DNA fragility at the KMT2A/ MLL locus: insights from old and new technologies

Abstract

The Mixed-Lineage Leukaemia (MLL/KMT2A) gene is frequently rearranged in childhood and adult acute leukaemia (AL) and in secondary leukaemias occurring after therapy with DNA topoisomerase targeting anti-cancer agents such as etoposide (t-AL). MLL/KMT2A chromosome translocation break sites in AL patients fall within an 8 kb breakpoint cluster region (BCR). Furthermore, MLL/KMT2A break sites in t-AL frequently occur in a much smaller region, or hotspot, towards the 3' end of the BCR, close to the intron 11/exon 12 boundary. These findings have prompted considerable effort to uncover mechanisms behind the apparent fragility of the BCR and particularly the t-AL hotspot. Recent genome-wide analyses have demonstrated etoposide-induced DNA cleavage within the BCR, and it is tempting to conclude that this cleavage explains the distribution of translocation break sites in t-AL. However, the t-AL hotspot and the centre of the observed preferential DNA cleavage are offset by over 250 nucleotides, suggesting additional factors contribute to the distribution of t-AL break sites. We review these recent genomic datasets along with older experimental results, analysis of TOP2 DNA cleavage site preferences and DNA secondary structure features that may lead to break site selection in t-AL MLL/KMT2A translocations.

Keywords: DNA topoisomerase II; KMT2A; MLL; chromosome translocation; etoposide; leukaemia.

Figure 1.

TOP2 strand passage activity and chromosomal translocation mechanism. (a) The TOP2 dimer binds to one DNA duplex (G segment, orange, 1–2) and cleaves both strands in a staggered cut with a 4-bp 5′overhang through which the enzyme protomers remain attached via a 5′-phosphotyrosyl linkage. A second DNA duplex (T segment, grey) passes through the transient enzyme-coupled break (highlighted by a red dashed circle, 2–3) driven by ATP-dependent enzyme conformational changes. The first duplex (G segment) is then re-ligated, and the products of the reaction are released from the enzyme (4–5). TOP2 poisons such as etoposide block the re-ligation step resulting in the accumulation of stabilized TOP2-DNA complexes know as cleavage complexes (CCs) which can be converted to protein-free DSBs with a 4 bp overhang by proteasomal action and TDP2. Repair of these breaks leads to the potential for chromosome translocation. (b) Molecular detail of TOP2 DNA cleavage. TOP2 generates a staggered break via nucleophilic attack by the active site tyrosine on the DNA sugar phosphate backbone of the G-segment. The resulting covalent TOP2-5′phosphotyrosyl–DNA complexes are normally transient and reversible but are stabilized by TOP2 poisons. Cellular processing generates protein free-ligatable breaks that are the substrate for NHEJ.

Figure 2.

Genomic and epigenetic landscape of the MLL BCR. (a) Intron-exon arrangement of the MLL/KMT2A gene, with the position of the major AL-associated BCR highlighted in red. (b) Exon 8 to 14 region of the MLL / KMT2A gene. t-AL break sites, mapped der(11) MLL translocation break sites; all AL break sites, mapped der(11) MLL translocation break sites from t-AL and de novo AL combined. Break site positions are from [,–25]. Genomic data (hg19) are from the following sources. CTCF K562, CTCF RAD21 and DNase CD34⁺, ENCODE [26]; End-seq ETO, Nalm-6 End-seq in the presence of etoposide, GEO GSE99194 [27]; End-seq NT, Nalm-6 End-seq not treated, GEO GSE99194 [27]; sBLISS data GEO121742 [28], hairpin [29]. (c) Enlargement of the t-AL hotspot and flanking region highlighting an inverted repeat coinciding with the hotspot. hg19 coordinate of the 1st nucleotide of the t-AL hotspot is indicated. For (b) and (c) break site counts were binned (binning = 2 nt) and plotted in histogram form.

Figure 3.

Comparison of TOP2 DNA cleavage site preferences and the t-AL hotspot sequence. (a) Illustration of a TOP2CC highlighting the 4 bp DNA overhang. Numbering is aligned with the numbering and sequences in part (b). Cleavage occurs between base −1 and +1 on both the Watson and Crick strands. (b) Base composition preferences or TOP2 DNA cleavage. Cleavage sites and base composition preferences are derived from the following sources: TOP2 CC-seq (Gittens et al.) [56], Ju et al. [57], Marsh et al. [58], Cornarotti et al. [59]. −1 and +5 positions are highlighted in red for clarity. Upper case lettering indicates over-representation at a given position, bold for more prominent overrepresentation; lower case indicates under-representation or absence at the given position. Dash represents no identified over/under representation. For TOP2 CC-seq the base preferences are derived from the average base composition of cognate TOP2 CC sites genome-wide in etoposide treated RPE1 cells [56]. (c) Partial alignment of the t-AL hotspot (shaded pink) and its adjacent 5′ flanking sequence (shaded grey) with cleavage site base composition preferences in (b), including the position of the inverted repeat (horizontal arrows). Black vertical arrows, in vitro etoposide-induced cleavage sites identified in [22]; blue arrows, in cell etoposide-induced cleavage sites identified in [36]; red arrows, apoptosis-associated DNA breaks identified in [36].

Figure 4.

Model for reciprocal MLL chromosome translocation upon exposure to TOP2 poisons. Chr11 segment (orange line) containing the MLL/KMT2A gene (thick section) depicted with CTCF/cohesin-mediated looping between the exon 12 CTCF site and an upstream region (CTCF depicted as green ovals, and cohesin as a purple ring). Elongating RNA pol II (dark grey circle) progressing towards the topological constraint of the loop base generates positive superhelical tension upstream of the exon 12 CTCF site that requires topoisomerase activity (blue) for resolution (a). In the presence of a TOP2 poison, normally transient TOP2-DNA complexes associated with strand passage activity are stabilized, and subsequently processed into protein-free DSBs (PFB, red arrow) (b), which in turn can lead to chromosome rearrangements via erroneous DNA repair (c). We suggest that the observed clustering of t-AL translocation break sites in the 11-bp hotspot (bounded by dark bars) results from a combination of DNA sequence permissive to TOP2 cleavage, DNA secondary structure features and factors such as nucleosome positioning, along with the constraint of generating a rearrangement favourable for AL development (a–c). In this model, balanced chromosome translocation is facilitated by the proximity of a translocation partner gene undergoing transcription in the same RNA polymerase cluster/transcription factory (large light grey circle) as MLL/KMT2A. In parallel, etoposide can also lead to DSBs associated with TOP2B at the exon 12 CTCF site (d), accounting for the peak of etoposide-induced breaks detected by End-seq and BLISS associated with this region (green arrow) (e). However, t-AL associated MLL/KMT2A translocations involving break sites at this location (f) are not frequently observed, presumably as they are less likely to be favourable for AL development.