We show that introns harboring translocation breakpoints in tumors are significantly longer than non-translocated introns of the same genes, but are not significantly enriched in sequence elements potentially involved in chromosomal rearrangements. Our findings provide evidence that double-strand breaks, the type of DNA damage that leads to translocations in tumors, are created at random in the genome, and that sequence elements do not play a general role in the localization of the breaks.
Translocation breakpoints in cancerChromosome aberrations are a common feature of cancer initiation and progression. Several types of aberrations are found in tumors, reciprocal translocations being one of the best characterized types of genetic alteration found in known cancer genes [1]. It is widely accepted that chromosome rearrangements in cancer are mediated by the repair of DNA double-strand breaks (DSB) through non-homologous end joining (NHEJ), which seems to be the major repair pathway in human somatic cells [2][3][4].Analysis of some translocations, especially in hematological malignancies, revealed that the breakpoints are some times distributed in a non-random fashion, either densely clustered in particular regions of some genes, or more diffusely clustered along one or a few specific introns in other genes [5]. This has led to suggestions that the presence of local chromatin features or specific sequence motifs can determine genomic domains that are particularly prone to sustain a DSB in the vicinity. In line with this view, various sequence elements have been reported to be associated with chromosomal rearrangements in tumors (reviewed in [6]). However, since clustering of breakpoints is not observed in most translocations identified to date, there is considerable debate as to whether the presence of DNA domains with a high risk of sustaining a DSB is important only in a few specific translocations, or it is a general feature common to all translocations in human tumors. The lack of a significant number of breakpoints cloned at the genomic level has delayed progress in this area, because the precise localization of the breakpoints and the sequence surrounding them is known only in a limited number of cases.
2In order to address this issue we have performed computational studies that enabled us to precisely locate a large number of cancer translocation breakpoints on the human genome reference sequence, and to analyze the genes involved with a view to unveil any potential factors accounting for the localization of breakpoints in human tumors.
Features of the genes involved in reciprocal translocations in tumorsWe collected 268 genes whose involvement in reciprocal translocations in various types of tumors has been reported in the literature, and extracted their sequence and annotations from the Ensembl database [7] using perl scripts and the Application Programming Interface (API) provided by the Ensembl project [8]. We compared various sequence features between these genes and a group of 9,406 genes tha...