What the only human trial actually showed
The strongest human evidence for base editing therapies comes from a single Phase 1 clinical trial that treated 11 patients (9 children and 2 adults) with relapsed/refractory T-cell acute lymphoblastic leukemia (T-ALL) using base-edited CAR-T cells [1]. The therapy was designed to knock out three genes simultaneously to prevent graft-versus-host disease, avoid the cells attacking each other, and evade a common chemotherapy drug. After treatment, all 11 patients achieved complete morphological remission (their bone marrow looked normal) at day 28, and 9 of 11 (82%) had no detectable minimal residual disease (MRD), meaning the cancer was cleared at a very sensitive level [1]. However, this was an open-label, single-arm trial—no comparison group—and it was designed primarily to test safety, not efficacy. The results are promising but far from definitive proof that base editing works better than existing treatments.
The durability of the response is encouraging but comes with important caveats. Seven of the 11 patients (63%) remained in remission 3 to 36 months after a subsequent stem cell transplant [1]. But two patients experienced leukemic escape where the cancer cells stopped expressing the target protein CD7, suggesting the cancer can evolve to evade the therapy. Also, viral reactivations were common after transplant, and three patients had significant virus-related complications. So while the trial shows base editing can be used to engineer human immune cells safely enough to move forward, it is a single small study, and the long-term benefits are still uncertain.
What animal studies reveal about the potential
The animal evidence for base editing is much more extensive and shows large, durable effects. In a study of 36 cynomolgus monkeys, a single intravenous infusion of a base-editing therapy targeting the PCSK9 gene (VERVE-101) reduced blood PCSK9 protein by 83% and LDL cholesterol by 69%, with effects lasting up to 476 days [2]. The editing was confirmed in liver biopsies, showing 70% of the target DNA bases were changed at the higher dose. Importantly, the therapy was well tolerated: liver enzyme elevations were temporary and resolved within 14 days, and no evidence of editing was found in sperm or offspring, addressing a key safety concern about unintended germline changes [2]. This study directly supported the launch of a first-in-human clinical trial for patients with familial hypercholesterolemia, though that trial's results are not yet published.
Another study achieved even higher editing rates in the retina of nonhuman primates, correcting an average of 75% of cone cells and 87% of retinal pigment epithelial (RPE) cells in vivo using a dual-vector base-editing system for Stargardt disease [3]. No off-target editing was detected in human retinal tissue samples. These results are striking because they suggest base editing could treat inherited retinal diseases that currently have no treatment. However, both studies are in animals, and the leap from primate retinas or monkey livers to human patients is substantial—many therapies that work spectacularly in animals fail in humans.
The gap between hype and human proof
The hype around base editing is understandable given the technology's elegance and the animal data, but the human evidence remains thin. Across the six studies here, only one involves human patients [1]; the rest are in animals, human cells in a dish, or technical improvements to the editing tools themselves [2][3][4][5][6]. For example, two papers describe new methods for editing mitochondrial DNA in human cells, achieving up to 77% efficiency in correcting disease-causing mutations in patient-derived cells [4][6]. These are important technical advances, but they are still in the lab—no patient has been treated with mitochondrial base editing yet. Similarly, a 2024 study improved base-editing efficiency in primary human T-cells by 82% for one version and 25% for another, but this was in cells grown in a dish, not in patients [5].
The honest answer is that base editing therapies have not yet accumulated enough human evidence to justify the full hype. The single human trial is small, uncontrolled, and focused on safety. The animal data are impressive and provide a strong rationale for ongoing trials, but the gap between animal success and human benefit is historically wide. The field is moving fast—the leukemia trial [1] and the PCSK9 monkey study [2] have both led to ongoing human trials—but until those larger, controlled human results are published, the hype remains ahead of the evidence.
About These Sources
This answer is built on 6 peer-reviewed studies — published from 2022 to 2025, 3 from 2024 or later, 6 in Q1 journals, collectively cited 533 times — selected as the most relevant from 6 studies that passed quality screening, drawn from 80 papers retrieved from a database of over 500 million.
Sources used in this answer
Universal base-edited CAR7 T cells for T-cell acute lymphoblastic leukemia
In a Phase 1 open-label trial of 11 patients with relapsed T-ALL, base-edited CAR-T cells led to complete remission in all patients at day 28, with 9 of 11 achieving MRD-negative remission and 7 of 11 remaining in remission 3–36 months after transplant; this is the only human trial among the studies.
Efficacy and Safety of an Investigational Single-Course CRISPR Base-Editing Therapy Targeting <i>PCSK9</i> in Nonhuman Primate and Mouse Models
In 36 cynomolgus monkeys, a single dose of VERVE-101 (base editing targeting PCSK9) reduced LDL cholesterol by 69% and blood PCSK9 by 83% for up to 476 days, with no detectable germline editing in sperm or offspring; this animal study directly supported a first-in-human trial.
High-efficiency base editing in the retina in primates and human tissues
In nonhuman primates, a dual-vector base-editing system for Stargardt disease corrected 75% of cone cells and 87% of RPE cells in vivo, with no off-target editing detected in human retinal tissue samples.
Targeted A-to-G base editing in human mitochondrial DNA with programmable deaminases
A new TALE-linked deaminase system (TALEDs) achieved targeted A-to-G base editing in human mitochondrial DNA at 17 target sites with up to 49% efficiency in human cells, enabling correction of mitochondrial mutations in the lab.
nCas9 Engineering for Improved Target Interaction Presents an Effective Strategy to Enhance Base Editing
Engineering nCas9 with specific point mutations improved base-editing efficiency by 82% for cytosine base editors and 25% for adenine base editors in primary human T-cells grown in culture, demonstrating a technical optimization strategy.
Strand-selective base editing of human mitochondrial DNA using mitoBEs
Mitochondrial base editors (mitoBEs) combining TALE-fused nickases and deaminases achieved up to 77% A-to-G or C-to-T editing efficiency in human mitochondrial DNA and corrected pathogenic mutations in patient-derived cells.