Oligonucleotide Microarray Analysis of Ocular Adnexal MALT Lymphoma with Trisomy 18.

Background: Recurrent karyotypic aberrations are shown in MALT-type lymphomas. It is unknown whether these heterogeneities are translated into difference of expression profiles creating distinct subtypes. Ocular adnexal MALT-type lymphoma (OAL) lacks contaminated extra-organ tissue, providing an opportunity to perform array analysis of pure tumor cell component. We have previously reported OAL cases with trisomy 18 might constitute a unique clinicopathological entity with shortened progression-free survival (

Patients and Methods: Archives of samples of OAL of our institute between 1995 and 2003 were utilized. Karyotypic informations were obtained by interphase fluorescence in situ hybridization (FISH) analysis. Oligonucleotide microarray with U133 Plus 2.0 Array (Affymetrix) were carried out. Statistical analysis was performed using GeneSpring GX 7.3.1 software. Biological networks were generated with Ingenuity Pathway Analysis version 3.1.

Results: A total of 26 cases of OAL were analyzed. Eighteen were male and 8 were female, with an overall median age of 56 years at the time of diagnosis (range, 30–90 years). Clinical stage was I (23 cases), II (1 case), and IV (2 cases). There were 1 case of IgH/MALT1 fusion, 16 cases of trisomy 3, 11 cases of trisomy 18, and 2 cases of trisomy 7 by FISH analysis. No cases had API2/MALT1 fusion or trisomy 12. In an unsupervised method, no clusters were identified within the categories including sex, age, anatomical site, trisomy3, 18, pathological features and progression. In a supervised method, a set of differentially expressed genes were identified which accurately stratified samples categories of trisomy 18 and others. Hierarchical clustering and principal component analysis of 26 OAL samples based on the differential expression of the 2506 genes, the two distinct subclasses accurately clustered together. The percentage of samples correctly predicted was 92% with leave-one-out cross validation using support vector machine, further confirming that our prior clustering very accurately predicted class assignment of OAL with trisomy 18 and that without trisomy 18 subset. The top three genes with high expression were HLA-DRB3, unidentified gene KIAA0924, and apoptosis related gene, MDM4.

A network analysis was performed to identify differently expressing genes related to apoptosis and proliferation. A case with t(14;18) showed over-expression of MALT1 gene and CARD11, as might be expected. Trisomy 18 cases had higher expression of Bcl-2 gene comparing to others. Other genes with higher expression shown in trisomy 18 in this network were anti-apoptotic genes including IKBKbeta and XIAP. A down-regulated gene was RTN4, functioning as down regulator of Bcl-2 gene. Cell cycle regulating genes including cyclin D2, CDK2, CDK6 and cyclin A2 were also highly expressed in OAL with trisomy18.

Conclusions: Our expression profile analysis again revealed uniqueness of trisomy 18 OAL. This might be attributed to the different levels of anti-apoptosis and proliferation gene-expressions.