Differential effects of itacitinib, fedratinib, and ruxolitinib in mouse models of hemophagocytic lymphohistiocytosis

Abstract

Hemophagocytic lymphohistiocytosis (HLH) comprises a severe hyperinflammatory phenotype driven by the overproduction of cytokines, many of which signal via the JAK/STAT pathway. Indeed, the JAK1/2 inhibitor ruxolitinib has demonstrated efficacy in preclinical studies and early-phase clinical trials in HLH. Nevertheless, concerns remain for ruxolitinib-induced cytopenias, which are postulated to result from the blockade of JAK2-dependent hematopoietic growth factors. To explore the therapeutic effects of selective JAK inhibition in mouse models of HLH, we carried out studies incorporating the JAK1 inhibitor itacitinib, JAK2 inhibitor fedratinib, and JAK1/2 inhibitor ruxolitinib. All 3 drugs were well-tolerated and at the doses tested, they suppressed interferon-gamma (IFN-γ)-induced STAT1 phosphorylation in vitro and in vivo. Itacitinib, but not fedratinib, significantly improved survival and clinical scores in CpG-induced secondary HLH. Conversely, in primary HLH, in which perforin-deficient (Prf1-/-) mice are infected with lymphocytic choriomeningitis virus (LCMV), itacitinib, and fedratinib performed suboptimally. Ruxolitinib demonstrated excellent clinical efficacy in both HLH models. RNA-sequencing of splenocytes from LCMV-infected Prf1-/- mice revealed that itacitinib targeted inflammatory and metabolic pathway genes in CD8 T cells, whereas fedratinib targeted genes regulating cell proliferation and metabolism. In monocytes, neither drug conferred major transcriptional impacts. Consistent with its superior clinical effects, ruxolitinib exerted the greatest transcriptional changes in CD8 T cells and monocytes, targeting more genes across several biologic pathways, most notably JAK-dependent proinflammatory signaling. We conclude that JAK1 inhibition is sufficient to curtail CpG-induced disease, but combined inhibition of JAK1 and JAK2 is needed to best control LCMV-induced immunopathology.

Graphical abstract

Figure 1.

Pharmacokinetic and pharmacodynamic studies of JAK1, JAK2, and JAK1/2 inhibition. (A) Plasma drug levels (μg/L) in naïve C57BL/6J mice after single oral dose of each inhibitor (120 mg/kg itacitinib, 120 mg/kg fedratinib, 90 mg/kg ruxolitinib). Mean values ± standard deviation (SD) are shown. (B) Intracellular STAT1 phosphorylation measured in bone marrow–derived macrophages (BMDM) cultured with vehicle (methyl cellulose) or itacitinib (2500, 500, or 50 nM), (C) vehicle (methyl cellulose) or fedratinib (1000, 300, or 30 nM), or (D) vehicle (captisol in citrate buffer) or ruxolitinib (1000, 50, or 1 nM) for 1 hour and then stimulated with IFN-γ (30 ng/mL) for 20 minutes. Representative overlayed histogram plots are shown below each graph. Experiments were performed using duplicates. Data are representative of 2 independent experiments. (E) Intracellular STAT1 phosphorylation measured in thioglycolate–elicited peritoneal macrophages after a single oral dose of JAK inhibitor and intraperitoneal (500 ng) injection of IFN-γ, shown as percentage of pSTAT1⁺ F4/80⁺ macrophages. F4/80^lo macrophages were excluded from the gating. (F) Representative histograms (left) and fold-change mean fluorescence intensity (MFI) (right) compared to isotype control. Each data point in panels E-F represents 1 mouse. Data are representative of 2 independent experiments. ∗P < .05, ∗∗∗P < .001, or ∗∗∗∗P < .0001 by pairwise comparison.

Figure 2.

JAK1 inhibition improves survival and clinical scores in CpG/αIL-10R-induced HLH. (A) Experimental schema for modeling secondary HLH, wherein WT C57BL/6J mice receive injections of CpG and αIL-10R antibody on days 0, 2, 4, and 7, and are then analyzed on day 9 or 10 after the first injection. (B) Probability of survival (B) and clinical scores (C) from naïve, or CpG/αIL-10R-treated mice who received vehicle, itacitinib (120 mg/kg twice daily [bid]), fedratinib (60 mg/kg bid), or ruxolitinib (90 mg/kg bid) from days 4 to 9 after first CpG/αIL-10R injection. Analysis was performed on day 10. Data are combined from 2 independent experiments (9-10 mice per group). (D) Spleen weights shown as percentage of final body weight in mice treated with vehicle, itacitinib (120 mg/kg bid), fedratinib (60 mg/kg bid), or ruxolitinib (90 mg/kg bid) on days 4 to 8 after the first CpG/αIL-10R injection. Analysis was performed on day 9. Peripheral blood hemoglobin (E), hematocrit (F), and platelet counts (G). Data points in panels D-G represent single mice from 3 pooled experiments. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, or ∗∗∗∗P < .0001 by pairwise comparison. For survival, log-rank test was performed to determine significance between treated groups and vehicle control. For clinical score, statistical significance was determined using two-way ANOVA.

Figure 3.

JAK1 and JAK2 inhibition differentially ameliorate CpG/αIL-10R-induced activation and accumulation of splenic monocytes and neutrophils, as well as serum cytokine levels. (A) Proportions and numbers of total Ly6C⁺CD11b⁺ splenic monocytes and (B) CD80+ splenic monocytes in naïve mice or CpG+aIL10R treated mice treated with vehicle, itacitinib (120 mg/kg bid), fedratinib (60 mg/kg bid), or ruxolitinib (90 mg/kg bid) from days 4 to 8 after the first CpG+aIL10R injection. Analysis was performed on day 9. Proportions and absolute numbers of total Ly6G⁺CD11b⁺ splenic neutrophils (C) and CD80⁺ splenic neutrophils (D). (E-L) Serum cytokine levels. Data points represent single mice from 3 pooled experiments. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, or ∗∗∗∗P < .0001 by pairwise comparison.

Figure 4.

JAK1 inhibition partially improves survival and disease features in LCMV-induced HLH. (A) Experimental schema for modeling primary HLH, wherein Prf1^−/− mice are infected with LCMV and analyzed day 9 or 30 PI. Probability of survival (B) and clinical scores (C) of naïve, or LCMV- infected Prf1^−/− mice treated with vehicle, itacitinib (120 mg/kg bid), fedratinib (60 mg/kg bid), or ruxolitinib (90 mg/kg bid) from days 4 to 29 PI (5 mice per group). Analysis was performed on day 30 PI. Peripheral blood hemoglobin (D), platelet counts (E), spleen weights (F) shown as percentage of final body weight, and (G-L) serum cytokine levels of mice analyzed on day 9 PI. Data points represent single mice from 5 pooled experiments. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, or ∗∗∗∗P < .0001 by pairwise comparison. For survival, the log-rank test was performed to determine statistical significance between treated groups and vehicle control. For clinical score, statistical significance was determined using two-way ANOVA.

Figure 5.

JAK1 inhibition is not sufficient to reduce CD8 T-cell number or capacity to produce IFN-γ. Proportions and numbers of total splenic CD8 T cells (A), CD44⁺ effector CD8 T cells (B), and gp33-specific CD8 T cells (C) in naïve mice or LCMV-infected Prf1^−/− mice treated with vehicle, itacitinib (120 mg/kg), fedratinib (60 mg/kg), or ruxolitinib (90 mg/kg) from days 4 to 8 PI. Analysis was performed on day 9 PI. (D-E) Representative contour plots and summary data of cytokine producing CD44⁺ CD8 T cells (D-E) and gp33⁺ CD8 T cells (F-G) after in vitro restimulation with gp33 peptide. Data points represent single mice from 5 pooled experiments. ∗P < .05, ∗∗P < .01, or ∗∗∗∗P < .0001 by pairwise comparison.

Figure 6.

Changes in transcriptional profiles of splenic CD8 T cells from Prf1^−/− mice infected with LCMV and treated with a JAK1, JAK2, or combined JAK1/2 inhibitor. (A) Unsupervised clustering using multidimensional scaling of transcriptional profiles of splenic CD8 T cells from mice infected with LCMV and treated with vehicle, itacitinib (120 mg/kg bid), fedratinib (60 mg/kg bid), or ruxolitinib (90 mg/kg bid) with 3 mice per group. (B) Venn diagram depicting the numbers of DEGs in itacitinib, fedratinib, or ruxolitinib vs vehicle-treated groups (adjusted P value <.05, |Log₂FC| >1). (C) Heat map across the significant DEGs in vehicle, itacitinib, fedratinib, or ruxolitinib treatment groups with numbered clusters labeled on the y-axis. To the right of each cluster the enriched hallmark pathways are shown. (D) Dot plot of significantly up- (red) and downregulated (blue) hallmark gene sets (MSigDB v7.5.1) in LCMV-infected mice treated with itacitinib, fedratinib, or ruxolitinib compared with vehicle control (adjusted P value <.05 and normalized enriched score [NES] >1). Gene sets in proliferation (purple), developmental (cyan), metabolic (pink), housekeeping (black), and immune (brown) pathways are colored respectively. (E-G) Gene set enrichment analysis (GSEA) enrichment plots of the top 2 pathways in LCMV-infected mice for itacitinib (E), fedratinib (F), or ruxolitinib (G) vs vehicle-treated controls. FC, fold change.

Figure 7.

Changes in transcriptional profiles of splenic monocytes from Prf1^−/− mice infected with LCMV and treated with JAK1, JAK2, or combined JAK1/2 inhibitor. (A) Unsupervised clustering using multidimensional scaling of transcriptional profiles of splenic CD11b⁺Ly6C⁺ monocytes from mice infected with LCMV and treated with vehicle, itacitinib (120 mg/kg bid), fedratinib (60 mg/kg bid), or ruxolitinib (90 mg/kg bid) with 3 mice per group. (B) Venn diagram depicting the numbers of DEGs in itacitinib, fedratinib, or ruxolitinib vs vehicle-treated groups (adjusted P value <.05, |Log₂FC| >1). (C) Heat map of the significant DEGs in vehicle, itacitinib, fedratinib, or ruxolitinib treatment groups with clusters labeled on the y-axis. To the right of each cluster the enriched hallmark pathways are shown. (D) Dot plot of significantly up- (red) and downregulated (blue) hallmark gene sets (MSigDB v7.5.1) in LCMV-infected mice treated with itacitinib, fedratinib, or ruxolitinib compared with vehicle control (adjusted P value <.05 and NES >1). Gene sets in proliferation (purple), metabolic (pink), housekeeping (black), DNA damage repair (cyan), and immune (brown) pathways are colored respectively. (E-G) GSEA enrichment plots of the top 2 pathways in CD11b⁺Ly6C⁺ monocytes for itacitinib (E), fedratinib (F), or ruxolitinib (G) vs vehicle-treated controls. FC, fold change.