Artificial Intelligence Boosted Evolution and Detection of Genetically Encoded Reporters for In Vivo Imaging

人工智能促进体内成像基因编码报告基因的进化和检测

• 批准号: 10533825
• 负责人: CHRISTIAN T FARRAR
• 金额: $ 68.14万
• 依托单位: MASSACHUSETTS GENERAL HOSPITAL
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-04-01 至 2024-12-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10533825
• 关键词: Algorithms; Amides; Artificial Intelligence; Biological Response Modifier Therapy; Cardiac; Cells; Chemicals; Clinic; Clinical; Clinical assessments; Collaborations; Detection; Development; Diagnosis; Discrimination; Disease; Drug Kinetics; Evolution; Fingerprint; Gene Transfer; Genes; Genetic Programming; Glioblastoma; Goals; Histologic; Image; Infection; Longitudinal Studies; Lysine; Machine Learning; Magnetic Resonance; Magnetic Resonance Imaging; Measurement; Medical Imaging; Metabolic; Methods; Modeling; Monitor; Mus; Oncolytic; Oncolytic viruses; Pathology; Performance; Physiologic pulse; Predisposition; Problem Solving; Productivity; Protein Engineering; Proteins; Protons; Reporter; Reporter Genes; Resistance; Resolution; Schedule; Sensitivity and Specificity; Signal Transduction; Specificity; Technology; Testing; Therapeutic; Therapeutic Agents; Time; Tissues; Tumor Cell Line; Viral; Virus Diseases; Virus Replication; Water; Work; cancer imaging; clinical translation; deep learning; deep neural network; detection method; detection sensitivity; gene therapy; imaging modality; improved; in vivo imaging; neoplastic cell; neural network architecture; novel; oncolytic virotherapy; pre-clinical; radio frequency; rapid technique; response; therapeutic gene; tool; trafficking; transgene expression; tumor; viral gene delivery

PROJECT SUMMARY Magnetic resonance (MR) reporter genes have the potential to monitor transgene expression non-invasively in real time at high resolution. These genes can be applied to interrogate the efficacy of gene therapy, to monitor viral therapeutics and viral gene delivery, to assess cellular differentiation, cell trafficking, and specific metabolic activity, and also assess changes in the microenvironment. Efforts toward the development of MR reporter genes have been made for over a decade, but, despite these efforts, the field is still in its early developmental stage. This reflects the fact that there are numerous complications, caused by the low sensitivity of detection, the need for substrates with their associated undesirable pharmacokinetics, and/or the difficult and, in some cases, delayed interpretation of signal changes. We have previously demonstrated that many of these challenges can be overcome with the use of a lysine rich protein (LRP) reporter gene, that is detectable by chemical exchange saturation transfer (CEST) MRI. However, to mature the CEST reporter gene technology and bring it towards clinical translation, its sensitivity and specificity need to be improved. In particular, the LRP reporter gene specificity is limited by the fact that the lysine amide exchangeable protons of LRP have the same chemical shift as amide protons from endogenous proteins. It is therefore difficult to distinguish the reporter CEST contrast from the background CEST contrast, both of which may be changing with time. The specificity is further limited by the sensitivity of the CEST contrast to intracellular pH where the qualitative CEST contrast cannot distinguish between exchange rate and concentration effects. Finally, a decrease in cytosolic pH, observed in many disease pathologies, reduces the amide proton exchange rate and hence the CEST reporter sensitivity. We therefore propose to develop improved MRI reporter genes and quantitative MRI detection methods that will facilitate the clinical translation of these methods for imaging biological therapeutics, such as oncolytic virotherapy. We hypothesize that CEST reporter genes with improved sensitivity and specificity along with improved quantitative CEST methods will enable viral infection and replication to be monitored longitudinally throughout OV tumor therapy. To test this hypothesis and establish the clinical potential of MRI reporter genes we will capitalize on two transformative technologies developed in our labs; (Aim 1) an artificial intelligence based genetic programming algorithm will be used for optimizing the sensitivity and specificity of the CEST reporter gene and (Aim 2) a CEST magnetic resonance fingerprinting (MRF) method will be used for the rapid quantification of both the reporter protein concentration and chemical exchange rate. (Aim 3) These methods will be validated for imaging oncolytic viral infection and replication in mouse glioblastoma tumor models.

项目摘要 磁共振（MR）报告基因具有非侵入性监测转基因表达的潜力 高分辨率实时。这些基因可以应用于询问基因治疗的功效，以监测 病毒疗法和病毒基因递送，以评估细胞分化，细胞运输和特定 代谢活性，还评估微环境的变化。为发展MR的努力 记者基因已经制造了十多年，但是，尽管有这些努力，但该领域仍处于早期状态 发展阶段。这反映了一个事实，即由于低的并发症而引起的许多并发症 检测的敏感性，对其相关的不良药代动力学的底物的需求和/或 困难，在某些情况下会延迟解释信号变化。 我们以前已经证明，使用赖氨酸富的可以克服许多这些挑战 蛋白质（LRP）报告基因，可通过化学交换饱和转移（CEST）MRI检测到。 但是，要使CEST记者基因技术成熟并将其带入临床翻译，其敏感性 并且需要提高特异性。特别是，LRP报告基因特异性受到以下事实的限制 LRP的赖氨酸酰胺可交换质子具有与酰胺质子相同的化学位移 内源性蛋白质。因此，很难区分记者的对比与背景 CEST对比，两者都可能随着时间而变化。特异性受到的灵敏度进一步限制 与细胞内pH对比的CEST在定性CEST对比无法区分的细胞内pH对比 汇率和集中效应。最后，在许多疾病中观察到的胞质pH值降低 病理学，降低酰胺质子汇率，从而降低了记者的灵敏度。 因此，我们建议开发改进的MRI报告基因和定量MRI检测方法 将促进这些方法的临床翻译，用于成像生物学疗法，例如溶瘤 病毒疗法。我们假设CEST报告基因具有提高灵敏度和特异性以及 改进的定量CEST方法将使病毒感染和复制能够纵向监测 整个OV肿瘤疗法。检验该假设并建立MRI报告基因的临床潜力 我们将利用实验室中开发的两种变革性技术。 （目标1）人工智能 基于遗传编程算法将用于优化CEST的灵敏度和特异性 记者基因和（AIM 2）将使用CEST磁共振指纹（MRF）方法用于快速 定量报告基因蛋白浓度和化学汇率。 （目标3）这些方法 将在小鼠胶质母细胞瘤肿瘤模型中进行成像肿瘤病毒感染和复制的验证。