Up to 1/5 of newly diagnosed CLL transform to an aggressive RS-DLBCL. The mechanisms underlying the event are very poorly defined. RS-DLBCL is often clonally related to CLL, as shown by the identity of their IgH rearrangement. Deletion of 13q14.3 is the most common lesion in unselected CLL. Notably, CLL cases showing no chromosome 13q14.3 deletion at diagnosis face a higher risk of transformation, while CLL samples with 13q14.3 deletion do not preserve the 13q deletion in the subclone transformed into RSDLBCL. With the aim of understanding the molecular mechanisms of the transformation from CLL to RS-DLBCL, we have performed a genome wide-DNA profiling. Samples derived from 12 patients were analyzed: 13 samples of RS-DLBCL and 8 matched CLL. In one patient, 2 separate RS-DLBCL sites were analyzed. For comparative purposes, 48 de novo DLBCL cases were also investigated. Tumor samples were analyzed with Affymetrix Human Mapping 250K SNP arrays, starting from frozen biopsies. IgH sequencing and multiple FISH analyzes were also performed. Samples were obtained in the course of routine diagnostics. RS-CLL samples were collected at the time of first diagnosis. The fraction of malignant cells in the pathologic specimen was >70%, as determined by morphologic and immunophenotypic studies. Chromosome 13 gains represented the only recurrent change in 2/8 clonally related CLL and DLBCL samples. RS-DLBCL presented recurrent gains at 1q44, 2p, 3q28, 4q35, 5p15-pter, 7q11-q21, 8q13-qter, 9q22, 13q14-q33, 14q32, 15q21-qter, 18q21-qter, 20q13-q13, 21q22. The whole chromosome 12 was gained in two cases and interstitial gains were observed on both arms of chromosome 7. Recurrent losses affected 1p36, 1q32, 3q28, 4p16, 4q24, 6q25, 7q11, 7q31-qter, 8p23, 8p11, 8q24, 9p21.3, 10q11, 10q11, 10q22, 11q13-q14, 13q14,13q34, 14q11, 15q14-q15, 17p11-pter, 19p13, 20q11. Recurrent regions of copy neutral LOH affected 1p21, 3q26, 6p21, 11q22 and 12q23. At least in one patient, the 1p, the 3q and the 12q regions of copy neutral LOH were already present in the germline DNA, excluding their somatic origin. In CLL, recurrent gains were observed in 2p25, 3q26, 8q21 and 8q22-q24, 9p24, the whole chromosome 12, 14q32 and 15q26. Losses affected 7q32 and 7q36, 8p23, 11q14-q23, 13q14, 14q24-q24, 15q14-q15, and 17p13-pter. Chromosome 13q showed a complex and heterogenous pattern of deletions and gains, including different breakpoints in matched RS-CLL and DLBCL samples, and gains taking place at the deletion breakpoints. Also, a case bearing a deletion in the CLL but not in the RS-phase, presented a heterozygosity of chromosome 13 in the DLBCL, indicating that the two chromosomes were not derived from the normal chromosome still present in the CLL. The comparison between de novo and RS-DLBCL showed differences. RS-DLBCL did not carry 6q deletions and gains of 3p and 19q. In conclusion, RS-DLBCL presented heterogenous genomic lesions, partially different from de novo DLBCL. Despite a common post-GCB/ABC morphology, the genomic profile of RS-DLBCL appeared intermediate between that reported for ABC- and GCB-DLBCL: RS-DLBCL carried only some of the lesions typically reported in ABC-DLBCL, since they lacked 6q (TNFAIP3 locus) deletions but carried 18q (BLC2 locus) gain. Despite being clonally related to the pre-existing CLL, the analyzed matched samples of RS-DLBCL and CLL showed complex lesions suggesting that a direct transformation from CLL to DLBCL might not always be the case.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author