Staying on the CAR-Twagon: Where and how cell therapies still retain their promise

“Hope for a one-time cure”. How the promise of cell therapy still burns bright — read on!

Source: European Molecular Biology Laboratory

In my previous post, I provided an overview of the cell therapy landscape and summarized some of the key challenges faced by both autologous and off-the-shelf cell therapy modalities.

To recap, the cell therapy space has lost some of its initial luster. Despite impressive clinical outcomes achieved in hematological malignancies, the manufacturing complexities surrounding autologous options, and the somewhat-disappointing clinical findings of off-the-shelf candidates, have curbed some of the early enthusiasm of investors and bio-entrepreneurs alike.

Here, I will look at the other side of the coin, and outline key advances that suggest the future of cell therapy remain bright — and how the next few years will determine whether we will realize the initial dream of cell therapy: “hope for a one-time cure”.

Looking to the future, we can find key innovations surrounding manufacturing, therapeutic design, and computational applications.

Looking to the future: Investing and innovating in manufacturing early to develop scalability by approval

Despite heterogeneity across all three cell therapy avenues, there has been a strong recognition that success in this space is contingent as much on manufacturing excellence as scientific innovation. TAT (turnaround time) reliability and minimal variability in cell material have been brought to the fore as physicians seek consistency in their patients’ cell therapy treatments. With a newfound focus on cGMP and production capabilities, cell therapy upstarts are rife with early investments in manufacturing to develop sufficient scalability and reproducibility, even by those with candidates still in preclinical stages. Such investments can be divided into:

• In-house capacity expansion — Recent investments in new facilities from TCR2 therapeutics, Nkarta, Artiva, and Lyell.
• External collaborations — Licensing and partnerships with external CMOs seen from Catamaran Bio and Triumvira .

Separately, a host of cell-therapy upstarts are focusing on innovations in the cell manufacturing process — new techniques to improve the efficiency, scalability, and precision of cell engineering and expansion. Some such examples are:

• SQZ biotech’s cell squeeze enhances material transduction into cells without genetic editing, enabling rapid cargo insertion into multiple cell types.
• Catamaran Bio’s TAILWIND platform utilizes a non-viral transposon system for CAR gene delivery into cells, for greater transfer efficiency and larger payloads.
• Poseida’s proprietary booster molecule improves cell expansion post CAR-engineering, expanding possible doses from a single manufacturing run.
• Ori Biotech aims to standardize, automate, and digitize the manufacturing process to develop reliable and reproducible processes typically plagued with variability and unpredictability.
• 64x Bio’s VectorSelect platform maps and develops cell lines optimized for viral vector production, addressing one of the key bottlenecks of the cell therapy supply chain.
• Berkeley Light’s Opto Cell Therapy workflow screens for desired phenotypes within heterogenous T-cell samples, reducing cell-to-cell variability and deterministically producing T-cells with the desired functionality.

Given recent manufacturing investments, the benefits of scalability, reliability, and consistency will only come to fruition post commercial approval in the next 3–5 years.

Such investments can assure scientists, physicians, and patients alike that these biotechs will have the necessary manufacturing prowess to hit the ground running once given the FDA green-light.

Looking to the future: Novel approaches combining protein and genetic engineering to enhance efficacy/safety

To circumvent hurdles around target specificity, off-target and immune toxicity, and therapeutic persistence, several novel developments combine both protein and genetic engineering innovations. Such “armored cell therapy” approaches have demonstrated promising preclinical data. However, given the continued challenges around translating clinical outcomes from preclinical models, a “wait and see” approach is recommended until further clinical results are published.

Below is a small selection of such innovations in development. It should be noted that biotechs pursuing novel approaches often select well-characterized targets (e.g., CD19, BCMA), as they seek to minimize biology risk while demonstrating platform viability / proof-of-concept.

“Programmable” cell therapy with administration of external agents for greater control over CAR-T activation, specificity, and safety

• Obsidian Therapeutics’ on-switch for TILs regulates IL-15 expression in its CytoTIL15 engineered cells via the administration of acetazolamide.
• UniCAR approaches from Cellectis (UCART19, UCART123) and Avencell (UniCAR and RevCAR) leverage soluble adaptors interacting between the CAR and the antigen, allowing for ‘switchable’ activation and specificity of cell therapy with improved scalability given simplified manufacturing of the flexible soluble adaptors. However, initial Phase 1 results from UCART19 had underwhelming results — demonstrating proof-of-concept for the UniCAR approach but with 2 patient deaths, 27% PFS at 6 months, and 55% OS.
• 2seventy bio’s DARIC (Dimerizing Agent Regulated Immunoreceptor Complex) platform regulates CAR formation (and therefore, anti-tumor activity) by engineering these receptors to be split into two components, only to combine upon the introduction of small molecule rapamycin.
• The dual-switch from Bellicum’s GoCAR technology — comprising of an activatory rimiducid-inducible iMC switch and inhibitory caspaCIDe iC9 — has shown to enhance proliferation, persistence, and safety in preclinical studies.

Innovative “gene circuits” create cell therapies that dynamically respond to its biological environment (e.g., an immunosuppressive TME):

• Logic-gated CAR from SentiBio has shown encouraging preclinical results, utilizing an “OR” gate (i.e., targeting both CD33 and FLT33) to improve tumor specificity and “NOT” gate (i.e., anti-EMCN, an antigen commonly expressed on healthy cells) to inhibit off-tumor activity.
• The Tmod platform from A2 biotherapeutics and iCAR technology from ImmPact Bio exploits the loss of genes in tumor cells by engineering both an activating receptor (targeting the tumor antigen) and a blocking receptor (targeting an antigen typically expressed on healthy not tumor cells), hoping to reduce off-target toxicity.
• Multiplexed gene modulation from Refuge Biotech conditionally activates or inhibits (via CRISPRa/CRISPRi) several genes within the engineered cell to enhance anti-tumor activity (e.g., PD-1 downregulation) based on extracellular signals that live only within the TME.
• A cellular switch developed by Catamaran Bio, dubbed “signal trap”, senses the immunosuppressive cytokine TGF-beta within the TME and responds by expanding the cell’s anti-tumor activity.

Receptor protein engineering for enhanced tumor specificity

• Adaptimmune’s SPEAR T-cell technology allows for more precise targeting of peptide fragments on antigens, demonstrated clinically with a 34% ORR from its afami-cel TCR-T candidate for sarcomas (soon to be submitted for FDA approval).
• Multi-antigen targeting by Fate Therapeutics’ FT596 (CD16/CD19) and ImmPact Bio’s bispecific (CD19/CD20) combines multiple receptors, to improve tumor specificity and reduce potential antigen escape.
• AffyImmune focuses on fine-tuning the affinity of its anti-ICAM-1 CAR, with the goal of improving specificity and reducing T-cell exhaustion.
• In a gene-therapy focused application, Moderna’s collaboration with Autolus for its CAT binder showcases the power of enhanced affinity binders to improve targeting on the surface of lipid nanoparticles, a potential drug delivery vehicle.

“Armoring” the cell via genetic engineering to reduce immunogenicity and enhance persistence

• Sana Biotech’s hypoimmune cell technology seeks to reduce immunogenicity of its engineered cell therapies by knocking out MHC I and II while overexpressing CD47, to enable improved persistence of its allogeneic candidates.
• Caribou Therapeutics’ chRDNA approach knocks out key genes (e.g., PD-1 for CB-010, B2M for CB-011) to enhance persistence and blunt T- and NK-mediated rejection by the host.
• Genetic reprogramming by Lyell seeks to overcome T-cell exhaustion by overexpressing the c-Jun gene.
• Nkarta has engineered a membrane-bound IL-15 to enhance its’ NKX019 candidate’s activity within the TME, while knocking out CISH and CD70 to provide resistance to TME inhibition.
• Genenta’s Temferon candidate comprises of an engineered tumor-associated myeloid cell that is designed to release IFN-alpha inside tumors

Such innovations in cell therapeutic design promise to transform efficacy/safety of future candidates, particularly for applications against solid tumors.

Looking to the future: Potential for computational approaches

More recently, there appears to be a novel sub-sector in the cell therapy landscape that could hold tremendous potential to unlock the space: computational (i.e., AI/ML-driven) applications to cell therapy.

For context, AI/ML applications in drug discovery has exploded over the last few years, demonstrating value creation at each step of the drug discovery value chain, from uncovering novel targets to ‘drugging the undruggable’ to enhancing patient selection for clinical trials. By utilizing an iterative ‘build-test-learn’ R&D cycle that continuously improves the underlying platform, new upstarts in this area are aiming to create candidates with higher probability of success and streamlined time to commercialization at lower cost (given lower failure rate).

Within cell therapy, such AI/ML applications are still nascent, but initial exemplars demonstrate the potential to create more optimized, more reproducible, and less costly cellular therapeutics. Exemplars include:

• More optimized — Identifying a cancer’s unique signature: The majority of AI/ML applications have been dedicated to discovering novel antigenic combinations that cell therapies can target to improve efficacy and specificity. Leading the pack is Repertoire Immune Medicine’s DECODE platform characterizing TCR-antigen pairs, with research ongoing that uses in silico screens to identify multi-antigen cancer signatures.
• More optimized — Receptor optimization: Select applications focus on computationally optimizing the binding and other properties (e.g., intracellular signaling) of CAR and TCR constructs, largely by testing a library of variants per receptor domain in closed loop systems. Single-domain optimization has been demonstrated by the ‘CARpooling’ technique from Arsenal Bio, while Serotiny aims for multi-domain optimization across the entire receptor.
• More reproducible, less costly — Manufacturing optimization: Novel approaches utilize AI/ML within manufacturing. Cellino, for example, is seeking to automate cell production, systematize cell sorting, and enhance stem cell control by utilizing computational techniques to phenotypically identify ‘sub-par’ cells and remove them via a laser-based editing system, parallel processing this across thousands of cell samples. Such approaches should minimize donor cell variability (key to allogeneic approaches) and improve manufacturing efficiency and cost
• More optimized, less costly — Smarter genetic engineering: Cutting-edge biotechs have begun to empower their full R&D process with an underlying AI/ML platform. For example, Modulus Therapeutics endeavors for ‘Convergent Design’ NK cells utilizing gene editing and screening capabilities to test a variety of potential engineering changes on a cell candidate’s properties (e.g., durability, specificity). AI/ML is leveraged to analyze these results and recommend the next set of smarter, more optimized genetic edits, iteratively creating a ‘build-test-learn’ flywheel that continuously improves their potential candidates.

To summarize, despite certain manufacturing and therapeutic challenges that remain today, cell therapy continues to hold great potential in transforming patient outcomes. Indeed by some accounts, CAR-T has also demonstrated a potential cure, with promising innovations in manufacturing, therapeutic design, and computational applications coming to the fore. What remains clear, however, is that the next 3–5 years, driven by the developments of today, will be critical in realizing the promise — or peril — of this cutting-edge space.

[Disclaimer: The views above represent my own, and not my current or previous employers. They reflect my understanding of the space, but may not be the latest, most comprehensive coverage of all companies, scientific advances, or clinical results.]