DNA repair genes are selectively mutated in diffuse large B cell lymphomas

Abstract

DNA repair mechanisms are fundamental for B cell development, which relies on the somatic diversification of the immunoglobulin genes by V(D)J recombination, somatic hypermutation, and class switch recombination. Their failure is postulated to promote genomic instability and malignant transformation in B cells. By performing targeted sequencing of 73 key DNA repair genes in 29 B cell lymphoma samples, somatic and germline mutations were identified in various DNA repair pathways, mainly in diffuse large B cell lymphomas (DLBCLs). Mutations in mismatch repair genes (EXO1, MSH2, and MSH6) were associated with microsatellite instability, increased number of somatic insertions/deletions, and altered mutation signatures in tumors. Somatic mutations in nonhomologous end-joining (NHEJ) genes (DCLRE1C/ARTEMIS, PRKDC/DNA-PKcs, XRCC5/KU80, and XRCC6/KU70) were identified in four DLBCL tumors and cytogenetic analyses revealed that translocations involving the immunoglobulin-heavy chain locus occurred exclusively in NHEJ-mutated samples. The novel mutation targets, CHEK2 and PARP1, were further screened in expanded DLBCL cohorts, and somatic as well as novel and rare germline mutations were identified in 8 and 5% of analyzed tumors, respectively. By correlating defects in a subset of DNA damage response and repair genes with genomic instability events in tumors, we propose that these genes play a role in DLBCL lymphomagenesis.

Figure 1.

DNA repair pathways affected by somatic and novel germline mutations in DLBCL. Somatic and novel germline mutations together were most frequent in MMR genes, followed by general DDR genes and NHEJ genes. To a lower extent, mutations were detected in FA, HR, BER, and NER pathways. The “Other” category refers to the p.R335W mutation in the DNTT gene, which encodes for a protein involved in the addition of nucleotides at junctions of rearranged IGH and TCR gene segments during V(D)J recombination. The percentage of DLBCL tumors affected by somatic or novel germline mutations in different DNA repair pathways is displayed. Mutations were discovered by SOLiD and validated by Sanger sequencing.

Figure 2.

Allelic imbalance at PARP1, EXO1, MDC1, and RPA1 loci in DLBCL samples and association with PARP1 expression. (A) Allelic imbalance events detected by SOLiD were validated by the use of polymorphic markers and fragment analysis in 22 DLBCL samples. Six additional samples that did not undergo SOLiD sequencing were also investigated. Some markers were not informative, because of either germline homozygosity or MSI in the tumor samples. Markers displaying allelic imbalance were analyzed in two independent experiments. (B) Expression of PARP1 and EXO1 was determined by quantitative real-time PCR in samples with and without allelic imbalance at the respective loci. Medians are depicted and the statistical difference between groups was assessed with the Mann-Whitney test (**, P < 0.01). Two independent experiments were performed and a representative experiment is shown.

Figure 3.

MSI and mutation profiles in MMR-mutated DLBCL. (A) Fragment analysis was performed with microsatellite markers in 28 DLBCL tumors. (B) MSI detection in five DLBCL cases with mutations in MMR genes. MSI detection was confirmed by one independent experiment. (C) The median of the total number of somatic mutations in the coding genome (exome) was higher in MMR-mutated DLBCL tumors that displayed MSI than in MSS tumors. (D) This difference was significant when looking specifically at the number of somatic indels detected by exome sequencing. (E and F) C:G→A:T transversions were significantly enriched in MMR-mutated DLBCL samples displaying MSI. The statistical difference between groups was assessed with the Mann-Whitney test (*, P < 0.05).

Figure 4.

FISH analysis in DLBCL tumors. The integrity of the IGH and BCL6 loci was assessed by FISH in 13 DLBCL tumors that underwent targeted sequencing of DDR and repair genes. Results from DL43 are depicted. (A) Rearrangement of one IGH locus was detected in two DLBCL cases (DL43 and DL48), both of which carried a mutation in a NHEJ gene (green, IGHV probe; red, IGH 3′ flanking probe). (B) Rearrangement of one BCL6 locus was detected in three tumors, including the two cases that displayed disruption of the IGH locus (green, BCL6 3′ flanking probe; red, BCL6 5′ flanking probe). (C and D) Both cases that displayed concurrent rearrangement of the IGH and BCL6 loci were investigated with an IGH/BCL2 dual fusion translocation probe and a MYC break-apart probe to determine the translocation partner of IGH. No fusion signal between IGH and BCL2 probes (C) or rearrangement of the MYC locus (D) was observed, thus supporting BCL6 as the most likely translocation partner of IGH in those tumors. The three signals detected by the IGH probe (green) of the IGH/BCL2 probe set (C) are consistent with the IGH rearrangement observed using the break-apart probe. Within each experiment, at least 100 intact, nonoverlapping nuclei were scored for each probe, and representative results are depicted.