Immunoglobulin variable gene diversity reflects distinct metabolic/signaling states and impacts measurable residual disease monitoring and clinical outcome in B cell acute lymphoblastic leukemia

Abstract

Abstract High-throughput sequencing (HTS) of clonal immunoglobulin (Ig) gene rearrangements has become a mainstay of B-lymphoblastic leukemia (B-ALL) measurable residual disease (MRD) monitoring; Ig HTS also provides insight into the widely documented leukemic clonal heterogeneity in Ig heavy chain (IgH) variable (V), diversity (D), and joining (J) gene rearrangements in B-ALL. B-ALL IgH clonal composition is canonically presumed to reflect the transformed cell's rearrangement state: B cells develop along a pathway with IgH D-J followed by IgH V gene rearrangement; thus, subclones comprised of a common DNJ sequence, but distinct alternative V gene involvement, would be assumed to derive from a cell stage at which only D-J recombination had taken place with persisting recombination machinery for subsequent V-DJ rearrangement. We and others have observed that this ‘V-DJ diversity’ may be associated with prognostic cytogenetic features, but its relevance as a supposed reflection of B cell developmental state has not been confirmed. Here, we define the biologic significance of IgH variable gene diversity in B-ALL and reveal its direct implications on HTS MRD and its potential impact on clinical outcome. We analyzed 148 B-ALL samples from patients enrolled on former Children's Oncology Group (COG) trials (N=99) and an institutional tissue bank (N=49) to test the functional implications of IgH clonal composition in B-ALL. We defined 3 cohorts: 1) No clonal IgH (N=7; 4.7%); 2) Diverse, with V-DJ ‘subclones’ containing unique V genes but a common DNJ stem (N=53; 35.8%), and 3) Non-diverse, with single VDJ rearrangement(s) per allele (N=88; 59.5%). We confirmed that the extent of V-DJ diversity in each B-ALL – quantified by Shannon entropy index – is inherently preserved across tissue sites (bone marrow (BM) vs. peripheral blood (PB) R2=0.99; P<.0001) and early induction therapy timepoints rather than merely a stochastic phenomenon. Chromosome microarray demonstrated no association between prognostic cytogenetics, V-DJ diversity, and other non-IgH breakpoint accumulation across 4 biologically distinct, prognostic cytogenetic categories. Using CyTOF, we found that regardless of the pattern of IgH rearrangement(s), all cases were enriched in pro-BII cell populations. However, by Gene Ontology and Hallmark gene set enrichment analysis (GSEA) and CyTOF, cases with V-DJ diversity had distinct transcriptional and intracellular proteomic features – including enriched gene and protein expression of pentose phosphate and mitochondrial metabolism genes and proteins, as well as activated mTOR signaling – compared to cases with non-diverse VDJ rearrangements. Further, by ATACseq, we observed distinct chromatin accessibility between diverse and non-diverse cohorts, including increased accessibility at the IgH locus in diverse cases. Integrated single-cell and clinical Ig HTS clonality data revealed that distinct V-DJ subclone sequences (derived from a similar progenitor population sharing a DNJ stem) comprise discrete cell populations over a variable range of abundances; therefore, all or some such sequences will not necessarily meet clinical HTS clonality/MRD assay criteria to be defined as ‘dominant/trackable’ for MRD determination. Both via single-cell RNA seq and custom Tapestri sequencing, we found that diverse V-DJ subclone sequences distinguish discrete cell subpopulations. Thus, we tested the impact of IgH rearrangement diversity on MRD detection and outcome using a dataset from 307 subjects with high risk (HR) pediatric B-ALL. Among the prognostically inferior cohort from COG AALL0232 which lacked any designated ‘trackable’ sequences (N=31/307) (Fries et al. Haematologica 2023), we found that 7 (of 31; 2% of the total 307) had subclonal IgH rearrangements with unique V/common DJ sequences of similar abundance such that none were designated 'trackable.' Despite treatment intensification based on positive end of induction (EOI) flow cytometry MRD on AALL0232, these patients had inferior 5-year EFS of 71.4% (95%CI 44.7-100%; P=.001). In 5 (of 7) cases, a subset of V-DJ 'subclones' were still detectable at the EOI timepoint, suggesting treatment resistance. In conclusion, IgH V-DJ diversity reflects distinct metabolic/transcriptional pathway activation in proB-like leukemia cells and defines discrete B-ALL cell populations with possibly distinct treatment responses, potentially impacting clinical outcome.