Transcriptome analysis of Sézary syndrome and lymphocytic-variant hypereosinophilic syndrome T cells reveals common and divergent genes

Abstract

Sézary syndrome (SS) 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 abnormal gene expression promoting T cell dysfunction, lymphoproliferation and transformation in SS, we first compared functional transcriptomic profiles of both resting and activated CD4 + CD45RO + T cells from SS patients and normal donors to identified differential expressed genes. Next, a meta-analysis was performed to compare our SS data to public microarray data from a novel benign disease control, lymphocytic-variant hypereosinophilic syndrome (L-HES). L-HES is a rare, clonal lymphoproliferation of abnormal memory T cells that produces similar clinical symptoms as SS, including severe pruritus and eosinophilia. Comparison revealed gene sets specific for either SS (370 genes) or L-HES (519 genes), and a subset of 163 genes that were dysregulated in both SS and L-HES T cells compared to normal donor T cells. Genes confirmed by RT-qPCR included elevated expression of PLS3, TWIST1 and TOX only in SS, while IL17RB mRNA was increased only in L-HES. CDCA7 was increased in both diseases. In an L-HES patient who progressed to peripheral T cell lymphoma, the malignant transformation identified increases in the expression of CDCA7 , TIGIT , and TOX , which are highly expressed in SS, suggesting that these genes contribute to neoplastic transformation. In summary, we have identified gene expression biomarkers that implicate a common transformative mechanism and others that are unique to differentiate SS from L-HES.

Gene Expression Comparison between Sézary Syndrome and Lymphocytic-Variant Hypereosinophilic Syndrome Refines Biomarkers for Sézary Syndrome

“…The public L-HES data set from Ravoet et al [30] was recently compared to gene expression data from SS memory T cells [22] (Figure 2). Importantly, both data sets were obtained on the same microarray platform.…”

“…The outcome of this meta-analysis approach was greater confidence in the identification of biomarker genes specific to the malignant phenotype of SS T cells, which eliminated Th2-and lymphoproliferation-associated genes inherent to L-HES. A common analysis workflow was used for both data sets to identify genes of interest, and changes in SS or L-HES gene expression compared to normal donors was based on a threshold of 2-fold with q ≤ 0.05 [22]. The outcome showed a highly significant degree of overlap between the abnormal gene expression profiles of SS and L-HES T cells compared to normal T cells ( Figure 2), suggesting that gene expression shared by SS and L-HES reflects benign lymphoproliferative and Th2 phenotypes rather than malignant processes.…”

“…Numerous studies have reported overexpression of TWIST1 in SS [8,14,16,22,72]. TWIST1 is a highly conserved developmental transcription factor with diverse pathological functions in solid and hematological tumors [103,104].…”

“…A remaining challenge to the understanding of SS is the lack of a well-defined, pre/early-neoplastic T-cell population suitable for studies of stage progression, and the evolution of neoplastic phenotypes. However, this role was recently filled by a benign T-cell lymphoproliferative disorder known as lymphocytic-variant hypereosinophilic syndrome (L-HES) [22], which has skin inflammation and hematologic abnormalities that resemble SS [23][24][25][26][27][28][29]. A dominant T-cell clone in the blood is detected in a large majority of L-HES cases [23,25], and atypical T cells often exceed 70% of circulating lymphocytes [30].…”

“…BIDs are characterized by polyclonal infiltrates of reactive T cells, and are not considered cutaneous lymphoproliferative disorders [31,32]. The combination of benign inflammatory and clonal lymphoproliferative phenotypes makes L-HES a unique disease control (Table 1) that can provide meaningful biological context that was previously lacking in transcriptomic studies of SS [22].…”