Different Effects of Selective Janus Kinase (JAK) Inhibitors in Murine Bone Marrow Failure

Introduction: Immune aplastic anemia due to T cell mediated destruction of bone marrow (BM) hematopoietic stem cells (HSCs) manifests as severe pancytopenia. Patients are treated with immunosuppressive therapy (IST) or hematopoietic cell transplant and require experienced tertiary care. Options are limited for relapsed/refractory bone marrow failure (BMF) syndromes and most currently used therapies require hospitalization. Previously, we found that ruxolitinib (RUX), a broad JAK inhibitor (JAKi), demonstrated efficacy and extended overall survival (OS) in our well-established murine BMF model (Groarke EM, Blood 2023). We hypothesized that more selective JAKi may show similar effects and provide alternatives in the event of treatment-limiting hematologic toxicities with RUX. We used selective JAKi to block single and combination pathways to study effects and toxicities. Systematic blockade of JAK/STAT pathways should also facilitate a systematic understanding of their role in murine immune BMF.

Methods: BMF was induced in CByB6F1 mice by a combination of 5 Gy total body irradiation plus injecting 5 × 10^6/mouse lymph node cells from C57BL/6 (B6) donors mismatched at major histocompatibility complex. BMF mice were either untreated or treated with 1 of 4 investigational JAKi by oral gavage for comparative analysis: baricitinib (BAR, selective JAK1/2 and TYK2 inhibitor) at 25 mg/kg once daily; tofacitinib (TOF, strong JAK3 and weak JAK1/2 inhibitor) at 15 mg/kg twice daily (BID); fedratinib (FED, selective JAK2 inhibitor) at 80 mg/kg BID; and solcitinib (SOL, selective JAK1 inhibitor) at 60 mg/kg BID. Animals were either bled and euthanized at day 14 to measure cellular composition in blood and BM, or monitored 4-10 weeks for survival. For toxicity analyses, each JAKi was administered to normal CByB6F1 mice, with initial bleeding 1 week prior to drug administration, followed by a 3-week treatment with JAKi, interim bleeding at week 2 (W2) and final bleeding and animal euthanasia at week 5 (W5).

Results: In BMF mice, BAR increased all peripheral cell counts (p < 0.001) and residual BM cells (p < 0.05), and reduced BM CD4+ (p < 0.05) and cytotoxic CD8+ cells (p < 0.001). TOF treatment increased RBC and platelet counts (p < 0.001) without effects on T-cells. FED-treated mice only showed improved RBC counts (p < 0.001), while SOL treatment increased CD4+ cells (p < 0.005) but had no effect on blood counts. The best survival data was observed in the BAR treated mice (p < 0.0001), followed by TOF (p < 0.0001). However, survival in the TOF group rapidly declined after day 30, with no live animals by day 36. Approximately 65% of the BAR-treated mice were alive by day 60, as part of the long-term survival study. No improvement in survival was noted in the FED and SOL groups with the majority of animals dead by day 20. Single-cell RNA sequencing of whole BM cells 10 days post-initiation of BMF showed all JAKi treated groups had increased proportions of HSCs, progenitor cells and myeloid cells. Reduced proportions of T cells were noted across all JAKi, with BAR showing the most prominent effects, consistent with flow cytometry results. For BAR, in comparison to control, there was no significant up or downregulation of assessed pathways, but the more selective JAKi showed upregulated cell cycle and inflammatory gene expression.

No JAKi resulted in toxicity, including BM cell numbers (total BM, myeloid progenitor, Lin-Sca-1+c-Kit+, and their subset cell populations) or HSPC function (by the colony-forming unit assay) in normal CByB6F1 mice at study termination (W5). Peripheral blood lymphopenia was noted: BAR, SOL, and TOF caused declines persisting from W2 to W5. Treatment with FED, SOL, and TOF reduced blood CD45R B cell percentage, most noticeable at W5. JAKi treatment also caused declines in blood CD11b myeloid cells, and CD11b+Ly6G+ G-MDSCs, which resolved on drug cessation by W5; effector memory (CD62L-CD44+) CD4 and CD8 T cells were similarly affected. Total BM CD4 and CD8 T cells, as well as BM G-MDSCs, were reduced in all JAKi-treated mice.

Conclusion: Not all JAKi were equally effective in treating murine BMF. Our results support the hypothesis that combined JAK1 and JAK2 blockade is essential to ameliorate the immune environment in BMF, whereas strong JAK3 combined with weak JAK1/2 inhibition (TOF), and very selective JAK1 (SOL) or JAK2 (FED) blockade, were inferior. None of the tested JAKi had apparent hematological toxicity.