Interstitial del(14)(q) involving IGH : a novel recurrent aberration in B-NHL

Deletions of chromosome 14 recurrently observed in B-cell malignancies are particularly frequent in chronic lymphocytic leukemia (CLL), multiple myeloma (MM), diffuse large B-cell lymphoma (DLBCL) and B-cell acute lymphoblastic leukemia.1 To characterize these aberrations, we initially mapped the deletions in 23 leukemia/lymphoma cases using a tiling path chromosome 14-array CGH (aCGH) platform (resolution of approximately 42 kb) comprising 838 bacterial artificial chromosome/P1 artificial chromosome (BAC/PAC) clones from the Chori 32K set (www.ensembl.org). Del(14)(q) have been successfully mapped in all 23 cases (Figure 1a) and subdivided into two main categories: (1) involving or (2) not involving 14q32.33. This latter category grouped eight cases with interstitial deletions ranging in size from 17 to 70 Mb and distributed along chromosome 14 (q13 ter) and one case with three dispersed deletions of 1.5, 7.5 and 2.5 Mb. The first category of del(14)(q) comprised 14 cases showing the common distal breakpoint mapped in the 105.26–105.41 Mb region harboring the IGH genes cluster. The size of the deleted region varied in six cases; their proximal breakpoints were bordered by RP11-164H3/95.19 Mb, CTD-2053J06/83.56 Mb, RP11-340F04/74.66 Mb, RP11-350H11/67.94 Mb, RP11-769O09/66 Mb and RP11-520H13/65 Mb. Particularly intriguing was the finding of exactly the same del(14)(q24.1q32.33) covering the region of approximately 36 Mb in eight cases (Figure 1b). The proximal border of this deletion was consistently flanked by RP11-35D12 (68.24 Mb)( 2) and RP11-720I19 (68.45 Mb)( 1), and the distal breakpoint was bordered by RP11-448N05 (105.26 Mb)( 1) and RP11-284A08 (105.43 Mb)( 2). Given that these terminal BACs are assigned to IGCH and IGVH, respectively, aCGH results were additionally validated by fluorescence in situ hybridization (FISH) with the LSI-IGH break-apart probe. As expected, all eight analyzed cases (as well as the remaining six cases with nonrecurrent IGH-involving deletions) showed a one-fusion one-green signal pattern due to loss of the 3'IGH flanking sequences (red signal). The proximal breakpoint of the del(14)(q24.1q32.33) validated with RP11-35D12 and RP11-720I19 showed lost or diminished signal from the latter clone (Figure 1b) and thereby pointed to a breakpoint in the region covered by the overlapping extremities of these BAC clones. The only gene located in this region was ZFP36L1. For FISH detection of the del(14)(q24.1q32.33) in other B-cell non-Hodgkin's lymphoma (B-NHL), a dual-color break-apart assay for the 14q24.1 breakpoint (RP11-35D12-SO/RP11-720I19-SG) and LSI-IGH were applied. The screening of 58 B-NHL cases with cytogenetic and/or FISH evidence of del(14)(q) led to identification of 13 additional cases with del(14)(q24.1q32.33), including one with the deletion masked by t(14;22)(q32;q11)/IGH-IGL (case 7). Additional FISH screening of 60 random CLL cases, seven MM cell lines, including four (RPMI-8226, L-363, OPM-1 and LP-1) with a previously described del(14),2 and 20 various B-NHL cases with structural aberrations of 14q21–q24, failed to identify ZFP36L1 and/or IGH rearrangements, indicating that these deletions are rare molecular events. The relevant clinical data of 21 cases with the del(14)(q24.1q32.33) (index cases) are summarized in Table 1. There were 14 male and 7 female patients (2.0 M/F ratio) ranging in age from 52 to 90 years (mean 71.6). These cases represented a spectrum of B-cell malignancies including typical CLL (n=8), atypical CLL showing either morphological or immunophenotypic discordances (for example, expressing CD22 (n=4) and/or FMC7 (n=2), and/or lacking expression of CD23 (n=2) or CD5 (n=1)) (n=7), low-grade B-NHL (marginal zone lymphoma (MZL), Waldenstrom macroglobulinemia (WM), not otherwise specified NHL (NOS-NHL)) (n=3) and MM (n=3). Case 17, diagnosed as MZL, transformed to a fatal DLBCL 2 years after diagnosis. Among CLL patients, five presented in Binet stage A, two in stage B and five in stage C (no available staging data in two cases). The majority of patients (18/21) required therapy and when given (alkylators), they responded initially (data not shown). Nine patients are alive after 4–88 months (median 35) from diagnosis. Twelve patients died (0.75–118 months, median 21 months) after diagnosis, either from progressive disease (n=5), infection (n=4), both disease and infection (n=1) or from unrelated reasons (other neoplastic disease, n=2). (Cyto)genetic features of index cases are summarized in Table 1. All but two cases (88%) showed a simple karyotype with 1–3 chromosomal changes per case (an average of 1.2). In five cases (30%), the del(14) was the sole abnormality. The most frequent chromosomal change associated with the del(14) was trisomy 12, evident in eight cases (47%) including five in which it appeared as the only additional change. IGVH mutation analysis performed in 15 index cases (Table 1) showed lack of somatic mutations (98–100% homology to germline VH genes) in 10 cases, somatic hypermutations (94.7–91.5% homology to germline VH genes) in 3 cases and presence of unmutated (100% each) as well as mutated (96.7 and 95.2%, respectively) VH sequences in 2 cases (nos. 18 and 19). In both these cases and in case 3 (unmutated CLL (umCLL)), usage of two different VH segments was identified. Altogether, 15 different VH segments were used, including 3–30 detected in 3 cases (including cases 3 and 18 with biclonal rearrangements) and 2–70 detected in 2 cases (cases 18 and 19, both with biclonal rearrangements). For a more precise mapping of the del(14)(q24.1q32.33) breakpoints, we used FISH with a pair of fosmid clones flanking ZFP36L1 (WI2-1846M8; WI2-1710N10) and two IGCH cosmid clones, cos3/64 covering JH, C and C3 and cos1 harboring S1/C1 sequences4 (Figure 1c). Moreover, we applied genomic qPCR with a set of primers selected for the following sequences: A and D, flanking ZFP36L1; B and C, coding ZFP36L1; E, representing IGHA2; and F and G flanking the 5'enhancer (E) of IGCH (Table 2) (Supplementary Table 1). Results of this compiled FISH and qPCR analysis are shown in Table 2. Briefly, the 14q24.1 breakpoints were mapped to the region proximal to ZFP36L1 (Ax2, B/Cx1) in three cases, within the gene (A/B-x2, C/D-x1) in six cases and distal to ZFP36L1 (A/B/Cx2, Dx1) in three cases. The 14q32.33 breakpoints were shown to be centromeric to sequences flanking E (Fx2/Gx2) in 10 cases, occurred within the region coding E (Fx1/Gx2) in 3 cases and were telomeric to all three analyzed IGCH sequences (Ex1/Fx1/Gx1) in 1 case (no. 3). The involvement of IGH in this particular del(14)(q24.1q32.33) (as well as six other deletions identified by aCGH; Figure 1a) suggested that this aberration may operate in a translocation-like manner and activate an unknown oncogene at 14q owing to juxtaposition to the E enhancer of IGH. To identify the affected gene, we analyzed the expression pattern of ZFP36L1 and 14 other genes mapped within the proximal 3 Mb (RAD51L1, ZFYVE16, RDH11, RDH12, ARG2, PLEKHH1, PLEK2, EIF2S1, ATP6V1D, MPP5, CN054, GPHN, NP_001004331) by qPCR analysis in 12 available cases (marked in Table 1) (Supplementary Table 1). MAX located 3.8 Mb from the breakpoint was included as an additional candidate gene. Four CLL and two myeloma cases without del(14) were used as controls. The results did not show any uniformly overexpressed gene and displayed a very inconsistent expression pattern of all evaluated genes in the index cases (Supplementary Table 2). The most frequently upregulated gene was RAD51L1 showing 2- to 18-fold increased expression when compared with controls in eight cases. Assuming that the gene targeted by del(14) (q24.1q32.33) escaped qPCR analysis, a global expression analysis using Human Genome U133 Plus 2.0 Affymetrix microarrays was performed on lymph nodes material from two cases (nos. 7 and 8; both umCLL). RNA extracted from involved lymph nodes from two umCLL cases without del(14) and two reactive lymph nodes was included as control. This global expression analysis identified 24 genes that were at least twofold upregulated, and 25 downregulated genes when compared to control CLLs (Supplementary Table 3 and Supplementary Figure 1). The former group included only one gene located on chromosome 14, CHURC1. This gene, mapped at 14q23.3 (64.45 Mb), was found to be 3.36-fold upregulated in the two analyzed cases. qPCR analysis, however, showed inconsistent expression pattern of CHURC1mRNA (and control RGPD4 and PLCD1 genes found to be upregulated in cases 7 and 8) in the remaining 10 analyzed cases (Supplementary Table 2). The proximal breakpoints of the remaining six IGH-involving del(14)(q) cases (Figure 1a) were mapped by aCGH in the vicinity of (or within) the following genes: TCL1A (95.24–95.25 Mb), SEL1L (81.00–81.06 Mb), TMED10 (74.66–74.71 Mb), RAD51L1 (6735–68.26 Mb), GPHN (66.04–66.71 Mb) and MAX (64.54–64.63 Mb). qPCR analysis of the first case with primers for TCL1A, TCL1B and TCL6 showed a low level of expression of these genes. In three other cases with the proximal breakpoints at 14q23.3–q24.1 (centromeric to ZFP36L1), we analyzed the expression status of candidate genes MAX and CHURC1. As in cases with the del(14)(q24.1q32.33), none of them showed a significant upregulation of these genes (data not shown). In summary, compiled aCGH and FISH analysis identified a novel type of interstitial del(14)(q) involving IGH in 27 B-NHL cases. In contrast to telomeric IGH deletions frequently detected by FISH in B-NHL5 reflecting somatic VH recombination events, these deletions are oncogenic. Notably, in 20 analyzed cases, the same del(14)(q24.1q32.33) with the IGH breakpoints proximal to the intrinsic E enhancer was found. This deletion was particularly associated with CLL (75%), trisomy 12 (47%) and unmutated status of VH (66%). Molecular cytogenetic findings led to the hypothesis that del(14)(q24.1q32.33) and similar nonrecurrent del(14) may upregulate expression of a gene(s) juxtaposed with E, like classical IGH-mediated t(14q32.33;var).6 Despite our extensive transcriptomic studies, however, the gene targeted by the recurrent del(14) (q24.1q32.33) has remained elusive. If the hypothesis is correct, our failure in identification of the affected gene might be caused either by its absence in the used Affymetrix GeneChip or the limited number of cases subjected to a global gene expression analysis. Alternatively, the del(14)(q24.1q32.33) may target for example noncoding sequences, like microRNAs (miRNAs), that will not be identified via a transcriptomic strategy. Several lines of evidence indicate that miRNAs play important roles in the pathogenesis of human cancers.7 Particularly interesting is the finding of a reduced expression of miR-15a and miR-16-1 (13q14) in del(13)(q)-positive CLLs.8 Browsing the chromosome 14 miRNA database, we found that only 3 of 49 identified miRNAs, mir-208, mir-624 and mir-625, are located respectively 41.4, 37.8 and 3.3 Mb centromeric to the proximal breakpoint of del(14)(q24.1q32.33) (http://microrna.sanger.ac.uk; released 25.10.06). Whether any of these miRNAs or other noncoding sequences are targeted by the del(14)(q24.1q32.33) needs further clarification. Although recurrent occurrence of the same del(14)(q24.1q32.33) suggests that this aberration is not simple tumor suppressor gene-associated deletion, we cannot exclude that cells harboring this aberration additionally (or mainly) benefit from the associated haploinsufficiency, or even homozygosity, of certain gene(s) located in the deleted 14q24.1q32.33 region, although such a gene(s) was not pinpointed by microarray analysis. Further investigations are needed to unravel the mechanism(s) and role of IGH-involving del(14)(q) in B-cell malignancies. This work was supported by the KULeuven Research Foundation (BIL03/12 and BIL05/59) and Grant G.0610.07 from the Fund for Scientific Research (FWO), Flanders, Belgium. Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)