From a bacterial defence to an approved medicine

CRISPR is a bacterial immune system repurposed as a tool. A guide molecule finds a chosen sequence of DNA and an enzyme cuts it. The cell repairs the cut, and the repair can disable a gene or, with more work, change it. It made gene editing cheap and accessible enough that a graduate student could do in a week what had previously taken a specialist laboratory a year.

Eleven years after the foundational paper, a therapy based on it was approved. By the standards of medicine that is extraordinarily fast.

Why sickle cell was first

Everything about the disease made it the easiest possible target. It is caused by a single well-characterised mutation. The affected cells are blood stem cells, which can be removed from the body, edited in a dish where delivery is not a problem, checked, and put back. And the therapeutic strategy did not even require fixing the mutation: switching a gene back on to produce fetal haemoglobin, which the body makes before birth and then stops, is enough.

Take any of those away and the problem becomes much harder. That is the honest frame for every other disease.

Delivery is the whole remaining problem

Editing cells outside the body works. Editing cells inside the body means getting the machinery to the right tissue, into the cells, without triggering an immune response and without editing everything else.

The liver is comparatively easy, because lipid nanoparticles naturally accumulate there, and the first in-body successes have been liver diseases. Muscle, brain, lung and eye are progressively harder. Viral vectors work and carry immunogenicity and cargo-size limits. Nanoparticles are safer and less precise about where they go.

Every headline about editing a new disease is really a headline about a delivery method, and that is the part to read carefully.

The tools got more precise

Cutting both strands of DNA and letting the cell repair is blunt and can produce unintended changes. Base editing changes a single letter without a double-strand break. Prime editing writes a short specified sequence, which in principle can correct most known pathogenic variants. Epigenetic editing switches genes on or off without altering the sequence at all, which is reversible and avoids the permanence question entirely.

These are substantially safer tools, and they are in early clinical use.

The price problem, which is not a detail

Roughly two million dollars for a one-time treatment. The therapy is manufactured per patient, in a process that takes weeks, with a hospital stay involving chemotherapy to clear the bone marrow first. That is the cost structure of a bespoke biological product.

The population that most needs this treatment lives largely in sub-Saharan Africa and India. A cure that exists and is unavailable to the majority of patients is a real outcome of the current model, not a rhetorical point, and it is why in-body editing, which would avoid the cell processing entirely, is the goal that matters most for access.

The line nobody has crossed properly

Editing somatic cells affects one patient. Editing embryos affects every descendant. The 2018 case of edited embryos in China resulted in a prison sentence and near-universal condemnation, and the scientific consensus remains that heritable editing is not acceptable with current knowledge of off-target effects.

That consensus holds and is not permanent, and the sequence in which it erodes is worth watching: serious disease first, then enhancement arguments.

What to watch

Watch in-body editing trial results outside the liver. Watch for a manufacturing approach that brings per-patient cost down by an order of magnitude. And watch base and prime editing trials, which are the safer tools and therefore the ones that determine how wide this goes.