DNA methylation instability as a biomarker to predict resistance or progression to advanced disease Phase following tyrosine kinase inhibitor therapy in chronic myeloid leukemia

Le résumé fourni par la source

Abstract Introduction DNA methylation occurs at defined CpG sites across the genome and plays a critical role in regulating gene expression. In young, healthy individuals, methylation patterns in peripheral blood cells are largely confined to stable methylation sites (SMSs). With aging, increased variability in these patterns has been observed, a phenomenon termed DNA methylation instability (DMI). Elevated DMI levels was found to reflect underlying epigenetic heterogeneity within the hematopoietic compartment and serve as a surrogate marker for clonal haematopoiesis. In our previous work we identified elevated DMI levels in lymphoid and myeloid malignancies (ASH 2023). Chronic myeloid leukemia (CML) arises from the clonal expansion of a hematopoietic progenitor cell harbouring the BCR::ABL1 fusion gene, but disease progression and treatment response are influenced by additional molecular and epigenetic alterations. We hypothesized that DMICML level can be employed as a prognostic biomarker in chronic phase (CP) CML. Specifically, DMICML level at diagnosis and during follow up may be used to predict response to tyrosine kinase inhibitors (TKI) therapy, achievement of treatment free remission and progression to blast phase CML. Patients and method Paired bone marrow samples collected at diagnosis and follow-up were obtained from 78 patients with CML undergoing treatment with TKIs. Samples from 41 healthy donors (HDs) were included as a control group to establish baseline methylation variability. Genome-wide DNA methylation profiling was performed using the Illumina MethylationEPIC BeadChip. To quantify DNA methylation instability specific to CML, we calculated a DMICML, defined as the deviation in methylation levels across a curated set of SMSs. Methylation at each site was measured using β-values representing the ratio of methylated probe intensity to the total signal intensity (methylated plus unmethylated). Results The CML cohort included 24 females (30%), with a median age of 55 years (range: 19–80). At diagnosis, 75 patients (96%) were in chronic phase, and 24 (31%) had high-risk Sokal scores. Most patients (n = 65, 84%) received imatinib as first-line therapy. After a median follow-up of 1,952 days, the 4-year progression-free survival (PFS) rate was 94% (95% CI: 86–98%). PFS differed significantly by response category: 97% among optimal responders, 91% in patients with resistance, and 50% in those with progressive disease (P = 0.002). DMICML was calculated based on 6,334 CpG sites. These sites were selected based on a cut-off with a standard deviation (SD) > 0.18, which provided the best discriminatory power for outcome prediction, producing an area under the curve (AUC) of 0.654 for outcome discrimination. No correlation was found between DMICML and age in the CML pts cohort. DMICML showed a significant decrease from diagnosis (0.248±0.015) to follow-up (0.210±0.016 p=1.075×10⁻13). Notably, both DMICML values were markedly higher compared to HDs (0.008±0.004 p=3.93×10⁻19), highlighting persistent epigenetic instability in CML despite treatment. DMICML at diagnosis distinguished optimal responders from the resistance/progression group: In optimal responders, the median DMICML at diagnosis was 0.248 (IQR 0.238-0.257), compared to 0.255 (IQR 0.250-0.259) in the resistance/progression group (p=0.021). In contrast, no correlation was found between DMICML at follow-up, and clinical outcomes between optimal responders and resistant/progression group (0.209 vs 0.212; p=0.49). DMICML at diagnosis was correlated with BCR::ABL1 transcript level (R = 0.38, p = 0.0015). However, no correlation was observed between DMICML and BCR::ABL1 transcript levels during follow-up. Conclusion Our study employs a genome-scale approach to characterize DNA methylation instability patterns and identify epigenetic drivers of CML. Elevated DMICML at diagnosis may reflect clonal heterogeneity and serve as a surrogate biomarker to predict resistance or disease progression during TKI therapy. Our findings contribute to improved patient risk and treatment stratification. Its correlation with BCR::ABL1 transcript level results warrants further refinement.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.