Promoter Shuffling by Sequential Genomic Rearrangements in Follicular Lymphoma Reveals an Ongoing Genomic Instability at the BCL6 Locus.

Abstract

BCL6 translocations are frequent in non-Hodgkin B-cell lymphoma. They are usually considered as primary abnormalities associated with Diffuse large B-cell lymphoma (DLBCL), and would be a consequence of the instability of the first intron of the gene where somatic mutations accumulate during normal B-cell maturation. Here, we describe three FL with complex BCL6 rearrangements and demonstrate that this gene can still remain unstable after a first translocation. Two cases were identified as part of our continuing assessment of t(3;14) translocations in lymphoma. The first presented with grade 3A FL. Reciprocal BCL6-Cg and Ig3-BCL6 fusion transcripts expressed from the der(14) and der(3) respectively were detected by RT-PCR. Cloning of both BCL6(intron1)-Sg3 and Sg3-BCL6(intron1) genomic breakpoints indicated a balanced translocation on both loci. More surprisingly, real time RT-PCR experiments also showed a low expression of Iμ-BCL6 transcripts, evoking a second t(3;14) involving the Sμ. Indeed, we could clone an additional and complex Sμ-Sg3-BCL6 junction. Notably, the Sg3-BCL6 breakpoint on this allele was identical to the one we had previously identified, to the notable exception of additional point mutations. These data suggest that two subclones were present in the tumor, one with the original t(3;14) and the other with an evolution of the der(3)t(3;14) where all DNA between the Sμ and Sg3 was deleted. The second patient presented with grade 3A FL. BCL6-Cg transcripts were identified and the corresponding der(14) junction was cloned. However, no reciprocal fusion transcript nor breakpoint were obtained. 5′Race experiments indicated that, instead of the expected IGH-BCL6 transcripts, the tumor expressed EIF4A2-BCL6 ones. Notably, the EIF4A2 gene is located 1Mb downstream from BCL6, in a tail to tail orientation. Cloning of Sg3-EIF4A2(intron1) and complex EIF4A2(intron1)-Sg3-BCL6(intron1) genomic junctions showed that a secondary paracentric inversion involving BCL6 and EIF4A2 had taken place on the der(3)t(3;14). The third patient was diagnosed with FL in 1993 and presented a DLBCL at first relapse in July 2002. Cytogenetic analysis revealed a t(14;18) and a t(3;7)(q27;q31). FISH study indicated that the t(3;7) breakpoint was in the distal 5′ alternative breakpoint region of BCL6. A new biopsy performed in June 2003 confirmed the transformation. Cytogenetic analysis again revealed both t(14;18) and t(3;7). However, EIF4A2-BCL6 and reciprocal BCL6-EIF4A2 fusion transcripts were detected, and molecular cloning of both genomic junctions allowed to verify the paracentric inversion. Retrospective investigation of the first biopsies by RT and genomic PCR confirmed the t(14;18) but excluded the inversion, which is therefore a secondary event. These complex rearrangements unambiguously show that after a first translocation, and even if the transformed phenotype is already gained, a dynamic process can still remodel the BCL6 locus. These cells thus remained in, or at least could still reach, a differentiation stage where an ongoing BCL6 instability led to a second promoter substitution. Of note, the sequential rearrangements, that took place on the same BCL6 allele for at least two cases, were not detectable by conventional cytogenetic or FISH and were only identified by extensive molecular analysis. Such acquisitions of cryptic secondary BCL6 abnormalities in FL are thus probably underestimated, and their incidence in progression deserves to be investigated.