Whole Transcriptome Studies of Blood to Predict Stroke Outcome

预测中风结果的血液全转录组研究

• 批准号: 10655229
• 负责人: Bradley Pearce Ander
• 金额: $ 64.19万
• 依托单位: UNIVERSITY OF CALIFORNIA AT DAVIS
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-04-15 至 2028-03-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10655229
• 关键词: 3-Dimensional; AMPA Receptors; Age; Anti-Inflammatory Agents; Atrial Fibrillation; Biological Markers; Biology; Blood; Blood Platelets; Blood Pressure; Brain; Brain-Derived Neurotrophic Factor; C-reactive protein; Clinical; Clinical Trials; Coagulation Process; Cytochrome P450; Data; Derivation procedure; Diabetes Mellitus; Engineering; Environment; Etiology; Future; Gene Expression; Genes; Genetic Markers; Genome; Glucose; Growth Factor; HMGB1 gene; Heat shock proteins; Hemoglobin concentration result; Human; Hyperlipidemia; IL6 gene; ITGB3 gene; Immune; Infection; Inflammatory; Interleukin-1; Interleukin-10; Ischemic Stroke; Knowledge; Leukocytes; Life Style; Light; Location; Logistic Regressions; Lymphocyte Count; Machine Learning; Measures; Molecular; Neurologic; Outcome; Outcome Measure; PARK2 gene; PTGS2 gene; Pathway Analysis; Pathway interactions; Patients; Pattern; Phosphoric Monoester Hydrolases; Platelet Count measurement; Prognosis; Proteins; RNA; Recovery; Regulator Genes; Sensitivity and Specificity; Severities; Stroke; Stroke Volume; System; TNF gene; United States National Institutes of Health; Validation; cohort; cytokine; functional outcomes; gene network; genetic risk factor; improved; indexing; kidney dysfunction; machine learning algorithm; machine learning prediction; molecular marker; monocyte; neurofilament; neuronal excitability; neutrophil; outcome prediction; patient stratification; peripheral blood; post stroke; presynaptic; repaired; response; sex; statistics; stroke outcome; stroke patient; success; support vector machine; transcriptome; transcriptome sequencing; vascular risk factor; whole genome

Abstract Several clinical variables are associated with outcomes following ischemic stroke (IS). However, clinical and demographic parameters account only for a portion of the outcome variance, thus it is difficult for clinicians to reliably predict long-term IS outcome. Hence, new biomarkers are needed. Molecules in blood are also associated with IS outcome including pro-inflammatory cytokines, anti-inflammatory cytokines and others. Genetic risk factors have also been associated with IS outcome. Unfortunately, combined blood and genetic biomarkers have not improved IS outcome predictions compared to clinical parameters since age, sex and initial NIHSS is said to predict outcome with a modest c-statistic approaching 0.7. Our preliminary data show gene expression in blood after IS can predict 90-day outcome better than age, sex and NIHSS. We hypothesize that a whole-genome approach of measuring RNA, which reflects the genome × environment × lifestyle interaction, to assess inflammatory, trophic and clotting genes will improve IS outcome prediction compared to clinical features alone. Thus, we propose the following aims: Aim #1a. Perform whole-genome RNA sequencing (RNAseq) of blood on a derivation cohort of IS patients at 1 day and 3 days after IS compared to matched vascular risk factor controls (VRFC). Aim #1b. Identify the most significantly regulated genes and pathways in blood at 1d/3d after IS that correlate with outcome, as measured by three outcome scales – mRS (modified Rankin Scale), NIHSS (NIH Stroke Scale) and Barthel Index at 90 days after IS. Aim #1c. Use Network Analysis to identify key hub genes after IS associated with outcome that might be causative. Aim #1d. Use Feature Engineering and Logistic Regression and/or other Machine Learning approaches, such as Support Vector Machines (SVM) and SVM Regression, to identify the least number of genes at 1 and/or 3 days that predict the three stroke 90-day outcome measures. Aim #1e. In a separate validation cohort of IS patients perform RNAseq to obtain the expression of the biomarker genes from Aim #1d. Input these into Machine Learning algorithms to predict patient 90-day outcomes (mRS, NIHSS, Barthel Index). Aim #2a. Demonstrate that gene expression is a better 90-day outcome predictor compared to each or a combination of clinical variables, such as stroke volume, initial NIHSS, location, etiology, age, sex, glucose levels, blood pressure, atrial fibrillation, and neutrophil, monocyte, lymphocyte, and platelet counts. Aim #2b. Delineate the underlying biology of these clinical outcome contributors by identifying genes and networks that correlate with them and pinpoint which of the genes/networks also correlate with each of the three outcomes. Significance: The findings of this study will develop biomarkers of ischemic stroke outcome to help clinicians predict IS outcome and aid future clinical trials in stratifying IS patients, thus significantly increasing chances for trial success. Equally important, the findings will provide much needed potential new treatment targets and unprecedented knowledge of how the peripheral blood transcriptome contributes to outcome and improve our understanding of the biology of repair and recovery after IS in humans.

摘要 几个临床变量与缺血性卒中（IS）后的结局相关。但临床 人口统计学参数仅占结果方差的一部分，因此临床医生很难 以可靠地预测长期IS结果。因此，需要新的生物标志物。血液中的分子也 与IS结果相关的细胞因子包括促炎细胞因子、抗炎细胞因子等。 遗传风险因素也与IS结局相关。不幸的是，结合血液和遗传 与临床参数相比，生物标志物没有改善IS结局预测，因为年龄、性别和 据说初始NIHSS预测结果的c-统计量接近0.7。我们的初步数据显示 IS后血中基因表达较年龄、性别和NIHSS评分更能预测90 d预后。我们 假设测量RNA的全基因组方法，它反映了基因组×环境× 生活方式的相互作用，以评估炎症，营养和凝血基因将改善IS的结果预测 相比于单独的临床特征。因此，我们提出以下目标：目标#1a。执行全基因组 IS后1天和3天IS患者衍生队列的血液RNA测序（RNAseq） 与匹配的血管危险因素对照（VRFC）相比。目标1b。确定最受监管的 IS后1d/3d血液中与结果相关的基因和途径，通过三种结果测量 IS后90天时的量表- mRS（改良兰金量表）、NIHSS（NIH卒中量表）和Barthel指数。目的 #1c。使用网络分析，以确定关键枢纽基因后，IS与结果可能是因果关系。 目标1d。使用特征工程和逻辑回归和/或其他机器学习方法，例如 作为支持向量机（SVM）和SVM回归，以在1和/或3处识别最少数量的基因 天，预测三个中风90天的结果措施。目标1 e。在IS的单独验证队列中 患者进行RNAseq以获得来自目标#1d的生物标志物基因的表达。把这些输入到 机器学习算法预测患者90天结局（mRS、NIHSS、Barthel指数）。目标2a。 证明基因表达是一个更好的90天的结果预测相比，每一个或组合 临床变量，如每搏输出量、初始NIHSS、部位、病因、年龄、性别、血糖水平、血液 血压、心房颤动和中性粒细胞、单核细胞、淋巴细胞和血小板计数。目标2b。划定 这些临床结果贡献者的基础生物学，通过识别与以下相关的基因和网络： 并找出哪些基因/网络也与三种结果中的每一种相关。 意义：本研究的发现将开发缺血性卒中结局的生物标志物，以帮助临床医生 预测IS的结果，并帮助未来的临床试验对IS患者进行分层，从而显著增加 为了审判的成功。同样重要的是，这些发现将提供急需的潜在新治疗靶点， 前所未有的知识如何外周血转录组有助于结果和改善我们的 了解人类IS后修复和恢复的生物学。