AUTHORS
Yongwen Zhuang, PhD, Helene Pelicano, PhD2, Rajyalakshmi Luthra, PhD3, Shashikant Kulkarni, MS, PhD, MBA, FACMG4, Ghayas C. Issa, MD5, Courtney DiNardo, MD, MSCE6, Carlos Bueso-Ramos, M.D., Ph.D.7, Sa Wang, MD8, Christiana Dobrzynski9, Danielle Goldberg10, Suzanne Rohrback, PhD11, James Baye, PhD12, Tevfik Dincer, PhD13, Erica Schnettler, PhD14, Pratheesh Sathyan, PhD15, Rashmi Kanagal-Shamanna, MD16
AFFILIATIONS
1 Illumina, Inc., San Diego, CA, United States
2 The University of Texas M. D. Anderson Cancer Center, Houston, TX, United States
3 The University of Texas M. D. Anderson Cancer Center, Houston, TX, United States
4 The University of Texas M. D. Anderson Cancer Center, Houston, Texas, United States
5 The University of Texas M. D. Anderson Cancer Center, Houston, Texas, United States
6 The University of Texas M. D. Anderson Cancer Center, Houston, Texas, United States
7 The University of Texas M. D. Anderson Cancer Center, Houston, TX, United States
8 The University of Texas M. D. Anderson Cancer Center, Houston, TX, United States
9 Illumina, Inc., San Diego, CA, United States
10 Illumina, Inc., San Diego, CA, United States
11 Illumina, Inc., San Diego, CA, United States
12 Illumina, Inc., San Diego, CA, United States
13 Illumina, Inc., San Diego, CA, United States
14 Illumina, Inc., San Diego, CA, United States
15 Illumina, Inc., San Diego, CA, United States
16 The University of Texas M. D. Anderson Cancer Center, Houston, TX, United States
ABSTRACT
Introduction: Menin inhibitor sensitivity in AML with KMT2A rearrangement (KMT2Ar) or NPM1 mutation is largely defined by dependence on the menin–KMT2A–HOX/MEIS transcriptional axis. Pharmacologic disruption of menin-KMT2A interaction represses HOXA/MEIS1 leukemogenic program and promotes leukemic differentiation. Despite meaningful clinical responses, patients relapse through acquired MEN1 mutations, clonal evolution, adaptive epigenetic escape and/or transcriptional reprogramming. However, the relationship between disease evolution and methylation-state changes at relapse is not well characterized. In this study, we performed longitudinal 5-base WGS of paired pre- and post-remission relapse AML samples from patients treated with menin inhibitors to evaluate paired genomic and 5-methylcytosine (5mC) methylation changes associated with treatment resistance.
Methods: Residual genomic DNA from paired pre-treatment and post-relapse AML samples treated with revumenib was used for library preparation (Illumina™ 5-base DNA Prep kits). Paired-end 2x150 bp sequencing was performed on the Novaseq™ X to a target average coverage of ~100X per sample. DRAGEN™ analysis was used for simultaneous assessment of genetic variants and genome-wide DNA 5mC signatures. Illumina Connected Multiomics was used to identify relapse-associated differentially methylated regions, followed by gene annotation and pathway-level assessment of menin inhibitor resistance programs.
Results: The study cohort included patients with AML (n=19) harboring genomic alterations associated with menin inhibitor sensitivity, including KMT2Ar or NUP98r, KAT6A::CREBBP or NPM1m. The median age was 37 years (4-77). All patients received revumenib and subsequently developed post-remission relapse. The median time from treatment start to relapse was 7 months. The primary paired analysis included 7 AML patients with KMT2Ar with available paired pre- and post-relapse samples. Genome-wide differential methylation analysis identified 10,811 autosomal candidate DMRs (DSS p-value<0.01 with paired Wilcoxon signed-rank test, FDR<0.05), among which 35 have methylation delta >0.05%. Most DMRs were hypomethylated at relapse (86%; mean delta-beta=-0.045). These findings support broad therapy-associated methylation remodeling of relapsed samples. In contrast to the predominant background of global hypomethylation, canonical menin/HOX target genes including MEIS1, HOXA, PBX3, MEF2C, and FLT3 showed selective hypermethylation (mean delta-beta=0.158). DMRs mapped to or near genes implicated in epigenetic escape, menin inhibitor resistance or AML progression. Overall, DMR methylation changes showed 73% directional concordance with COSMIC pan-cancer methylation patterns. Pathway enrichment analysis (KEGG) identified significant involvement of focal adhesion (FDR=2.4X10⁻⁷), Rap1 signaling (FDR=1.6X10⁻⁶), PI3K-Akt signaling (FDR=4.5X10⁻⁴), and AML pathway (FDR=5.0X10⁻³). Principal component analysis of study-relevant methylation features showed shared post-relapse epigenetic trajectory, with top loadings enriched for AML-associated driver genes. No well-characterized MEN1 resistance mutations were identified at the current sequencing depth. Potential relapse-emergent indels in a 3.5 kb regulatory region were identified in 5/7 patients (71%). chr11:64813288 emerged as a potential hotspot in 3/5 patients which could suggest a potential novel regulatory resistance mechanism. Among post-relapse samples, patients with vs without relapse-emergent indels showed differential methylation affecting genes implicated in niche-mediated resistance. Pathway enrichment analysis revealed significant involvement of integrin signaling (FDR<0.05), focal adhesion (FDR<0.05), and ECM-receptor interaction (FDR<0.05). Differential methylation was observed in HOXA cluster genes and MEIS1, along with PI3K-Akt pathway components. Further correlation of mutation status in pathway-relevant genes, flow cytometry immunophenotypic characterization, and clonal evolution is ongoing.
Conclusions: Relapse after menin inhibitor therapy in KMT2A-r AML showed recurrent therapy-associated methylome remodeling, characterized by global hypomethylation with selective hypermethylation of canonical HOX/MEIS pathway genes. The enrichment of DMRs involving AML resistance-associated genes and signaling pathways supports adaptive epigenetic escape and/or clonal evolution as additional mechanisms contributing to menin inhibitor resistance.
DISCLOSURES
Rashmi Kanagal-Shamanna, MD: No relevant conflicts of interest to declare.