Biological and Clinical Relevance of Surrogate Markers of IgVH Mutational Status in B-Cell Chronic Lymphocytic Leukemia.

B-cell chronic lymphocytic leukemia (B-CLL) is a heterogeneous disease; some patients have a rapidly progressing disease and others exhibit an indolent course and survive for many years without treatment. Mutation status of IgVH genes utilized by CLL cells represents a very reliable predictor of clinical outcome in B-CLL, but its analysis is expensive and beyond the capacities of most diagnostic laboratories. To identify surrogate markers we performed a gene expression profiling analysis of CD19+ purified cells from 80 B-CLL untreated patients in Binet stage A, by means of Affymetrix GeneChip® HGU133A arrays. The comparison of 46 IgVH-unmutated versus 34 mutated samples using the Prediction Analysis of Microarrays software identified 78 differentially expressed probes, specific for 59 well-characterized genes. Specifically, 43 genes had a higher and 16 genes a lower average expression in the IgVH unmutated group. These genes are involved in cellular functions, including cell cycle regulation (SEPT7, SEPT10, CDK2AP1), cell proliferation (SLAMF1, LDOC1), apoptosis (CD63, IFT57, P2RX1, RNF130, TNFRSF1B), cell adhesion (CNTNAP2, C1orf38, PCDH9), immune response (ZAP70, IFI44), signal transduction (AKAP13, RASGRP1, USP6NL, TGFBR3, AKAP12), lipid metabolism and fatty-acid degradation (FADS3, LPL, LASS6), cell-cell signalling (FCRL2), phospholipid biosynthetic process (AYTL2), regulation of circadian rhythm (EGR3, CRY1, OPN3), DNA-dependent regulation of transcription (MYBL1, NR4A2, NRIP1, ZBTB20), muscle development (VAMP5, SRI, DMD). The expression signature identified in the proprietary database was then validated by means of a meta-analysis of a publicly available gene expression dataset of 100 B-CLL (Haslinger et al., 2005), showing classification accuracy measures leading to a global classification rate of 82.93% of the test set and thus suggesting the strength of the identified expression signature. The expression levels of 11 genes (LPL, ZBTB20, ZAP70, CRY1, COBLL1, SEPT10, LDOC1, TNFRSF1B, DMD, SRI, NRIP1) were confirmed by means of quantitative real-time PCR (Q-RT-PCR) in a subset of 40 CLL patients. The prognostic impact for Time To Treatment (TTT) of the 59 candidate genes of our classifier model was investigated in 77 patients. Forty-nine (36.4%) of these received treatment after a median follow up of 4 years. As expected, patients with unmutated IgVH genes had a risk of therapy requirement that was about 3 times higher (HR: 3.1,95% C.I. 1.6–5.8, p<0.0001) than those with mutated IgVH. Based on microarray expression levels, 43/59 genes significantly predicted TTT with a HR ranging from 1.5 (LPL gene) to 4.2 (SRI gene) (value for ZAP-70 = HR: 1.9, 95% C.I. 1.0–3.4, p=0.039). The same analysis performed in the panel of the 11 genes validated by Q-RT-PCR revealed 4 candidate genes which significantly predicted TTT. Specifically, Cox univariate analysis confirmed ZAP-70 as a predictor of disease outcome and underscored the prognostic role of the LPL, TNFRSF1B and CRY1 genes. The predictive power of the novel putative surrogate markers for the IgVH mutation status is now being further validated at protein expression level.