Abstract
T-cell prolymphocytic leukemia (T-PLL) is rare and presents with widespread disease. Indolent presentations are seen but eventually progress. The disease shows marked chemoresistance and is best treated with the monoclonal anti-CD52 antibody (CAMPATH). Prolymphocytes show a post-thymic phenotype and are CD4+CD8− (65%), CD4−CD8+ (10%) or CD4+CD8+ (25%). This double positive phenotype, raises questions about the putative ontology of T-PLL. Morphological heterogeneity, with typical (75%), small cell (20%) and cerebriform/sezary-like variants (5%) is described. Inversions or reciprocal translocations of chromosome 14 involving breakpoints at q11 (TCR a/d) and q32.1 (TCL1 and TCL1b) are seen (~ 80%). Other common abnormalities involve chromosome 8, translocation (X;14)(q28;q11) and, ATM (11q23). We investigated the clinico-pathological heterogeneity in T-PLL, at the level of the transcriptome and evaluated the ability of gene expression profiling to sub-classify T-PLL. Total RNA was extracted from blood prolymphocytes (>92% purity) of 22 patients. cDNA synthesis followed by biotin-labelled cRNA synthesis was carried out as per Affymetrix protocols. Fragmented cRNA was hybridized to the Human U133 PLUS2 GeneChip array (54K probes). Microarray services were provided by MRC geneservice (UK HGMP Resource Centre). Hierarchical clustering of samples was performed using a filtered gene set (12,456) and >4 different algorithims. Prediction analysis for micoarray (PAM) and significance analysis of microarray (SAM) were used to evaluate class performance, and partition genes using pre-defined labels of immunophenotype, karyotype, response and morphology. Validation was performed by RT-PCR in a subset of genes.Unsupervised analysis robustly and reproducibly partitioned samples into 2 groups; A (n=8) and B (n=14). SAM analysis identified 4487 differentially expressed transcripts (false discovery rates <1%), >40% of which showed >2-fold difference in expression between the groups. There was no statistical difference in age, immunophenotype or karyotype betweeen groups, however, differential response to CAMPATH was seen. PAM analysis refined a sub-group of ~123 genes which most efficiently differentiated these groups. Group A showed significantly higher rates of non-response and progressive disease as compared to group B (n=14, p=0.036). Key differences related to apoptosis and cell-cycle associated gene expression. Down regulation of caspases (CASP1, CASP2,CASP4, CARD8 and CASP8AP2), cyclins (CCNC, CCND2, CCND3, CCNG1, CCNI, CCNT2), bcl-2, HDAC1, HIPK2, IL6R and ATM were frequent in group A with upregulation of genes implicated in NF-kB (TRAF4, SQSTM1) and TNF pathways (LMNA, ARTS-1), as well as transcription factors such as ATF-3. CD52 expression was ~2-fold higher in group B and may explain in part, differential responses to CAMPATH. RT-PCR validated gene expression data for LMNA and ATF-3. Despite the small numbers, algorithim-independent segregation into 2 consistent groups, in conjunction with the magnitude of gene differences, presence of many mutually exclusive divisions, and low prediciton errors, imply that the 2 identified profiles arise from fundamental differences at a regulatory level and thus likely represent a generalisable classification for T-PLL. Differential responses to CAMPATH may be a sub-feature of this grouping.
Is geographical variation driving the transcriptomic responses to multiple stressors in the kelp Saccharina latissima?