To the editor:

Acute myeloid leukemia (AML) is an aggressive hematologic malignancy characterized by clonal proliferation and impaired differentiation of myeloid progenitor cells [1]. AML predominantly affects older adults and exhibits substantial genetic heterogeneity, including structural chromosomal alterations and recurrent somatic mutations [2]. In the 2022 WHO/ICC classifications, AML entities are defined by recurrent genetic alterations [3, 4]. Among these, mutations in isocitrate dehydrogenase 1 (IDH1) and IDH2 occur in approximately 10–20% of cases and define a metabolically and epigenetically distinct subgroup [5]. The IDH1 p.R132H mutation produces the oncometabolite 2-hydroxyglutarate, which inhibits α-KG-dependent dioxygenases such as TET enzymes and induces DNA hypermethylation and transcriptional repression, thereby blocking hematopoietic differentiation and promoting leukemogenesis [6]. In addition to cell-intrinsic mutations, AML progression is influenced by extrinsic factors. Among these, the bone marrow niche contributes to a permissive and pro-inflammatory microenvironment that fosters leukemic proliferation and suppresses normal hematopoiesis [7]. A key mediator of this inflammatory milieu is IL-1/IL1R1 signaling, which disrupts hematopoietic stem-cell function and promotes early leukemic expansion [8, 9]. While the metabolic and epigenetic consequences of IDH1 mutations are well defined, it remains unclear whether IDH1-mutant (mut) AML (i) drives distinct intracellular signaling programs and (ii) shows an altered sensitivity to pro-inflammatory conditions in the bone marrow niche, thereby influencing treatment response and disease progression. These findings raise the question of how genetically defined AML subtypes - such as IDH1-mut AML - are shaped by inflammatory signals in the bone marrow niche.

To identify genotype-specific signaling pathways of IDH1-mut AML, we analyzed gene expression profiles from two independent AML cohorts (BeatAML2.0 and GSE146173; Fig. 1A, S1). Based on differential gene expression analysis (Table S2), gene set enrichment analysis (GSEA) identified pathway-level alterations between IDH1-mut and IDH1-wt AML samples. Among 39 significantly enriched pathways, 32 (82%) were downregulated in IDH1-mut AML, with 22 (69%) related to immune or inflammatory signaling (Fig. S2A). The top three downregulated hallmark pathways - inflammatory response, IL6/JAK/STAT3 signaling and interferon-gamma response - were consistently reduced in both cohorts (Fig. S2B/C). Similarly, in IL1R1-low AML cases, a suppression of inflammatory Hallmark and KEGG pathways was observed compared to IL1R1-high AML (Fig. S3, Table S3), mirroring the transcriptional profile of IDH1-mut AML. Based on the GSEA results from the IDH1-mut vs. IDH1-wt comparison, integration of the top three hallmark pathways with downregulated genes from the BeatAML2.0 dataset identified 43 overlapping candidates (Fig. S4, Table S4). Protein-protein interaction network analysis (STRING) revealed four functional clusters, including one centered on IL1R1 and related genes (IL1R2, IL18R1, IL18RAP), highlighting dysregulation of IL1R-family signaling, alongside modules linked to pattern recognition, TNF signaling and interferon responses (Fig. 1B). Comparison across genotypes confirmed that IL1R-family genes were specifically downregulated in IDH1-mut AML, but not in IDH2-mut cases (Figs. 1C, S5A). In contrast, genes associated with TNF signaling, pattern recognition and interferon-stimulated pathways were reduced in both IDH1- and IDH2-mut AML (Fig. S5B–D). Given the downregulation and shared chromosomal locus of IL1R-family genes (Fig. S6A), we hypothesized coordinated epigenetic regulation of this specific locus. DNA methylation profiling confirmed significant hypermethylation of the IL1R locus in IDH1-mut AML, with 20 of 73 CpG sites (27%) differentially methylated (Fig. S6B–E, Table S5). To assess clinical relevance of IL1R1 expression, AML patients were stratified into IL1R1 expression tertiles. For patients that received intensive induction therapy (n = 402; Fig. S7A) genotype-stratified analysis showed significant higher complete remission (CR) rates in the IL1R1-low group for both IDH1-wt and IDH1-mut AML (Fig. 1D). Patients with low IL1R1 expression also had significantly longer overall survival (OS) (low: median 1373 vs. medium 475 vs. high 439 days; p < 0.001) (Fig. 1E). IDH1 mutations were significantly enriched in the IL1R1-low group (14.9% vs. 7.5% vs. 3%; p = 0.002), however IDH1 mutation status alone did not affect survival (Fig. S7B). In a multivariable Cox model including age and ELN 2022 risk groups, IL1R1 expression as continuous variable was independently associated with OS (p < 0.001; Fig. 1F). The same prognostic association was observed when the analysis was extended to the entire cohort of intensive and non-intensively treated AML patients (n = 583), confirming IL1R1 as an independent marker of favorable outcome in AML (Fig. S7C/D). Together, these findings link the IDH1-mut genotype to IL1R-family gene downregulation and hypermethylation. Low IL1R1 expression, enriched in IDH1-mut AML, was associated with improved treatment response and survival, suggesting that reduced IL1R-mediated signaling is clinically relevant in IDH1-mut AML.

Fig. 1: Dysregulated inflammatory signaling and clinical impact of IL1R1 expression in AML.
figure 1

A Overview of cohort composition and RNA-seq-based differential gene expression and pathway analyses. B Network analysis of candidate genes from downregulated immune response pathways in IDH1-mut vs. IDH1-wt AML. Yellow nodes represent individual genes, highlighting functional modules involving IL1R-family members, TNF signaling mediators, pattern recognition and innate immune response, and interferon-stimulated antiviral pathways. C IL1R1 mRNA expression levels (RNA-seq) in IDH1/2-wt, IDH1-mut, and IDH2-mut AML samples from the BeatAML2.0 cohort. Each dot represents one sample (IDH1/2-wt, n = 497; IDH1-mut, n = 49; IDH2-mut, n = 69). p-values from t-tests; **p < 0.01; n.s., not significant. D–F Analyses restricted to intensively treated patients (n = 402) from the BeatAML2.0 cohort. D Complete remission (CR) rates across IL1R1 expression tertiles (low, medium, high) for IDH1-wt and IDH1-mut AML. For IDH1-wt and IDH1-mut genotype groups, the number of patients (n), number achieving CR (n), CR rates (%) and corresponding p-values from Fisher’s exact test comparing the low expression group versus the combined medium and high groups are shown. E Kaplan-Meier analysis of OS by IL1R1 expression tertiles. P-value was calculated using the log-rank test. Tables below indicate median survival, number of patients (n) and IDH1 mutation distribution in each expression group. Differences in IDH1 mutation frequency across expression groups were assessed using the chi-square test. F Multivariable Cox regression for OS including IL1R1 expression (continuous), age (≥60 vs. <60 y) and ELN 2022 risk classification. Hazard ratios with 95% confidence intervals are shown. For this model, the number of events was 237; the global p-value (log-rank) was 4.3515e−08; AIC was 2495.27; and the concordance index was 0.62.

Full size image

To explore the functional consequences of reduced IL1R1 expression in IDH1-mut AML, primary blasts (Table S6) were stimulated with IL-1β. Under basal conditions, IDH1-mut AML blasts showed reduced IL1R1 expression (Fig. S8, Table S7). Upon IL-1β stimulation, IDH1-mut blasts displayed markedly reduced induction of inflammatory mediators, including TNF, IL6, and CCL20, compared with IDH1-wt (Fig. S9A–D, Table S7), consistent with impaired NF-κB pathway activation. GSEA confirmed downregulation of TNF-α signaling via NF-κB and inflammatory response pathways (Fig. S10). For functional validation, we used the modified KG-1a cell line heterozygous for the IDH1 R132H mutation (IDH1-het) [10]. Consistent with patient data, IL1R1 expression was significantly reduced in IDH1-het compared to IDH1-wt cells, both under basal conditions and after IL-1β stimulation (Fig. S11A). Inflammatory protein profiling using Olink® showed significantly lower secretion of IL-1β-induced NF-κB target chemokines in IDH1-het cells (Fig. 2A), but selective upregulation of TRAIL (TNFSF10) and other immune modulators. Protein validation by ELISA confirmed distinct secretion patterns: IL-8 was strongly induced in IDH1-wt but not in IDH1-het cells, PLAU in both and TRAIL selectively in IDH1-het cells (Fig. S11B/C). Given its pro-apoptotic role, we tested whether IL-1β induces apoptotic programs in IDH1-het cells. GSEA revealed significant enrichment of apoptosis-related signatures in IL-1β-stimulated IDH1-het compared to IDH1-wt KG-1a cells (Fig. 2B), indicating that IDH1-mut reprograms IL-1β-responsive signaling towards a pro-apoptotic phenotype. Cross-cohort transcriptomic analyses revealed reduced basal TRAIL mRNA expression in IDH1-mut compared with IDH1-wt AML in BeatAML2.0 and a similar but nonsignificant trend in GSE146173 (Fig. S12A/B). These findings indicate that TRAIL expression is not constitutively elevated in IDH1-mut AML but becomes selectively induced under inflammatory conditions, consistent with the IL-1β-dependent response observed in our experimental models. To further test the functional relevance of IL1R1-dependent TRAIL induction under inflammatory conditions, IDH1-wt and IDH1-het KG-1a cells were exposed to pro-inflammatory stromal conditioned medium (HS-5 CM). HS-5 CM significantly increased TRAIL secretion and apoptosis in IDH1-het cells, both of which were abrogated by the IL1R1 antagonist Anakinra (Figs. 2C, D, S13A). Time-resolved caspase-3/7 activation confirmed accelerated apoptotic kinetics in IDH1-het compared with IDH1-wt cells under inflammatory conditions (Figs. 2E/F, S13B/C). In primary AML blasts, HS-5 CM induced higher TRAIL and lower TNF levels in IDH1-mut samples (Fig. S14A/B) and led to increased caspase-3/7 activation, which was reversed by IL1R1 blockade (Fig. S14C/D). Together, these results demonstrate that IDH1 mutations are associated with dysregulated IL1R1 signaling, characterized by reduced receptor expression, impaired downstream responses and increased susceptibility to stromal inflammatory signals. This identifies a genotype-linked vulnerability of IDH1-mut AML to inflammation-induced cell death through IL1R1 signaling.

Fig. 2: Altered IL1R1 signaling and apoptosis in IDH1-mut AML cells.
figure 2

A Volcano plot depicting significantly differentially secreted proteins in IDH1-het vs. IDH1-wt KG-1a cells following IL-1β stimulation (10 ng/mL, 18 h), based on n = 6 biological replicates and measured using the Olink® Target 96 Inflammation panel. Dashed horizontal line indicates a -log10 adjusted p-value threshold of 1.3 (p = 0.05); vertical dashed lines represent log2 fold change cutoffs of ±1. B Bar plot showing GSEA results based on transcriptomic profiles of IL-1β–stimulated IDH1-wt and IDH1-het KG-1a cells. Selected hallmark gene sets are ranked by NES. Yellow bars indicate positive enrichment, green bars negative enrichment. Asterisks denote significance based on FDR q-values: *q < 0.05. C ELISA measurement of TRAIL levels in cell supernatants of IDH1-wt and IDH1-het KG-1a cells following stimulation with HS-5 CM for 48 h, with or without Anakinra (10 µg/ml) treatment. Data represent mean ± SD from three independent biological replicates. Statistical analysis: unpaired two-tailed t-tests with Šidák correction for multiple comparisons; *p < 0.05, **p < 0.01; n.s. = not significant. D Flow cytometric quantification of apoptosis (% Annexin V+ cells) in IDH1-wt and IDH1-het cells after 48 h of stimulation with 10 ng/ml IL-1β or HS-5 CM, with or without Anakinra (10 µg/ml) treatment. Data represent mean ± SD from three independent biological replicates. Statistical analysis: unpaired two-tailed t-tests with Šidák correction for multiple comparisons; *p < 0.05, **p < 0.01, ***p < 0.001; n.s. = not significant. E, F Caspase-3/7 activation in IDH1-wt and IDH1-het KG-1a cells under control conditions or with HS-5 stromal cell–derived conditioned media (CM) over 72 h, assessed by live-cell fluorescence imaging without (E) or with 10 µg/ml Anakinra (F). Fluorescence was measured at 0, 24, 48 and 72 h. Data represent mean ± SD of n = 3 biological replicates. Statistical analysis was performed using two-way repeated measures ANOVA; Geisser-Greenhouse correction was applied in (E) due to violation of sphericity. **p < 0.01, n.s. = not significant.

Full size image

Inflammatory signaling is increasingly recognized as a determinant of AML biology, contributing to disease progression and hematopoietic dysfunction [11], yet, it remains unclear how genetic drivers modulate these pathways. Here we show that IDH1-mut AML is characterized by downregulation of inflammatory signaling pathways, consistent with prior reports linking IDH1 mutations to immune suppression [12]. Low expression levels of IL1R1, a central regulator of inflammatory pathways, correlated with improved chemotherapy response and survival in both IDH1-wt and IDH1-mut AML, with significant enrichment of IDH1-mut samples among IL1R1-low cases. This extends previous studies identifying high IL1R1 expression as a negative prognostic marker [13, 14] by showing that IL1R1-low status is significantly enriched in IDH1-mut AML. Functionally, IDH1-mut AML cells showed an apoptosis-prone response to IL-1β or stromal-derived inflammatory stimuli, contrasting the general pro-survival effects of IL-1β in AML. Reduced IL1R1 expression underpinned this IDH1-specific phenotype, as IL1R1 blockade with Anakinra reversed apoptosis in IDH1-mut but not IDH1-wt cells. At the molecular level, IDH1-mut AML showed impaired IL1R1-mediated signaling, reduced NF-κB activation, selective TRAIL induction, and enrichment of pro-apoptotic transcriptional programs. Although additional 2-HG-dependent epigenetic alterations may contribute to this phenotype, the consistent association of low IL1R1 expression with improved response and survival indicates a biologically relevant vulnerability in IDH1-mut AML. Importantly, this does not imply that IL1R1 should be therapeutically inhibited in this subgroup. On the contrary, our data suggest that intact IL1R1 signaling facilitates apoptosis induction in IDH1-mut cells under inflammatory conditions, indicating that its activity may in fact contribute to therapeutic efficacy. Accordingly, IL1R antagonists such as Anakinra or canakinumab, currently evaluated in hematologic malignancies (NCT04239157) [7, 15], may require genotype-specific consideration. These findings may also have implications for IDH1-targeted therapies such as Ivosidenib, which lower 2-HG levels and promote differentiation.

In conclusion, our data indicate that IDH1 mutations remodel IL1R-mediated signaling toward an inflammatory-anergic but pro-apoptotic state, revealing a previously unrecognized vulnerability of IDH1-mut AML to microenvironmental inflammatory stress. Reduced IL1R signaling may not only predict improved chemotherapy response but also enhance responsiveness to differentiation therapies. Exploiting this vulnerability may guide the development of more effective, genotype-adapted treatment strategies to improve outcomes in IDH1-mut AML.