DataSheet_6_Meta-Analysis of HTLV-1-Infected Patients Identifies CD40LG and GBP2 as Markers of ATLL and HAM/TSP Clinical Status: Two Genes Beat as One.csv

Human T-lymphotropic virus 1 (HTLV-1) was the first recognized human retrovirus. Infection can lead to two main symptomatologies: adult T-cell lymphoma/leukemia (ATLL) and HTLV-1 associated myelopathy/tropical spastic paraparesis (HAM/TSP). Each manifestation is associated with distinct characteristics, as ATLL presents as a leukemia-like disease, while HAM/TSP presents as severe inflammation in the central nervous system, leading to paraparesis. Previous studies have identified molecules associated with disease development, e.g., the downregulation of Foxp3 in Treg cells was associated with increased risk of HAM/TSP. In addition, elevated levels of CXCL10, CXCL9, and Neopterin in cerebrospinal fluid also present increased risk. However, these molecules were only associated with specific patient groups or viral strains. Furthermore, the majority of studies did not jointly compare all clinical manifestations, and robust analysis entails the inclusion of both ATLL and HAM/TSP. The low numbers of samples also pose difficulties in conducting gene expression analysis to identify specific molecular relationships. To address these limitations and increase the power of manifestation-specific gene associations, meta-analysis was performed using publicly available gene expression data. The application of supervised learning techniques identified alterations in two genes observed to act in tandem as potential biomarkers: GBP2 was associated with HAM/TSP, and CD40LG with ATLL. Together, both molecules demonstrated high sample-classification accuracy (AUC values: 0.88 and 1.0, respectively). Next, other genes with expression correlated to these genes were identified, and we attempted to relate the enriched pathways identified with the characteristic of each clinical manifestation. The present findings contribute to knowledge surrounding viral progression and suggest a potentially powerful new tool for the molecular classification of HTLV-associated diseases.

人类T淋巴细胞白血病病毒1型（Human T-lymphotropic virus 1, HTLV-1）是首个被确认的人类逆转录病毒。该病毒感染可引发两种主要临床症候群：成人T细胞淋巴瘤/白血病（Adult T-cell lymphoma/leukemia, ATLL）与HTLV-1相关脊髓病/热带痉挛性截瘫（HTLV-1 associated myelopathy/tropical spastic paraparesis, HAM/TSP）。两种病症各具独特特征：ATLL表现为白血病样疾病，而HAM/TSP则以中枢神经系统重度炎症为主要表现，最终导致痉挛性截瘫。既往研究已鉴定出多种与疾病发生相关的分子：例如调节性T细胞（Regulatory T cell, Treg）中叉头框P3（Foxp3）的表达下调与HAM/TSP发病风险升高相关；此外，脑脊液中新蝶呤（Neopterin）、CXC趋化因子配体9（CXCL9）与CXC趋化因子配体10（CXCL10）的水平升高也会增加发病风险。但上述分子仅与特定患者群体或病毒毒株相关。此外，绝大多数研究未同时对比所有临床表型，而严谨的分析需同时纳入ATLL与HAM/TSP样本；同时样本量不足也为开展基因表达分析以明确特异性分子关联带来了挑战。为解决上述局限并提升表型特异性基因关联分析的效能，本研究利用公开的基因表达数据开展了荟萃分析。通过应用监督学习技术，本研究鉴定出一对协同发挥作用的潜在生物标志物：鸟苷酸结合蛋白2（GBP2）与HAM/TSP相关，CD40配体（CD40LG）则与ATLL相关。二者联合使用时展现出极高的样本分类准确率（受试者工作特征曲线下面积[AUC]值分别为0.88与1.0）。随后，本研究进一步鉴定出与上述基因表达相关的其他基因，并尝试将富集得到的通路与各临床表型的特征相关联。本研究结果丰富了人们对HTLV致病进程的认知，并为HTLV相关疾病的分子分类提供了一种潜在的高效新工具。