Ruxolitinib displays a narrow therapeutic window when treating established HLH and inhibition of JAK2, but not JAK1, is potentially toxic during HLH. (A) Survival and global clinical score of prf^−/− mice treated after onset of HLH, on day 11, with continuously administered ruxolitinib in the drinking water at various doses. (B) Survival and global clinical score of prf^−/− mice with established HLH treated with ruxolitinib, anti-IFN-γ, or combinations thereof. Anti-IFN-γ antibody was administered intraperitoneally (XMG1.2, 40 mg/kg, every 3-4 days) and ruxolitinib was given via oral gavage twice daily at the indicated doses. (C) Impact of ruxolitinib on marrow and peripheral blood cellularity. Bone marrow and peripheral blood neutrophil counts were assessed 16 days after infection, with the indicated treatments starting on day 6 (ruxolitinib was administered at 90 mg/kg per dose orally twice daily). (D) Survival and global clinical score of prf^−/− mice after LCMV infection treated with the JAK1-selective inhibitor AZD4205, with or without anti-IFN-γ antibody. AZD4205 was administered as 25 mg/kg twice daily via oral gavage. (E) Assessment of JAK2 blockade in vivo by ruxolitinib and AZD4205. Mice were treated with ruxolitinib (90 mg/kg, IP) or AZD4205 (25 mg/kg, IP) and injected with granulocyte macrophage colony-stimulating factor (200 ng, IP) 1 hour later. Twenty minutes later, animals were euthanized and peritoneal macrophages were harvested via peritoneal lavage with 1% paraformaldehyde. After methanol permeabilization, phospho-STAT5 was measured in F4/80⁺ cells by flow cytometry. N = 8-37 mice/group (29 mice total treated with high- or low-dose ruxolitinib in panel A), combined from 2 to 12 independent experiments. *P < .05, **P < .01, ***P < .001.