The mutational landscape of ocular marginal zone lymphoma identifies frequent alterations in TNFAIP3 followed by mutations in TBL1XR1 and CREBBP

Abstract

Ocular marginal zone lymphoma is a common type of low-grade B-cell lymphoma. To investigate the genomic changes that occur in ocular marginal zone lymphoma, we analyzed 10 cases of ocular marginal zone lymphoma using whole-genome and RNA sequencing and an additional 38 cases using targeted sequencing. Major genetic alterations affecting genes involved in nuclear factor (NF)-κB pathway activation (60%), chromatin modification and transcriptional regulation (44%), and B-cell differentiation (23%) were identified. In whole-genome sequencing, the 6q23.3 region containing TNFAIP3 was deleted in 5 samples (50%). In addition, 5 structural variation breakpoints in the first intron of IL20RA located in the 6q23.3 region was found in 3 samples (30%). In targeted sequencing, a disruptive mutation of TNFAIP3 was the most common alteration (54%), followed by mutations of TBL1XR1 (18%), cAMP response element binding proteins (CREBBP) (17%) and KMT2D (6%). All TBL1XR1 mutations were located within the WD40 domain, and TBL1XR1 mutants transfected into 293T cells increased TBL1XR1 binding with nuclear receptor corepressor (NCoR), leading to increased degradation of NCoR and the activation of NF-κB and JUN target genes. This study confirms genes involving in the activation of the NF-kB signaling pathway is the major driver in the oncogenesis of ocular MZL.

Keywords: RNA sequencing; marginal zone lymphoma; mutation; ocular; whole-genome sequencing.

Figure 1. Copy number alterations in ocular MZL

a. GISTIC analysis of genomic regions with copy number loss.q-values (-log10 transformed) from GISTIC analysis (y-axis) are plotted across the genome (x-axis). Regions with q-values of <0.25 (red line) were considered significant. TNFAIP3 copy numbers in significantly deleted regions per sample are shown in the top right inset. b. The effects of copy number alternations on gene expression. The x-axis denotes samples grouped by copy number status, while the y-axis denotes expression in transcripts per million (TPM). P-values were derived from one-sided t-tests.

Figure 2. Landscape of somatic alternations in ocular MZL

Somatic alternations detected by whole genome (n=10) and targeted sequencing (n=38) are represented as a heatmap. Each column and row represents an affected individual and a gene, respectively. Genes altered in at least two patients were selected, and the fraction of affected individuals per gene is shown on the right. Mutated genes were clustered into four groups according to signaling pathway or molecular function.

Figure 3. Schematic representation of mutations in TBL1XR1

a. Location of missense mutations in the TBL1XR1 protein. Frequency of missense mutations (y-axis) is shown by amino acid (x-axis). b. Protein structure of TBL1XR1 with mutations. Mutant residues are highlighted in cyan. Mutations are incorporated into TBL1XR1 model using swapaa tool of Chimera version 1.10.2 considering appropriate rotamer.

Figure 4. Gene expression signature in ocular MZL

Differentially expressed pathways between ocular MZL (n=10) and normal marginal zone B-cell samples (n=10) are shown (p-value <0.01 from a two-sided t-test). The pathway activity level calculated by ssGSEA is a transformed z -score and is shown as a heatmap.

Figure 5. Functional characterization of TBL1XR1 WD40 domain alternations

a. Schematic diagram of the structure of TBL1XR1. Mutants occurred in the WD40 domain as indicated with circles. b. Co-immunoprecipitation of TBL1XR1 mutants. TBL1XR1 physically interacts with NCoR repressor complexes. Mutant TBL1XR1 interacts with NCoR and HDAC3 more frequently than wild-type TBL1XR1. c. Expression of NCoR protein levels in cells transfected with wild-type or mutant TBL1XR1. d. Fluorescence images of TBL1XR1 and Hoechst staining and merged images in 293T cells transfected with each TBL1XR1 mutant. e. NF-κB activity assay measuring the effects of each TBL1XR1 mutant. f. Cell proliferation comparison of TBL1XR1 mutants expressed in cells over 48 h using a CCK-8 assay.

Figure 6. Effects of TBL1XR1 mutations on gene expression

a. Hierarchical clustering of gene expression microarray data. An unsupervised hierarchical clustering algorithm was used to group cell lines expressing mutant (L282P and H348Q), wild-type, and siRNA knockdown based on the expression patterns of NF-κB (n=304) and JUN target genes (n=155). b. GSEA of NF-κB and JUN target gene sets in mutant (n=2; L282P and H348Q) versus wild-type TBL1XR1 (n=1) based on microarray expression data. Enrichment scores (ES) and p-values are shown in the top right of the enrichment plot. Genes contributing to core enrichment are shown as a heatmap. c. GSEA of NF-κB and JUN target gene sets in mutant (n=1) versus wild-type TBL1XR1 (n=9) based on RNA-seq data.