Acute lymphoblastic leukemia (ALL) arising in infants less than 1 year of age is characterized by genetic rearrangements of the KMT2A gene (previously MLL) and an exceedingly poor prognosis. We have previously shown that infant KMT2A -rearranged (KMT2A -R) ALL has one of the lowest numbers of somatic mutations of any sequenced cancer with a mean of only 1.3 non-silent mutations being present in all leukemia cells per patient (Andersson et al., Nat Genet, 2017). Despite the paucity of mutations, activating mutations within the PI3K/RAS signaling pathway were present in about half of cases, most of which were subclonal with a mutant allele frequency (MAF) <0.30. In addition, some patients harbored multiple activating mutations at varying MAFs, suggesting the presence of multiple low-frequency leukemia clones. Based on these findings, we hypothesized that the genetic landscape of infant KMT2A -R ALL is more clonally heterogeneous than other pediatric leukemias, which might contribute to its poor prognosis as current treatment regimens may not eliminate all clones.

To test our hypothesis, high coverage bulk and single-cell genomics were used to study the mutation profiles of subclones in diagnostic samples from four infants younger than 6 months of age with KMT2A -R ALL. This included patients with t(4;11)(q21;q23) KMT2A-AFF1 (n=3) of which two relapsed and a patient with t(11;19)(q23;p13.3) KMT2A - MLLT1 (n=1) that remains in remission. DNA was isolated from bulk leukemia cells and high-depth paired-end whole-genome sequencing was performed on paired diagnostic and remission samples at an average haploid coverage of 160x and 35x, respectively. In addition, whole-exome sequencing was performed on the diagnostic samples at an average exon coverage of 375x. Somatic alterations, including single-nucleotide variations (SNVs), insertions-deletions (indels), structural variations (SVs), and copy number alterations, were detected using multiple analytical pipelines. In parallel, approximately 75 viable single cells from each diagnostic sample were isolated and the DNA from the cells underwent whole genome amplification using the Fluidigm C1 system.

Analyses of the whole-genome sequencing data from the four infant cases revealed an average of 715 putative sequence variants (range of 54-1199) per case across the genome. Using the combined data from whole-genome- and exome sequencing, an average of 55 (range 51-65) non-silent mutations affecting coding genes or non-coding RNAs were identified, with 83% having support by both sequencing methods, validating their presence. An average of 63% and 92% of the non-silent mutations detected by whole-genome and exome sequencing, respectively, were subclonal, demonstrating the enhanced ability to detect low frequency variants with higher sequencing depth.

Subclonal mutations in genes within the PI3K/RAS signaling pathway were detected in 2/4 cases with each patient harboring both a KRASG12D and a FLT3V491L as well as an additional NRASG13D for one of the cases. The recurrent FLT3V491L mutation is present in COSMIC and has been previously detected in leukemia, but to the best of our knowledge, it has not been experimentally verified to be activating.

We are now performing high-throughput amplicon-based sequencing at the single cells from those patients using 150 targets/cell to determine the number of distinct clones into which the bulk mutations segregate. The targets include a mixture of coding, intergenic variants and SVs. In addition, we are determining if each cell contains a mutation at a selected number of loci that are leukemia-associated mutational hotspots. After acquiring the data, we will reconstruct the clonal architecture of the samples, enabling us to assess the clonal heterogeneity and mutational histories of the leukemias.

Taken together, these data will provide a unique look into the underlying biology of infant KMT2A -R ALL by providing new details about the population genetic diversity and temporal genetic changes that occur during KMT2A -R infant leukemogenesis. These data may also provide insight into the universally poor prognosis of patients with KMT2A -R infant ALL.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution