CRISPR against cancer cells: an enzyme that shreds DNA

CRISPR against cancer cells works differently in this study: not as a gene editor, but as a programmable self-destruct switch. The Cas12a2 enzyme recognizes RNA carrying a cancer-associated mutation and then begins cutting chromatin throughout the same cell.

That is a major departure from conventional CRISPR editing. A standard editor is guided to one defined stretch of DNA to make a localized change. Cas12a2 first searches for a matching RNA sequence and, once activated, switches into a destructive mode that damages DNA at many sites and triggers the cell’s response to severe genomic injury.

Recognize the cell instead of repairing the mutation

Cas12a2 comes from a bacterial antiviral defence system. For an infected bacterium, destroying its own genome can be an effective last resort: the cell dies, but viral replication is stopped with it. The researchers sought to convert this natural emergency mechanism into a programmable tool for eliminating tumour cells.

Cas12a2 is paired with a guide RNA complementary to the intended target. Before the enzyme encounters the matching transcript, its destructive activity remains off. Once the sequences match, Cas12a2 becomes activated and cuts chromatin, the complex of DNA and its associated proteins. In mammalian cells, this produced numerous double-strand DNA breaks, cell-cycle arrest and cell death.

The strategy is particularly relevant to mutations that are difficult to attack with conventional drugs. Many medicines work by fitting into a defined pocket on a protein. Mutant tumour-suppressor proteins often lack such accessible pockets, and restoring their lost function has proved difficult. A prominent example is TP53, which is altered in roughly 40–50% of cancers.

Rather than attempting to repair mutant p53, the team programmed Cas12a2 to recognize RNA transcribed from the mutant gene and use it as a marker identifying the cell to be destroyed.

What happened in cells and mice

The system was tested against several cancer-associated RNA signatures in cultured human cells. These included a characteristic deletion in EGFR and several point mutations in TP53. With an appropriately selected guide RNA, Cas12a2 strongly restricted the growth of cells carrying the target mutation, while cells expressing the corresponding wild-type sequence were much less affected under the experimental conditions.

This distinction is central to the method. Cas12a2 does not cut only the mutant genomic site. The mutant RNA acts as the key that activates widespread DNA destruction across the cell. Selectivity therefore depends not on the location of every DNA break, but on how accurately the system distinguishes tumour RNA from the normal transcript.

The researchers then moved to two mouse cancer models. In a MYC-driven liver tumour model, Cas12a2 messenger RNA and its guide were delivered in lipid nanoparticles. With nine animals in each group, targeting MYC markedly reduced the area occupied by tumour tissue at the experimental endpoint compared with a non-targeting guide.

A second model used human lung-cancer cells carrying a TP53 mutation. The early-stage experiment included eight mice per group. Delivery of Cas12a2 programmed against the mutant transcript reduced the tumour’s bioluminescent signal. An effect was also observed when treatment began later, although the tumours were not completely eliminated.

These findings establish proof of concept: an RNA molecule can serve as an intracellular identity signal that triggers the death of the cell producing it. They do not yet establish a clinically usable treatment.

Why this is still far from a therapy

Delivery is the most immediate obstacle. To control a tumour, Cas12a2 and its guide RNA would need to reach a sufficient fraction of cancer cells without becoming activated in essential healthy tissues. Because the enzyme causes extensive genomic damage after activation, the required safety margin is unusually demanding.

Cas12a2 also cannot recognize every possible sequence. Target selection is constrained by a neighbouring activation motif, and performance depends on the abundance and location of the target RNA inside the cell. A clinically important mutation might therefore be technically unsuitable for targeting or might not be expressed in every tumour cell.

Tumours are also heterogeneous. Some cells may lack the selected mutation, produce too little of the relevant RNA or escape delivery altogether. Those survivors could repopulate the tumour. Longer studies will also be needed to assess immune responses, systemic toxicity, lasting genomic consequences and the possibility that similar transcripts in normal cells could activate the enzyme unintentionally.

The work represents an unusual shift in strategy. CRISPR is not being used to repair a gene or inhibit a single protein. It reads a molecular signature and, when the signature matches, turns the cell’s own genome into the route to its destruction. Whether that process can be controlled with sufficient completeness and safety remains a question for further preclinical development.