Selective Janus kinase inhibitors exhibit different effects in murine immune aplastic anemia

Aplastic anemia (AA), the prototype immune bone marrow failure (BMF) syndrome, is characterized by immune-mediated destruction of hematopoietic stem and progenitor cells (HSPCs) in the bone marrow (BM) by activated T lymphocytes, leading to pancytopenia and resultant clinical sequelae.1, 2 The immunopathogenesis of AA has been well established through both clinical studies and murine models.3-5 Activated CD4+ T helper cells initiate aberrant immune responses that promote expansion and activation of CD8+ cytotoxic T cells. These effector cells release proinflammatory cytokines such as interferon-gamma (IFN-γ) and TNF-alpha (TNF-α), amplifying the immune response and upregulating Fas/FasL interactions, ultimately triggering apoptosis of HSPCs.6-8 Acutely, regulatory T cells (Tregs) are reduced, and M1 macrophages exhibit polarization defects, both reversed following effective immunosuppressive therapy (IST).9-11 Standard options for severe AA are IST or hematopoietic cell transplant, depending on age, comorbidities, and donor availability. Current IST regimens with antithymocyte globulin are burdensome, often involving hospitalization and intravenous infusions, emphasizing the need for less toxic, orally administered alternatives. We previously demonstrated that the oral Janus kinase (JAK)1/2 inhibitor ruxolitinib (RUX) achieved robust efficacy in murine models of immune AA/BMF.12 In the current study, we investigated four additional JAK inhibitors (JAKi) chosen strategically to explore selective pathways: baricitinib (BAR, JAK1/2 inhibitor),13-15 solcitinib (SOL, JAK1 inhibitor),16 fedratinib (FED, JAK2 inhibitor),17 and tofacitinib (TOF, JAK1/3 inhibitor).16, 18, 19 Our objectives were (1) to explore the relationship between JAK specificity, pathway inhibition, and immune modulation, and (2) to assess the therapeutic efficacy of alternative JAKi in immune BMF. Using our previously established murine model: immune AA/BMF was induced in pre-irradiated hybrid (BALB/cByxC57BL)F1 recipient mice via infusion of 5–6 × 106 lymph node cells from B6 donors. Mice were treated with oral JAKi: BAR (25 mg/kg daily), FED (80 mg/kg twice daily), SOL (60 mg/kg twice daily), or TOF (15 mg/kg twice daily), beginning on Day 2 post-infusion for 7–10 days in short-term studies or 3 weeks for long-term survival assessments. Peripheral blood counts, cytokine levels, and BM cellularity were analyzed at Days 10 and 14. Histology, colony-forming cell assays, and RNA sequencing were performed to characterize hematopoietic and immune effects. Data were analyzed using JMP or GraphPad Prism; P < 0.05 was considered significant (Methods detailed in Supporting Information). Selective JAK inhibition (Figure 1A) improved hematologic parameters but with differing potency. At Day 10, all four inhibitors increased peripheral neutrophils (NEU), red blood cells (RBC), hemoglobin (HGB), and platelets (PLT), except FED, which failed to significantly elevate NEU. BAR and TOF, but not FED or SOL, improved residual BM cell numbers (RBM, excluding T cells). Among the JAKi, BAR produced the most pronounced recovery of hematological indices (Figure 1B). Histological analysis confirmed superior preservation of BM architecture and cellularity in BAR-treated mice compared to untreated controls and other inhibitors (Figures 1C and S1A–C). At Day 14, BAR maintained its superior performance across all parameters, with further increases in BM cellularity and RBM cells (Figure S2A). In contrast, SOL and FED yielded only modest improvements in RBC and HGB (without significant benefit in NEU or RBM cells), while TOF only improved RBC, HGB, and PLT counts without BM restoration (Figure S2B,C). At the cellular level, all inhibitors reduced BM infiltration of CD4+ and CD8+ T cells (Figure 1D). BAR and TOF increased splenic CD4+ T cells, though only BAR expanded Tregs in the spleen compared to untreated mice (Figure S2D). All inhibitors decreased apoptosis of RBM at Day 10, with JAK1/2 exhibiting the strongest suppression (Figure S3A). BAR also increased colony-forming units and myeloid progenitors (MPs) while reducing KSL (c-Kit+Sca1+Lin−) cell population (Figure S3B,C). By Day 14, mice receiving JAK1/2 inhibition (BAR) had sustained suppression of T cells and apoptosis, whereas those receiving JAK1 (SOL) or JAK2 (FED) inhibition showed only minimal suppression. Potent JAK1/3 inhibition (TOF) reduced CD8+ T cells but failed to enhance BM cell viability (Figure S2A–C). Plasma cytokine profiling revealed that inhibition of JAK1/2 significantly reduced proinflammatory mediators CCL2, CCL5, IL-5, IL-10, IFN-γ, and TNF-α at Day 10 (Figure 1E). At Day 14, CCL2, CCL5, and TNF-α remained suppressed, but IL-10 and IFN-γ reverted towards baseline (Figure S4A). In contrast, selective JAK1 or JAK2 inhibition failed to attenuate cytokine levels; JAK2 inhibition increased IL-1α (Figure S4B). Inhibition of JAK1/3 reduced IL-10 and CCL5 but paradoxically elevated IFN-γ (Figure S4C). Together, these findings demonstrate that dual JAK1/2 blockade most effectively suppresses inflammatory cytokines in immune BMF. Survival analysis further supported the superior efficacy of dual JAK1/2 inhibition (Figure 1F). Mice treated with JAK1/2 inhibitor (BAR) achieved 70% long-term survival (7/10) similar to RUX as previously reported,12 whereas JAK1 inhibition (SOL) and JAK2 inhibition (FED) resulted in no survival benefit. JAK1/3 inhibition (TOF) extended the median survival by only 8 days compared to controls, and long-term survival was not achieved. Longitudinal blood counts confirmed durable hematologic recovery in JAK1/2 treated mice (Figure 1G), while those treated with JAK1/3 inhibition exhibited transient improvement followed by decline in blood counts (Figure 1H), correlating with our reported survival data. Therefore, dual JAK1/2 inhibition was required for sustained therapeutic efficacy and improved survival. Single-cell RNA sequencing (scRNA-seq) demonstrated that all JAKi preserved HSPCs, erythroblasts, B cells, and granulocytes, while reducing CD4+ and CD8+ T cells by 10 days, with BAR producing the most significant effects (Figure 2A). Inhibition of JAK1/2 downregulated inflammatory response genes in hematopoietic subsets, whereas selective JAK1, JAK2, or JAK1/3 inhibition yielded weaker or inconsistent effects (Figures 2B and S5C). Of note, JAK1/3 inhibition induced higher expression of immune response genes including Stat1, 2, and 3, and Zbp1, Irf7, Ifit1, Ifit3, and Oas3 in HSPCs (Figure 2B). BAR modulated IL2–STAT5 and TNF-α/NFKB pathways in a cell-type-specific manner, underscoring the selective downstream signaling effects of JAK inhibition (Figure S6). Further analysis of sorted Lin−c-Kit+ (LK) cells revealed that BAR maintained the highest proportion of HSC/MPP, while FED, SOL, and TOF favored more differentiated LMPP retention, suggesting that JAK inhibition differentially preserves progenitor hierarchy (Figure S5F). In contrast to other JAKi, BAR induced elevated expression of Stat5a and reduced expression of Stat5b in LK cells (Figure 2C), which may explain their different impact on HSPCs. Further bulk RNA sequencing of BM CD8+ T cells identified time-specific transcriptomic effects of BAR. At Day 10, CD8+ T cells from BMF and BAR-treated mice showed overlapping transcriptomic profiles; however, by Day 14, they diverged markedly (Figure 2D). Early BAR exposure suppressed apoptosis and TNF-α/NFKB signaling while transiently activating interferon responses. Later at Day 14, BAR primarily suppressed T-cell proliferation pathways (Figure 2E), resulting in significant downregulation of selected T-cell activation and function genes including Ly6a, Gzmb, Stat4, Gzma, Ifngr1, Tyk2, and Gzmk, and moderate reductions in Fas, Il2rg, and Il2rb. However, upregulation of Stat1, Stat3, Lag3, and Ifngr2 (Figure 2F) was observed. In toxicity studies (Figure S7A–C), normal mice treated with each JAKi for 3 weeks exhibited minimal hematologic toxicity. Mild reductions in WBC and lymphocytes occurred, accompanied by increased PLT counts. No significant changes in BM histology, cellularity, or immune composition were observed, confirming favorable safety profiles. Collectively, these data demonstrate that dual JAK1/2 inhibition with BAR represented the most effective strategy for mitigating immune-mediated BM injury in a murine BMF model. BAR's therapeutic benefit appears to derive from suppression of activated T cells, reduction of proinflammatory cytokines, and preservation of Tregs and BM progenitors, resulting in improved long-term outcomes. These results parallel prior findings with RUX, another JAK1/2 inhibitor.12 In graft-versus-host disease murine models, dual JAK1/2 inhibitors RUX and BAR have also demonstrated superiority in suppressing T-cell activation, inflammatory cytokines, and survival.14, 20 Both BAR and RUX downregulated Ifngr1, implicating the IFN-γ/IFN-R1 signaling pathway as a key mediator of immune-mediated marrow destruction.8 Additionally, BAR upregulated Stat1, Stat3, and Ifngr2, suggesting divergent downstream JAK–STAT signaling responses that may enhance immune regulation while possibly supporting hematopoietic recovery. Time- and cell-type-specific transcriptomic differences underscore the complex biology of JAK inhibition. Dual JAK1/2 inhibition's immunosuppressive effects were most pronounced at later disease stages, consistent with delayed transcriptional downregulation of T-cell activation genes. In contrast, selective JAK1 or JAK2 inhibition produced transient improvements, possibly due to compensatory signaling by the uninhibited kinase. Despite potent JAK 1/3 inhibition, TOF demonstrated inferior long-term efficacy compared to RUX or BAR. We speculate that a broad inhibition across the JAK family may be needed for optimal immune suppression and hematopoietic recovery in immune BMF, but that JAK3 signaling may have an important role in sustaining hematopoiesis, a hypothesis further supported by our scRNA-seq data on HSPCs (Figure S3B,C). However, the scope of this study did not include functional analysis to directly evaluate the role of JAK3 preservation in hematopoietic recovery and therefore does not provide definitive evidence of such an association. Study limitations included the absence of pharmacokinetic data, though minimum doses were based on previously published animal studies. scRNA-seq of BM cells was restricted to Day 10 in untreated mice due to limited cell yield at Day 14. Nevertheless, bulk RNA-seq of CD8+ T cells at both time-points enabled temporal profiling. And finally, murine data may not fully extrapolate to human disease, and the efficacy of JAKi in BMF patients will need to be confirmed in clinical trials. In summary, this study underscores that selective JAK1/2 inhibition (BAR) offers superior immunomodulation and hematopoietic preservation compared with more selective or pan-JAK inhibition. The findings refine our understanding of the JAK–STAT axis in immune BMF and support further exploration of targeted JAK 1/2 inhibition. Flow cytometry sorting and analyses were performed in the Flow Cytometry Core of the National Heart, Lung, and Blood Institute. RNA sequencing was performed in the DNA Sequencing Core of the National Heart, Lung, and Blood Institute. Jibran Durrani: Conceptualization; writing—original draft; methodology; investigation; formal analysis; writing—review and editing. Xingmin Feng: Conceptualization; writing—original draft; methodology; investigation; formal analysis; writing—review and editing. Zhijie Wu: Software; formal analysis. Shouguo Gao: Software; formal analysis. Ash Lee Manley: Investigation; formal analysis. Liangliang Wu: Investigation; formal analysis. Joshua Glass: Investigation; formal analysis. Bhavisha A. Patel: Conceptualization; formal analysis; writing—review and editing. Jichun Chen: Conceptualization; writing—original draft; methodology; investigation; formal analysis. Neal S. Young: Conceptualization; formal analysis; writing—review and editing. Emma M. Groarke: Conceptualization; formal analysis; writing—review and editing. The authors declare no conflicts of interest. Funding for this study was provided by the Intramural Research Program of the National Heart, Lung, and Blood Institute. The authors confirm that the data supporting the findings of this study are available within the article and its Supporting Information. Single-cell gene expression data and bulk-RNA sequencing data were submitted to the National Center for Biotechnology Information Gene Expression Omnibus (GEO accession numbers GSE293533 and GSE293645). Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.