nChroma Bets a No-Cut Editor Can Silence Hepatitis B

nChroma Bio dosed its first hepatitis B patient on 28 January with CRMA-1001, a CRISPR tool that silences viral DNA with methyl tags and never cuts the genome.

Mouse and monkey results behind that infusion were published on 21 September. A Phase 1/2 trial is listed at hospitals in Hong Kong, Auckland, London and Lyon. The Boston company is betting those tags can shut the virus’s hidden DNA archive and let people stop the daily pills that rarely clear it.

Daily Pills Leave a Viral Archive in the Liver

Oral drugs such as tenofovir and entecavir can slow cirrhosis and cut the risk of liver cancer, and most people who start them must stay on them for life. They block new copies of the virus. They do not erase the template that keeps making those copies.

Once inside a liver-cell nucleus, hepatitis B forms a small circular DNA, a mini-chromosome that virologist and hepatologist Fabien Zoulim at the University of Lyon calls “a genomic archive of the virus.” Stop the pills, and that archive can restart a severe rebound. Zoulim says less than 10% of people treated for at least ten years can cease daily medication without that risk. That stop-without-rebound figure is a different measure from the roughly 1% who meet a formal functional cure on pills alone.

The virus can also stitch pieces of itself into the cell’s own DNA. Those inserts can keep making surface antigen, blunt the immune response, and help seed liver cancer. John Tavis, a molecular virologist at Saint Louis University in Missouri, called the effect of a single experimental injection in animals “quite impressive.”

This is really where we need to go to end up being able to address this pandemic of HBV around the world.

John Tavis, molecular virologist, Saint Louis University

A functional cure, as late-stage trials now define it, means hepatitis B surface antigen is gone and viral DNA stays undetectable for at least 24 weeks after a finite course, with no need to restart pills. That is the finish line nChroma has written on its label. It has not yet run that race in people.

A Dead Cas9 That Only Stamps Methyl Tags

CRMA-1001 is a lipid nanoparticle packed with RNA. The RNA tells liver cells to build a fusion protein and a guide strand that tows it to a conserved stretch of the hepatitis B genome, a site the company says matches more than 94% of genotypes A through H.

WHAT THE EDITOR IS BUILT FROM

• Dead Cas9: A version of Streptococcus pyogenes Cas9 that can bind DNA but has had its cutting activity switched off.
• KRAB domain: A transcriptional repressor that helps shut gene reading at the bound site.
• Methylation pair: DNMT3A and DNMT3L enzymes that deposit methyl groups on the viral DNA, the chemical tags that keep the genes off.

The same particle is meant to mark both the free-ranging mini-chromosomes and the fragments already woven into human DNA. In the mouse work published on 21 September, a single dose produced more than a 3 log10 drop in viral biomarkers, tied to new methylation on hepatitis B DNA. After three doses, up to 90% of animals had undetectable surface antigen and viral DNA at six months. In human liver cells, surface antigen and e-antigen fell 98% versus controls. Monkeys showed only a brief rise in liver enzymes, and only at the highest dose tested. Profiling found no detectable unintended targets in the human genome.

Study co-author Angelo Lombardo, a molecular biologist at the San Raffaele Telethon Institute for Gene Therapy in Milan, said the animal package “supported clinical advancement of this approach.” He co-founded the company and has no official role there. nChroma declined to comment on the paper. At a liver meeting in November 2025, the firm also said the editor worked alongside entecavir, with deeper DNA suppression than either agent alone, and that the same chemistry silenced a surrogate liver gene in monkeys for more than a year.

Why Cutting That Archive Can Raise Cancer Risk

Nuclease editors go after the same archive with scissors. Precision BioSciences is already infusing PBGENE-HBV, an ARCUS nuclease designed to cut covalently closed circular DNA and integrated viral DNA, and has reported tenfold drops in circular-DNA transcripts on liver biopsy after two infusions at 0.4 mg/kg. Cutting DNA that is stitched into a hepatocyte chromosome can scramble the host genome. That is the cancer worry that pushed nChroma toward tags instead of breaks. The trade is plain: a cutter can destroy the template, and a methyl editor can only silence it, and only for as long as the tags remain.

GSK’s 19% Cure Rate Is Now the Bar

The pills nChroma wants to replace already have a late-stage challenger that does not edit DNA at all. GSK’s bepirovirsen is an antisense drug licensed from Ionis that binds viral RNA. In two phase 3 trials, B-Well 1 and B-Well 2, it was given as weekly shots for 24 weeks on top of standard nucleoside treatment. People who held a response could stop those pills at week 48. The primary readout was a functional cure at week 72.

B-Well 1 recorded a functional cure in 127 of 650 patients on bepirovirsen (20%) and in none of 328 on placebo. B-Well 2 recorded it in 106 of 570 patients (19%) and in none of 286 on placebo. GSK says functional cure rates remain low, typically only 1% on nucleoside analogues. Among people who started with surface antigen at or below 1,000 IU/mL, the functional-cure gap versus placebo was 24.6 percentage points in B-Well 1 and 27.8 points in B-Well 2. Tony Wood, GSK’s chief scientific officer, said those rates are a first for the disease. The company planned global filings from the first quarter of 2026 and has Fast Track status from the US FDA.

That is the number a one-shot silencer now has to beat, and it was generated in more than 1,800 patients across 29 countries, not in mice. Bepirovirsen still leaves the circular DNA in place. It buys a finite course for about one in five people. nChroma’s wager is that tagging the archive itself can push past that ceiling without a double-strand break.

FIVE BETS ON THE SAME VIRUS

Jeff Walsh, nChroma’s chief executive, called existing drugs “rarely curative” when the company said the first patient has been dosed. He is right about the pills. He is now running behind both a phase 3 antisense and another epigenetic editor that has already read out in people.

Human Data From a Rival Epigenetic Silencer

Tune Therapeutics presented TUNE-401 at the European Association for the Study of the Liver congress in early June, calling it the first clinical evidence of epigenetic silencing in hepatitis B. The early trial enrolled about 20 people at sites in New Zealand, Hong Kong and Moldova, with single-dose cohorts and a three-dose group. John McHutchison, Tune’s chief executive, said four people became biomarker negative and that reductions reached 98% in one patient over 250 days. The firm said it saw dose-dependent drops across surface antigen, pregenomic RNA, e-antigen and core-related markers, which it reads as direct silencing of the circular DNA. It plans a larger phase 2 study as early as late 2026.

Edward Gane, professor of medicine at the University of Auckland, is principal investigator of Tune’s New Zealand arm. He is also the listed Auckland contact for nChroma’s trial. Man-Fung Yuen at Queen Mary Hospital in Hong Kong is nChroma’s Hong Kong contact and an investigator on Precision’s cutting-editor study. The same two hepatologists will see a nuclease, Tune’s silencer and nChroma’s silencer in their own clinics.

The development of a finite HBV cure has long been stalled by a central challenge: how to safely and specifically target and permanently silence cccDNA.

Edward Gane, hepatologist, University of Auckland, on Tune’s TUNE-401

nChroma’s published package is still animal data. Tune’s is already a human signal, even if it is small and still on background pills. That gap is the part of the bet the 21 September paper does not close.

Four Hospitals Began Infusing the Editor

nChroma itself is a 2024 construction. Chroma Medicine and Nvelop Therapeutics merged on 11 December 2024 and closed an oversubscribed $75 million financing led by Cormorant Asset Management, ARCH Venture Partners, Atlas Venture and Newpath Partners, with Walsh as chief executive. The lead cargo was CRMA-1001, aimed at chronic hepatitis B and hepatitis D.

FROM MERGER TO FIRST INFUSION

1. 11 December 2024: Chroma and Nvelop combine as nChroma Bio and raise $75 million.
2. 16 December 2025: Hong Kong issues a certificate for a first-in-human trial of CRMA-1001.
3. 28 January 2026: The first participant receives an intravenous infusion; New Zealand and UK trial authorizations are in hand.
4. 15 June 2026: The public registry lists four recruiting hospitals after its last update.
5. 21 September 2026: The mouse, cell and monkey study is published.

The registry entry, NCT07200193, describes an open-label single and multiple ascending dose study of safety, pharmacokinetics, pharmacodynamics and efficacy in adults with chronic hepatitis B. Up to four dose levels are planned. Participants stay on antiviral therapy during dosing. Sites listed as recruiting are Queen Mary Hospital in Hong Kong (Yuen), New Zealand Clinical Research in Auckland (Gane), King’s College Hospital in London, and Hospices Civils de Lyon, where Zoulim is the listed contact. Estimated primary completion is 31 December 2032.

Lombardo called the move into people “a huge step” and “a realization of many, many years of work.” Walsh said in January that early clinical data could arrive in late 2026 or early 2027. That readout, if it comes, will be safety and biomarker movement on background pills, not a B-Well-style off-treatment cure rate.

Most of the 240 Million Never Get the Pills

The World Health Organization estimates that 240 million people were living with chronic hepatitis B in 2024, about 2.9% of the world’s population, with 0.9 million new infections that year. The infection caused an estimated 1.1 million deaths, mostly from cirrhosis and hepatocellular carcinoma. GSK notes that chronic hepatitis B accounts for about 56% of liver cancer cases.

THE 2024 BURDEN IN FOUR FIGURES

• On treatment: Only 10 million people, 4.3% of those living with the infection, were taking antivirals.
• Diagnosed: 65 million people, 27%, knew their status.
• Eligible: Updated 2024 treatment guidance puts more than 50% of those infected in line for pills, depending on setting.
• Deaths: 1.1 million in 2024, while new infections fell 32% from 2015.

WHERE THE INFECTIONS SIT

• Western Pacific: 102 million people, the largest regional total.
• Africa: 64 million people; the region accounted for 68% of new infections in 2024.
• South-East Asia: 43 million people.
• Eastern Mediterranean, Europe, Americas: 16 million, 9.7 million and 5 million.

A lipid-nanoparticle CRISPR infusion given at Queen Mary Hospital, Auckland, King’s and Lyon is a long way from the clinics that hold most of that burden. Cheap daily pills already fail to reach 95.7% of the people who have the virus. A one-course genetic medicine will not close that gap on a Boston trial map. If the tags work, the next problem is who can get the drip.

Can the Methyl Tags Survive Liver Cell Turnover?

The scientific objection is not that the mouse drops are fake. It is that a methyl mark is a software change, and liver cells keep dividing. If the tags fade, the archive can be read again. If the nanoparticle misses a fraction of hepatocytes, those cells can keep feeding virus into the blood. Coverage of every infected cell is still a delivery problem, not a paper problem. nChroma’s monkeys only flared liver enzymes at the top dose, which is a start on safety, not a proof that silencing holds for a decade in a human liver full of integrated DNA laid down in childhood.

GSK has already shown that a 24-week RNA drug can take about one in five people off pills. Tune has already dropped several viral markers in an early human cohort with a different epigenetic editor. Precision is already cutting the circular DNA in people, cancer risk and all. nChroma’s $75 million bet is that a dead Cas9 with two silencing domains, given as one or a few infusions, can outlast all three without breaking chromosomes.

The first person received that infusion on 28 January. The registry’s primary completion date is 31 December 2032. The methylation has to hold in dividing human livers long before then, in the biomarker data Walsh said could arrive in late 2026 or early 2027.

Disclaimer: This article is news reporting and analysis of published research, company statements and a registered clinical trial. It is informational only and is not medical advice, a treatment recommendation, or a judgment on whether any person should join a study or stop prescribed antiviral medication. Readers who have hepatitis B, or who are considering experimental therapy, should consult a qualified physician or hepatologist before making any change to their care. Figures, trial sites, enrolment status and regulatory plans reflect the named sources as of the dates cited and may change as the studies proceed.