Identification of New Risk Loci and Regulatory Mechanisms Influencing Genetic Susceptibility to Acute Lymphoblastic Leukaemia

There is increasing evidence for an inherited genetic basis of susceptibility to acute lymphoblastic leukaemia (ALL) in children. In particular, genome-wide association studies (GWAS) have identified nine genomic regions with common polymorphisms associated with ALL risk. However, these loci cumulatively only explain a minority of the genetic risk of childhood ALL, and differences in the etiology amongst molelcular subtypes of ALL remain poorly understood. In this study, we performed one of the largest ALL susceptibility GWAS, totalling 5,321 cases and 16,666 controls of European descent, to comprehensively map ALL risk loci. With this large sample size, we also sought to identify risk genes for two major subtypes of ALL, namely those with ETV6-RUNX1 fusion and hyperdiploid ALL.

This meta-analysis included four previously published GWAS datasets with European cases from the UK ALL cohorts and the German Berlin-Frankfurt-Münster ALL trials, and the US cases enrolled on ALL trials in Children's Oncology Group and at St. Jude Children's Research Hospital (Nat Genet2009, Blood 2012, Nat Genet 2013, Blood 2013, Nat Commun 2018). The replication series included 2,237 ALL cases of non-European ancestry from the COG and St. Jude GWAS cohorts. For each GWAS dataset, the genotypes of ~10 million SNPs in each study were imputed. After filtering out SNPs on the basis of minor allele frequency and imputation quality, we assessed associations between ALL status and SNP genotype in each study using logistic regression. Risk estimates were combined through an inverse variance weighted fixed-effects meta-analysis. The meta-analysis identified 16 risk loci above genome-wide significance (P < 5 x 10^-8), of which ten confirmed previously published associations. Of the six new candidate risk loci, four were validated in the replication series (P < 0.05): for all B-ALL at 9q21.3 (nearest gene TLE1), for hyperdiploid ALL at 5q31.1 (C5orf56) and 6p21.31 (BAK1), and for ETV6-RUNX1 positive ALL at 17q21.32 (IGF2BP1). As well as providing further evidence for the 21q22.3 (ERG) association for all B-ALL, we also identified novel risk variants with the hyperdiploid subtype-specific association.

To gain insight into the biological basis of association signals, we examined the epigenetic landscape of risk regions in B-cells. Of particular note, hyperdiploid ALL-specific risk variant at 6p21 is located within an intron 1kb downstream of the BAK1 transcription start site and possess histone marks characteristic of active promoter activity and open chromatin accessibility. The top SNP at this locus falls within a transcription factor binding cluster, and the C-risk allele is associated with reduced BAK1 expression (P<2×10^-16). Mendelian randomization analysis confirmed a significant association with BAK1 expression and ALL consistent with a causal relationship. Similarly, the ETV6-RUNX1-specific risk variant at 17q12 is located in the second intron of IGF2BP1 and is predicted to influence topological associated domain structure in cis. Interestingly, ETV6-RUNX1 positive ALL leukemia cells showed a dramatic overexpression of IGF2BP1 compared to those without this fusion gene (P=3.7×10^-23), suggesting its potential role in the pathogenesis of this ALL subtype.

Analysis of disruption of transcription factor binding sites at ALL risk loci genome wide identified over-representation of PBX1 (P=0.007), TCF3 (P=0.007), ETS1 (P=0.009), RUNX1 (P=0.012) and ERG (P=0.030), BRD4 (P=0.007), and NR3C1 (P=0.009). Many of these transcription factors are also somatically mutated in ALL, pointing to multiple mechanisms by which they can contribute to leukemogenesis. Using the Linkage Disequilibrium Adjusted Kinships method, we estimated that ALL risk loci identified so far accounted for 31% of the total variance in genetic risk of this cancer. Polygenic risk score analysis indicates that an individual in the top 1% of genetic risk would have a 4.7-fold increased risk of ALL when compared to an individual with median genetic risk.

In summary, our study provides further evidence for inherited susceptibility to ALL and support for subtype specificity at risk loci. Our findings also highlight a model of ALL pathogenesis relying on transcriptional deregulation, particularly of genes involved in B cell differentiation.