Advances in Immunoengineering

In vivo generation of Chimeric Antigen Receptor (CAR)—expressing immune cells represents a novel development in the field of immunotherapy. Preclinical and early clinical data supports the efficacy and safety of viral and RNA-based in vivo CAR engineering technologies, heralding a novel era. This progress opens a new avenue towards multi-modal engineering of the immune system, to address medical needs intractable with alternate treatment modalities.

In vivo generation of CAR-T cells offers a scalable alternative to ex vivo therapy but remains limited by delivery efficiency, toxicity, and durability. We describe an in vivo CAR-T platform using CD8-targeted LNPs to transiently deliver T cell-restricted CAR and modular effector mRNAs. This enables selective CAR expression, tunable local immune amplification, enhanced antigen-dependent cytotoxicity, and robust functional activity, establishing a programmable framework for immune-cell engineering.

Alaya.bio has developed a proprietary platform combining the gene transfer efficiency of lentivectors with advanced biodegradable polymer and lipid engineering to improve targeting, specificity, and reduced immunogenicity. Using scalable GMP-compliant bioprocesses, Alaya produces transduction-deficient lentivectors functionalised with shielding polymers and targeting ligands. In NSG mice engrafted with human leukaemia cells, a single intravenous injection controlled tumor burden and prolonged their survival for over 28 days without acute toxicity. CAR-T cells were detected in bone marrow at necropsy. This versatile platform shows strong potential for universal in situ CAR-T therapies and broader cell reprogramming applications beyond immuno-oncology, including genetic disorder treatments.

SleepingBeauty100X represents a cost-effective alternative to lentiviral vectors when generating CAR-T. Current regulation requires the absence of transposase within infused CAR-T products, which was demonstrated for standard manufacturing procedures validated in clinical trials such as LION-1 and CARAMBA. However, protein-stability displays a concern for fast-manufacturing or in vivo approaches, harbouring a theoretical risk of re-transposition which may cause genetic instability, or modification of non-T cells after infusion. Thus, we have developed a modified SB-Transposase with a significantly reduced half life, and compared it's efficacy and safety to SB100X. 

CAR-T cells are living drugs capable of long-term proliferation and persistence in patients. However, once administered, they act autonomously, limiting clinicians’ ability to modulate activity and increasing risks such as cytokine release syndrome, ICANS, and on-target, off-tumour toxicities without precise, real-time control. To address this, we engineered drug-responsive protein switch systems and demonstrated that they enable reversible functional control of CAR-T cells in vivo via oral drug administration.

Orthogonal split-antibody-based Interleukin-12 receptors exhibit high affinity for the inert small molecule chelator DOTAM. DOTAM-inducible Interleukin-12 receptors (D12Rs) result in attenuated agonism which is potentiated during concurrent CAR signaling. CAR D12R T cells showed enhanced cytotoxicity under repeated antigen exposure and mediated durable in vivo efficacy in human xenograft models. This effect is mediated by increased tumour-infiltration of a rapidly proliferating progenitor-rich T cell subset.

Targeting intracellular antigens, particularly tumour specific neoantigens in solid tumours, with adoptive T cell therapies comes to the forefront of clinical interest. However, different limitations should be considered, such as the loss of neoantigens, the modification of antigen peptide presentation, tumour heterogenicity, the immunosuppressive activity of the tumour environment, graft-versus-host disease and exhaustion of T cells. This perspective highlights three interdependent pillars that together define the next generation of adoptive immune therapies: T memory cell–based strategies, control of graft-versus-host disease (GVHD), and modulation of the tumour microenvironment (TME). These three interlocking strategies form a coherent, conceptual framework for future immunotherapy development. Their integration linking cellular longevity, immune tolerance, and microenvironmental control offers a roadmap toward effective, safe, and durable cancer immunotherapy. Looking forward, progress will depend on integrating existing strategies rather than pursuing innovations in isolation. Clinical success will require harmonising Tm enrichment, immune tolerance, and TME reprogramming into unified therapeutic protocols.

This talk will showcase how Coding Bio integrates high-throughput functional immune screening with machine learning to design next-generation, modular T cell engagers (TCEs). By training models on immune activation rather than binding affinity, we’re redefining how efficacy and safety are optimised in silico. I’ll discuss our lead trispecific program now advancing toward Phase 1, our transition into solid tumours, and how we’re building the foundation model for immunotherapy design.

All approved T cell engagers engage the TCR complex through CD3ε, but this is not the only therapeutically exploitable site. We developed VHH bispecifics targeting a novel quaternary epitope at the TCRα/β constant region interface, providing pan-αβ T cell redirection with full TRBC1/TRBC2 allele coverage and inherent human/cynomolgus cross-reactivity. In matched cytotoxicity assays, these anti-TCR bispecifics achieve picomolar potency despite moderate monovalent affinities, raising testable questions about how TCR complex geometry influences T cell activation.