CPS: Medium: An AI-enabled Cyber-Physical-Biological System for Cardiac Organoid Maturation

确定和控制人诱导多能干细胞（hiPSC）衍生组织的成熟的能力对于组织工程、再生医学、药理学和合成生物学是至关重要的，这需要在组织的三维（3D）体积上以及在组织发育的时间过程中以细胞分辨率询问和干预细胞活性。该提案旨在建立一个人工智能支持的网络物理生物系统，以监测和控制hiPSC衍生的心肌细胞（hiPSC-CM）类器官在发育过程中的成熟。拟议的研究将开发“类组织”纳米电子器件，可以集成到正在发育的心脏类器官中，将电子传感器和致动器网络分布在整个组织的3D体积中，并实现组织水平的记录和控制在整个时间过程中以单细胞分辨率发展。原位单细胞RNA测序将用于整合基因表达数据与连续物理传感数据。将建立机器学习和统计模型来解释在线传感数据，并将开发网络控制方法来闭环在线控制心脏类器官成熟。开发的硬件和软件可以应用于几乎任何当前的生物系统，其中细胞状态的变化可以通过电子传感器和致动器可靠地记录和控制。该提案的成功将进一步融合人工智能，纳米电子学和生物学领域，为生物和生物医学工程的访问和控制带来无限的机会。多学科团队合作将代表一个成功的案例，来自不同领域的思想流派，包括生物工程，机器学习，统计学，控制理论等，相互启发和补充，以创造每个领域最先进的研究成果。研究团队还将与内部和外部合作伙伴合作，为来自不同背景的学生开展教育和社会活动，例如为本科生提供先进纳米电子制造的电子研讨会，研讨会和新课程，为当地K-12学生提供研讨会和图尔斯参观，以探索干细胞培养，在线视频传播基因组学，数学和计算建模，我们建议开发网络物理系统与生物系统的无缝集成，实现闭环控制，能够以单细胞分辨率实时，双向和长期稳定地询问和干预组织网络的3D体积中的细胞活动。作为示范，我们将把这种网络-物理-生物系统应用于hiPSC-CM类器官，促进和加速它们的成熟。我们将通过以下4项技术创新来实现我们的目标：（A）开发技术，将可拉伸网状纳米电子器件与多功能传感器和致动器集成到心脏类器官中，从而实现对类器官发育的实时监测和控制;（B）在原位单次测量期间精确地记录电子传感器，细胞RNA测序，以确定心脏类器官的分子成熟，并将空间基因表达谱与单细胞分辨率的传感数据相关联;（C）开发新的机器学习模型和工具，以识别基因表达与类器官范围的电和机械记录之间的统计干扰，并建立在线预测模型，以实时确定心脏类器官的成熟;（D）开发有效和可扩展的强化学习（RL）确定优化的电激活模式以促进心脏类器官成熟并测试其在患者中的性能的方法，该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The ability to determine and control the maturation of human-induced pluripotent stem cell (hiPSC) derived tissues is critical to tissue engineering, regenerative medicine, pharmacology, and synthetic biology, which requires the interrogation and intervention of cellular activities across the three-dimensional (3D) volume of tissues and over the time course of tissue development at cellular resolution. This proposal aims to build an AI-enabled cyber-physical-biological system to monitor and control the maturation of hiPSC derived cardiomyocyte (hiPSC-CM) organoids during development. The proposed research will develop “tissue-like” nanoelectronics that can be integrated into the developing cardiac organoids, distributing the electronic sensor and actuator network throughout the entire 3D volume of the tissue and enabling tissue-level recording and control over the entire time course of development at single-cell resolution. In situ single-cell RNA sequencing will be used to integrate gene expression data with continuous physical sensing data. Machine learning and statistical models will be built for interpreting the online sensing data, and cyber-control methods will be developed for the closed-loop online control of the cardiac organoid maturation. The developed hardware and software can be applied to virtually any current biological systems, in which the change of cellular states can be reliably recorded and controlled through the electronic sensors and actuators. The success of this proposal will further merge the field of AI, nanoelectronics, and biology, bringing unlimited opportunities for access and control to biological and biomedical engineering. The multidisciplinary teamwork will represent a successful case that schools of thought from diverse fields including bioengineering, machine learning, statistics, control theory, etc. inspire and complement each other to create state-of-the-art research results in each field. The research team will also collaborate with internal and external partners to launch educational and societal activities for students from diverse backgrounds, such as providing e-seminars, workshops and new courses for undergraduate students on advanced nanoelectronics fabrication, and workshops and tours for local K-12 students to explore stem cell culture, online videos to disseminate new research in genomics, mathematical and computational modeling, integration of AI, nanoelectronics, and biology.We propose to develop a seamless integration of cyber-physical systems with biological systems, enabling a closed-loop control, capable of real-time, bidirectionally, and long-term stably interrogating and intervening cellular activities across the 3D volume of tissue networks at single-cell resolution. As a demonstration, we will apply this cyber-physical-biological system to the hiPSC-CM organoids, promoting and accelerating their maturation. We will achieve our goal through the following 4 technical innovations: (A) developing technologies to integrate stretchable mesh nanoelectronics with multifunctional sensors and actuators to the cardiac organoids, enabling real-time monitoring and control of organoid development; (B) precisely registering electronic sensors during in situ single-cell RNA sequencing to determine the molecular maturation of cardiac organoids and correlate spatial gene expression profiling with sensing data at single-cell resolution; (C) developing novel machine learning models and tools to identify the statistical interference between gene expression and organoid-wide electrical and mechanical recording and also building online predictive models to real-time determine the maturation of cardiac organoids; (D) developing effective and scalable Reinforcement Learning (RL) methods to determine optimized electrical activation patterns to promote the maturation of cardiac organoids and to test its performance in patient-specific cardiac organoids.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.