Proteomic Biomarkers Prognostic for Diabetic Wound Healing

项目摘要/摘要 伤口愈合受损是糖尿病中一个令人担忧的问题，原因是超过750,000人 在美国，每年有70,000例糖尿病足溃疡(DFU)和70,000例下肢截肢。DFU是一种 具有不同临床病程的异质性疾病，迫切需要确定受试者 伤口愈合受损的风险更高。我们对循环的初步合作研究 蛋白质组指向DFU病程中有趣的候选生物标志物蛋白。在我们的试点中，学习 我们已经鉴定出与预期的伤口愈合结果相关的蛋白质属于这些类别。 作为黏附分子、羧基肽酶和炎症蛋白等。路径网络 以这些蛋白质为种子的分析显示，β连环蛋白和细胞Myc(c-Myc)是联系最紧密的 节点。有趣的是，Wnt/β连环蛋白通路和转录因子c-myc被广泛地涉及到 在DFU中发挥作用，包括糖尿病足联盟(DFC)调查人员的工作。DFU课程是 异质性和多因素，因此我们假设多生物标志物组合将提供最理想的 预言。由蛋白质水平反映的生物过程通常是相互关联的，因此我们假设 通过使用并行生物统计和机器学习方法，将为我们提供工具来构建健壮的签名。 这项提议有三个目标：目标1(R61阶段)将专注于提炼候选蛋白质样本 糖尿病受试者的预期伤口愈合。我们将对我们的 DFC队列受试者中的候选蛋白质(在我们的初步研究中确定)跟踪3个月 病程。我们将使用由机器学习方法支持的生物统计工具，足够用于 最终确定关键蛋白质样本的相关数据。在目标2(R61阶段)中，我们将开发一种专注的、 蛋白质组学生物标记物特征，由10-12个蛋白质组成的靶向定量小组组成。为此， 我们将执行广泛的分析验证，随后我们将建立一个专注和量化的多 生物标记物面板。最后，在目标3(R33阶段)将开始确定我们的预后的临床效用 DFU课程的生物标志物签名。为此，我们将在 DFC队列。我们将使用严格的生物统计指标来评估生物标志物签名的性能 (综合辨别能力、阿卡克标准等)。该项目的进展将确定和 初步确定生物标记物标志对糖尿病创面愈合前景的预测价值 当然了。这些努力将有助于提炼具有感兴趣表型的目标人群，通过提供目标 可以增强2/3期研究的临床试验登记标准的可量化指标。

PROJECT SUMMARY / ABSTRACT Impaired wound healing is an alarming problem in diabetes, attributed to the development of over 750,000 diabetic foot ulcerations (DFU) and 70,000 lower extremity amputations per year in the USA. DFU is a heterogeneous disease with a variable clinical course and there is an urgent, unmet need to identify subjects who are at higher risk of an impaired wound healing. Our preliminary collaborative study of the circulating proteome points to the interesting candidate biomarker proteins prognostic of the DFU course. In our pilot, study we have identified proteins associated with the prospective wound healing outcome belonging to such classes as adhesion molecules, carboxypeptidases, and inflammatory proteins among others. Pathway network analyses seeded with these proteins revealed β catenin and cellular Myc (c-Myc) among the most connected nodes. Interestingly, Wnt/β catenin pathway and transcription factor, c-Myc have been extensively implicated to play a role in the DFU including work of the Diabetic Foot Consortium (DFC) investigators. The DFU course is heterogeneous and multi-factorial, thus we hypothesize that multi-biomarker panel will offer the most optimal prognostication. Biological processes reflected by protein levels are often connected, thus we hypothesize that by employing parallal biostatistical and machine learning approach will offer us tools to build a robust signature. This proposal has the three aims: Aim 1 (R61 phase) will focus on refining candidate protein exemplars for a prospective wound healing among subjects with diabetes. We will perform semi-targeted measurements of our candidate proteins (identified in our preliminary study) in subjects of the DFC cohort followed for a 3-month disease course. We will employ biostatistical tools supported by machine learning methods, adequate for correlated data to finally determine key protein exemplars. In Aim 2, (R61 phase), we will develop a focused, proteomics biomarker signature comprising of a targeted quantitative panel of 10-12 proteins. For that purpose, we will perform an extensive analytical validation and subsequently we will build a focused and quantitative multi- biomarker panel. Finally, in Aim 3 (R33 phase) will initiate determination of the clinical utility of our prognostic biomarker signature for the DFU course. To this end, we will perform targeted biomarker measurements in the DFC cohort. We will employ rigorous biostatistical metrics to evaluate the biomarker signature’s performance (integrated discrimination ability, Akake criterion among others). Advancements in this project will identify and initially determine a value of the biomarker signature prognostic for the prospective diabetic wound healing course. These efforts will aid in refining a target population with a phenotype of interest by providing objective quantifiable metrics that can enhance clinical trial enrollment criteria for Phase 2/3 studies.