Attention:NSF PAR will be unavailable due to scheduled facility maintenance July 10th, 5 PM EDT through July 13th. We apologize for any inconvenience.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Reinforcement learning‐guided control strategies for CAR T‐cell activation and expansion
Abstract Reinforcement learning (RL), a subset of machine learning (ML), could optimize and control biomanufacturing processes, such as improved production of therapeutic cells. Here, the process of CAR T‐cell activation by antigen‐presenting beads and their subsequent expansion is formulated in silico. The simulation is used as an environment to train RL‐agents to dynamically control the number of beads in culture to maximize the population of robust effector cells at the end of the culture. We make periodic decisions of incremental bead addition or complete removal. The simulation is designed to operate in OpenAI Gym, enabling testing of different environments, cell types, RL‐agent algorithms, and state inputs to the RL‐agent. RL‐agent training is demonstrated with three different algorithms (PPO, A2C, and DQN), each sampling three different state input types (tabular, image, mixed); PPO‐tabular performs best for this simulation environment. Using this approach, training of the RL‐agent on different cell types is demonstrated, resulting in unique control strategies for each type. Sensitivity to input‐noise (sensor performance), number of control step interventions, and advantages of pre‐trained RL‐agents are also evaluated. Therefore, we present an RL framework to maximize the population of robust effector cells in CAR T‐cell therapy production.  more » « less
Award ID(s):
2042503
PAR ID:
10574184
Author(s) / Creator(s):
; ; ;
Publisher / Repository:
Wiley
Date Published:
Journal Name:
Biotechnology and Bioengineering
Volume:
121
Issue:
9
ISSN:
0006-3592
Page Range / eLocation ID:
2868 to 2880
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ; ; ; ; ; ; et al (, Advanced Materials)
    Abstract Chimeric antigen receptor (CAR) T cell therapy has achieved remarkable clinical success in the treatment of hematological malignancies. However, producing these bespoke cancer‐killing cells is a complicated ex vivo process involving leukapheresis, artificial T cell activation, and CAR construct introduction. The activation step requires the engagement of CD3/TCR and CD28 and is vital for T cell transfection and differentiation. Though antigen‐presenting cells (APCs) facilitate activation in vivo, ex vivo activation relies on antibodies against CD3 and CD28 conjugated to magnetic beads. While effective, this artificial activation adds to the complexity of CAR T cell production as the beads must be removed prior to clinical implementation. To overcome this challenge, this work develops activating lipid nanoparticles (aLNPs) that mimic APCs to combine the activation of magnetic beads and the transfection capabilities of LNPs. It is shown that aLNPs enable one‐step activation and transfection of primary human T cells with the resulting mRNA CAR T cells reducing tumor burden in a murine xenograft model, validating aLNPs as a promising platform for the rapid production of mRNA CAR T cells. 
    more » « less
  2. ; ; ; ; ; ; ; ; ; ; et al (, Small)
    Abstract With six therapies approved by the Food and Drug Association, chimeric antigen receptor (CAR) T cells have reshaped cancer immunotherapy. However, these therapies rely on ex vivo viral transduction to induce permanent CAR expression in T cells, which contributes to high production costs and long‐term side effects. Thus, this work aims to develop an in vivo CAR T cell engineering platform to streamline production while using mRNA to induce transient, tunable CAR expression. Specifically, an ionizable lipid nanoparticle (LNP) is utilized as these platforms have demonstrated clinical success in nucleic acid delivery. Though LNPs often accumulate in the liver, the LNP platform used here achieves extrahepatic transfection with enhanced delivery to the spleen, and it is further modified via antibody conjugation (Ab‐LNPs) to target pan‐T cell markers. The in vivo evaluation of these Ab‐LNPs confirms that targeting is necessary for potent T cell transfection. When using these Ab‐LNPs for the delivery of CAR mRNA, antibody and dose‐dependent CAR expression and cytokine release are observed along with B cell depletion of up to 90%. In all, this work conjugates antibodies to LNPs with extrahepatic tropism, evaluates pan‐T cell markers, and develops Ab‐LNPs capable of generating functional CAR T cells in vivo. 
    more » « less
  3. (, ICML)
    Agents that are aware of the separation between themselves and their environments can leverage this understanding to form effective representations of visual input. We propose an approach for learning such structured representations for RL algorithms, using visual knowledge of the agent, such as its shape or mask, which is often inexpensive to obtain. This is incorporated into the RL objective using a simple auxiliary loss. We show that our method, Structured Environment-Agent Representations, outperforms state-of-the-art model-free approaches over 18 different challenging visual simulation environments spanning 5 different robots. 
    more » « less
  4. (, 70th International Astronautical Congress)
    Reinforcement learning (RL) has recently shown promise in solving difficult numerical problems and has discovered non-intuitive solutions to existing problems. This study investigates the ability of a general RL agent to find an optimal control strategy for spacecraft attitude control problems. Two main types of Attitude Control Systems (ACS) are presented. First, the general ACS problem with full actuation is considered, but with saturation constraints on the applied torques, representing thruster-based ACSs. Second, an attitude control problem with reaction wheel based ACS is considered, which has more constraints on control authority. The agent is trained using the Proximal Policy Optimization (PPO) RL method to obtain an attitude control policy. To ensure robustness, the inertia of the satellite is unknown to the control agent and is randomized for each simulation. To achieve efficient learning, the agent is trained using curriculum learning. We compare the RL based controller to a QRF (quaternion rate feedback) attitude controller, a well-established state feedback control strategy. We investigate the nominal performance and robustness with respect to uncertainty in system dynamics. Our RL based attitude control agent adapts to any spacecraft mass without needing to re-train. In the range of 0.1 to 100,000 kg, our agent achieves 2% better performance to a QRF controller tuned for the same mass range, and similar performance to the QRF controller tuned specifically for a given mass. The performance of the trained RL agent for the reaction wheel based ACS achieved 10 higher better reward then that of a tuned QRF controller 
    more » « less
  5. ; ; ; ; ; ; ; ; ; ; et al (, Journal for ImmunoTherapy of Cancer)
    null (Ed.)
    Background Adoptive cell therapy based on the infusion of chimeric antigen receptor (CAR) T cells has shown remarkable efficacy for the treatment of hematologic malignancies. The primary mechanism of action of these infused T cells is the direct killing of tumor cells expressing the cognate antigen. However, understanding why only some T cells are capable of killing, and identifying mechanisms that can improve killing has remained elusive. Methods To identify molecular and cellular mechanisms that can improve T-cell killing, we utilized integrated high-throughput single-cell functional profiling by microscopy, followed by robotic retrieval and transcriptional profiling. Results With the aid of mathematical modeling we demonstrate that non-killer CAR T cells comprise a heterogeneous population that arise from failure in each of the discrete steps leading to the killing. Differential transcriptional single-cell profiling of killers and non-killers identified CD137 as an inducible costimulatory molecule upregulated on killer T cells. Our single-cell profiling results directly demonstrate that inducible CD137 is feature of killer (and serial killer) T cells and this marks a different subset compared with the CD107a pos (degranulating) subset of CAR T cells. Ligation of the induced CD137 with CD137 ligand (CD137L) leads to younger CD19 CAR T cells with sustained killing and lower exhaustion. We genetically modified CAR T cells to co-express CD137L, in trans, and this lead to a profound improvement in anti-tumor efficacy in leukemia and refractory ovarian cancer models in mice. Conclusions Broadly, our results illustrate that while non-killer T cells are reflective of population heterogeneity, integrated single-cell profiling can enable identification of mechanisms that can enhance the function/proliferation of killer T cells leading to direct anti-tumor benefit. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Text Encoded Conversion
  • XML
  • Save / Share this Record
  • Send to Email