synNotch-programmed iPSC-derived NK cells usurp TIGIT and CD73 activities for glioblastoma therapy

Kyle B. Lupo, Xue Yao, Shambhavi Borde, Jiao Wang, Sandra Torregrosa‐Allen, Bennett D. Elzey, Sagar M. Utturkar, Nadia A. Lanman, MacKenzie McIntosh, Sandro MatosevicNature Communications 2024 · DOI

TL;DR

Researchers have engineered induced pluripotent stem cell-derived natural killer (iPSC-NK) cells to overcome glioblastoma's immunosuppressive environment. By "usurping" the inhibitory TIGIT-CD155 interaction and simultaneously blocking CD73, these engineered NK cells effectively eliminate tumors and reprogram the tumor microenvironment, offering a promising new allogeneic therapy.

What Problem Does This Paper Address?

Glioblastoma (GBM) is an aggressive brain cancer notoriously resistant to current immunotherapies. A major hurdle is the highly immunosuppressive tumor microenvironment (TME), which cripples the effectiveness of immune cells like Natural Killer (NK) cells. One key player in this suppression is the TIGIT receptor on NK cells interacting with its ligand CD155 on GBM cells. While targeting TIGIT alone has shown limited success, this paper investigates a more sophisticated approach: not just blocking this interaction, but actively hijacking it to promote anti-tumor activity. Furthermore, GBM also produces immunosuppressive adenosine via CD73, which further dampens immune responses. The authors aim to address the combined challenge of TIGIT/CD155-mediated immunosuppression and CD73-driven adenosine production in GBM.

Methods

The core innovation lies in the engineering of iPSC-derived NK cells using a synNotch system. This system acts as a "decoy" receptor:

  • TIGIT-SynNotch Construct: The engineered NK cells express a construct that includes a TIGIT extracellular domain. When this TIGIT domain binds to CD155 on GBM cells, it triggers a conformational change within the synNotch receptor.
  • Localized CD73 Blockade: This conformational change activates a downstream signaling cascade, leading to the transcription and release of a single-chain variable fragment (scFv) antibody that specifically targets and blocks CD73.
  • "Usurping" TIGIT Signaling: Crucially, this synNotch activation redirects the inhibitory TIGIT-CD155 signal towards an activating pathway, effectively turning an immunosuppressive interaction into an anti-tumor one.

The researchers also utilized:

  • iPSC-derived NK cells: These cells offer an allogeneic source, meaning they can be derived from healthy donors and used in multiple patients, overcoming limitations of autologous (patient-derived) NK cells.
  • CRISPR/Cas9: Used to generate knockout GBM cell lines for CD155 and CD73 to validate the specificity of the engineered NK cells.
  • In vitro assays: Cytotoxicity, degranulation, and cytokine production assays were used to assess NK cell function.
  • In vivo models: Orthotopic and flank xenograft models in immunocompromised mice were used to evaluate tumor growth inhibition, immune cell infiltration, and survival.
  • Immunohistochemistry (IHC) and Flow Cytometry: Standard techniques for analyzing immune cell populations and marker expression in tumors and blood.
  • TCGA Data Analysis: Used to correlate CD155 and CD73 expression with patient outcomes.

Datasets

  • TCGA-GBM (The Cancer Genome Atlas - Glioblastoma Multiforme) dataset: Used for transcriptional gene expression analysis to assess the correlation between CD155 (PVR) and CD73 (NTSE) expression and patient survival.
  • Human GBM cell lines: Various cell lines (e.g., GBM43 WT, GBM43 CD155KO, GBM43 CD73KO, GBM43 CD155/CD73KO, GBM10, SJ-GBM2, U87MG, GL261, HCN-2) were used for in vitro experiments.
  • Patient-derived xenograft models: Orthotopic GBM models were established using GBM43-luciferase cells in NSG mice.
  • Immunocompetent mouse models: GL261 intracranial and flank glioma models in C57BL/6 mice were used for in vivo studies with antibody blockade.

Key Insights and Ideas

  • Beyond Simple Blockade: The most significant insight is that simply blocking TIGIT isn't enough. The strategy of "usurping" the TIGIT-CD155 interaction by redirecting its signaling pathway towards activation is a novel and powerful approach.
  • Dual Targeting is Crucial: Co-targeting both the TIGIT/CD155 axis and CD73 is significantly more effective than targeting either alone, highlighting the complex interplay of immunosuppressive mechanisms in GBM.
  • Localized Therapy: The synNotch system ensures that the CD73 blockade is localized to the NK cell-tumor interface, minimizing off-target effects and maximizing therapeutic impact.
  • Reprogramming the TME: The engineered NK cells don't just kill tumor cells; they also recruit beneficial immune cells (like T cells) and reduce suppressive cells (M2 macrophages), fundamentally altering the tumor's defense mechanisms.
  • Allogeneic Potential: Using iPSC-derived NK cells provides a scalable and potentially off-the-shelf solution, overcoming the challenges of obtaining and engineering autologous NK cells.

Key Results

  • GBM Exhibits High CD155 and CD73 Expression: TCGA data revealed that both CD155 and CD73 are significantly upregulated in GBM and their co-expression is a negative prognostic factor.
  • Engineered iPSC-NK Cells are Effective: TIGIT.SynNotch.aCD73 engineered iNK cells demonstrated significantly higher lysis of GBM43 WT cells compared to WT iNK cells in vitro.
  • Specificity of Engineered NK Cells: The engineered NK cells showed specific killing of CD155+CD73+ GBM cells, with minimal impact on cells lacking these targets.
  • Tumor Growth Inhibition in Vivo: In orthotopic GBM models, treatment with engineered iNK cells led to a drastic reduction in tumor growth and significantly improved survival compared to PBS or WT iNK cells.
  • Enhanced Immune Infiltration: In vivo studies showed increased infiltration of NK cells and CD8+ T cells into tumors treated with the engineered NK cells, along with a downregulation of immunosuppressive M2 macrophages.
  • Reduced Exhaustion Markers: Engineered NK cells harvested from tumors showed reduced expression of exhaustion markers like PD-1 and TIM-3, suggesting improved functionality.

Limitations and Future Work

  • In Vitro vs. In Vivo Efficacy: While in vitro functional assays showed modest improvements in GBM killing, the in vivo tumor growth was dramatically reduced. The authors suggest this discrepancy might be due to TME-specific factors not present in in vitro systems.
  • Potential for Antigen Escape: While the engineered NK cells don't solely rely on target expression for lysis, the authors acknowledge the potential for antigen escape mechanisms within GBM. However, their approach aims to mitigate this by targeting multiple pathways.
  • Allogeneic Compatibility: The authors mention that further modifications to generate fully allogeneic and patient-compatible, hypoimmunogenic iPSCs through HLA editing are possible for broader clinical application.
  • Clinical Translation: The study presents a preclinical demonstration, and future work will focus on translating these findings into clinical immunotherapies for GBM. The authors believe this platform has the potential for allogeneic adoptive transfer immunotherapies.

Summary updated

Abstract

Severe heterogeneity within glioblastoma has spurred the notion that disrupting the interplay between multiple elements on immunosuppression is at the core of meaningful anti-tumor responses. T cell immunoreceptor with Ig and ITIM domains (TIGIT) and its glioblastoma-associated antigen, CD155, form a highly immunosuppressive axis in glioblastoma and other solid tumors, yet targeting of TIGIT, a functionally heterogeneous receptor on tumor-infiltrating immune cells, has largely been ineffective as monotherapy, suggesting that disruption of its inhibitory network might be necessary for measurable responses. It is within this context that we show that the usurpation of the TIGIT - CD155 axis via engineered synNotch-mediated activation of induced pluripotent stem cell-derived natural killer (NK) cells promotes transcription factor-mediated activation of a downstream signaling cascade that results in the controlled, localized blockade of CD73 to disrupt purinergic activity otherwise resulting in the production and accumulation of immunosuppressive extracellular adenosine. Such "decoy" receptor engages CD155 binding to TIGIT, but tilts inhibitory TIGIT/CD155 interactions toward activation via downstream synNotch signaling. Usurping activities of TIGIT and CD73 promotes the function of adoptively transferred NK cells into intracranial patient-derived models of glioblastoma and enhances their natural cytolytic functions against this tumor to result in complete tumor eradication. In addition, targeting both receptors, in turn, reprograms the glioblastoma microenvironment via the recruitment of T cells and the downregulation of M2 macrophages. This study demonstrates that TIGIT/CD155 and CD73 are targetable receptor partners in glioblastoma. Our data show that synNotch-engineered pluripotent stem cell-derived NK cells are not only effective mediators of anti-glioblastoma responses within the setting of CD73 and TIGIT/CD155 co-targeting, but represent a powerful allogeneic treatment option for this tumor.
BibTeX
@inproceedings{lupo2024synnotch,
  author = {Kyle B. Lupo and Xue Yao and Shambhavi Borde and Jiao Wang and Sandra Torregrosa‐Allen and Bennett D. Elzey and Sagar M. Utturkar and Nadia A. Lanman and MacKenzie McIntosh and Sandro Matosevic},
  title = {{synNotch-programmed iPSC-derived NK cells usurp TIGIT and CD73 activities for glioblastoma therapy}},
  booktitle = {Nature Communications},
  year = {2024},
  doi = {10.1038/s41467-024-46343-3},
}
Figure