Heart failure proteomics: an epidemiology study

心力衰竭蛋白质组学：流行病学研究

• 批准号: 10699754
• 负责人: Veronique Roger
• 金额: $ 166.48万
• 依托单位: NATIONAL HEART, LUNG, AND BLOOD INSTITUTE
• 依托单位国家: 美国
• 资助国家: 美国
• 起止时间: 至
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10699754
• 关键词: American Heart Association; Attenuated; Bioinformatics; Biological Markers; Books; Cardiology; Cessation of life; Characteristics; Clinic; Clinical; Clinical Data; Collaborations; Communities; Complex; Confidence Intervals; Congestive Heart Failure; Data; EFRAC; Electronic Health Record; Equipment and supply inventories; Etiology; European; Foundations; Framingham Heart Study; Heart failure; Immune response; Individual; Inflammation; Infrastructure; International; Kidney Diseases; Machine Learning; Manuscripts; Measurement; Meta-Analysis; Pathway Analysis; Pathway interactions; Patients; Phenotype; Plasma Proteins; Play; Population Heterogeneity; Population Study; Preparation; Proteins; Proteomics; Renal function; Risk; Role; Sampling Studies; Societies; Syndrome; Technology; Therapeutic; Validation; Work; angiogenesis; aptamer; base; biobank; cohort; comorbidity; design; effective therapy; epidemiology study; imaging program; meetings; mortality; mortality risk; novel; predictive signature; prognostic; prognostic value; proteomic signature; risk stratification

We studied proteomic signatures associated with death in a HF community cohort and quantified 7,335 plasma proteins using an aptamer-based technology. With machine learning, we identified proteomics cluster signatures associated with mortality and examined how these signatures predicted death while adjusting extensively for clinical parameters included in the Meta-analysis Global Group in Chronic Heart Failure (MAGGIC) score, and comorbidity burden. We applied 10-fold cross-validation to optimize the generalizability of our clustering results, and explored mechanistic pathways with bioinformatics enrichment and pathway analyses. Among 1388 individuals with HF, machine learning identified 2 clusters based on 405 proteins associated with death. Cluster assignment was univariately associated with the risk of death with a large increased risk in Cluster 2 patients (HR 2.41, 95% confidence intervals CI 2.12 - 2.75; p <0.0001). This association was only slightly attenuated after adjustment for the MAGGIC score and comorbidity burden (HR: 1.82; 95% CI, 1.58 - 2.08, p<0.0001) and did not differ by ejection fraction. Mechanistic pathways were mainly related to matrix remodeling, immune response, inflammation, and angiogenesis. In summary, taken collectively, these initial results indicate that proteomics provides important information on the phenotypes of the HF syndrome and the proteomic signatures play an important role in risk stratification. Using machine learning, our initial studies led to the identification of proteomic signatures associated with the risk of death independently of clinical factors. Key mechanistic pathways were identified laying the foundation for mechanistic therapeutic approaches. To pursue this work on deep phenotyping of HF using proteomics, we are currently examining how proteomic signature differ by the presentation of the HF syndrome including specific etiologies (ischemic versus nonischemic) and critical comorbidities (kidney disease). Studies planned for year 2023 include collaboration with the Framingham Heart Study and deployment of the requisite infrastructure for the urgently needed expansion to diverse populations through partnership with the imaging program in place at MedStar. This work has been or will be presented at several national and international meetings listed below. Corresponding manuscripts are in preparation. Annual Meeting of the Society for Epidemiology Research in June 2022 Proteomic Signatures in Heart Failure: a population-based study Kayode O Kuku Hoyoung Park, Suzette J. Bielinski, Nicholas B. Larson Jungnam Joo, Vronique L. Roger 2022-Abstract-Book.pdf (epiresearch.org). European Society of Cardiology meeting in August 2022. Proteomic Signatures of Heart Failure Mortality in the Community Kayode O Kuku, Hoyoung Park, Brittany Dulek, Suzette J. Bielinski, Jungnam Joo, Vronique L. Roger American Heart Association Scientific Sessions in November 2022 o Proteomic Assessment of Novel Kidney Function Biomarkers in Heart Failure: A Community Study Joseph J. Shearer, Hoyoung Park, Jungnam Joo, Kayode O Kuku, Suzette J. Bielinski, Sheila M. Manemann, Vronique L. Roger o Proteomic Signatures Of Ischemic And Non-ischemic Heart Failure In A Community Cohort Kayode O Kuku, Hoyoung Park ,Joseph J. Shearer, Jungnam Joo, Brittany Dulek, Suzette, J. Bielinski MEd, Vronique L. Roger, MD, MPH

我们在一个心力衰竭社区队列中研究了与死亡相关的蛋白质组特征，并使用基于适体的技术量化了7,335种血浆蛋白。通过机器学习，我们确定了与死亡率相关的蛋白质组群特征，并检查了这些特征是如何预测死亡的，同时广泛调整了Meta分析全球组慢性心力衰竭(MAGGIC)评分中包括的临床参数和共病负担。我们应用10倍交叉验证来优化我们的聚类结果的概括性，并通过生物信息学丰富和路径分析来探索机制路径。 在1388名心力衰竭患者中，机器学习基于405种与死亡相关的蛋白质识别出两个簇。在第2组患者中，分组分配与死亡风险和大量增加的风险是单变量相关的(HR2.41，95%可信区间CI2.12-2.75；p&lt；0.0001)。在调整MAGGIC评分和合并症负担后，这种相关性仅轻微减弱(HR：1.82；95%CI，1.58-2.08，p&lt；0.0001)，并且不受射血分数的影响。其机制主要与基质重塑、免疫反应、炎症和血管生成有关。 综上所述，这些初步结果表明，蛋白质组学提供了有关心力衰竭综合征表型的重要信息，蛋白质组学特征在风险分层中发挥着重要作用。使用机器学习，我们的初步研究导致识别出与死亡风险相关的蛋白质组特征，而不受临床因素的影响。确定了关键的机械性途径，为机械化治疗方法奠定了基础。为了利用蛋白质组学对HF进行深入的表型鉴定，我们目前正在研究HF综合征的蛋白质组特征如何不同，包括特定的病因(缺血性与非缺血性)和严重的共病(肾脏疾病)。2023年计划的研究包括与弗雷明翰心脏研究中心合作，以及通过与MedStar现有的成像计划合作，部署必要的基础设施，以满足向不同人群扩展的迫切需要。 这项工作已经或将在下列几次国家和国际会议上介绍。相应的手稿正在准备中。 2022年6月召开的流行病学研究学会年会：心力衰竭中的蛋白质组学签名：基于人群的研究Kayode O Kuku HoYoung Park，Suzette J.Bielinski，Nicholas B.Larson Jungnam Joo，Vronique L.Roger 2022-Abstract-Book.pdf(epearch.org)。 2022年8月召开的欧洲心脏病学会会议。社区心力衰竭死亡率的蛋白质组特征：Kayode O Kuku，HoYoung Park，Brittany Dulek，Suzette J.Bielinski，Jungnam Joo，Vronique L.Roger 美国心脏协会2022年11月科学会议 O心力衰竭中新的肾功能生物标志物的蛋白质组学评估：社区研究约瑟夫·J·希勒、霍扬·帕克、Jungnam Joo、Kayode O Kuku、Suzette J.Bielinski、Sheila M.Manemann、Vronique L.Roger O社区队列中的缺血性和非缺血性心力衰竭的蛋白质组学特征：Kayode O Kuku，HoYoung Park，Joseph J.Sheeller，Jungnam Joo，Brittany Dulek，Suzette，J.Bielinski Med，Vronique L.Roger，MD，Mph