Targeted Next-Generation Sequencing and Genome-Wide High-Resolution Copy Number DNA Arrays Allow the Identification of Five Novel RUNX1 Fusions In Hematological Malignancies.

Abstract 1193

RUNX1 is a crucial transcription factor involved in cell lineage differentiation during hematopoiesis. It contains a “Runt homology domain” (RHD; exons 3–5, amino acids 50–177) and a transactivation domain (TAD; exon 8, amino acids 291–371). RUNX1 can act as an activator or repressor of target gene expression and thus far two different mechanisms of somatically acquired alterations have been recognized: intragenic mutations and translocations. Most of the translocations involving RUNX1 lead to the formation of a fusion gene consisting of the 5` part of RUNX1 fused to sequences on partner chromosomes. We here present data on 5 cases, 4 acute myeloid leukemias (AML) and 1 chronic myelomonocytic leukemia (CMML) patient, respectively, where previous cytogenetic and FISH analyses revealed reciprocal translocations involving RUNX1. However, even sophisticated molecular diagnostic work-up failed to identify the corresponding RUNX1 fusion partners. Therefore, we used a combination of 454 shotgun pyrosequencing and long-oligonucleotide sequence capture microarrays to reveal these unknown RUNX1 partner genes in four cases. In detail, we performed DNA sequence enrichment using microarrays containing capture probes that were covering a contiguous region on chr. 21 (36,160,098 – 36,421,641), thereby allowing a specific enrichment by hybridization for genomic DNA where the RUNX1 gene is located (Roche NimbleGen 385K chip, Penzberg, Germany). This targeted next-generation sequencing (NGS) assay enabled to capture and sequence single reads mapping to both RUNX1 and other genomic regions (Burrows-Wheeler Aligner's Smith-Waterman algorithm). In median, 324 bp per patient (170,000 reads) with an 18-fold coverage were sequenced and in all cases chimeric reads were detectable, thereby confirming the presence of RUNX1 translocations and, moreover, identifying and characterizing 4 novel fusions on a molecular level. In one AML case, KCNMA1 was fused to RUNX1. KCNMA1, a potassium large conductance calcium-activated channel family member on chromosome 10q22.3, had recently been described to play a role in breast cancer invasion and metastasis to brain. In our case, as confirmed by RT-PCR and Sanger sequencing, the chimeric RUNX1-KCNMA1 fusion led to the disruption of the RHD of RUNX1. In the three additional cases, RUNX1 was fused to genomic regions on chromosomes 10q22, 17q21, and 5q13.3, respectively. The RUNX1-10q22 and the reciprocal 10q22-RUNX1 fusion were confirmed by PCR from genomic DNA and subsequent Sanger sequencing. According to its genomic structure the translocation RUNX1-chr.10q22 will result into the translation of a truncated RUNX1 protein with an intact RHD, but without TAD. Notably, in the remaining two cases, chr.17q21-RUNX1 and chr.5q13.3-RUNX1, only the reciprocal fusion events were detectable by PCR. In case chr.17q21-RUNX1 the translocation would disrupt RUNX1 after the RHD. In chr.5q13.3-RUNX1 the predicted fusion would not impact the RHD and TAD domains because the breakpoint is located before exon 1. In the fifth patient, we performed an analysis using a high-resolution genome-wide cytogenetic copy number DNA microarray to resolve a novel t(X;21)(p11;q22). In this case, the derivative chromosome × was duplicated, leading to a partial trisomy 21q and a partial trisomy X. On chr. 21 the breakpoint was mapped to be located in intron 6–7 within the RUNX1 gene. The breakpoint on the X-chromosome mapped to Xp11.23, thus leading to a truncated RUNX1 protein without the TAD domain. In summary, RUNX1 rearrangements either led to RUNX1 with an intact RHD and TAD (n=1), RUNX1 with an intact RHD but without TAD (n=3, dominant negative effect; similar to RUNX1-RUNX1T1), or to RUNX1 with a disrupted RHD and without TAD domains, leading to haploinsufficiency (n=1). In conclusion, the RUNX1 recombinome is an interesting target to understand pathogenetic heterogeneity in hematological malignancies. Here, we demonstrated that NGS and copy number DNA microarrays allow the identification of novel RUNX1 fusion partners not detectable by standard molecular techniques and reveals that cytogenetic reciprocal translocations lead to different types of RUNX1 alterations.