Simultaneous inhibition of JAK and SYK kinases ameliorates chronic and destructive arthritis in mice

Abstract

Introduction: Despite the broad spectrum of antirheumatic drugs, RA is still not well controlled in up to 30-50 % of patients. Inhibition of JAK kinases by means of the pan-JAK inhibitor tofacitinib has demonstrated to be effective even in difficult-to-treat patients. Here, we discuss whether the efficacy of JAK inhibition can be improved by simultaneously inhibiting SYK kinase, since both kinases mediate complementary and non-redundant pathways in RA.

Methods: Efficacy of dual JAK + SYK inhibition with selective small molecule inhibitors was evaluated in chronic G6PI-induced arthritis, a non-self-remitting and destructive arthritis model in mice. Clinical and histopathological scores, as well as cytokine and anti-G6PI antibody production were assessed in both preventive and curative protocols. Potential immunotoxicity was also evaluated in G6PI-induced arthritis and in a 28-day TDAR model, by analysing the effects of JAK + SYK inhibition on hematological parameters, lymphoid organs, leukocyte subsets and cell function.

Results: Simultaneous JAK + SYK inhibition completely prevented mice from developing arthritis. This therapeutic strategy was also very effective in ameliorating already established arthritis. Dual kinase inhibition immediately resulted in greatly decreased clinical and histopathological scores and led to disease remission in over 70 % of the animals. In contrast, single JAK inhibition and anti-TNF therapy (etanercept) were able to stop disease progression but not to revert it. Dual kinase inhibition decreased Treg and NK cell counts to the same extent as single JAK inhibition but overall cytotoxicity remained intact. Interestingly, treatment discontinuation rapidly reversed such immune cell reduction without compromising clinical efficacy, suggesting long-lasting curative effects. Dual kinase inhibition reduced the Th1/Th17 cytokine cascade and the differentiation and function of joint cells, in particular osteoclasts and fibroblast-like synoviocytes.

Conclusions: Concurrent JAK + SYK inhibition resulted in higher efficacy than single kinase inhibition and TNF blockade in a chronic and severe arthritis model. Thus, blockade of multiple immune signals with dual JAK + SYK inhibition represents a reasonable therapeutic strategy for RA, in particular in patients with inadequate responses to current treatments. Our data supports the multiplicity of events underlying this heterogeneous and complex disease.

Fig. 1

Efficacy of Janus kinase (JAK) + spleen tyrosine kinase (SYK) inhibition as a preventive treatment in glucose-6-phosphate isomerase (G6PI)-induced arthritis. DBA/1 mice were immunized with G6PI + complete Freund’s adjuvant on day 0, after T regulatory cell depletion with anti-CD25 Ab (days −8 and −11). Treatments (20 mg/kg of tofacitinib - JAK inhibitor (JAKi) - and/or 30 mg/kg of PRT062607 – SYK inhibitor (SYKi)) were orally administered daily from day 0. Samples were obtained on day 31 unless otherwise specified. a Time course for arthritis scores and resulting areas under these score curves. b Mean weight of both inguinal lymph nodes at day 6 and day 31 post-immunization. c Numbers of CD11c^hiMHCII⁺CD86⁺ activated dendritic cells (DCs) in the spleen. d Cytokine secretion by splenocytes stimulated in vitro with 20 μg/ml recomninant human G6PI for 72 h. e Numbers of CD19⁺MHCII⁺CD86⁺ activated B cells in the spleen. f Levels of G6PI-specific IgGs in plasma. g Histological features of G6PI-induced arthritis, including severe inflammation and bone/cartilage destruction of tarsal and phalangeal joints from hind paws (H&E staining). Arrows cell aggregates (left, bar 500 μM), infiltration of lymphocytes, macrophages and neutrophils in the synovium (middle, bar 200 μM) and osteoclasts (right, bar 50 μM). h and i Data and representative histological sections of tarsal joints (bar 500 μM). Graphs show mean (+ standard error of the mean) for 3–5 mice/group, *p <0.05 versus vehicle control, ^# p <0.05 versus non-immunized and untreated group (naive). N naive, V/Veh vehicle, IFN interferon

Fig. 2

Efficacy of Janus kinase (JAK) + spleen tyrosine kinase (SYK) inhibition as a curative treatment in glucose-6-phosphate isomerase (G6PI)-induced arthritis. DBA/1 mice were immunized with G6PI + complete Freund’s adjuvant on day 0, after T regulatory cell depletion with anti-CD25 Ab (days −8 and −11). Treatments were orally administered daily from day 12 (red dotted line). Samples were obtained on day 31. a Time course for arthritis scores and resulting areas under these score curves (AUC). b Detailed clinical score by paw segment at day 31. c Cytokine secretion by splenocytes stimulated in vitro with 20 μg/ml recombinant human G6PI for 72 h. d Cytokine expression in splenocytes. e Numbers of CD19⁺MHCII⁺CD86⁺ activated B cells in spleens. f Levels of G6PI-specific IgGs in plasma. g and h Data and representative histological sections of tarsal joints (bar 200 μM). Graphs show mean (+ standard error of the mean) for 3–7 mice/group, *p <0.05 versus vehicle control. N naive, V/Veh vehicle, Pred/Predni prednisolone

Fig. 3

Therapeutic comparison of Janus kinase (JAK) + spleen tyrosine kinase (SYK) inhibition and anti-TNF therapy in glucose-6-phosphate isomerase (G6PI)-induced arthritis. DBA/1 mice were immunized with G6PI + complete Freund’s adjuvant on day 0, after T regulatory cell depletion with anti-CD25 Ab (days −8 and −11). Treatments were administered starting on day 12 (red dotted line). Samples were obtained on day 36. a Time course for arthritis scores. b Mean weight of both inguinal lymph nodes. c Mean ankle thickness of both hind paws measured with a manual caliper. d Histopathological scores for tarsus and pannus formation in both phalanges and tarsus. Representative histological sections of tarsal joints are also shown (bar 500 μM). e Time course for arthritis scores for mice with mild, moderate and severe arthritis at the time of treatment initiation. Graphs show mean (± standard error of the mean) for 2–10 mice/group, *p <0.05 versus vehicle control. N naive, Veh vehicle

Fig. 4

Evaluation of potential immunotoxicity in in vivo assays. a-c DBA/1 mice were immunized with glucose-6-phosphate isomerase (G6PI) + complete Freund’s adjuvant and exposed to treatment daily for 31 days. a Red blood cell counts and leucocyte counts in peripheral blood. b Splenocyte numbers. c Numbers of CD4⁺ T cells (CD3⁺CD4⁺CD8⁻) and B cells (CD3⁻CD19⁺MHCII⁺) in spleens. d-f CD-1 mice were treated from day −14 to 14 and immunized with keyhole limpet hemocyanin (KLH) on day 0 (T cell-dependent antibody response (TDAR) assay). d Red blood cell counts in peripheral blood. e Numbers of CD4⁺ T cells (CD3⁺CD4⁺CD8⁻) and B cells (CD3⁻CD19⁺MHCII⁺) in the spleen. f Levels of KLH-specific IgM and IgGs in plasma at day 7 and 14, respectively. Graphs show mean (+ standard error of the mean) for 3–6 mice/group, *p <0.05 versus vehicle control, ^# p <0.05 versus non-immunized and untreated group (naive). N naive, Veh vehicle, JAKi Janus kinase inhibitor, SYKi spleen tyrosine kinase inhibitor, Pred prednisolone

Fig. 5

Evaluation of potential immunotoxicity in in vitro/in vivo assays and effects of treatment withdrawal. a Numbers of T regulatory (Treg) cells (CD3⁺CD4⁺CD25^hiFoxp3⁺), CD8⁺ T cells (CD3⁺CD8⁺CD4⁻) and natural killer (NK) cells (CD3⁻CD49b^hi) in DBA/1 mice immunized with glucose-6-phosphate isomerase (G6PI) + complete Freund’s adjuvant and exposed to treatment daily for 31 days. b Cell-mediated cytotoxicity assay with splenocytes either treated in vivo (left) or in vitro (right) and YAC-1 as target cells. c Phagocytic activity of peripheral blood mononuclear cells (PBMC) treated in vitro. d Time course for arthritis scores and numbers of Treg cells and NK cells in the spleen of G6PI-immunized mice treated for 35 days and sacrificed on day 47. Graphs show mean (+ standard error of the mean) for 2–6 mice/group, *p <0.05 versus vehicle or media control, ^# p <0.05 versus non-immunized and untreated group (naive). N naive, Veh vehicle, Predni prednisolone, TC target cells, Med media, JAKi Janus kinase inhibitor, SYKi spleen tyrosine kinase inhibitor

Fig. 6

Effects of Janus kinase (JAK) + spleen tyrosine kinase (SYK) inhibition on fibroblast-like synoviocytes and osteoclasts. a Cytokine and matrix metalloproteinase (MMP) production by fibroblast-like synoviocytes (FLS) from arthritic mice stimulated with 10 ng/ml rmTNF and 50 ng/ml rmIL-17A and treated in vitro with kinase inhibitors (0.5 μM) for 24 h. b Representative images after crystal violet staining and data showing the invasive potential of FLS from arthritic mice treated with kinase inhibitors (0.5 μM) in transwell collagen-coated chambers for 48 h. c Numbers of CD11b^hiGr1⁻cells, which contain most osteoclast precursors, and d rank ligand (RANKL) expression in spleens of DBA/1 mice immunized with glucose-6-phosphate isomerase + complete Freund’s adjuvant and exposed to JAK inhibitor (JAKi) + SYK inhibitor (SYKi) daily for a month. e Representative images after tartrate-resistant alkaline phosphatase staining showing the impact of JAKi/SYKi on osteoclast differentiation from bone marrow cells after 5 days in a RANKL + M-CSF enriched culture. f Activity of mature osteoclasts on cortical bone slices treated with kinase inhibitors (0.5 μM) for 48 h. Graphs show mean (+ standard error of the mean) of a representative experiment of 2–4 with 3–7 samples/group, *p <0.05 versus dimethyl sulfoxide (DMSO) or naive control. Veh vehicle

Fig. 7

Effects of Janus kinase (JAK) + spleen tyrosine kinase (SYK) inhibition on effector and memory T and B cells. a Proliferation of purified T cells stimulated with 1 μg/ml anti-CD3 and 0.5 μg/ml anti-CD28 Abs in the presence of kinase inhibitors (0.5 μM) for 72 h. b Cytokine secretion by splenocytes from arthritic mice in vitro re-exposed to 20 μg/ml recombinant human glucose-6-phosphate isomerase (G6PI) and treated with JAK inhibitor (JAKi) + SYK inhibitor (SYKi) (0.5 μM) for 72 h. c-f BALB/c mice were immunized and challenged with ovalbumin (OVA) + alum; 2 weeks after challenge, mice were treated with kinase inhibitors for 11 days. c-e Numbers of memory CD4⁺ T cells (CD3⁺CD4⁺CD44^hi), antigen-specific (CD3⁺CD4⁺CD44^hiCD40L⁺) and cytokine-producing (CD40L⁺cytokine⁺) memory CD4⁺ T cells upon OVA re-exposure in vitro. f Numbers of OVA-specific memory plasma cells (CD138⁺κλ^hiOVA⁺ 5-Bromo-2′-deoxyuridine⁺). Graphs show mean (+ standard error of the mean) for 3–5 mice/group, *p <0.05 versus media or vehicle control. Med media, Veh vehicle