Identification of progression markers in B-CLL by gene expression profiling

• Published in: Experimental hematology 2005-08, Vol.33 (8), p.883-893
• Main Authors: Fält, Susann, Merup, Mats, Gahrton, Gösta, Lambert, Bo, Wennborg, Anders
• Format: Article
• Language: English
• Subjects: Aged; Biomarkers, Tumor - genetics; Biomarkers, Tumor - metabolism; Female; Gene Expression Profiling; Gene Expression Regulation, Leukemic - genetics; Humans; Leukemia, Lymphocytic, Chronic, B-Cell - genetics; Leukemia, Lymphocytic, Chronic, B-Cell - metabolism; Male; Middle Aged; Oligonucleotide Array Sequence Analysis; Phosphoprotein Phosphatases - biosynthesis; Phosphoprotein Phosphatases - genetics; Protein Phosphatase 2; Proteins - genetics; Proteins - metabolism; Retinoblastoma-Like Protein p130
• ISSN: 0301-472X; 1873-2399
• DOI: 10.1016/j.exphem.2005.05.007

B-cell chronic lymphocytic leukemia is a heterogeneous disease with a pronounced variation in the clinical course. With the purpose of identifying genes that could be related to disease progression, we have performed gene expression profiling on B-CLL patients with an indolent disease and patients with a progressive disease with need for therapy. we applied the Affymetrix GeneChip technique to 11 B-CLL patients with stable and 10 patients with clinically progressive disease. Supervised and unsupervised clustering methods with different algorithms were used to identify genes that tend to give a distinction between stable and progressive disease. The supervised learning procedures identified groups of genes with a combined power to discriminate samples from progressive and stable disease with 70–90% accuracy. The gene for protein phosphatase 2 regulatory subunit B′ (B56) γ isoform ( PPP2R5C) and the gene for retinoblastoma-like 2 (p130) ( RBL2) were included among the best discriminators; both genes were downregulated in progressive as compared to stable B-CLL. In a hierarchical clustering analysis based on gene expression pattern three clinical subcategories could be identified: one with a more severe clinical outcome, a second one with good prognosis, and a third one that was intermediate between the other two groups. Our application of microarray analysis on a clinically well defined material has identified a number of genes with combined expression patterns related to stable or progressive disease in general. Unsupervised clustering suggested the existence of subclasses of samples in the progressive group that may be identifiable through gene expression patterns.