The enemy is an archive. Chronic hepatitis B affects more than 250 million people worldwide, according to Nature. When the virus infects a liver cell, it repairs its genome into a tight loop called cccDNA, a free-floating mini-chromosome that sits in the nucleus and keeps printing viral RNA. One Lyon hepatologist calls it "a genomic archive of the virus." It is the reason the disease is chronic.

Today's standard drugs, nucleos(t)ide analogues, stop the virus from copying itself, but they do not touch the archive. Stop the pills and the virus comes back. Fewer than 10% of people treated for at least ten years can safely stop, Nature reports. Tune Therapeutics puts the rate of functional cure on current therapy at roughly 1% per year.

A second hiding place. HBV also stitches fragments of itself into the host's own chromosomes. That integrated DNA keeps making viral proteins that confuse the immune system, and the molecular damage it causes is linked to liver cancer.

Why not just cut? CRISPR-style nucleases can chop viral DNA, and several programs are trying. But cutting DNA in a liver already studded with integrated viral fragments means making many breaks at once, and breaks are exactly what cancer biology feeds on. Nature notes that cutting therapies "can increase cancer risks for people with hepatitis B." The scalpel is sharp in the wrong direction.

The felt, not the saw. Epigenetic editing leaves the sequence alone. It uses a CRISPR protein whose cutting blades have been deliberately disabled ("dead" Cas9), guided by RNA to a chosen stretch of DNA and bolted to enzymes that attach methyl groups. Methylated DNA gets packed away and goes quiet. The idea borrows from nature: a rare few chronic HBV patients silence the virus on their own through exactly these repressive marks.

Continued overleaf →

01 · Tune: seventeen months of quiet

The first human proof arrived at the European liver congress (EASL) in May. Tune Therapeutics' TUNE-401 is an mRNA-plus-guide payload in a lipid nanoparticle that the liver translates into an HBV-targeting silencer. It methylates both the cccDNA archive and the integrated viral DNA.

In a Phase 1b/2a study across New Zealand, Hong Kong and Moldova, doses ran from 0.2 to 0.85 mg/kg, with a multi-dose arm of up to three infusions. Every participant at dose levels 2 through 4 showed repression of HBV biomarkers. Among HBeAg-negative patients at the two highest doses, 4 of 7 lost detectable pregenomic RNA, the direct readout of an active archive. Among HBeAg-positive patients, 3 of 5 lost HBeAg. After a single dose, repression has held for up to 17 months.

That CEO spent years running R&D at Gilead, where, he told The Scientist, his team screened millions of small molecules looking for one that could silence cccDNA safely. None could. The safety profile so far is mild to moderate infusion reactions and brief liver-enzyme bumps that resolve within days. Tune says Phase 2 could start as early as late 2026.

02 · nChroma: a two-front war, now peer-reviewed

On 21 September, nChroma Bio and collaborators at Milan's San Raffaele Telethon Institute published the preclinical case for CRMA-1001 in Nature Biomedical Engineering. The design is the same: dead Cas9, guide RNA, methylating effectors, lipid nanoparticle. It attacks both hiding places at once.

In mouse models the company reported more than a 3-log (1,000-fold) drop in surface antigen, with undetectable antigen and viral DNA in 90% of mice at the top dose. In monkeys, the same effector and delivery system silenced a stand-in liver gene, PCSK9, for more than a year after one dose, with only minimal, temporary side effects. The first human received CRMA-1001 in January, in a trial in Hong Kong and New Zealand.

03 · The machinery is now a platform

Look at that monkey experiment again. nChroma proved safety with a cholesterol gene, using the same silencer and the same nanoparticle. Only the guide changed. For an engineer, that is the headline: one chassis, many addresses. The virus is the first customer, not the last.