Gene expression profiling of resting and stimulated CD4+CD45RO+ T cells from Sezary syndrome patients and normal donors

Sezary syndrome is an aggressive cutaneous T cell lymphoma with pruritic skin inflammation and immune dysfunction, driven by neoplastic, clonal memory T cells in both peripheral blood and skin. To gain insight into how abnormal gene expression in Sezary syndrome promotes T cell dysfunction, lymphoproliferation and transformation, we first compared functional transcriptomic profiles of both resting and activated memory T cells from Sezary syndrome patients and normal donors. To differentiate gene expression associated with malignancy vs. benign inflammation and proliferation, we performed a within-platform meta-analysis of our data for Sezary syndrome and a GEO data set (GSE12079) for lymphocytic variant hypereosinophilic syndrome (L-HES). L-HES is a benign lymphoproliferation of clonal memory T cells that produces skin symptoms very similar to Sezary syndrome. This approach revealed gene expression changes unique to either Sezary syndrome or L-HES, and a subset of genes dysregulated in both SS and L-HES. L-HES patient 1 progressed to peripheral T cell lymphoma, and acquired Sezary-like gene expression during progression, suggesting that these genes contribute to neoplastic transformation. Original Sezary data: CD4+CD45RO+ T cells from three Sezary syndrome patients and three normal donors. To obtain functional gene expression, T cells were stimulated with PMA + A23187 for 0, 2, and 6 hours. Total RNA was then extracted and hybridized on Affymetrix HG U133 Plus 2.0 microarrays (1 microarray per sample). L-HES data: 23 samples from GEO series GSE12079. One replicate of L-HES patient 3 (GSM304966) was excluded. This series includes single microarrays for normal donor samples and 2-3 microarray replicates for L-HES samples. Normalized signal intensities from replicate L-HES microarrays were averaged prior to determining fold changes. Meta-analysis: Genes differentially expressed between cases and controls (within each study), or between stimulated and unstimulated conditions were identified using RankProd. The threshold for differential expression was log2FC ≥ |1|, and percentage of false prediction (pfp) < 0.05. Genes differentially expressed in Sezary only, L-HES only, or both diseases were identified using GeneVenn by comparing the lists of differentially expressed genes from each study.