What exactly is an in vivo CRISPR therapy, and how is it different from other gene therapies?
In vivo CRISPR therapy means the gene-editing machinery is delivered directly into the patient's body, typically via an intravenous infusion, to edit cells inside the body — most often liver cells. This is different from ex vivo approaches, where cells are removed from the patient, edited in a lab, and then infused back. The delivery vehicle is often a lipid nanoparticle (LNP), the same technology used in mRNA COVID-19 vaccines, which encapsulates the CRISPR components (Cas9 protein or its mRNA and a guide RNA) and safely transports them to target cells [3][10]. Once inside the cell, the CRISPR system makes a precise cut in the DNA to permanently disable a disease-causing gene or correct a mutation.
What does the human data show about safety and durability so far?
The most direct human evidence comes from a phase 1 trial of NTLA-2001 for transthyretin amyloidosis, a life-threatening disease where misfolded protein accumulates in nerves and the heart. In six patients, a single infusion caused only mild side effects and led to a dose-dependent reduction in the target protein: 52% at the lower dose and 87% at the higher dose by day 28 [11]. This is the first-in-human study of an in vivo CRISPR therapy, and while the follow-up is short, it demonstrates that the approach can work safely in people. The authors note that preclinical studies showed durable knockout of the TTR gene after a single dose [11].
For other conditions, human trials are just beginning. VERVE-101, which targets the PCSK9 gene to lower LDL cholesterol, has completed primate studies and was set to start human trials in mid-2022 [1][5]. In those primate studies, a single dose led to 68% lower LDL cholesterol that lasted for a full year, with only a temporary, self-resolving rise in liver enzymes [1][5]. Importantly, there was no evidence of the edit being passed to offspring — sperm samples from treated male primates showed no editing, and none of 436 offspring from treated female mice carried the edit [1]. This addresses a key safety concern about germline editing.
What are the current limits of the evidence, and what risks remain?
The biggest gap is long-term human data. The NTLA-2001 trial only reported 28-day results, and while the primate data for VERVE-101 extends to 476 days, that's still not a human lifetime [1][11]. The durability of effect over many years is unknown. For example, in the related field of AAV-based gene therapy for hemophilia, the FDA required 2-year data before considering approval because the therapeutic effect declined after 12 months in some patients [3]. This highlights that even promising early results need longer follow-up.
Another concern is off-target editing — unintended changes to other parts of the genome. The studies here used comprehensive genomic assessments and found no signs of unintended edits in cell and animal models [4], but detecting rare off-target events in humans is challenging. The authors of a review note that 'better research tools are needed for detecting off-target effects' [3]. Additionally, the delivery method itself can cause temporary side effects: transient liver enzyme elevations were seen in multiple primate studies [1][2][4], and while they resolved without intervention, this requires monitoring in humans. Finally, the ex vivo CRISPR therapies (where cells are edited outside the body) for sickle cell disease and beta-thalassemia have shown durable benefits for up to 6 years, but these involve a different risk profile, including the need for chemotherapy-like conditioning [6][7][8][9].
About These Sources
This answer is built on 11 peer-reviewed studies — published from 2021 to 2025, 5 from 2024 or later, 11 in Q1 journals, collectively cited 1,745 times — selected as the most relevant from 15 studies that passed quality screening, drawn from 59 papers retrieved from a database of over 500 million.
Sources used in this answer
Efficacy and Safety of an Investigational Single-Course CRISPR Base-Editing Therapy Targeting <i>PCSK9</i> in Nonhuman Primate and Mouse Models
In non-human primates, a single dose of VERVE-101 (in vivo CRISPR base editing targeting PCSK9) led to 70% gene editing in the liver, 83% reduction in blood PCSK9 protein, and 69% reduction in LDL cholesterol, with effects lasting up to 476 days; no germline editing was detected in sperm or offspring.
Abstract 17013: CTX320: An Investigational in vivo CRISPR-Based Therapy Efficiently and Durably Reduces Lipoprotein (a) Levels in Non-Human Primates After a Single Dose
In non-human primates, a single dose of CTX320 (in vivo CRISPR targeting Lp(a)) reduced plasma Lp(a) by up to 94% from baseline, with effects persisting through day 224; transient liver enzyme elevations resolved within 14 days.
Clinical gene technology in Australia: building on solid foundations
A review of gene therapy in Australia notes that while in vivo CRISPR has entered human trials (e.g., for transthyretin amyloidosis), long-term safety and durability data are still lacking, and better tools for detecting off-target effects are needed.
Abstract 16908: CTX310: An Investigational <i>in vivo</i> CRISPR-Based Therapy Efficiently and Durably Reduces ANGPTL3 Protein and Triglyceride Levels in Non-Human Primates After a Single Dose
In non-human primates, a single dose of CTX310 (in vivo CRISPR targeting ANGPTL3) achieved up to 71% liver editing, 86% reduction in ANGPTL3 protein, and 64% reduction in triglycerides, with effects durable past 32 weeks; no unintended genomic changes were detected.
Abstract 11274: VERVE-101—An Investigational Single-Course Gene Editing Medicine Targeting PCSK9—Durably and Potently Lowers PCSK9 and LDL-C Concentrations in Non-Human Primates
In 43 non-human primates, VERVE-101 (targeting PCSK9) produced 89% lower blood PCSK9 and 68% lower LDL cholesterol at one year after a single dose; no editing was found in sperm samples from treated males.
Correction of ineffective erythropoiesis and durable clinical benefit with exagamglogene autotemcel for transfusion-dependent β-thalassemia
In a phase 3 trial of exa-cel (ex vivo CRISPR for transfusion-dependent beta-thalassemia), 98% of 56 patients achieved transfusion independence for a mean of 41.4 months, with stable gene editing and normalization of iron metabolism; safety was consistent with the conditioning regimen.
Long-term follow-up demonstrates durable clinical benefits of exagamglogene autotemcel for sickle cell disease with recurrent vaso-occlusive crises: Final results of climb SCD-121
In a completed phase 3 trial of exa-cel for sickle cell disease, 91% of 45 evaluable patients were free of severe vaso-occlusive crises for at least 12 months, with a mean VOC-free duration of 35.3 months; stable gene editing and improved quality of life were observed.
Durable Clinical Benefits with Exagamglogene Autotemcel for Severe Sickle Cell Disease
In a phase 3 trial of exa-cel for sickle cell disease, 90% of 40 evaluable patients achieved freedom from severe vaso-occlusive crises for at least 12 months, with a mean VOC-free duration of 29.3 months; stable gene editing and improved hemolysis markers were observed.
Durable Clinical Benefits with Exagamglogene Autotemcel for Transfusion-Dependent β-Thalassemia
In a phase 3 trial of exa-cel for transfusion-dependent beta-thalassemia, 94% of 52 evaluable patients achieved transfusion independence for up to 5 years, with stable gene editing and normalization of hemoglobin levels; no deaths or malignancies occurred.
In vivo LNP-CRISPR Approaches for the Treatment of Hemophilia
A review of LNP-CRISPR approaches for hemophilia concludes that non-viral LNP delivery is safer than viral vectors and that preclinical studies show sufficient and sustainable therapeutic effects, supporting its potential as a valid treatment option.
CRISPR-Cas9 In Vivo Gene Editing for Transthyretin Amyloidosis
In the first-in-human phase 1 trial of NTLA-2001 (in vivo CRISPR for transthyretin amyloidosis), six patients had only mild adverse events and dose-dependent reductions in serum TTR protein of 52% (low dose) and 87% (high dose) at day 28.