7 General discussion | 171 chromosome 6 at the Ig loci, which we used as a positive control. In our study we did not identify RAG1/2-dependent NBS1 binding sites at the genes or breakpoints frequently seen translocated or deleted in B-ALL patients, such as ETV6/RUNX1, BCR/ABL1 or E2A, early B-cell factor 1 (EBF1), lymphoid enhancer-binding factor-1 (LEF1), IKAROS family zinc finger 1 (IKZF1) and IKAROS family zinc finger 3 (IKZF3)11–13. We hypothesize that differences between murine and human chromatin configuration might account for this observation. In addition, the employed cell lines already harbor a transforming element (i.e. v-Abl), which could change the accessibility of genes. Though the most “obvious suspects” were not implicated in our study, we have detected RAG1/2-dependent DNA breaks in the proximity of several other genes previously associated with B-cell malignancies in mouse and/ or human. For instance, in the proximity of Klf4, previously shown to be a putative tumor suppressor in pre-B cells and B-cell malignancies14, in the proximity of Abl2 which, next to the frequently rearranged ABL1, has also been reported to be rearranged in some cases of high-risk B-ALL15; or in the proximity of MADS box transcription enhancer factor 2 (Mef2c). The enforced overexpression of MEF2C resulted in lymphoid tumors co-expressing CD3 and CD19, resembling the human phenotype of mixed ALL16. The group of David Schatz (Yale University, New Haven, CT) has previously demonstrated that RAG1 and RAG2 can bind DNA throughout the genome and suggested that their DNA cleaving activity may not be restricted to canonical RSS sequences. In one of their studies, RAG1 was found to bind to approximately 3400 RAG1 binding sites in mouse pre-B cells, and around 18300 RAG2 binding sites throughout the genome of mouse pre-B cells. The binding of RAG1 and RAG2 however, did not correlate with the presence of spontaneous RAG-dependent translocations, concluding that in vivo such events must be rare17. When we compared our data with the publicly available data from the aforementioned study, we found only around 3% overlap between the RAG1 binding sites, as identified by the Schatz lab, and our RAG1/2-dependent NBS1 binding sites. This rather low overlap can be attributed to the use of different cell systems– primary mouse pre-B cells as compared to the v-Abl-transformed pre-B cells used in our study. Also, it is important to emphasize that the study of the Schatz lab mapped RAG1 and RAG2 binding to DNA, while our study mapped the DSBs resulting from RAG1/2 activity on the genome-wide scale. It is unknown to what extent the binding of RAG1 and RAG2 to chromatin results in the formation of DSBs. The overlap between the RAG1 or RAG2 binding profile and the RAG-dependent NBS1 binding profile within one cell system has not been studied yet, opening up possibilities for follow-up investigations of actual the extent, to which the binding of RAG to DNA effectively results in DSBs. Next, we examined the NBS1 binding pattern in more detail and aimed to identify common denominators of the RAG-mediated NBS1 binding sites. What makes these sequences RAG (off) targets? In the mouse v-Abl pre-B cell line used in our study, v-Abl transformation was shown to arrest the developing B cells at the pro-to-pre B-cell stage,
RkJQdWJsaXNoZXIy MTk4NDMw