Two-gene knockout in CAR-T cells strengthens tumour control

Two-gene knockout in CAR-T cells produced an effect that could not readily be predicted from either gene modification alone. Among 2,485 gene pairs examined, researchers identified a combination that substantially improved the antitumour performance of engineered immune cells. The benefit extended beyond their ability to kill cancer cells: it also involved changes in how the immune cells generated and used energy.

The findings, published in Nature on 4 November 2026, point to a potentially valuable approach to improving cell-based cancer immunotherapy by targeting interacting regulatory pathways rather than individual genes.

Why disabling one gene may not be enough

CAR-T therapy relies on T lymphocytes, immune cells capable of recognising and destroying other cells. Scientists engineer these lymphocytes to express a synthetic receptor that recognises a particular molecule on the surface of cancer cells.

Once equipped with this receptor, the modified T cells can identify and attack tumour cells carrying the corresponding target.

CAR-T therapies have achieved remarkable clinical results in certain blood cancers. Their performance against many solid tumours, however, remains much less consistent.

The tumour microenvironment creates multiple obstacles. Immune cells face limited nutrients, unfavourable metabolic conditions and signals that suppress their activity. Prolonged stimulation can also undermine their ability to survive, proliferate and maintain effective tumour killing.

One way to address these limitations is to remove genes that constrain T-cell activity.

But immune cells are controlled by interconnected regulatory networks. Disabling one gene may produce a useful improvement, while disabling two can generate an outcome that is difficult to anticipate from their individual effects.

The researchers set out to identify precisely such interactions.

Screening 2,485 gene pairs inside living organisms

The team used CRISPR gene editing to disrupt pairs of genes in human CAR-T cells.

They investigated 71 genes associated with T-cell regulation, examining 2,485 combinations.

A central feature of the experiment was that the screening was carried out in mice bearing experimental tumours. The researchers assessed which genetically modified CAR-T cells became enriched within tumour tissue.

This allowed the selection process to take place in a living tumour environment, where immune cells encounter challenges that cannot be fully reproduced in conventional cell cultures.

Promising combinations were subsequently tested in separate experiments against unmodified controls and CAR-T cells carrying single-gene knockouts.

The strongest candidate was the NR4A1–SOCS3 combination.

These genes regulate different aspects of T-cell biology.

NR4A1 encodes a transcriptional regulator involved in controlling gene expression. Among its functions, it has been associated with programmes that limit T-cell performance during persistent stimulation.

SOCS3 encodes a protein that restrains signalling downstream of certain cytokines, the molecular messengers immune cells use to coordinate their responses.

Removing both genes produced a synergistic effect: the benefit exceeded what would be expected by simply adding together the effects of the two individual knockouts.

The broader screen revealed an important feature of this biology.

Across the tested combinations, the outcome of knocking out two genes was only weakly related to the sum of their individual effects. Knowing which single-gene modifications worked best was therefore insufficient to reliably identify the strongest combinations.

The most promising targets were not necessarily those that looked most effective in isolation.

A metabolic advantage, not just a stronger attack

Further experiments revealed that NR4A1–SOCS3 knockout affected more than conventional markers of T-cell activation.

Gene-expression profiling, including single-cell analysis of T cells recovered from experimental tumours, revealed changes in pathways involved in cellular metabolism.

Mitochondria emerged as an important part of the response.

These structures help cells convert nutrients into usable energy. Their function matters particularly for immune cells that must continue working under demanding conditions.

The researchers found that the double-knockout CAR-T cells showed improved mitochondrial activity and metabolic fitness.

That distinction matters because successful immunotherapy requires more than a single encounter with a cancer cell.

T cells must survive repeated stimulation, retain their functional capacity and continue eliminating tumour cells over time. The ability to sustain these demands depends partly on their metabolic state.

In laboratory experiments involving repeated exposure to tumour targets, the double-knockout CAR-T cells maintained stronger antitumour activity than comparison cells.

The advantage also appeared in animal experiments.

Across several tumour models, NR4A1–SOCS3-deficient cells achieved stronger and more durable tumour control. The researchers tested models of solid tumours and haematological cancers, including systems established using patient-derived malignant cells.

In several comparisons, the combined knockout outperformed conventional CAR-T cells, either single-gene knockout and selected genetic modifications previously proposed to improve T-cell therapy.

The team also examined safety-related measures, including inflammatory signals, tissue changes and genomic alterations. The enhanced antitumour activity did not produce an obvious deterioration in the safety measures assessed in these experimental systems.

That finding has important limits.

This was a preclinical study, not a clinical trial. Although human T cells and patient-derived cancer cells were used, no patients received the experimental double-knockout CAR-T therapy as part of this work.

The results therefore do not establish clinical efficacy or safety. Human T cells growing and acting in laboratory mice cannot fully reproduce the interactions that take place in patients with an intact immune system and complex tumour biology.

The study nevertheless provides a different way of thinking about genetic engineering for immunotherapy.

Instead of identifying one inhibitory gene at a time, researchers can systematically search for pairs whose combined disruption produces unusually large benefits.

For NR4A1 and SOCS3, that interaction appears to influence both immune-cell activity and the metabolic machinery needed to sustain it.

Neither knockout alone adequately predicted the strength of their combined effect. The key discovery came from studying the interaction between them.