In APL, noncoding mutations and SNP converge on WT1

In this issue of Blood, Song et al¹ show that in acute promyelocytic leukemia (APL), focal noncoding enhancer mutations disrupt MYB binding and WT1 expression, mimicking a germline single nucleotide polymorphism (SNP) that favors APL emergence.

APL is a model disease not only because of its exquisite sensitivity to the curative effect of retinoic acid and arsenic² but also because of its uncommon epidemiology. APL incidence is almost constant across all ages, pointing to a single key rate-limiting step. The initiating events are clearly translocation-driven RARA fusions, but extensive genomic studies, primarily exome sequencing, have revealed frequent activation of cooperating events during APL progression, primarily FLT3 mutations, but also that of many classic acute myeloid leukemia (AML) oncogenes, including WT1.^3,4 Whole-genome sequencing (WGS) of tumors has revealed a very large number of alterations, allowing detailed classification of the mutagenic processes. However, only a small fraction of these alterations in tumors are believed to be actual drivers.⁵ The most common alterations include mutations in splice sites that yield truncation/destabilization of tumor suppressors.^5,6 Gain of function alterations in enhancer sequences also exist, as exemplified by mutations in the telomerase gene promoter that increase telomerase expression by allowing de novo binding of transcription factors, as found in familial melanoma and other cancers.⁷ Yet, the impact of noncoding somatic mutations affecting enhancer function and directly contributing to oncogenesis appears to be limited in many cancers. Then, could WGS reveal a broader repertoire of genetic alterations that contribute to APL pathogenesis?

In this study, Song and colleagues investigated the mutational landscape of APL-associated noncoding mutations by performing WGS in 24 paired APL and germline samples. They subsequently focused on the cis-regulatory regions controlling gene expression and found that noncoding mutations were enriched in active enhancers bound by key myeloid pioneer factors, particularly MYB. When coupling analysis of these regulatory region mutations to RNA sequencing analyses, the authors discovered 38 mutated zones associated with deregulation of oncogene expression. Remarkably, in 3 patients among the 24 explored, tightly clustered mutations were identified in an intron of WT1, a gene with a complex relationship with myeloid leukemogenesis. By extending the analysis to 169 APL patients, the authors ultimately identified 6 somatic mutations in a hyperfocal 3 basepair region. Eight patients exhibited a germline polymorphism in the same sequence. As the frequency of this polymorphism in the Chinese population is 0.82%, this SNP exhibits a significant enrichment in APL patients, suggesting that it might be a risk factor for APL development. The mutant alleles opposed enhancer function, and patients displayed low WT1 expression transcribed from the mutant allele. Out of the 14 patients with this noncoding variant/mutant among the 169 APLs explored, 4 presented with biallelic functional inactivation, pointing to WT1 as a tumor suppressor in APL. Two patients displayed biallelic enhancer inactivation with homozygote variants or mutations; 2 others exhibited sharply reduced expression of one allele, while the other allele expressed a mutant WT1 protein. Mechanistically, these variants strongly reduced the binding of MYB to this intronic enhancer. MYB binding to this intronic enhancer is required for efficient interaction with the WT1 proximal promoter (see figure). These findings constitute a mirror image of T-cell acute lymphoblastic leukemia, in which gain of function intronic mutations create an MYB binding site in an enhancer of the TAL1 gene to promote its expression.⁸ It will be interesting to investigate the prevalence of these WT1 noncoding mutations and germline polymorphism in other AMLs.

First identified as a master gene whose loss of function drives abnormal kidney development, subsequent studies demonstrated that WT1 is mutated or overexpressed in many AMLs. Enhanced WT1 expression or mutations confers a poor prognosis, and WT1 was explored as a possible tumor antigen in AML. Mechanistically, WT1 recruits TET2 to modulate DNA methylation of WT1 target genes.⁹ In animal models, WT1 haploinsufficiency contributes to transformation.¹⁰ WT1 modulates the all-trans retinoic acid signaling pathway, which may explain the high incidence of WT1 alterations (coding or noncoding) in APL when compared with non-APL AML.

Collectively, these new results highlight a previously unrecognized molecular mechanism for loss of WT1 expression, possibly linked to the high prevalence of copy-neutral loss of heterozygosity in relapsing patients.³ Once again, APL serves as a paradigm for precision oncology.