Novel Small Molecule Tyrosine Kinase 2 Pseudokinase Ligands Block Cytokine-Induced TYK2-Mediated Signaling Pathways

Abstract

A member of the Janus kinase (JAK) family, Tyrosine Kinase 2 (TYK2), is crucial in mediating various cytokine-signaling pathways such as interleukin-23 (IL23), interleukin-12 (IL12) and type I Interferons (IFN) which contribute to autoimmune disorders (e.g., psoriasis, lupus, and inflammatory bowel disease). Thus, TYK2 represents an attractive target to develop small-molecule therapeutics for the treatment of cytokine-driven inflammatory diseases. Selective inhibition of TYK2 over other JAK isoforms is critical to achieve a favorable therapeutic index in the development of TYK2 inhibitors. However, designing small molecule inhibitors to target the adenosine triphosphate (ATP) binding site of TYK2 kinase has been challenging due to the substantial structural homology of the JAK family catalytic domains. Here, we employed an approach to target the JAK homology 2 (JH2) pseudokinase regulatory domain of the TYK2 protein. We developed a series of small-molecule TYK2 pseudokinase ligands, which suppress the TYK2 catalytic activity through allosteric regulation. The TYK2 pseudokinase-binding small molecules in this study simultaneously achieve high affinity-binding for the TYK2 JH2 domain while also affording significantly reduced affinity for the TYK2 JAK homology 1 (JH1) kinase domain. These TYK2 JH2 selective molecules, although possessing little effect on suppressing the catalytic activity of the isolated TYK2 JH1 catalytic domain in the kinase assays, can still significantly block the TYK2-mediated receptor-stimulated pathways by binding to the TYK2 JH2 domain and allosterically regulating the TYK2 JH1 kinase. These compounds are potent towards human T-cell lines and primary immune cells as well as in human whole-blood specimens. Moreover, TYK2 JH2-binding ligands exhibit remarkable selectivity of TYK2 over JAK isoforms not only biochemically but also in a panel of receptor-stimulated JAK1/JAK2/JAK3-driven cellular functional assays. In addition, the TYK2 JH2-targeting ligands also demonstrate high selectivity in a multi-kinase screening panel. The data in the current study underscores that the TYK2 JH2 pseudokinase is a promising therapeutic target for achieving a high degree of biological selectivity. Meanwhile, targeting the JH2 domain represents an appealing strategy for the development of clinically well-tolerated TYK2 inhibitors that would have superior efficacy and a favorable safety profile compared to the existing Janus kinase inhibitors against autoimmune diseases.

Keywords: JAK homology 2 (JH2) pseudokinase regulatory domain; Janus Kinases (JAK); Tyrosine Kinase 2 (TYK2); cytokine pathways; selectivity; small molecule inhibitors.

Figure 1

Small molecule ligands bind to the TYK2 JH2 pseudokinase domain. (A) Domain structural illustration of TYK2 and schematic model of cytokine-mediated receptor/TYK2 activation: TYK2 protein contains 4 components: the kinase domain (JH1), the pseudokinase domain (JH2), the FERM (Four-point-one protein, Ezrin, Radixin, Moesin) domain, and the Src homology (SH2) domain; In the quiescent state, TYK2 kinase domain (JH1) activity is autoinhibited by its pseudokinase JH2 domain through intermolecular regulation. Upon cytokine (e.g., IL12, IL23, IFNα) stimulation, TYK2/STAT pathway is initiated by extracellular binding of cytokines with their cognate receptors, which results in apposition of receptor-associated TYK2 and triggering their enzymatic activity. Activated TYK2 further mediates phosphorylation of STATs on their tyrosine residues, leading to the cascade of down-stream cytokine signaling, including STAT dimerization, nuclear translocation, DNA binding and target gene induction. Small molecule TYK2 JH2 ligands block TYK2-associated cytokine signaling by binding to the TYK2 pseudokinase domain and allosterically inhibiting its kinase activities. This prevents TYK2 from phosphorylating STATs and other substrates, so that cytokine-induced intracellular signals cannot be transduced; (B) Model of the TYK2 pseudokinase domain bound to TYK2 ligands, SHR2396, 2178 and 9332. TYK2 pseudokinase domain residues corresponding to those of protein kinases normally involved in catalytic machinery are shown in stick; (C) Chemical structure of small molecule TYK2 JH2 ligands and JAK kinase inhibitors described in this study [SHR2178, 2396, 8751, 9332, 2915, 3290, 3110, 3591, Deucravacitinib (BMS986165, TYK2 pseudokinase ligand), Ropsacitinib (PF647, TYK2-IN-8, TYK2 kinase inhibitor), Brepocitinib (PF841, TYK2/JAK1 inhibitor) and Tofacitinib (JAK1/JAK3 inhibitor)].

Figure 2

Small molecule TYK2 pseudokinase ligands showed high affinity-binding for TYK2 JH2 pseudokinase domain and low affinity for isolated TYK2 JH1 kinase domain. (A–F) Binding affinity of TYK2 and JAK compounds to TYK2 JH2 pseudokinase domain in KdELECT Competition Binding Assay (DiscoverX). Representative dose-response curves of: (A) SHR2178; (B) SHR8751; (C) SHR9332; (D) SHR2396; (E) SHR0936; (F) Deucravacitinib; (G) Ropsacitinib; (H) Brepocitinib. Dissociation constants (Kds) for test compound-TYK2 JH2 interactions are calculated by measuring the amount of isolated TYK2 JH2 pseudokinase captured on the solid support as function of the test compound concentrations; (I-P) Binding affinity of TYK2 and JAK compounds to TYK2 JH1 catalytic domain in DiscoverX KdELECT Competition Binding Assay. Representative dose-response curves of: (I) SHR2178; (J) SHR8751; (K) SHR9332; (L) SHR2396; (M) SHR0936; (N) Deucravacitinib; (O) Ropsacitinib; (P) Brepocitinib. Dissociation constants (Kds) for test compound-TYK2 JH1 interactions are calculated by measuring the amount of isolated TYK2 JH1 kinase captured on the solid support as function of the test compound concentrations. Confirmatory studies were performed for the compounds. The figures are representative of three independent experiments with duplicates performed for each experiment (A–F, I–N) unless stated otherwise. The average Kd values and standard deviations of each compound from three independent experiments are calculated and summarized in the table.

Figure 3

Small molecule TYK2 pseudokinase ligands block IFNα- and IL23- mediated TYK2/STAT phosphorylation in human T cells. (A, B) Representative dose-response curves of TYK2 JH2 ligands, SHR8751, 2396 and 9332 on IFNα-induced phosphorylation (Tyr701) of STAT1 (A) and IL23-induced phosphorylation (Tyr705) of STAT3 (B) in Kit225 cells. Cells were seeded in 384-well plate at 1×10⁵ cells/well and treated with serially diluted compounds (30 μM, 1:3) for 1 hour and stimulated with human recombinant IFNα or IL23 for 20 minutes. The treated cells were then lysed and cellular phosphorylated-STAT1 (Tyr701) or STAT3 (Tyr705) levels were measured by AlphaLISA; Inhibition data were calculated by comparison to vehicle control wells for 0% inhibition and non-stimulated control wells for 100% inhibition. Dose response curves were generated to determine the concentration required to suppress 50% of cellular response (IC₅₀) as derived by non-linear regression analysis using GraphPad Prism. Confirmatory studies were performed for the compounds. The figures are representative of three independent experiments with duplicates performed for each experiment unless stated otherwise. The average IC₅₀ values and standard deviations of each compound from three independent experiments are calculated and summarized in the table. (C, D) Suppression of STAT5 phosphorylation in human CD3+ T cells by a TYK2 JH2 ligand, SHR1759. After 1hour incubation with serially diluted compounds, human whole blood samples were stimulated with recombinant human IFNα for 15 minutes and then lysed for removal of red blood cells, fixed, and permeabilized. White blood cells were stained with anti-CD3 FITC antibody, AF647 anti-Stat5 (pY694) antibody and later analyzed by flow cytometry. (C) Representative histogram of AF647 showed that IFNα-induced phosphorylation of STAT5 (Tyr694) in CD3+ T cells of human whole blood was suppressed by SHR1759 in a dose-dependent manner; (D) After gating on the CD3-positive population, the phosphorylation of cellular STAT5 in each treatment condition was quantitated by median fluorescence intensity (MFI). Inhibition data were calculated by comparison to vehicle control wells for 0% inhibition and non-stimulated control wells for 100% inhibition. Each bar represents the % inhibition at indicated compound concentration. Confirmatory studies were performed for the compounds. The figures are representative of two independent experiments with duplicates performed for each experiment.

Figure 4

TYK2 pseudokinase ligands demonstrate significant functional activity in suppressing IL12- or IFNα- mediated TYK2/STAT pathways in human PBMCs and whole blood. (A) Dose dependent response of TYK2 JH2 ligands, SHR2915, 0936, 1039 and 1213 on IL12-induced IFNγ production in human PBMCs. 1.0 × 10⁶/ml PBMCs were treated by serially diluted (30 μM, 1:3) compounds and incubated with hIL12 for overnight in complete RPMI. The cultured supernatant was collected for detection of IFNγ secretion using AlphaLISA method. (B-D) Dose dependent effect of TYK2 JH2 ligands, SHR8751 and 9332 (B), SHR2915 and Ropsacitinib (C) and SHR0936 (D) on IFNα-stimulated CXCL10 production in human whole blood. Blood samples were incubated with serially diluted compounds (30 μM, 1:3) and stimulated with recombinant human IFNα overnight. The plasma was collected for detection of CXCL10 production by AlphaLISA method. Inhibition data were calculated by comparison to vehicle control wells for 0% inhibition and non-stimulated control wells for 100% inhibition. IC₅₀ values were derived by non-linear regression analysis using GraphPad Prism. The figures are representative of the experiments on specimens from three different donors with duplicates performed for each donor. The average IC₅₀ values and standard deviations of each compound from experiments on three donors are calculated and summarized in the table.

Figure 5

Small molecule TYK2 JH2 pseudokinase ligands demonstrate differentiated binding pattern against JAK family members compared to JH1 kinase inhibitors. (A–C) Representative dose-response curves for binding affinity of TYK2 JH2 ligand, SHR9332 to JH1 kinase domain of JAK1 (A), JAK2 (B) and JAK3 (C) assessed by KdELECT Competition Binding Assay (DiscoverX); (D–F) Representative dose-response curves for binding affinity of a TYK2 kinase inhibitor, Ropsacitinib to JH1 kinase domain of JAK1 (D), JAK2 (E), and JAK3 (F) assessed by KdELECT Competition Binding Assay (DiscoverX); (G–H) Representative dose-response curves for binding affinity of SHR9332 to JH2 pseudokinase domain of JAK1 (G), and JAK2 (H) assessed by KdELECT Competition Binding Assay (DiscoverX); (I–J) Representative dose-response curves for binding affinity of Ropsacitinib to JH2 pseudokinase domain of JAK1 (I) and JAK2 (J) assessed by KdELECT Competition Binding Assay (DiscoverX). Dissociation constants (Kds) for test compound-JH1 or -JH2 domain interactions are calculated by measuring the amount of isolated JH1 or JH2 protein segments captured on the solid support as function of the test compound concentrations. Confirmatory studies were performed for the compounds. The figures are representative of three independent experiments with duplicates performed for each experiment unless stated otherwise. The average Kd values and standard deviations of each compound from three independent experiments are calculated and summarized in the table.

Figure 6

TYK2 pseudokinase small molecule ligand is highly selective across a multi-kinaseScan panel. (A) Chemical structure of TYK2 JH2 ligand, SHR1274 and representative dose-response curves of SHR1274 binding to TYK2 JH2 pseudokinase and JH1 kinase domains assessed by DiscoveX KdELECT Competition Binding Assay. The figures are representative of the technical duplicates in the experiment. Dissociation constants (Kds) for test compound-JH1 or -JH2 domain interactions are calculated by measuring the amount of isolated JH1 or JH2 protein segments captured on the solid support as function of the test compound concentrations. (B) Kinase selectivity profile of SHR1274 at 1 µM across 98 kinases in the kinome screening panel, with kinase binding inhibited by > 99% (large red circles) or 65-90% (small red circles). Screening “hits” are identified by quantifying the amount of kinase captured in test compounds versus control by using quantitative PCR that detects the associated DNA label; (C) Selectivity scores at S (1): > 99% competition, S (10): 90-99% competition or S (35): 65-90% competition.

Figure 7

TYK2 pseudokinase ligands showed limited potency in affecting JAK1-3 functional activities. (A–C) Effect of TYK2 JH2 ligands SHR8751, 9332, 2396, Deucravacitinib and JAK inhibitors Tofacitinib and Ruxolitinib in suppressing JAK1 (A), JAK2 (B) and JAK3 (C) kinase activities assessed by Z’LYTE kinase bio-chemical assay. Serially diluted test compounds were added into a kinase reaction buffer that contains JAK kinases (JAK1, JAK2 or JAK3, respectively) and Z9-LYTE™ Tyr6 Peptide for measurement of JAK kinase-induced peptide phosphorylation. The curve fit was generated, and statistical analysis was performed using GraphPad Prism. IC₅₀ values were derived by non-linear regression analysis and defined as the compound concentration at which the response level was reduced to half of its maximum relative to a DMSO control; (D, G) Effect of TYK2 JH2 ligands and TYK2/JAK kinase inhibitors SHR8751, 9332, 2396, Deucravacitinib and Ruxolitinib (D) and SHR2915, PF841 and Tofacitinib (G) on JAK1-dependent IL6-induced phosphorylation of STAT3 (Tyr705) in TF1 cells. TF1 cells were treated by serially diluted compounds (30 μM, 1:3) for 1 hour and stimulated by recombinant human IL6 for 20 mins. Phosphorylation of STAT3 (Tyr705) was measured in cell lysates using AlphaLISA assay; (E, H) Effect of TYK2 JH2 ligands and TYK2/JAK kinase inhibitors SHR8751, 9332, 2396, Deucravacitinib and Ruxolitinib (E) and SHR2915, PF841 and Tofacitinib (H) on JAK2-dependent EPO-induced phosphorylation of STAT5 (Tyr694/699) in TF1 cells. TF1 cells were treated by serially diluted compounds (30 μM, 1:3) for 1 hour and stimulated by recombinant human EPO for 20 mins. Phosphorylation of STAT5 (Tyr694/699) was measured in cell lysates using AlphaLISA assay; (F, I) Effect of TYK2 JH2 ligands and TYK2/JAK kinase inhibitors SHR8751, 9332, 2396, Deucravacitinib and Ruxolitinib (F) and SHR2915, PF841 and Tofacitinib (I), on JAK1/3-dependent IL2-induced phosphorylation of STAT5 (Tyr694/699) in Kit225 cells. Kit225 cells were treated by serially diluted compounds (30 μM, 1:3) for 1 hour and stimulated by recombinant human IL2 for 20 mins. Phosphorylation of STAT5 (Tyr694/699) was measured in cell lysates using AlphaLISA assay. For each assay, inhibition data were calculated by comparison to vehicle control wells for 0% inhibition and non-stimulated control wells for 100% inhibition. Dose response curves were generated to determine the IC₅₀ values as derived by non-linear regression analysis using GraphPad Prism. Confirmatory studies were performed for the compounds. The figures are representative of three independent experiments with duplicates performed for each experiment unless stated otherwise. The average IC₅₀ values and standard deviations of each compound from three independent experiments are calculated and summarized in the table.