Therapy-induced mutations drive the genomic landscape of relapsed acute lymphoblastic leukemia

Abstract

Li and colleagues report the genomic landscape of over 100 patients with relapsed acute lymphoblastic leukemia. Analysis of diagnosis-relapse-remission trios suggest that whereas early relapse is mediated by retained subclones, late relapse is driven by mutations induced by and conferring resistance to chemotherapy.

Feasibility of monitoring peripheral blood to detect emerging clones in children with acute lymphoblastic leukemia†

“…Notably, p.H352D was detected at a fractional abundance of 29% in the relapse sample (WES VAF 32%) and in two separate aliquots of an on‐therapy maintenance sample drawn 116 days prior to relapse at 0.0095% and 0.0068% for an average of 0.008% (Figure 1B; Figure S2A,B). Overexpression of the previously uncharacterized NT5C2 p.H352D in an engineered Reh cell line led to a ∼4‐fold increase in IC50 of 6‐TG (Figure S2C), which was corroborated in a recent report 9,10 …”

“…At the time of our study, NGS methods could accurately detect mutations at a VAF of 1%, but consistent and accurate detection below 1% was difficult due to inherent error rates of sequencing platforms 11 . However, a recent report detected NT5C2 variants down to 0.007% VAF in bone marrow 154 days prior to relapse using ultra‐deep sequencing and a novel bioinformatics approach to control for site‐specific sequencing errors 10 . We successfully generated seven of nine ddPCR probes, however only two achieved LODs < 0.1%.…”

“…We were able to most effectively backtrack mutations in genes encoding two proteins with a role in the metabolism of nucleosides, NT5C2 and PRPS1 , for which there are known relapse‐specific hotspot mutations thereby facilitating ddPCR. These mutations have been shown to confer resistance to thiopurines, the backbone of maintenance therapy 9,10,14,15 . The detection of such clones while on therapy indicates a competitive advantage and impending clinical relapse 16 .…”

Mutational signatures: experimental design and analytical framework

“…1b). Cancer cell lines derived from patients with relapsed disease are likely to be even more pathophysiologically awry, with effects on mutational outcome [17]. Such lines will have been exposed to a multitude of natural and iatrogenic insults, may have highly disordered genomes, and been subjected to extensive rounds of selection pressure promoting evolvability within the cell population.…”

“…This could culminate in increased mutagenesis. Counterintuitively, it could also result in reduced mutagenesis if the physiological compensation to overcome selective pressure leads to physiological shifts that tend to suppress DNA damage [17]. The chosen biological model must also be amenable to clonal expansion following single-cell bottlenecking, and here, cancer cell lines tend to fare better than immortalised normal cells.…”