Tunable Universal OR-gated CAR T cells for AML

Menna Siddiqui, Jingyao Chen, Madeline Loffredo, Seung‐Hee Lee, Han Deng, Yongshuai Li, Nelia Leemans, Tim Lu, Brian S. Garrison, Marcela Guzmán Ayala, Nicholas Frankel, Wilson W. WongbioRxiv 2024 · DOI

TL;DR

This paper introduces a novel "OR-gated" Chimeric Antigen Receptor (CAR) T cell therapy platform (SUPRA) designed to overcome the challenges of treating Acute Myeloid Leukemia (AML). By targeting multiple antigens (CD33 and FLT3) in a tunable and programmable way, this approach aims to improve efficacy, prevent antigen escape, and reduce off-tumor toxicity, offering a potentially safer and more effective treatment for AML patients.

What Problem Does This Paper Address?

Acute Myeloid Leukemia (AML) is a particularly aggressive blood cancer with a poor prognosis. A major hurdle in treating AML with CAR T-cell therapy is the heterogeneity of the disease, meaning cancer cells can have different surface markers. This heterogeneity can lead to antigen escape, where some cancer cells survive treatment because they don't express the targeted antigen, causing relapse. Furthermore, many antigens targeted in AML are also present on healthy hematopoietic stem cells (HSCs), leading to dangerous on-target, off-tumor toxicity (like myeloablation). Conventional CAR T-cell therapies, which typically target a single antigen, struggle to address both these issues simultaneously. This paper aims to solve this by developing a CAR system that can simultaneously target multiple antigens in a controlled manner.

Methods

The core innovation is the Split, Universal, and Programmable (SUPRA) CAR platform. This system decouples the antigen-binding part from the CAR T-cell itself. It consists of:

  • zipCAR: A universal CAR expressed on the T cell, featuring a "zipper" domain for binding.
  • zipFv: Soluble adapter molecules, each containing a complementary "zipper" domain and a single-chain variable fragment (scFv) that binds to a specific antigen.

This "split" design allows for:

  • OR-gate logic: The T cell is activated if either antigen A or antigen B is present. This is achieved by combining zipFvs for CD33 and FLT3.
  • Tunability: The concentration of zipFvs can be adjusted to control T cell activation and specificity.
  • Multiplexed targeting: Multiple zipFvs can be used to target several antigens simultaneously.

The researchers designed and tested various zipFv variants targeting CD33 and FLT3. They used NFAT-GFP Jurkat T cells as an activation reporter system to screen for on-target and off-target activity. In vitro cytotoxicity assays were performed using engineered K562 cell lines and human AML cell lines (MV4-11, THP-1). Bio-layer interferometry (BLI) was used to characterize the binding affinities of the zipFvs. Finally, in vivo xenograft mouse models were employed to validate the system's efficacy and safety.

Datasets

The study utilized several cell lines for experiments:

  • HEK293T cells: For lentivirus production.
  • Jurkat T cells: Engineered with NFAT-GFP reporter for activation screening.
  • K562 cell lines: Engineered to express single or dual antigens (CD33, FLT3, or both) to test SUPRA CAR specificity and efficacy.
  • MV4-11 and THP-1 cell lines: Human AML cell lines with varying antigen expression levels.
  • Nalm6 cell line: A B-cell precursor ALL line engineered to express CD33 or FLT3 (or both) for in vivo studies.
  • Primary human CD3+ T cells: Isolated from peripheral blood for functional assays.
  • Primary human CD34+ enriched hematopoietic stem and progenitor cells (HSPCs): Obtained from bone marrow or cord blood to assess off-target toxicity.

No new benchmark datasets were introduced; rather, existing cell lines and primary cells were used to evaluate the novel SUPRA CAR system.

Key Insights and Ideas

  • Decoupling antigen binding from the CAR: This fundamental design choice is key to SUPRA's flexibility and tunability, allowing for independent optimization of the CAR and its targeting adapters.
  • OR-gate logic for AML: Targeting multiple antigens with an "OR" gate is a smart strategy for AML, where antigen expression can vary, helping to prevent antigen escape and ensure T cell activation against a broader range of cancer cells.
  • ZipFv dose control is crucial: The concentration of zipFvs directly impacts T cell activation. This allows for fine-tuning the therapy to maximize efficacy against cancer cells while minimizing toxicity to healthy cells.
  • Differential binding affinities matter: The varying affinities of CD33 and FLT3 zipFvs influence their competitive binding and dose-dependent behavior, highlighting the need for careful characterization and optimization of each adapter.
  • SUPRA offers a higher activation threshold: Compared to conventional CARs, SUPRA showed lower basal activation, suggesting increased specificity and reduced risk of off-target effects, though potentially requiring higher doses for potent anti-tumor activity.

Key Results

  • Effective dual-antigen targeting: The SUPRA OR-gate system successfully targeted both CD33 and FLT3 on AML cell lines, demonstrating comparable efficacy to conventional CARs in eliminating cancer cells.
  • Tunable dose response: By adjusting zipFv concentrations, researchers identified optimal doses that maximized T cell activation against cancer cells. For example, 100 ng of FLT3 zipFv and 25 ng of CD33 zipFv were found to be effective.
  • Differential affinity characterized: Bio-layer interferometry revealed that the CD33 zipFv has a significantly higher binding affinity (lower IC50) than the FLT3 zipFv, explaining observed differences in dose-dependent activity and potential competition.
  • In vivo efficacy demonstrated: In a xenograft mouse model using Nalm6 cells, treatment with both FLT3 and CD33 zipFvs significantly reduced tumor burden and increased survival compared to single-target treatments or no treatment.
  • HSPC sparing achieved: In vitro studies showed minimal toxicity to healthy hematopoietic stem and progenitor cells (HSPCs) at therapeutic doses, indicating a promising therapeutic window that spares essential healthy cells.

Limitations and Future Work

  • Activation threshold vs. potency: While the higher activation threshold of SUPRA enhances specificity, it might require higher doses to achieve potent anti-tumor responses. The authors suggest that dose optimization can compensate for this.
  • Competition and affinity: The differential affinities of zipFvs can lead to competition, impacting efficacy. Further work is needed to fully understand and leverage these interactions for optimal combinatorial dosing.
  • Complexity of AML heterogeneity: While targeting CD33 and FLT3 is a significant step, AML is a complex disease with other potential target antigens. The SUPRA platform's flexibility allows for the addition of more zipFvs to target additional antigens.
  • "AND" and "NOT" logic: The authors mention prior work on "AND" and "NOT" logic gates within the SUPRA system and suggest combining these with the OR gate for even greater specificity.
  • Clinical translation: While promising, the study is preclinical. Future work will involve further validation and clinical trials to assess the safety and efficacy of SUPRA CAR T cells in human patients.

Summary updated

Abstract

Abstract/Summary Acute myeloid leukemia (AML) is a hematopoietic malignancy characterized by antigen heterogeneity and poor prognosis. A potential therapeutic approach to address this heterogeneity is targeting multiple surface antigens to prevent antigen escape and relapse. Chimeric antigen receptor (CAR) T cells are an adoptive cell therapy that have demonstrated remarkable clinical success in the treatment of B cell malignancies, and many efforts are underway to adapt them to myeloid malignancies. To tackle the heterogeneity of AML, logically targeting multiple antigens through an “A OR B” gated CAR circuit would be desirable. Here we combined FLT3 antigen targeting with the well characterized CD33 myeloid marker as a combinatorial OR gate approach using our split, universal, programmable (SUPRA) CAR platform. The split platform affords tunability over activation levels and multiplexed targeting that cannot be achieved through a tandem bispecific approach. We systematically characterized the specificity and sensitivity of different SUPRA CAR adapters against each target individually and in combination against a panel of target cell lines. Our results demonstrate that this CAR system can effectively target two antigens with equivalent efficacy to conventional CARs while reducing the engineering burden associated with designing CAR T cells against multiple antigens. Furthermore, we can characterize an effective dose range where off-target cytotoxicity against hematopoietic stem and progenitor cells is minimized. With the recent clinical advances in universal CAR designs, our SUPRA OR gate has the potential to provide an effective and safer solution to treating AML.
BibTeX
@misc{siddiqui2024tunable,
  author = {Menna Siddiqui and Jingyao Chen and Madeline Loffredo and Seung‐Hee Lee and Han Deng and Yongshuai Li and Nelia Leemans and Tim Lu and Brian S. Garrison and Marcela Guzmán Ayala and Nicholas Frankel and Wilson W. Wong},
  title = {{Tunable Universal OR-gated CAR T cells for AML}},
  year = {2024},
  doi = {10.1101/2024.04.13.589307},
  howpublished = {bioRxiv},
}
Figure