Deletion 20q12 is associated with histological transformation of nodal marginal zone lymphoma to diffuse large B-cell lymphoma

Abstract

The genetic and molecular abnormalities underlying histological transformation (HT) of nodal marginal zone lymphoma (NMZL) to diffuse large B-cell lymphoma (DLBCL) are not well known. While del(20q12) is commonly deleted in myelodysplastic syndrome it has not previously been associated with DLBCL. We recently described a case of DLBCL harboring del(20q12) in a patient with a history of MZL involving lymph nodes and skin. Here we report eight matched cases of transformed MZL(tMZL): six from nodal MZL (tNMZL) and two from splenic MZL (tSMZL). We found >20% del(20q12) in 4/6 tNMZL, but not in tSMZL, nor in unmatched DLBCL, MZL with increased large cells (MZL-ILC), or MZL cases. To examine whether transformation is associated with a specific gene signature, the matched cases were analyzed for multiplexed gene expression using the Nanostring PanCancer Pathways panel. The differential gene expression signature revealed enrichment of inflammatory markers, as previously observed in MZL. Also, tMZL and de novo DLBCL were enriched for extracellular matrix proteins such as collagen and fibronectin, vascular development protein PDGFRβ, DNA repair protein RAD51, and oncogenic secrete protein Wnt11. A subset of genes is expressed differentially in del(20q12) tMZL cases vs non-del(20q12) tMZL cases. These results suggest a specific pathway is involved in the histological transformation of NMZL, which could serve as an indicator of aggressive clinical course in this otherwise indolent neoplasm.

FIGURE 1

Del(20q12) assessed by FISH in MZL, tNMZL, tSMZL and DLBCL. (A) Percent del(20q12) in DLBCL from (i) 8 MZL (seven matched cases, one unmatched case), (ii) 12 MZL-ILC (5 matched cases, seven unmatched cases), (iii) six tNMZL (matched cases), (iv) two tSMZL (matched cases), and (v) eight DLBCL (unmatched cases). Box represents 25/75th percentile, whiskers 5/95th percentile, black line indicates median, red dot indicates arithmetic mean. Significance (P value) was calculated by two-tailed, unpaired heteroscedastic t test. (B) Progression of del(20q12) for four NMZL cases (patients 1, 3–5) with matching MZL, MZL-ILC and tNMZL specimens. (C,D) Custom FISH probe for del(20q12) with flanking control probes covering RH65460 and RH47082. (C) Cell positive for del(20q12). (D) Cell negative for del(20q12). Note: the two control probes are not typically resolved on an individual chromosome

FIGURE 2

Differential gene expression of matched MZL/tMZL cases. (A) Violin plot of fold change in expression log₂(tMZL/MZL) of all genes with average fold change >1.5. (B) Volcano plot of −log₂ of false discover rate (FDR) P value vs log₂ of fold change. Genes highlighted in red have uncorrected P value (t test) < .05. Note: due to large number of genes (hypotheses) vs cases (samples), no gene expression changes have a significant change following FDR correction

FIGURE 3

Gene expression changes associated with del(20q12) in tNMZL. (A) The z-score transformed heatmap of fold changes clusters the 17/770 genes with the del(20q12)-specific differential expression in tMZL. The z-score transformed fold changes are calculated gene by gene: each gene’s mean fold changes and SD are calculated and then used to calculate its z-scores. The tMZL cases with >20% del(20q12) and tMZL cases with <20% del(20q12) can be differentiated from each other according to the opposite signs of z-score fold change values in 17 genes within the PanCancer Array. (B) Comparisons of four representative before-after plots for matched MZL/tMZL cases. Two plots per gene illustrate the differential gene expressions in cases with del(20q12) < 20% and del(20q12) > 20% for 4 of the 17 genes selected based on differential expression z-score. Blue solid lines correspond to NMZL/tNMZL cases, while red dashed lines correspond to SMZL/tSMZL cases