LAUSR

TO THE EDITOR Single base substitutions (SBSs) and insertions/deletions (indels; IDs) arise through several mechanisms such as errors during DNA replication/repair and exposures to mutagens, with the different mutational processes occasionally generating specific mutational signatures. SBS signatures (SBSsigns) result from recurring trinucleotide patterns of the transition/transversion types of somatic single nucleotide variants (SNVs) and their flanking nucleotides, whereas ID signatures (IDsigns) are defined according to size, nucleotides affected, and the presence of repetitive/ microhomology regions (https://cancer.sanger.ac.uk/signatures/). Some signatures are associated with underlying etiologic factors, e.g. SBS7 and ID13 in UV-associated melanoma and SBS4 and ID3 in smoking-induced lung cancer [1, 2], whereas others are linked to inherent defects of DNA recombination, replication, and repair (SBS6 and ID1) or caused by spontaneous or enzymatic deamination of 5-methylcytosine to thymine (SBS1) (https://cancer.sanger. ac.uk/signatures/). Based on whole genome sequencing (WGS) of pediatric acute myeloid leukemia (AML), we recently reported that SBS18 – a signature characterized by frequent C > A transversions – is enriched in t(8;21)(q22;q22)/RUNX1::RUNX1T1-positive cases [3] (gene fusion designation according to recent guidelines [4]). Brandsma et al. [5] subsequently also showed that SBS18 is common in childhood AML, including cases with RUNX1::RUNX1T1. Considering that SBS18 has been associated with DNA damage caused by reactive oxygen species (ROS) (https://cancer.sanger.ac. uk/signatures/sbs/sbs18/) and that the RUNX1::RUNX1T1 chimeric protein is known to downregulate the expression of the OGG1 gene encoding a DNA glycosylase that excises oxidized guanines [6], we hypothesized that ROS could be involved in the genesis of childhood AML with RUNX1::RUNX1T1 [3]. Whether SBS18 is overrepresented also in adult RUNX1:: RUNX1T1-positive AML is unknown. In fact, our knowledge of SBSsigns is rudimentary—and non-existing as regards IDsigns—in adult core binding factor (CBF) AML, which consists of cases positive for either RUNX1::RUNX1T1 or CBFB::MYH11 [inv(16) (p13q22)/t(16;16)(p13;q22)] [7]. The only publication to date addressing SBSsigns in adult CBF AML reported a high frequency of SBS1 [8], a clock-like signature that accumulates with age (https://cancer.sanger.ac.uk/signatures/). To ascertain if SBS18 is a common mutational signature in adult CBF AML, we performed WGS of ten cases with RUNX1::RUNX1T1 and ten with CBFB::MYH11, focusing not only on SBSsigns but also on IDsigns. All patients had de novo AML, thus excluding those previously exposed to chemoand/or radiotherapy that could have affected the mutational signatures. The cases were selected based on the availability of good quality DNA from both diagnosis and remission. The median age of the patients was 51.5 years (range 19-74 years) and the female/male ratio was 1:1.5. All genetic analyses were performed at the Department of Clinical Genetics and Pathology, Office for Medical Services, Region Skåne, Lund, Sweden. The basic clinical and genetic features of the CBF AMLs are summarized in Supplementary Table 1 and data on WGS of paired diagnostic/ remission samples and bioinformatic analyses are provided in Supplementary Information. The average sequencing depths of the WGS varied from 29× to 57× per sample (median 40×) and the Q30 value was 96.23%, with 2 × 150 bp read length. The WGS analyses confirmed the RUNX1:: RUNX1T1 and CBFB::MYH11 gene fusions in all cases and also revealed that the genomic breaks clustered within introns 6 of RUNX1 and 1 of RUNX1T1 and within introns 5 of CBFB and 33 of MYH11, respectively (Supplementary Table 1). No other chimeric genes were detected. All chromosomal gains and losses previously found by conventional G-banding were identified by WGS except for two subclonal trisomies in one case (Supplementary Tables 1 and 2). WGS also identified 32 copy number abnormalities (≤10 Mb) and five uniparental isodisomies, all of which undetectable by chromosome banding analyses (Supplementary Table 2). None of the cases displayed any signs of chromothripsis. A median of 1437 (range 22–1834) and 1049 (561–1369) SNVs was identified in the RUNX1::RUNX1T1and CBFB::MYH11-positive cases, respectively, corresponding to 0.01-0.61 SNVs/indels per Mb. Comparing the transition and transversion types between the two gene fusion groups revealed highly similar frequency distributions, except for a slight excess of C > T transitions in the cases with CBFB::MYH11 (49% versus 47%; P= 0.012; Mann–Whitney U test; Supplementary Fig. 1). Overall, the most