What are TCRs, and what can we do with them?

T-cell receptors (TCRs) are found on T-cells - a type of white blood cell (lymphocyte) that are produced in the bone marrow and mature in the thymus. Together with B-cells, which produce antibodies, T-cells make up the adaptive immune system .

TCRs and antibodies both recognise antigens, which are proteins produced by infectious agents like viruses and bacteria or aberrant, harmful cells such as cancer. This ability empowers the adaptive immune system to learn and adapt to threats to our health and specifically target them – unlike the innate immune system , which can only respond in a pre-programmed, nonspecific way.

However, TCRs and antibodies interact with antigens in different ways. Antibodies recognise and bind directly to antigens expressed on the surface of a cell or pathogen. TCRs recognise antigen peptides selected and processed from within the cell and presented on the surface by the major histocompatibility complex (MHC), often referred to as the human leukocyte antigen (HLA) complex in humans.

As a result, TCRs can also recognise intracellular antigens that are normally concealed within cells but revealed as peptide fragments by the HLA complex. This opens up a whole new universe of targets that previous antibody-based therapeutic approaches could not explore.

The rise of T cell therapies

Once a TCR engages its target peptide:HLA complex it activates the T-cell, which starts to rapidly multiply and attack, as well as secreting cytokines that stimulate other cells in the immune system. This triggers a targeted and coordinated immune response against a pathogen or cancer.

These characteristics have made T-cells and TCR-based therapeutics a promising new modality, particularly for cancer.?

The first use of T-cells for cancer therapy was back in 1988, when patients with metastatic melanoma were treated with infusions of T-cells that had been extracted from their tumours and expanded in the lab. In over 50% of the patients , the tumours started to shrink.

It wasn’t until nearly three decades later that the US FDA approved the first genetically engineered T-cell therapies: tisagenlecleucel (Kymriah ) to treat B-cell acute lymphoblastic leukaemia, and axicabtagene ciloleucel (Yescarta ) to treat large B-cell lymphoma.?

These are both CAR T-cell therapies , which work by genetically engineering T-cells to express an antibody-based chimeric antigen receptor (CAR) that recognises a target antigen and triggers the immune response.?

More recently, we’ve seen the rise of TCR T-cells , which use engineered TCRs rather than CARs to target intracellular tumour antigen peptides. Approaches such as afami-cel demonstrate particular efficacy in solid tumours, where CAR-T cell therapies have faced previous challenges.

From T-cell therapies to bispecific immune engagers

While these T-cell-based therapies are promising, they are challenging to develop, manufacture and deliver to patients . To solve this problem, researchers have turned to soluble bispecific T-cell engagers (TCEs) .?

Most TCEs developed to date are double-headed antibodies that bring the patient’s own immune cells directly into contact with cancer cells and trigger the immune response. One antibody domain recognises a target antigen, and another binds to the CD3 protein on the surface of T-cells and initiates activation.?

Then, in January 2022, Immunocore announced FDA approval of tebentafusp-tebn (KIMMTRAK ) – a TCR-based bispecific immune engager targeting the intracellular antigen gp100 presented by HLA-A2 – for the treatment of uveal melanoma.

The approval of tebentafusp-tebn paves the way for high potency bispecific immune engagers based on TCRs, which can access a much wider range of intracellular targets than conventional antibody-based therapeutics to treat previously undruggable diseases. However, TCR discovery and engineering is not a simple process.

The slow process of TCR discovery and engineering

TCR discovery and engineering can be a very lengthy and technically demanding process. Unlike well-established antibody discovery methods , traditional TCR discovery relies on labour-intensive techniques. This usually involves collecting TCRs from donors, carefully stimulating them in vitro, and close monitoring and analysis to find the rare few TCRs that specifically target a desired antigen.

But simply finding a TCR that recognises a target antigen is not enough. Naturally occurring TCRs have evolved to maintain a balance between recognising foreign antigens and avoiding self-antigens. This is achieved through multiple specific interactions between the TCRs on T cells and the peptide-MHC complexes on target cells. As a result, TCRs have inherently low binding affinity to their target.?

In order to develop effective soluble bispecific TCR therapeutics, natural TCRs need significant further engineering to increase their affinity for the target antigen. At the same time, low cross-reactivity with healthy tissues must be ensured to avoid side effects. Striking this delicate balance can be a very complex and time-consuming process.

Etcembly’s approach to TCR discovery and optimisation

To accelerate the TCR discovery and engineering processes, Etcembly has developed the EMLy?machine learning platform, which uses generative AI and structural modelling to discover and engineer TCRs to picomolar affinity. We then use these TCRs to build bispecific immune engagers (ETCers) for treating conditions such as cancer and autoimmune diseases (Figure 1).

Figure 1: Etcembly’s T-cell engaging receptor (ETCer) interacting with an antigen:HLA complex on the target cell via the TCR and activating CD3 receptor on the T-cell via an antibody-based effector module. The two components are linked together by a silenced Fc domain.

EMLy? scans publicly available TCR sequence and structural data, along with proprietary data from our in-house single-cell sequencing microfluidics technology. This comprehensive approach identifies TCR repertoires likely to bind to a target, dramatically narrowing down the pool of candidates to allow for a more targeted approach.

We then test the TCRs in the lab and feed the resulting data back into EMLy?. Through a combination of structural modelling, generative AI large language models, and binding affinity modelling, EMLy? learns from the lab results and refines the candidate TCRs. The iterative process of testing in the lab, feeding back data into EMLy? and returning to the lab allows us to discover and optimise TCRs for specific targets quickly and effectively.

This innovative approach led to the development of ETC-101, the world's first AI generated bispecific T-cell engager with picomolar affinity . ETC-101 targets PRAME , an antigen associated with many types of cancer, including melanoma, lung cancer, and ovarian cancer.?

This high affinity TCR was developed in just 11 months — a significant reduction compared to the conventional 2+ year timeline for TCR discovery and engineering pipelines. As EMLy? continues to evolve and learn, future programs are expected to be even faster and more efficient.

TCRs for all

TCR-based therapeutics hold huge promise for treating a range of diseases, including cancers and autoimmune conditions. But the promise of these treatments will only be realised if we can transform the complex, time-consuming and labour-intensive process of TCR discovery and engineering.

Our groundbreaking approach aims to overcome the barriers holding back the discovery and engineering of TCR candidates, accelerating the development of high-quality, potent and safe immunotherapies and bringing powerful new treatments to the patients who need them.