As of October 2026, in vivo gene editing, where the editor is infused into the body rather than applied to cells in a dish, has produced one completed phase 3 trial, several early-stage cholesterol and rare-disease datasets, and one serious safety setback. The picture is encouraging and uneven. Most of what follows comes from company announcements, and only some of it has passed peer review.

The lead example is hereditary angioedema, where a single infusion of an editing therapy cut attacks sharply against placebo. The cautionary example is a transthyretin amyloidosis program that was paused after a patient died of liver injury. Both belong in the same honest summary.

What “in vivo” means, and why it is harder

Ex vivo therapies, such as Casgevy, remove a patient’s cells, edit them in the lab and return them. That route allows quality checks before anything goes back in. In vivo therapies skip the lab step. The editor, usually packed in a lipid nanoparticle that the liver readily takes up, is infused and the cells inside the patient do the rest. A single mistake cannot be sorted out beforehand, and the edit is permanent. Regulators and clinicians treat the early data accordingly.

The upside is reach. Many diseases are not accessible through blood stem cells, and some patients cannot tolerate the chemotherapy-style conditioning that ex vivo treatments require. A one-time infusion that reprograms liver cells could be simpler to deliver at scale, which explains why so many companies have moved in this direction over the past few years. The same simplicity concentrates the risk in a single step that cannot be undone, which is why the liver findings in several of these programs draw such close attention.

Scorecard: the main programs

Program Target and editor Latest data I could verify Evidence level
Lonvo-z (Intellia) Hereditary angioedema; CRISPR-Cas9 delivered by lipid nanoparticle Phase 3 HAELO, 80 patients: 87% fewer attacks than placebo over weeks 5 to 28 Randomized; company release plus NEJM publication
Nex-z (Intellia) Transthyretin amyloidosis Phase 3 trials paused after liver injury and a death; enrollment since reinitiated Company reports; trial ongoing
VERVE-102 (Lilly) PCSK9 in familial hypercholesterolemia; base editor Phase 1b interim, 35 adults: LDL-C down 9% to 62% by dose Open-label, company-reported, NEJM publication
CTX310 (CRISPR Therapeutics) ANGPTL3 for high LDL and triglycerides; CRISPR-Cas9 Phase 1, 15 adults: LDL down nearly 50%, triglycerides about 55% Peer reviewed (NEJM), very short follow-up
BEAM-302 (Beam) Alpha-1 antitrypsin deficiency; base editor Phase 1/2, 29 treated: mutant protein down 79% to 84% at the doses highlighted Company release

Hereditary angioedema: the first phase 3 result

Intellia announced the HAELO phase 3 topline on April 27, 2026. The trial randomized 80 people aged 16 and older, 52 to a single 50 mg dose of lonvo-z and 28 to placebo. Over weeks 5 to 28, the monthly attack rate averaged 0.26 with the therapy and 2.10 with placebo, an 87 percent reduction. In the treated group 62 percent were attack-free and off preventive therapy for the six-month window, against 11 percent on placebo (Intellia release via BioSpace).

Fuller data followed on June 13 at the EAACI congress in Istanbul, with a simultaneous New England Journal of Medicine paper. By the company’s account, no serious adverse events occurred in the treated arm, and the side effects were infusion reactions, headache, fatigue, back pain and upper respiratory infections, all mild or moderate (Intellia release via BioSpace).

As of its second-quarter report, Intellia said it was still waiting for the FDA to accept its rolling biologics license application, expected in the second half of 2026, and was aiming for a US launch in the first half of 2027 if approved (Intellia, Q2 2026 results). Nothing is approved yet.

The setback: nex-z and a liver death

Intellia’s nex-z targets transthyretin amyloidosis. In late 2025 a man in his early 80s, dosed on September 30 in the cardiomyopathy trial MAGNITUDE, developed severe liver injury and died. Intellia reported lab values consistent with a pattern called Hy’s Law, and the company said the treating physician pointed to complicating conditions. Dosing was paused in MAGNITUDE and its sister trial MAGNITUDE-2, and the FDA placed clinical holds (GEN, November 7, 2025).

In late January 2026 the hold on MAGNITUDE-2, in the nerve-damage form of the disease, was lifted with enhanced liver monitoring and a larger planned enrollment of roughly 60 patients (Clinical Trial Vanguard, January 29, 2026). By the August Q2 update, enrollment had been reinitiated in both trials. Intellia also reported a genomic analysis of more than 600 patient samples in which the highest liver enzyme elevations clustered in carriers of one particular HLA allele, and said it was discussing the finding with regulators.

That last detail is a hypothesis about risk, not a settled explanation. It does suggest what a safer path might look like: screening for the variant before dosing.

Cholesterol: two routes to one big market

Heart disease drives interest because it affects far more people than any rare disease. Two approaches have reported human data.

PCSK9 base editing. Lilly’s May 25, 2026 release described an interim analysis of 35 adults with heterozygous familial hypercholesterolemia or premature coronary disease, dosed in six cohorts from 0.3 to 1.0 mg/kg. PCSK9 fell by 51 percent to 88 percent, and LDL-C by 9 percent at the lowest dose up to 62 percent at the highest. Median follow-up was about nine months at the February 27, 2026 cutoff, with 15 participants followed for at least a year. No treatment-related serious adverse events were reported. The response was not perfectly dose-ordered, since the 0.7 mg/kg group (33 percent) fell below the 0.6 mg/kg group (45 percent). Lilly plans a phase 2 study by the end of 2026 (Eli Lilly release, May 25, 2026).

ANGPTL3 knockout. CTX310 was reported at the American Heart Association meeting in 2025 and in the New England Journal of Medicine (doi 10.1056/NEJMoa2511778). Fifteen adults received a single infusion across five dose levels. LDL cholesterol fell by nearly 50 percent overall and triglycerides by about 55 percent, and at the top dose both fell by roughly 60 percent. Two participants (13 percent) had serious adverse events, and the investigators described the sample as small with short follow-up (American College of Cardiology summary).

Alpha-1 antitrypsin deficiency

Beam’s BEAM-302 corrects the PiZ mutation that causes this lung and liver disease. In its March 25, 2026 update, with a February 10 data cutoff, 29 patients had been treated. At 60 mg, mean total AAT was about 16.1 micromolar and the harmful Z-AAT protein was down 84 percent. At 75 mg, the figures were 14.4 micromolar and 79 percent. Beam chose 60 mg for pivotal development. One patient in a repeat-dose group had a Grade 4 rise in the liver enzyme ALT, which the company described as asymptomatic and untreated (Beam Therapeutics release, March 25, 2026). Updated data were scheduled for a September 8 presentation at the European Respiratory Society congress, which I could not review.

Where regulators and the calendar stand

Lonvo-z holds US orphan drug and regenerative medicine advanced therapy (RMAT) designations, UK Innovation Passport status and EU PRIME designation, according to Intellia’s June release. Designations speed interaction with regulators; they do not predict approval. The rolling FDA submission began in April 2026, and acceptance of the full application is still pending as of the company’s August update.

Lilly’s PCSK9 program has a Fast Track designation for lowering LDL-C in people with high lifetime cardiovascular risk. Beam has set a pivotal dose. These are milestones on a long road, and the cholesterol programs in particular will need much larger and longer studies before anyone can weigh benefit against risk for a drug aimed at millions of people rather than thousands.

It is also worth remembering the commercial backdrop. The first approved editing medicine, Casgevy, reported $43 million in first-quarter 2026 revenue according to Vertex, a reminder that approval and uptake are separate hurdles (Vertex, Q1 2026 results).

How to read all of this

  1. Check who ran the study and who wrote the headline. Most numbers above come from the sponsors.
  2. Look at the control. Only the angioedema trial compared against placebo. The others are open-label, so every comparison is against the patient’s own baseline.
  3. Look at follow-up. Months of data cannot speak to decades, and an edit is meant to last decades.
  4. Watch the liver. The liver takes up lipid nanoparticles, which is why it is the easiest organ to reach, and also why enzyme spikes recur across programs.

Nothing here is medical advice, and none of these investigational therapies is approved for use.

Sources