were pre-incubated for 1 h with 10 μM of Cpd A or with 1 μM of the pan-JAK inhibitor CP690550. To induce STAT3 activity, the cells were stimulated for 1.5 h with recombinant human IL-21 (3 nM), followed by two washing steps using buffer containing PBS/0.5% BSA. The cells were fixed with 2% paraformaldehyde for 15 min at room temperature, washed and resuspended in ice-cold 90% methanol solution and stored overnight at −20°C. The samples were washed and stained for 1 h with an Alexa 647-labeled anti-phospho (pY705)-STAT3 antibody (BD biosciences). After two washes the cells were analyzed using a LSR Fortessa (BD Bioscience) and analyzed using FlowJo software (Tree Star).

In order to detect RORγt binding to its cognate DNA binding sites, biotinylated oligonucleotide probes containing putative 4 × RORE or mutated-RORE DNA sequences were synthesized (Microsynth, Switzerland) and annealed in order to form a double stranded DNA probe. The probes contained the following sequences (where the bold nucleotides represent a putative RORγt binding site or the mutated form of the RORE):

RORE sense strand: 5'GGTAAGTAGGTCATGGTAAGTAGGTCATGGTAAGTAGGTCATGGTAAGTAGGTCATCGTGAC-3' Biotin

RORE antisense strand: 3'-CCATTCATCCAGTACCATTCATCCAGTACCATTCATCCAGTACCATTCATCCAGTAGCACTG-5'

RORE mutated sense strand: 5'-GGTAAGTACCTCATGGTAAGTACCTCATGGTAAGTACCTCATGGTAAGTACCTCATCGTGAC-3' Biotin

RORE mutated antisense strand: 3'-CCATTCATGGAGTACCATTCATGGGAGTACCATTCATGGAGTACCATTCATGGAGTAGCACTG-5'

Nuclear extracts originating from HUT78 cells stably expressing eGFP-RORγt were prepared as previously described (45). Briefly, cells were incubated with Cpd A (10 μM final concentration) for 0.5 h followed by a 2 h-stimulation period using CD3 antibody (2.5 μg/ml, clone OKT-3, BioXcell) plus PMA (8 ng/ml, Sigma). Nuclear lysates were generated using a nuclear extraction kit (Pierce, Thermo scientific), followed by a pre-clearing step with streptavidin-sepharose beads (GE healthcare Amersham) for 1 h at 4°C. The supernatants were incubated with biotinylated double-stranded RORE or with mutated RORE probes (10 nM final concentration) for 1 h at 4°C, followed by addition of a 50% slurry of streptavidin-sepharose beads for 1 h. The bound complexes were washed five times with 1 ml buffer consisting of 50 mM Tris-HCl, 1 mM EDTA, 1% NP-40, 150 mM NaCl and protease inhibitors (Roche). After the final wash, the pull-downs were resuspended in 2 × SDS-PAGE sample buffer (Sigma), boiled and used for SDS-PAGE and Western blotting. The samples were immunoblotted with a polyclonal RORγ-specific antibody (Novus Biologicals). As a loading control, nuclear extracts containing 2 × 10⁶ cell equivalent were blotted with RORγ-specific antibody and with GAPDH-specific antibody (clone 14C10, Cell Signaling). Protein bands were visualized by enhanced chemiluminescence (Western bright Quantum, Advansta) followed by detection of the band using a Western blot imaging system (Fusion FX, Vilber Lourmat).

The HaCaT keratinocyte cell line was purchased from AddexBio and was maintained in RPMI 1640 medium supplemented with Glutamax, Penicillin/Streptomycin (all from Gibco) and 10% FCS (PAA). Normal primary human epidermal keratinocytes (NHEKs) from a male and female caucasian adult were purchased from Lonza and PromoCell, respectively. Cells were cultured in Keratinocyte basal growth medium (KBM; Lonza) containing KGM2 growth supplements including insulin, human epidermal growth factor, bovine pituitary extract, hydrocortisone, epinephrine, transferrin, and gentamicin/amphotericin B (Lonza). HaCaT cells were incubated in a 96-well-microtiter plate (2 × 10⁴ cells/well) with Th17 cytokine-containing supernatants obtained from total CD4⁺ T-cells that were polarized for 72 h in the presence of anti-CD3/CD28 antibodies and a Th17 polarization cocktail as described above. The conditioned supernatants were generated in the presence of a pharmacologically active Cpd A concentration (10 μM) and the keratinocyte responses were compared with an inactive Cpd A concentration (0.13 nM). Alternatively, HaCaT and NHEKs (1 × 10⁴ cells/well) were used in spike-in experiments using recombinant human cytokines at the concentrations determined in the Th17 supernatants exposed to bioactive concentration (10 μM) and inactive Cpd A concentration (0.13 nM). Cytokine concentrations in supernatants were determined from representative experiments and they contained approximately 150 pg/ml or 600 pg/ml of IL-17 in 10 μM or in 0.13 nM treated samples, respectively. The TNF-α concentration was similar in supernatants exposed to low or high Cpd A concentrations and was 4 ng/ml. To study the impact of the IL-17 cytokine, IL-17 in the supernatants and in the spike-in cocktails was neutralized by a 0.5 h pre-incubation with an IL-17-specific antibody [0.75 μg/ml; Secukinumab (Cosentyx®, Novartis)]. All cultures were incubated in triplicates and after a 24 h incubation period, HaCaT cells (6 × 10⁴ cells/sample) and NHEKs (1 × 10⁵ cells/sample) were harvested and further processed for qRT-PCR analysis.

Full-thickness skin punch biopsies (8 mm diameter) were taken from surgical discards. Briefly, subcutaneous fat was removed from the skin ranging in size between 150 and 200 cm² and 8 mm punches were obtained. Samples were incubated in IMDM medium supplemented with Glutamax, 10% FCS, 1% non-essential amino acids, 1 mM Sodium pyruvate, Penicillin/Streptomycin and Amphotericin B (2.5 μg/ml). Incubations were done in the presence of various concentrations of Cpd A or DMSO vehicle control together with a Th17-inducing cocktail consisting of IL-6 (20 ng/ml), TGF-β1 (5 ng/ml), IL-1β (10 ng/ml), IL-23 (10 ng/ml), anti-IL-4/IFNγ (5 μg/ml), and anti-CD3/CD28 antibodies (1 μg/ml). On day 4, supernatants were harvested and analyzed for IL-17 cytokine expression, whereas skin punch biopsies were snap frozen, stored at −80°C and further processed for qRT-PCR analysis.

Total RNA from human Th17 cells (6 × 10⁵ cells), RORγt-RORE HUT78 cells (4 × 10⁵ cells) and HaCaT cells (6 × 10⁴ cells) was extracted using RNeasy kit including a DNAse I digestion step according to the manufacturer‘s protocol (Qiagen). Frozen skin biopsies (8 mm punches) were resuspended in RLT buffer (Qiagen) containing 1% β-mercaptoethanol and homogenized using a Precellys-Cryolys device (Bertin Instruments) according to the manufacturer's instructions. cDNAs were prepared with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Quantitative RT-PCR analysis was performed using a TaqManViiA7 (Applied Biosystems). All runs were accompanied by the internal control gene human β-glucoronidase (Gus) (431088E). The samples were run in triplicate or in hexaplicate and normalized to Gus using the ΔΔCt method to provide arbitrary units representing relative expression levels. The following Taqman probes were used for RT-PCR (Thermo Scientific): RORC (Hs01076112_m1), IL17A (Hs00936345_m1), IL22 (Hs01574154_m1), IL26 (Hs00218189_m1), IL23R (Hs00332759_m1), CCR6 (Hs00218189_m1), NFKBIZ (Hs00355476_m), DEFB4A (Hs00823638_m1), IL36G (Hs00219742_m1), and CCL20 (Hs00355476_m1).

The means were compared using both ANOVA followed by Dunnett's test for multiple comparisons (GraphPad Software Corporation Inc. CA). A p < 0.05 was considered statistically significant.

In a previous communication, we reported the discovery of a series of imidazopyridine RORγt inhibitors by a structure-based hit-to-lead optimization effort (34). Cpd A, which is an advanced representative of this series (Figure 1A), was chosen for investigating further in vitro RORγt-related pharmacology. Nuclear hormone receptors interact with co-activators and co-repressors to control gene transcription (46). We therefore examined in a biochemical TR-FRET assay, the effect of Cpd A to interfere with the RIP140 co-activator peptide binding to the human RORγt-LBD. Consistent with our previous studies, Cpd A disrupted the interaction of the RIP140 peptide with the RORγt-LBD in a concentration-dependent manner with an IC₅₀ value of 5.7 nM. Similar IC₅₀ values in the RIP140 co-activator displacement assay were obtained when the published RORγt inhibitor SR2211 was used (Figure 1B).

To confirm that Cpd A represses the transcriptional activity of RORγt in a cellular context, we performed a RORγt-LBD/GAL4 based transactivation assay in Jurkat T-cells where luciferase reporter gene transcription is driven by the GAL4-RORγt-LBD fusion construct. We could demonstrate that Cpd A possessed good activity in a T-cell environment resulting in complete inhibition of GAL4 promoter activity with an IC₅₀ value of 95 nM, while the SR2211 compound was less potent (Figure 1C). To address the potency of Cpd A against the closely related ROR family members RORα and RORβ we performed LBD-GAL4 reporter gene assays and could show that the inhibitor lacked activity toward the RORα and RORβ isoforms, indicating that the compound acted selectively against RORγt (Figure 1D). To confirm that Cpd A inhibits full-length RORγt, we expressed the full-length version of RORγt along with a 4 × RORE response element that induces luciferase gene expression once the elements are occupied by RORγt. For these assays, we used HUT78 T-cells that stably expressed human RORγt, as previously described (45), which were transfected with plasmids containing the consensus 4 × RORE DNA binding elements (AGGTCA) driving the luciferase reporter gene activity. As a control, HUT78-RORγt expressing cells were stably transfected with non-functional, mutated forms of the RORE (ACCTCA). HUT78 cells that did not express RORγt that contained the 4 × RORE did not have any detectable luciferase activity, whereas in RORγt expressing cells RORE-driven reporter gene activity was significantly up-regulated (Figure 2A). Similarly, RORγt expressing cells transfected with the mutated version of the ROREs failed to show any luciferase reporter activity (Figure 2A). Expression of RORγt mRNA was not altered in cells transfected with the functional or mutated RORE versions ruling out that lack of transcriptional activity in the clone carrying the mutated ROREs occurs as a result of reduced RORγt expression (Figure 2B). Treatment with Cpd A significantly repressed 4 × RORE-driven luciferase activity with an IC₅₀ value of 64 nM (Figure 2C). To further examine the activity of the inhibitor in a more natural setting, we determined whether Cpd A could inhibit endogenous IL-17A secretion in the HUT78 T-cell line that stably expressed full-length RORγt. When these cells are stimulated with anti-CD3 Ab plus PMA, IL-17A secretion is dependent on RORγt expression (45). Pharmacological inhibition of RORγt by Cpd A resulted in complete attenuation of IL-17A production in a concentration-dependent manner with an IC₅₀ value of 27 nM, whereas the IC₅₀ value for SR2211 was ~10-fold higher (Figure 2D). The observed inhibition could be due to non-specific suppression of CD3/PMA-induced T-cell signaling. However, IL-2 cytokine production which is not controlled by RORγt was not affected by Cpd A (Figure 2E) indicating that the inhibitor interfered exclusively with the IL-17 pathway.

In summary, Cpd A showed potent and selective inhibition of RORγt-mediated reporter gene transcription and of IL-17A production in various T-cell assays expressing either full-length RORγt or the LBD of RORγt. By contrast, SR2211 was in all our cellular assays less potent than Cpd A.

In order to gain mechanistic insight of how Cpd A inhibits RORγt transcriptional activity it was further investigated whether the compound interfered with the ability of RORγt to bind to its cognate RORE DNA binding sites. Based on the results in the RORE-dependent transactivation assay described above (Figures 2A–C), the compound may inhibit the transcriptional activity of RORγt by disrupting binding to the ROREs located at the target DNA regulatory sites or by interfering with co-factor assembly crucial for target gene transcription without affecting binding to the ROREs.

We explored whether Cpd A may adversely affect the DNA-binding activity of RORγt by assessing the interaction between this nuclear hormone receptor and the ROR elements. In a previous communication, we reported that recruitment of RORγt to the ROREs occurs after stimulation of RORγt HUT78 T-cells with anti-CD3/PMA (45). Using this cellular model, nuclear extracts were prepared from stimulated RORγt HUT78 T-cell transfectants and pull-down studies were conducted using the DNA oligonucleotides containing the RORE sequences (AGGTCA) or the non-functional RORE sequences (ACCTCA). Western blot analysis of the pull-down experiments revealed that RORγt occupied only the functional RORE sequences, whereas RORγt binding was not detectable when the mutated versions of the RORE oligonucleotides were used as a specificity control (Figure 2F). RORγt association with the ROR elements was still intact in cells that were treated with the inhibitor and was similar compared to DMSO treated samples (Figure 2F). These results indicate that Cpd A inhibits the activity of RORγt through interference with co-factor assembly and/or other transcription factor interactions rather than interfering with the DNA-binding capacity of RORγt.

We next addressed whether Cpd A impacted Th17 differentiation in primary human T-cells. Total CD4⁺ T-cells originating from healthy volunteers were isolated and stimulated with anti-CD3 plus anti-CD28 specific antibodies together with a Th17 skewing cytokine cocktail in the presence of Cpd A. The inhibitor showed better potency than SR2211 and potently suppressed IL-17A secretion in a concentration-dependent manner with an IC₅₀ value of 39 nM (Figure 3A). Likewise, intracellular staining followed by flow cytometric analysis revealed that Cpd A reduced the frequencies of IL-17A⁺ cells within the CD4⁺ T-cell population during the 3-day Th17 differentiation culture period while the frequencies of INF-γ⁺ cells were not affected (Figure 3B). When naive or memory human CD4⁺ T-cells were differentiated for 7 days toward the Th17 cell phenotype, Cpd A was potent and impaired IL-17A cytokine secretion in both CD4⁺ T-cell subsets with IC₅₀ values of 15 and 27 nM for naive and memory CD4⁺ T-cells, respectively (Figures 3C,D). While Cpd A impaired IL-17A secretion in CD4⁺ T-cells that were skewed toward the Th17 phenotype, it did not have any impact on Th0, Th1, or Th2 signature cytokine production (IL-2, IFN-γ, and IL-13, respectively), indicating that the compound inhibited selectively the Th17 pathway (Figures 3E–G).

Regulatory T-cells (Tregs) and Th17 cells are reciprocally regulated, depending on the balance between the expression of RORγt and FoxP3 and on the cytokine environment (47, 48). It is possible that RORγt inhibition mediated by Cpd A may skew the Th17/Treg cell ratio toward the Treg pathway. We determined whether our RORγt inhibitor had an impact on the expression of the transcriptional regulator FoxP3 which is crucial for the development and inhibitory function of regulatory T-cells (Tregs) (49). PBMCs were cultured for 96 h under Treg inducing conditions in the presence of Cpd A and FoxP3 expression among CD4⁺CD25high T-cells was monitored. Flow cytometric analysis of gated CD4⁺CD25high T-cells revealed high expression of FoxP3 in these cells and incubation of our RORγt inhibitor did not result in altered FoxP3 expression as shown by similar mean fluorescence intensities between DMSO or inhibitor treated cells (Figure 3H), indicating that Cpd A does not potentiate Treg development.

Since STAT3 transcriptional activity is critical for Th17 differentiation (50) we confirmed that our compound lacked any activity against the JAK/STAT3 pathway. We monitored the phosphorylation state of STAT3 in DMSO or Cpd A treated Ba/F3 cells. These cells stably express the human IL-21 receptor and are responsive to IL-21 stimulation by upregulating STAT3 phosphorylation. IL-21-induced STAT3 activity was monitored by flow cytometry. Inhibition of the JAK pathway by the pan-JAK inhibitor CP690550 blocked IL-21-induced STAT3 activity as evidenced by impaired phosphorylated (pY705)-STAT3 (Figure 3I). In contrast, Cpd A did not affect STAT3 phosphorylation compared to DMSO treatment (Figure 3I), suggesting that the JAK/STAT3 pathway is not directly targeted by Cpd A.

We next explored whether expression of genes that are known to be regulated by RORγt were downregulated by the compound during the Th17 polarization. Using total CD4⁺ T-cells that were cultured for 72 h with a Th17 skewing cytokine mix we could show, that consistent with the effect of inhibition of IL-17A protein expression, Cpd A reduced RORγt signature gene expression in a concentration-dependent manner. In particular, pro-inflammatory Th17 pathway genes including IL17A, IL17F, IL22, IL26, IL23R, and CCR6 were significantly attenuated by Cpd A (Figures 4A–F).

In summary, the RORγt low molecular weight compound reported here consistently and selectively suppressed not only IL-17A production by CD4⁺ T-cells, but also Th17-associated pro-inflammatory gene expression in human primary Th17 cells.

Human skin harbors a heterogeneous pool of immunocompetent cells, including DC subsets, Langerhans cells, CD4⁺, CD8⁺ T-cells, γδ T-cells, and innate lymphoid cells, that are able to produce pro-inflammatory cytokines/chemokines that can contribute to the pathogenesis of skin inflammatory disorders. Because of the good inhibitory profile of our compound we next investigated the impact of Th17 pathway inhibition on human keratinocyte responses triggered by Th17-derived cytokines present in supernatants originating from compound treated Th17 cells. For this, we selected cell supernatants originating from total CD4⁺ Th17 cells that were incubated with Cpd A for 72 h at a pharmacologically inactive concentration (0.13 nM) and at an active concentration (10 μM), containing 600 and 150 pg/ml of IL-17, respectively as depicted in Figure 3A. To determine the impact of IL-17 to induce keratinocyte responses, we depleted this cytokine from the supernatants using Secukinumab, a neutralizing IL-17-selective antibody prior to the co-culture experiment. IL-17 responsive transcripts that are upregulated in psoriasis were analyzed in keratinocytes and included NFKBIZ, DEFB4, IL36G, and CCL20. To evaluate the contribution of the polarizing Th17 cytokines present in the Th17-derived supernatants keratinocyte cultures were incubated with CD3/CD28-specific antibodies together with IL-6, IL-1β, TGF-β, and IL-23 at the same concentrations used in the Th17 polarization assays. Alternatively, recombinant human cytokines at the respective concentrations determined in the supernatants from Cpd A-treated cells were spiked into the HaCaT cultures to determine the contribution of other cytokines present in the Th17 supernatants on keratinocyte responses. A feature of IL-17 is that it acts in a synergistic way with other cytokines, particularly TNFα, to induce downstream gene/protein expression (51, 52). One abundant cytokine present in the Th17 supernatants was TNFα (4 ng/ml) whose level was not affected by RORγt inhibition. The Th17 supernatants contained similar IFN-γ concentrations and spike-in experiments revealed that this cytokine either alone or in combination with IL-17 and TNFα did not modulate expression of the IL-17 activated genes and therefore IFN-γ was omitted (data not shown). Upon 24 h activation with Th17-derived supernatants containing biologically inactive concentrations of Cpd A (0.13 nM), keratinocytes up-regulated expression of the IL-17-induced transcription factor NFKBIZ, the antimicrobial peptide DEFB4A and the chemokine CCL20 and IL36G (Figure 5A). In contrast, supernatants obtained from Th17 samples that contained pharmacologically active concentrations of the RORγt inhibitor (10 μM) showed a marked reduction of these IL-17 signature genes. Consistent with the observation that the inactive and biologically active Cpd A-treated samples contained 150 and 600 pg/ml of IL-17, respectively, neutralization of IL-17 by Secukinumab led to further downregulation of these transcripts to the levels observed when the keratinocytes were incubated with the Th17 cytokines only (Figure 5A). The effect of Secukinumab to reduce these IL-17 activated genes was less pronounced in supernatants originating from cells treated with high Cpd A concentration. These results demonstrate that IL-17 was the main cytokine present in the Th17 supernatants that induced the expression of NFKBIZ, DEFB4, and CCL20 in the keratinocyte cultures. Furthermore, the inhibitory capacity of Cpd A, when used at 10 μM, to attenuate IL-17 production and IL-17-induced gene activation was almost as pronounced as antibody-mediated neutralization of IL-17. IL36G expression was partially dependent on IL-17 because Secukinumab led to a moderate inhibition of IL36G indicating that other cytokines present in the Th17 supernatants or present in the Th17 cytokine mix induced IL36G gene expression. Similar results were obtained when recombinant IL-17A and TNFα cytokines were spiked into the keratinocyte cultures. NFKBIZ, DEFB4, IL36G, and CCL20 expression was induced following stimulation of HaCaT cells with IL-17A/TNFα and the upregulation of these transcripts correlated directly with the IL-17A concentrations (600 pg/ml vs. 150 pg/ml) that were used for keratinocyte activation (Figure 5B). Co-stimulation of keratinocytes with 150 pg/ml of IL-17 which was the level determined in Th17 supernatants from cells treated with 10 μM of Cpd A induced only minimal NFKBIZ, DEFB4, IL36G, and CCL20 expression. Pre-incubation of the cytokines with Secukinumab led to significantly impaired IL-17A/TNFα-mediated gene expression similar to the levels observed with TNFα-induced activation (Figure 5B). Because IL-17-mediated responses elicited by immortalized HaCaT cells may differ from those induced by primary keratinocytes, additional spike-in experiments using IL-17A/TNFα cytokines similar to those depicted in Figure 5B were performed using NHEKs. NHEKs responded in a similar way to IL-17A/TNFα stimulation by upregulating NFKBIZ, DEFB4A, CCL20, and IL36G gene expression (Figure 5C), although the IL-17A/TNFα stimulation index was lower in NHEKs mostly due to the higher basal gene expression levels in these cells. This was particular evident when CCL20 transcript levels were analyzed. Upregulation of these transcripts correlated with the IL-17 concentrations that were used for the spike-in experiments (600 pg/ml vs. 150 pg/ml). Like in the HaCaT cells, neutralization of IL-17 by Secukinumab in NHEKs resulted in reduced gene expression similar to those observed induced after TNFα stimulation.

In summary, these results demonstrate that Th17 cell-derived IL17 is the key cytokine displaying pro-inflammatory activity in keratinocytes and that pharmacological inhibition of RORγt was efficacious to attenuate IL-17-driven keratinocyte responses.

In the next set of experiments we wanted to translate these findings and investigated whether addition of Cpd A was able to alter IL-17 pathway responses in freshly prepared full-thickness skin cultures. To recapitulate the predominant Th17/IL-23 cytokine signature occurring in psoriasis, immunocompetent skin-resident cells were activated with anti-CD3/CD28 antibodies together with a Th17 skewing cytokine cocktail, analogous to our blood-derived T-cell polarization studies. The full-thickness skin punch biopsies were incubated in the presence of Cpd A and IL-17 cytokine production and IL-17-induced DEFB4A expression in the skin was analyzed. Stimulation of skin-resident immune cells under Th17-polarizing conditions elicited significant IL-17A cytokine secretion in ex vivo human skin tissue, whereas skin samples that were left unstimulated barely produced IL-17A (Figure 6A). Addition of the RORγt inhibitor at the beginning of the activation period resulted in a concentration-dependent inhibition of IL-17A secretion by skin-resident immunocompetent cells (Figure 6A). In additional experiments, we investigated the impact of pharmacological inhibition of RORγt on secondary responses that may have been altered in the skin. Skin punches were homogenized and subjected to gene expression analysis by PCR. Correlating with the IL-17A inhibition pattern, h-BD2 gene expression was significantly down-regulated in Cpd A-treated skin punches (Figure 6B). Abundancy of NFKBIZ, CCL20, and IL36G transcripts were very low and results were inconclusive (data not shown).

These results suggest that selective RORγt inhibition during Th17 skewing conditions leads to attenuated IL-17 cytokine expression by immunocompetent skin-resident cells resulting in impaired secondary gene activation responses in keratinocytes and in ex vivo skin immersion cultures.

All authors work for Novartis. CG, KK, DO, and SH own shares and/or options from Novartis.