Loss-of-function alterations in JAK family genes predict outcomes of immune checkpoint therapy across multiple cancer types

Dear Editor, Immune checkpoint inhibitors (ICIs) have shown remarkable success in cancer treatment, yet their efficacy varies widely, with many patients failing to respond despite established biomarkers like tumor mutational burden (TMB), microsatellite instability-high (MSI-H), and programmed death-ligand 1 (PD-L1)[1]. The role of Janus kinase (JAK) family genes – key regulators of interferon (IFN)-γ signaling – in ICI response remains complex; though JAK alterations were blamed for resistance[2,3], they are also tied to heightened immunity[4]. Clinical trials in non–small-cell lung cancer and Hodgkin lymphoma now show JAK blockade plus ICI can convert IFN-driven suppression into augmented response[5,6]. Methods To clarify the paradox, we first analyzed genomic data from 76 911 cancer patients across 224 studies to catalog loss-of-function (LOF) alterations in JAK1, JAK2, JAK3, and TYK2 in over 30 cancer types, excluding amplifications and the gain-of-function JAK2 V617F mutation. We then investigated the impact of JAK LOF alterations on ICI efficacy in a discovery set of 778 exome-sequenced tumors and a pan-cancer validation cohort of 1660 MSK-IMPACT profiles. Prognostic relevance was further assessed in 10 967 TCGA non-ICI cases, while mechanistic insights were sought by contrasting DNA damage repair (DDR) alterations and CIBERSORT-deconvoluted immune infiltration (relative fractions) between JAK LOF and non-LOF tumors. Additional study particulars are delineated in the Supplemental Digital Content Methods, available at: https://links.lww.com/JS9/H403. Results Across 76 911 patients, JAK LOF alterations averaged 5.2% (Supplemental Digital Content Figure S1, available at: https://links.lww.com/JS9/H398), peaked in nonmelanoma skin (27.4%), endometrial (21.6%), and melanoma (12.4%) tumors; also, they were consistently linked to elevated TMB in both ICI and TCGA cohorts (Supplemental Digital Content Table S1, available at: https://links.lww.com/JS9/H399), especially within microsatellite stable (MSS) disease. In the ICI discovery cohort (n = 778), baseline characteristics of the two groups were well-balanced (Supplemental Digital Content Table S2, available at: https://links.lww.com/JS9/H400). JAK LOF correlated with improved overall survival (median OS: 34.8 vs 15.1 months, P = 0.002; Fig. 1A) and progression-free survival (median PFS: 23.7 vs 10.6 months, P = 0.014; Fig. 1B), raised the objective response rate to 51.6% (P < 0.001; Fig. 1C) and the disease control rate to 71.0% (P = 0.002; Fig. 1C), with significance maintained after covariate adjustment (Table 1), and benefited even MSS tumors (P = 0.032; Supplemental Digital Content Figure S2, available at: https://links.lww.com/JS9/H398). These survival benefits were validated in an independent MSK-IMPACT cohort (n = 1660), where JAK LOF independently predicted better median OS (33.0 vs 17.0 months, adjusted P = 0.004; Fig. 1D and Table 1). Notably, no survival differences were observed in the TCGA cohort without ICI treatment (Fig. 1E–H), supporting its predictive rather than prognostic role. Figure 1.: Correlation between JAK loss-of-function alterations and efficacy of immune checkpoint inhibitors with underlying mechanisms. (A) Overall survival, (B) progression-free survival, and (C) treatment response ratios between JAK LOF and non-LOF patients in the ICI treatment discovery cohort. (D) Overall survival between JAK LOF and non-LOF alteration patients in the ICI treatment validation cohort. (E) Overall survival, (F) progression-free survival, (G) disease-free survival, and (H) disease-specific survival between JAK LOF and non-LOF alteration patients in the non-ICI treatment cohort. (I) Frequency of altered events in the DDR pathway between JAK LOF and non-LOF alteration patients. Others refer to the remaining 65 genes that are linked to more than one DDR pathway or coordinate cellular and molecular responses to DNA damage. (J–M) mRNA expression of the immune-related gene signature between JAK LOF and non-LOF alteration patients. Comparison of relative fractions of (N) antitumor effect immune cells and (O) immunosuppressive cell infiltration between JAK LOF and non-LOF alteration groups. BER, base excision repair; CI, confidence interval; CR, complete response; DCR, disease control rate; DDR, DNA damage response; DR, direct damage reversal/repair; FA, Fanconi anemia; HR, hazard ratio; HR, homology-dependent recombination; ICI, immune checkpoint inhibitor; LOF, loss-of-function; MMR, mismatch repair; NER, nucleotide excision repair; NHEJ, non-homologous DNA end joining; NP, nucleotide pool maintenance; ORR, objective response rate; PD, progression disease; PR, partial response; SD, stable disease; TLS, translesion DNA synthesis. Table 1 - COX regression analysis of survival in patients treated with immune checkpoint inhibitors in discovery and validation cohorts.. Parameter Univariable analysis Multivariable analysis HR (95% CI) P-Value HR (95% CI) P-Value Age (< 65 vs ≥ 65) 1.242 (0.867–1.780) 0.238 1.343 (0.923–1.954) 0.124 Sex (male vs female) 0.859 (0.692–1.067) 0.169 0.864 (0.937–1.388) 0.746 Drug type (mono vs combo) 0.692 (0.222–2.157) 0.526 0.679 (0.218–2.116) 0.604 TMB (< 10 vs ≥ 10) 0.566 (0.459–0.699) < 0.001 0.669 (0.456–0.982) 0.040 JAK (non-LOF vs LOF) 0.532 (0.354–0.801) 0.002 0.580 (0.346–0.915) 0.021 Age (< 65 vs ≥ 65) 0.763 (0.523–1.112) 0.159 0.713 (0.481–1.055) 0.091 Sex (male vs female) 0.998 (0.765–1.301) 0.988 0.991 (0.672–1.460) 0.963 Drug type (mono vs combo) 1.357 (0.990–1.860) 0.058 1.306 (0.951–1.792) 0.099 TMB (< 10 vs ≥ 10) 0.761 (0.580–0.999) 0.050 0.844 (0.555–1.285) 0.429 JAK (non-LOF vs LOF) 0.570 (0.362–0.900) 0.014 0.477 (0.238–0.953) 0.036 Age (< 65 vs ≥ 65) 1.004 (0.876–1.152) 0.951 1.018 (0.887–1.168) 0.799 Sex (male vs female) 1.134 (0.986–1.303) 0.078 1.125 (0.978–1.293) 0.099 Drug type (mono vs combo) 0.561 (0.453–0.695) < 0.001 0.545 (0.439–0.675) < 0.001 TMB (< 10 vs ≥ 10) 0.568 (0.479–0.673) < 0.001 0.574 (0.482–0.684) < 0.001 JAK (non-LOF vs LOF) 0.651 (0.466–0.909) 0.010 0.615 (0.440–0.860) 0.004 CI, confidence interval; HR, hazard ratio; LOF, loss of function; OS, overall survival; PFS, progression-free survival; TMB, tumor mutational burden. Mechanistically, JAK LOF was linked to impaired DDR (Fig. 1I; Supplemental Digital Content Table S3, available at: https://links.lww.com/JS9/H401), elevated expression of immune-related genes (T-cell receptors, checkpoints, T-effector and IFN-γ pathways, and the tumor microenvironment; Fig. 1J–M; Supplemental Digital Content Table S4, available at: https://links.lww.com/JS9/H402), and favorable immune cell infiltration (Fig. 1N and O) – increased CD8+ T cells, follicular helper T cells, M1 macrophages, and decreased Tregs and M2 macrophages. Discussion Whereas TMB, PD-L1, and DNA mismatch repair status are hindered by standardization and heterogeneity[1], JAK LOF – detected by next-generation sequencing as a binary, cost-effective biomarker – predicts ICI benefit across tumors through enhanced DNA repair defects and immune activation, and should be added to genomic panels to refine patient selection and personalize immunotherapy. Despite a 5.2% average incidence, JAK LOF is enriched in skin, endometrial, colorectal, and cervical cancers and predicts ICI benefit even in MSS tumors. Far from fostering immune evasion[2,3], it is accompanied by activated T-effector/IFN-γ signatures, dense CD8+ T cell infiltration, and high immune checkpoint expression, suggesting a crosstalk with IFN-γ-independent pathways such as cGAS-STING that remodel the microenvironment and recruit effector cells. This creates a “hot” yet functionally constrained phenotype – preexisting immunity with adaptive resistance – providing a mechanistic rationale for combining JAK inhibitors with ICIs[4–6]. Although the retrospective design and sample size limit some conclusions, consistent results across cohorts support further prospective validation. Conclusions JAK LOF marks pan-cancer ICI sensitivity, couples DNA-repair defects to hot immunity, and warrants JAK-blockade plus ICI combos to blunt resistance.