Introduction
Effective treatments of human autoimmune diseases, which are complex and heterogeneous by nature, require therapeutic perturbation or restoration of multiple redundant and distinct mechanisms, or a master regulator of such pathways. In the case of rheumatoid arthritis (RA) pathogenesis, the critical role of the adaptive immune response and proinflammatory cytokines has been unequivocally established by the efficacy of marketed biologics targeting tumor necrosis factor (TNF) alpha, interleukin (IL)-6, CD20 (B-cell depletion) and CD80/86 (modulation of T-cell co-stimulation). However, their efficacy are capped by limited efficacy, with 40% of patients never responding to treatments and only 20% of patients experiencing a major reduction in disease activity [
1]. There thus remains a tremendous unmet clinical need for more effective therapeutic strategies, with a goal of sustained remission for a greater number of patients with RA.
Current therapeutic strategies pursued by the biopharmaceutical industry include those that target the janus kinase (JAK)-mediated signaling pathway, lymphocyte migration using chemokine CCR1 antagonist, and B cells using either depleting antibodies that recognize common cell surface antigens, such as CD22 and CD19, or blocking antibodies to B-cell survival factor such as B-cell activating factor (BAFF) or a proliferation-inducing ligand (APRIL). Although the pan-JAK inhibitor tofacitinib was recently approved by the US Food and Drug Administration for the treatment of RA, it is still not superior to the biologics in terms of efficacy and safety. For other autoimmune diseases that are in dire need of safer and/or more effective therapies, the anti-BAFF antibody belimumab, despite showing marginal efficacy in clinical trials, was approved for treatment of systemic lupus erythematous (SLE). Disappointingly, another anti-BAFF antibody (tabalumab) also did not show adequate efficacy in a phase 3 RA trial (Elli Lilly news release, 13 December 2012). Whether an agent that neutralizes both BAFF and APRIL will produce better results remains to be seen.
Other emerging approaches target key enzymes involved in mediating multiple signal transduction pathways. One such enzyme is the spleen tyrosine kinase (SYK), which is a master regulator in coupling activated immunoreceptors to the mobilization of downstream signal transduction cascades that affect diverse biological functions. One of the best characterized modules in the transmission of B-cell receptor (BCR) activating signals within B cells is the SYK–Bruton’s tyrosine kinase (BTK) axis, where BTK acts as an essential downstream effector of SYK in regulating both the maturation and survival of the B-cell lineage. Given the central role of SYK in transmission of antigen receptor signals that are critical for autoantibody production and the various innate immune effector functions, pharmacological inhibition of the catalytic function of SYK is expected to have pleiotropic anti-inflammatory effects and to impact multiple steps in the pathogenesis of autoimmune disorders [
2]. This could result in greater and/or broader therapeutic efficacy as well as increased coverage of the patient population, and perhaps a decreased propensity to lose therapeutic efficacy over time.
Here, we describe the discovery and characterization of a novel ATP-competitive inhibitor of SYK: 6-[(1R,2S)-2-amino-cyclohexylamino]-4-(5,6-dimethyl-pyridin-2-ylamino)-pyridazine-3-carboxylic acid amide, designated RO9021. The inhibitor RO9021 has reasonable kinase selectivity profile, potency and oral bioavailability and is capable of suppressing various innate and adaptive immune responses in vitro, as well as disease progression in the mouse collagen-induced arthritis (mCIA) model. RO9021 could thus serve as a lead candidate for further development of selective SYK inhibitors for the potential treatment of immunological disorders.
Materials and methods
Experimental animals
C57BL/6 and DBA/1J adult mice were purchased from Charles River Laboratories (Wilmington, MA, USA). All animal procedures were approved by and conducted in accordance with the Institutional Animal Care and Use Committee guideline at Hoffmann-La Roche (Nutley, NJ, USA).
Chemical compounds and reagents
SYK inhibitor RO9021 was designed and synthesized at Hoffmann-La Roche. Tofacitinib citrate (CP-690,550) was acquired from Selleck Chemicals LLC (Houston, TX, USA). All chemical reagents were purchased from Sigma-Aldrich (St Louis, MO, USA), antibodies for flow cytometric analysis were acquired from BD Biosciences (San Jose, CA, USA), cytokines were acquired from R&D Systems (Minneapolis, MN, USA) and antibodies for western blots were acquired from Cell Signaling Technologies (Danvers, MA, USA), unless indicated otherwise.
SYK inhibitor, kinase selectivity and kinase activity assays
RO9021 was designed and synthesized at Hoffmann-La Roche, Inc. Specificity to SYK was assessed by an ATP competitive binding assay at 1 μM compound concentration at KINOMEscan Inc. (San Diego, CA, USA) [
3].
The inhibitory potency to SYK was determined in a radiometric assay using inactive SYK kinase. Briefly, SYK protein (Invitrogen, Carlsbad, CA, USA) was dephosphorylated by PTP1B phosphatase (Invitrogen) and then the reaction was initiated by the addition of substrate cocktail that contained 20 μM ATP, 0.025 μCi ATP-γ-
33P (Perkin Elmer, Waltham, MA, USA) and 10 μM biotinylated synthetic peptide (Biotin-EPEGDYEEVLE; Biomer Technology, Pleasanton, CA, USA) [
4‐
6]. The reaction was carried out for 30 minutes and resulting
33P incorporation was determined by top counter.
Co-crystallization of SYK and RO9021
SYK (356 to 635) containing a kinase domain was cloned, expressed, and purified, and co-crystallization of SYK and RO9021 was carried out following the protocol as reported previously by our group [
7,
8]. The structure has been deposited in the Protein Data Bank [PDB:4GFG].
Calcium influx fluorometric imaging plate reading assay
Human B-cell lymphoma cell line Ramos (CRL-1596; American Type Culture Collection, Manassas, VA, USA) or T-cell lymphoma cell line Jurkat (CRL-2063; American Type Culture Collection) were loaded with calcium dye (BD Biosciences) for the assay. Baseline fluorescence was recorded for about 20 seconds followed by stimulation with mouse anti-human IgM (10 μg/ml, clone M2E6; Antibody Solutions Inc., Mountain View, CA, USA) for Ramos cells or mouse anti-human CD3 (10 μg/ml; BD Biosciences) for Jurkat cells, and the maximal fluorescent counts in each well were recorded.
Detection of the phosphorylation of BTK, PLCγ2, ERK and AKT
Ramos cells were pretreated with RO9021 followed by stimulation with goat F(ab′)2 anti-human IgM (10 μg/ml; Southern Biotech, Birmingham, AL, USA). The protein phosphorylation was detected with rabbit antibodies of anti-phospho-BTK(Y223) (Epitomics Inc., Burlingame, CA, USA), anti-phospho-PLCγ2(Y1217), anti-phospho-ERK(T202/Y204) or anti-phospho-AKT(S473).
Flow cytometric analysis of CD69 upregulation in B cells
Heparinized blood was collected from healthy volunteers and pre-incubated with RO9021 followed by stimulation with goat F(ab′)2 anti-human IgM (50 μg/ml) overnight. The samples were stained with PE mouse anti-human CD20 and APC mouse anti-human CD69. The percentage of activated (CD69hi) B cells was determined using unstimulated (negative control) and stimulated (positive control) samples as references.
Peripheral blood mononuclear cells (PBMCs) were isolated by centrifugation from heparinized blood in a Vacutainer CPT Cell Preparation Tube (BD Biosciences). PBMCs were cultured for 1 to 2 hours to allow monocytes to adhere, and nonadherent cells were washed away. The monocytes were stimulated with human IgG-coated (Jackson Immunology, West Grove, PA, USA) copolymer microsphere beads (Thermo Scientific, Fremont, CA, USA) or lipopolysaccharide (1 ng/ml; Sigma-Aldrich) for 4 hours. TNFα levels in supernatants were determined by enzyme-linked immunosorbent assay kits (BD Biosciences).
IgE-induced histamine release in human mast cells
The method has been reported previously by our group [
9]. Briefly, human cord blood-derived CD34
+ hematopoietic stem cells (AllCells, Emeryville, CA, USA) were cultured in a serum-free StemPro-34 medium (Invitrogen) with stem cell factor (100 ng/ml) and IL-6 (50 ng/ml) for 8 weeks followed by 5-day stimulation with IL-4 (10 ng/ml). For measuring histamine release, cells were sensitized with 0.1 μg/ml anti-4-hydroxy-3-nitrophenylacetyl hapten IgE (Serotec, Raleigh, NC, USA) overnight, and then cross-linked with 1 μg/ml NP(30)-BSA (Biosearch Technologies, Novato, CA, USA) for 30 minutes. Supernatants were collected and assayed for histamine release using a histamine enzyme immunoassay (Oxford Biomedical Research, Rochester Hills, MI, USA). The percentage of histamine release was calculated by comparing various treatments with positive control.
Flow cytometric analysis of phosphorylated STAT1 and STAT5
Human PBMCs were pre-incubated with compound for 30 minutes followed by 20 minutes stimulation with IL-2 (100 ng/ml) for signal transducers and activators of transcription 5 (STAT5) phosphorylation or IFNγ (100 ng/ml) for STAT1 phosphorylation. For IL-2-induced STAT5 phosphorylation, cells were stained with FITC anti-human CD3 and Alexa Fluor 647 anti-STAT5(pY694), and quantitated pSTAT5 fluorescence intensity gated on the CD3+ T-cell population. For IFNγ-induced STAT1 phosphorylation, cells were stained with PE anti-human CD14 (Beckman Coulter, Indianapolis, IN, USA) and Alexa Fluor 647 anti-STAT1 (pY701), and quantitated pSTAT1 fluorescence intensity gated on CD14+ monocytes/macrophages.
Mouse bone marrow macrophage-derived osteoclastogenesis
Bone marrow cells were obtained from C57BL/6 mouse tibiae and suspended in culture medium supplemented with monocyte colony-stimulating factor (100 ng/ml) for 16 hours. Nonadherent cells were harvested and further cultured with monocyte colony-stimulating factor (M-CSF) (100 ng/ml) and receptor activator of nuclear factor kappa-B ligand (100 ng/ml) for 3 days to induce the formation of multinuclear osteoclasts [
10‐
13]. The cells were stained using a tartrate-resistant acid phosphate (TRAP) staining kit (Sigma, St Louis, MO, USA). TRAP
+ multinuclear cells were counted for each well under a microscope.
Human B cells were enriched using RosetteSep® human B-cell enrichment cocktail (StemCell Technologies, Vancouver, Canada), followed by stimulation with ODN2006 (1 μM; InvivoGen, San Diego, CA, USA) and IFNα (20 ng/ml) for 3 days. The IL-6 production in the supernatant was measured by AlphaLISA kit (Perkin Elmer). The live cells were quantitated by the CellTiter-Glo® luminescent Cell Viability Assay kit (Promega, Madison, WI, USA).
Human B cells were differentiated with ODN2006 (50 nM) and IL-2 (10 ng/ml) for 6 days. The differentiated cells were stained with V450 anti-CD38, FITC anti-CD20, PE anti-CD19 and APC intracellular IgM. The plasmablasts were identified as CD19+CD38+CD20–IgM+ cells. The production of IgG and IgM was quantitated by AlphaLISA.
Human plasmacytoid dendritic cells (pDCs) were isolated by negative selection from PBMCs with the human pDC Isolation Kit (Miltenyi Biotec, Auburn, CA, USA). The purity was confirmed with CD303 (BDCA-2; Miltenyi Biotec) staining and stimulated with ODN2216 (1 μM) for 2 days. The production of IFNα and TNFα was measured by AlphaLISA.
Murine collagen-induced arthritis model
The mCIA model has been reported previously [
9]. Briefly, DBA1/J male mice were injected intradermally with 0.1 ml bovine type II collagen (100 μg; Chondrex Inc., Redmond, WA, USA) and complete Freund’s adjuvant (200 μg; Difco, Detroit, MI, USA) followed by second immunization on day 21 with bovine type II collagen and incomplete Freund adjuvant. RO9021 was administered orally, randomized into different groups (14 mice/groups), every day for 14 days starting on the day after second immunization. Clinical arthritis scores (1 to 4) of individual paws were assessed and the arthritic index for each mouse was determined by adding the individual scores of all four paws. The level of cytokines in serum was determined by Luminex analysis (Millipore, Billerica, MA, USA).
Histopathological analysis
Hind paws from CIA mice were collected into 10% neutral buffered formalin. After decalcification in 10% formic acid, paws were embedded in paraffin, sectioned at 8 μm and stained with toluidine blue. Inflammation (infiltration of inflammatory cells), pannus, cartilage damage, and bone resorption were scored in a double-blinded fashion by a board-certified pathologist at Boulder BioPATH, Inc. (Boulder, CO, USA) using standard criteria, with 0 being normal and 5 being the most severe.
Half-maximal inhibitory concentration determination and statistical analysis
Half-maximal inhibitory concentration (IC50) values and dose/concentration response curves were determined by sigmoidal dose–response curve fitting using XLFit (IDBS, Alameda, CA, USA) or Prism (GraphPad Inc., La Jolla, CA, USA). In most studies, the IC50 values reported were the average from at least two studies conducted with samples in replicate. For in vivo studies, one-factor and two-factor comparisons were performed, respectively, using one-way or two-way analysis of variance plus Dunnett’s post test.
Discussion
Current biologic agents have not been able to break the ceiling in terms of delivering better and broader efficacy for treatment of autoimmune diseases. Given this persistent unmet need and the promise of multipathway inhibition to deliver breakthrough efficacy, pharmacological modulation of intracellular signaling components with small molecule agents offers an attractive alternative therapeutic strategy, provided the risk/benefit profile is acceptable. In this regard, the SYK–BTK axis is an attractive target because it is critical for antigen receptor signaling, abnormal regulation of which has been implicated in the pathogenesis of several autoimmune diseases, including RA and SLE [
28].
Among the reported agents targeting the SYK, the inhibitor fostamatinib (R788) has demonstrated reduced clinical efficacy compared with other therapeutic agents. In our hands, however, R788 is not a very selective kinase inhibitor, inhibiting one-half of the kinome in the KinaseScan assay, including JAK and vascular endothelial growth factor receptor (data not shown), which is consistent with previous reports [
15,
29], suggesting that the clinical activities of R788 are not solely attributed to SYK inhibition. Some of the off-target activities might also account for the observed adverse effects in clinical trials, including high blood pressure, which is due to vascular endothelial growth factor receptor inhibition. R788 is also a relatively weak SYK inhibitor in whole blood assays, which is potentially attributed to high plasma protein binding. A more selective, potent SYK inhibitor will thus be necessary to address the mechanism of action and evaluate the efficacy as well as any potential on-mechanism toxicity associated with SYK inhibition in clinical trials.
To this end, we have developed an alternative chemical scaffold of SYK inhibitor, designated RO9021. The protein kinase selectivity profile of RO9021 was assessed by the widely accepted KinomeScan method, which utilizes a proprietary active site-directed competition binding assay to quantitatively measure interactions between test compounds and more than 450 human kinases and disease-relevant mutant variants. As shown in Figure
1 and Additional file
1: Figure S1, beside SYK with 99% competition only six protein kinases, including JAK1 and JAK3, have more than 90% competition, indicating that RO9021 has superb selectivity. Since truncated forms of recombinant JAK1 and JAK3 were utilized in the KinomeScan assay, we examined the ability of RO9021 to inhibit JAK-mediated signaling in cell-based assays and found the compound had weak or no activity (Figure
3). In contrast, RO9021 inhibited phosphorylation of SYK downstream effectors, namely PLCγ2 and BTK, in response to BCR engagement (Figure
2A), consistent with the known biology of SYK in BCR signaling. Taken together, these data strongly indicate that the compound effect in cells is mediated by SYK inhibition. Furthermore, RO9021 has reasonable oral bioavailability profiles and thus can be used to interrogate the various reported biological roles of SYK in preclinical disease models. In addition to suppression of BCR signaling in human PBMCs and whole blood (Figure
2C,D), FcγR signaling in human monocytes (Figure
2E), and FcϵR signaling in human mast cells (Figure
2F), we showed that RO9021 also blocked osteoclastogenesis of mouse bone marrow macrophages
in vitro (Figure
4).
Certain autoimmune diseases, such RA and SLE, arise from an inappropriate immune response of the body against self-antigens [
1,
30,
31]. SLE, for instance, is characterized by the loss of tolerance to self-nuclear antigens, the deposition of immune complexes in tissues, and multiorgan involvement [
32]. Studies have shown that nuclear-acid sensing pathways implicated in the subversion of the innate immune response to discriminate between self-antigen and foreign antigens are those mediated by the TLRs in the context of SLE pathogenesis [
32‐
34]. BTK, which is a downstream kinase of SYK, has been implicated in TLR signaling recently [
24,
25], whereas the role of SYK in TLR signaling is not well appreciated. It has been reported that the TLR9 agonist CpG could induce TLR-9 independent SYK phosphorylation and activation through actin cytoskeleton reorganization, leading to activation of Src family kinases [
35]. Recruitment of SYK to TLR9 and phosphorylation of TLR9 are required for CpG-induced cytokine production. We therefore used RO9021 to study the role of SYK in TLR9 signaling. Interestingly, the kinase function of SYK is essential for TLR9-mediated responses in human B cells (Figure
5A,B,C). Inhibition of SYK kinase function resulted in a decreased level of plasmablasts, IgM, IgG and IL-6 upon B-cell differentiation in the presence of TLR9 ligand. In addition, RO9021 also potently inhibited IFNα production by human pDCs upon TLR9 activation (Figure
5D). Importantly, the effects on TLR9 responses are specific because RO9021 did not inhibit TLR4-dependent TNFα production by human monocytes (Figure
2E) or activation of the JAK–STAT pathway stimulated with either IL-2 or IFNγ (Figure
3). Our study showed for the first time that kinase activity of SYK is critical for TLR9 signaling pathway in B cells and pDCs.
The role of SYK in TLR signaling in B cells and pDCs could have significant implication for SYK inhibitors as therapeutic agents for SLE since the development and progression of the disease are believed to be driven by the inappropriate activation of TLR7, TLR8 and TLR9. Other studies have indicated that autoimmunity in RA and psoriasis also is mediated through one or more of these TLRs [
36‐
41]. In this regard, it is noteworthy that TLR antagonists such as IRS-954 (TLR7 and TLR9 antagonist) [
42,
43] and IMO-8400 (TLR7, TLR8 and TLR9 antagonist) are currently undergoing clinical trials in SLE.
Consistent with its inhibitory activities in various innate and adaptive immune responses, oral administration of RO9021 inhibited arthritis progression in the mCIA model (Figure
6). Importantly, there was correlation between pharmacokinetics analysis of compound exposure and pharmacodynamics analysis based on anti-IgD-induced CD69 expression on B cells (B220
+), indicating on-target mode of action. Furthermore, the pharmacodynamics analysis suggests that 5-hour compound coverage was sufficient to ameliorate the disease in this model.
Competing interests
All authors were employees of Hoffmann-La Roche during the preparation of this manuscript, and declare that they have no competing interests.
Authors’ contributions
CL designed and coordinated the study, carried out the experiments and interpretation of the results, and wrote the manuscript. JH, JZ, AP, JM, TW, NR, TT, MS, CL, YK, D-QH, and DX carried out or participated in the experiments. MZ worked on pharmacokinetics data. JH and ML designed the compound. SN and JD supervised the study. S-LT supervised the study and wrote the manuscript. All authors read and approved the final manuscript.