IDH2 Mutations from Clonal Hematopoiesis and Acquired IKZF1 Deletions Define a Novel B-ALL Subtype after Lenalidomide Therapy for Multiple Myeloma

Background: Lenalidomide (Len), approved in 2017 as maintenance therapy for multiple myeloma (MM), is known to increase the cumulative incidence of secondary malignancies, including Len-associated B-cell precursor ALL (LenB-ALL). However, the molecular basis of LenB-ALL leukemogenesis is not yet fully understood.

Methods: We present the results of a comprehensive molecular characterization of consecutive diagnostic samples from patients (pts) with secondary B-ALL following Len therapy for MM. Samples were sent in for GMALL (German Multicenter Study Group for Adult ALL) reference ALL diagnostics and were analyzed by gene panel sequencing (seq), IGH rearrangement profiling, SNParray / MLPA and DNA methylation profiling, bulk RNA-seq, and, in selected cases, single-cell RNA-seq.

Results: A total of 58 consecutive LenB-ALL pts (median 64.7 years, range 47-78) were included. The mutational profile showed a strong enrichment of clonal hematopoiesis (CH) genes with three mutually exclusive mutation patterns: (1) TP53 mutated in 17/58 pts (29 %), (2) IDH2 p.R140Q mutated in 13/58 pts (22 %), which is rare in primary B-ALL (<1%), and (3) other mutations including DNMT3A, NRAS and KRAS.

Unsupervised gene expression analysis of LenB-ALL within the context of our large primary B-ALL cohort (n = 596) grouped TP53 mutated LenB-ALL cases together with low hypodiploid B-ALL cases which frequently harbor TP53 mutations. Consistent with its more heterogeneous genomic makeup, the DNMT3A, N/KRAS cases showed a less distinct gene expression signature. However, IDH2-mutated LenB-ALL cases formed a highly distinct cluster, which could be predicted with 99% accuracy, compared to other LenB-ALL and primary B‑ALL cases, establishing a novel IDH2-mutated LenB-ALL subtype. Epigenomic profiling using DNA methylation arrays independently confirmed this cluster separation and showed hypermethylation of top variably methylated CpGs in IDH2 and DNMT3A mutated cases, consistent with the epigenetic functions of these regulators.

Interestingly, IKZF1 – encoding Ikaros, a target of Len-induced proteasomal degradation – harbored intragenic deletions in 7/13 (54 %) IDH2-mutated LenB-ALL pts. A loss of Ikaros may lead to a B-cell maturation arrest at the proliferative large pre-B cell stage culminating in (oligo)clonal B-ALL expansion (I Joshi et al Nat Immunol 2014). Consistently, we observed evidence of IGH clonal evolution / oligoclonality in 6/12 IDH2-mutated LenB-ALL (50 %) using amplicon-based IGH NGS, compared to only 24 % in 355 adult primary pre/c-B-ALL.

To elucidate the clonal architecture of IDH2-mutated LenB-ALL, we analyzed diagnostic and follow-up samples by IDH2 R140Q- and IKZF1del-specific ddPCR (sensitivity 0.1 %) and compared the results with IGH-based MRD kinetics: IKZF1-specific ddPCR suggested a late onset of IKZF1-deletions during leukemogenesis with sub- and oligoclonality and restriction to the B-ALL compartment. In contrast, the IDH2 mutation was detected as clonal hematopoiesis in 4 of 13 patients during MRD-negativity, where it was confirmed in both, lymphoid and myeloid compartments by ddPCR on FACS-sorted hematopoietic compartments. Single-cell proteotranscriptomics of a diagnostic and remission bone marrow sample evidenced a preleukemic, multilineage IDH2-mutated clone in one patient. In one patient with an available bone marrow aspirate from the time of MM diagnosis, the IDH2 mutation was not detected in the MM.

Conclusions: We describe a novel molecular subset of LenB-ALL driven by IDH2-mutated CH. We hypothesize that IDH2-mutated CH expands in the presence of Len and that Len-induced degradation of Ikaros leads to a maturation arrest of IDH2-mutated B-cell progenitors. Subsequent IKZF1 deletions render Ikaros inactivation independent of Len providing the framework for full IDH2-driven leukemic transformation. Prospectively, the prevalence of IDH2-mutated CH in multiple myeloma and its impact on the risk of LenB-ALL are to be evaluated.