Single-cell multiomic mapping of genetic predisposition to childhood B-cell acute lymphoblastic leukemia

Technical Summary: Single-cell multiomic mapping of genetic predisposition to childhood B-cell acute lymphoblastic leukemia

Problem Statement
Childhood B-cell acute lymphoblastic leukemia (B-ALL) is the most common pediatric malignancy, with inherited germline genetic variation playing a substantial role in disease susceptibility. Genome-wide association studies (GWAS) have identified at least 16 replicable risk loci; however, the underlying mechanisms remain poorly understood. A primary challenge is that GWAS signals typically contain numerous linked variants due to linkage disequilibrium (LD), making it difficult to distinguish causal variants from non-causal neighbors. Furthermore, while computational fine-mapping can prioritize candidate single nucleotide polymorphisms (SNPs), these predictions remain probabilistic. Crucially, there is a lack of systematic approaches to functionally resolve these risk variants within the precise cellular contexts—specifically rare and transient B-cell progenitors—where they exert their effects during hematopoietic development.

Methodology
The authors employed a systematic, integrative framework combining statistical genetics with high-resolution single-cell multiomics to map risk variants to their functional mechanisms:

1. Statistical Fine-Mapping: The study utilized three complementary algorithms (FINEMAP, SuSiE, and CARMA) on a large trans-ethnic meta-GWAS of childhood B-ALL to define 95% credible sets for 16 risk loci. The union of these sets yielded 1,054 likely causal variants, the vast majority of which are non-coding.
2. Single-Cell Multiomic Mapping (scMultiome): The authors constructed a high-resolution reference map of human hematopoiesis by integrating joint chromatin accessibility (ATAC-seq) and RNA expression (scRNA-seq) data from 190,309 cells across 34 distinct cell types. This map was used to assess the enrichment of risk variants in specific cellular states using the SCAVENGE algorithm, which propagates genetic association signals across a cell-to-cell network to calculate trait-relevance scores (TRS).
3. Allele-Specific Chromatin Accessibility (caQTL) Analysis: Leveraging a bulk ATAC-seq dataset from 156 B-ALL patients, the authors identified chromatin accessibility quantitative trait loci (caQTLs). They overlapped these with fine-mapped risk variants to detect allele-specific effects on chromatin accessibility, specifically looking for "caSNPs" (chromatin accessibility SNPs) that alter regulatory element activity.
4. Transcription Factor (TF) Motif Disruption: The study integrated caQTL effect sizes with predicted TF binding affinity changes (using motifbreakR) to identify risk alleles that disrupt or create TF binding sites. This allowed for the inference of which TFs mediate the regulatory impact of specific risk variants.
5. Variant-to-Target Gene Mapping: To link regulatory variants to target genes, the authors combined two orthogonal approaches: (i) correlation of chromatin accessibility and gene expression within the scMultiome map, and (ii) the Activity-By-Contact (ABC) model applied to Hi-C and ATAC-seq data from B-ALL blasts.
6. Functional Validation: The study focused on the ELK3 locus and the risk variant rs2268508. Validation included:
   • Reporter Assays: Lentiviral reporter constructs in REH and Nalm-6 cells to test enhancer activity of the risk vs. non-risk alleles.
   • CRISPR Activation (CRISPRa): Activation of the endogenous enhancer in primary CD34+ hematopoietic stem and progenitor cells (HSPCs) to measure ELK3 upregulation.
   • Overexpression and Knockout: RNA-seq of ELK3-overexpressing cells and scRNA-seq of HSPC-derived B cells with ELK3 knockout to assess transcriptional consequences and differentiation defects.

Key Results

• Cellular Context of Risk: SCAVENGE analysis revealed that B-ALL risk variants are significantly enriched in early B-cell developmental stages, specifically V(D)J pro-B and small pre-B cells. This enrichment was not observed in other hematopoietic lineages, suggesting that inherited susceptibility acts primarily by perturbing regulatory programs during the window of active V(D)J recombination.
• TF Network Perturbation: The study identified 74 transcription factors whose motif activity or expression correlates with risk variant enrichment. These TFs (including EBF1, PAX5, RUNX1, and IRF4) are critical for B-cell lineage commitment. Analysis of caQTLs revealed that nearly half of the fine-mapped regulatory variants cause strong alterations in the binding of these key TFs (e.g., RUNX1, SPI-B, PU.1).
• High-Confidence Susceptibility Genes: By integrating CRE-gene correlations and ABC model contacts, the authors identified 34 high-confidence B-ALL susceptibility genes, including known drivers (e.g., ARID5B, IKZF1, GATA3) and novel candidates.
• Mechanism of rs2268508 and ELK3:
  • The risk allele (C) of rs2268508, located in an intronic enhancer of ELK3, creates a stronger binding motif for PU.1 (SPI1) and increases chromatin accessibility in B-cell progenitors.
  • Functional assays confirmed that the risk allele acts as a potent enhancer, selectively upregulating ELK3 expression.
  • ELK3 overexpression suppressed NF-κB signaling and lymphocyte activation pathways while promoting a core leukemia stem cell signature and recapitulating transcriptional changes driven by ERG overexpression.
  • Conversely, ELK3 knockout in primary HSPCs impaired B-cell maturation, reducing the proportion of cells transitioning to immature and mature B-cell states and downregulating B-cell progenitor drivers.

Significance and Claims
The paper claims to establish a comprehensive "variant-to-function" map that deciphers the regulatory disruptions underlying inherited B-ALL predisposition. By integrating statistical fine-mapping with single-cell multiomics and functional validation, the authors provide mechanistic insights into how common germline variants rewire B-cell progenitor regulatory networks.

Specifically, the study identifies ELK3 as a previously unrecognized regulator of B-cell development and leukemogenesis. The authors posit that the rs2268508 risk allele promotes B-ALL susceptibility by enhancing ELK3 expression, which simultaneously disrupts normal B-cell maturation (via NF-κB suppression) and promotes a leukemia stem cell state. This dual role mirrors the behavior of ERG, another ETS-family TF.

The authors conclude that their integrative multiomic framework offers a scalable platform to translate GWAS discoveries into mechanistic understanding, potentially paving the way for future targeted prevention approaches by identifying the specific regulatory circuits and developmental windows where genetic risk converges to initiate malignancy.