Gene Editing in Human Embryos: Base Editing's Promise and Peril

Introduction

Gene editing has moved from science fiction to laboratory reality faster than most people imagined. The latest step: a team at Columbia University led by geneticist Dieter Egli has used a technique called base editing to make precise, single-letter changes in the DNA of early human embryos, targeting genes linked to heart disease and blood disorders. The result, published as a preprint on bioRxiv in May 2026 and not yet peer-reviewed, is the most accurate embryo gene editing reported to date.

This matters for two distinct reasons. Scientifically, it demonstrates that base editing avoids the catastrophic chromosomal damage that plagued earlier CRISPR attempts in human embryos. Ethically, it moves the world one step closer to a technology that could eliminate inherited disease before birth, but that could also be misused to select or enhance traits in children.

Neither side of that picture is simple, and the gap between laboratory result and clinical application remains very wide.

This article is for general information only and does not constitute medical advice. The study discussed here is a preprint that has not yet undergone peer review.

What is Base Editing

From CRISPR to Base Editing: a better scalpel

Genome editing began in earnest with CRISPR-Cas9, a system that cuts both strands of the DNA double helix at a target site. The cut is powerful but imprecise at the repair stage: human cells, especially in early embryos, often repair the break incorrectly, deleting large segments of DNA or even losing entire chromosomes. A 2020 study by Egli's own team demonstrated this catastrophic outcome when CRISPR was used to repair a gene linked to hereditary blindness in human embryos.

Base editing, developed by David Liu at Harvard University in 2016 and 2017, takes a different approach. Instead of cutting both strands, a base editor chemically converts one DNA letter into another, directly and without breaking the double helix. There are two main types:

• Cytosine Base Editors (CBE): convert cytosine (C) to thymine (T)
• Adenine Base Editors (ABE): convert adenine (A) to guanine (G)

Think of CRISPR as scissors that cut a word out of a sentence, with unpredictable results when the sentence is reassembled. Base editing is more like a spell-checker that changes one letter in place without disturbing the surrounding text.

The Columbia study used an Adenine Base Editor (ABE), which changes A to G. This type of edit can mimic naturally occurring protective mutations or silence disease-causing versions of a gene.

The Columbia Study

What the Columbia Study Did

Egli's team targeted three genes in early-stage human embryos:

PCSK9 , a gene that regulates low-density lipoprotein (LDL) cholesterol levels in the blood. People who carry naturally occurring loss-of-function variants in PCSK9 have substantially lower LDL cholesterol and a reduced lifetime risk of coronary heart disease. The researchers used the adenine base editor to introduce a single A-to-G change that switched PCSK9 off, mimicking this protective natural variant.

HBG1 and HBG2 , genes that control the production of fetal hemoglobin, a form of hemoglobin present before birth that declines after birth and is replaced by adult hemoglobin. People with sickle cell disease or thalassemia suffer because their adult hemoglobin is defective. Reactivating fetal hemoglobin production is a known therapeutic strategy already used in approved treatments. The A-to-G change in HBG1 and HBG2 introduced a mutation that mimics a naturally protective variant linked to higher fetal hemoglobin levels in adulthood.

Delivery method: The base editor was delivered as a protein, injected at the time of fertilization or at the pronuclear stage (shortly after fertilization, when egg and sperm nuclei are still separate). This was critical: when the editor was delivered as messenger RNA, embryo development stopped early. Protein delivery allowed embryos to develop normally to the blastocyst stage, where stem cell lines could be derived and analyzed.

Results and the Mosaicism Problem

Results: A Step Forward, with a Significant Limitation

What worked: - Base editing achieved an overall on-target editing efficiency of approximately 70% across 304 targetable alleles analyzed - No chromosomal abnormalities or large deletions were detected, in direct contrast to CRISPR-Cas9 results in the same type of experiment - Small insertions or deletions (indels) after base editing were rare - Edited embryos developed normally to the blastocyst stage - In some embryos, both PCSK9 and HBG genes were edited simultaneously

The central limitation: mosaicism In a significant proportion of embryos, the editing was not uniform. Some cells within the same embryo received the intended letter change while others retained the original sequence. This phenomenon, called mosaicism, is a serious obstacle to clinical use. An embryo in which only some cells carry the corrected gene would not provide the uniform genetic change needed to prevent disease. Worse, a mix of edited and original cells in a developing body could potentially create unpredictable health effects.

The study does not report the precise fraction of mosaic embryos, but both Egli and independent commentators acknowledge it is a substantial proportion.

What this means for safety: Bioethicist Ana Iltis of Wake Forest University told The New York Times that the absence of chromosomal damage is only the first safety question. Harmful effects of DNA editing in embryos might not become visible until after birth, during childhood development, or even in adulthood. Current safety data cannot address these long-term questions.

Evidence quality: Moderate , single-lab preprint study, not yet peer-reviewed. Results consistent with prior base editing data in other species but require independent replication in humans.

Base Editing vs CRISPR vs Preimplantation Genetic Testing

The Ethical Dimension

The Ethical Dimension

Base editing in human embryos occupies one of the most contested territories in all of modern biology. Any change made at this stage is heritable: it will be passed to all cells of the resulting person, and potentially to that person's children and their descendants. This is what scientists call germline editing, and it is subject to strict regulation or outright prohibition in most countries.

The concern is not only about safety. It is also about where the technology leads.

Fyodor Urnov, a geneticist at the University of California, Berkeley, warned that research of this kind risks becoming what he called "a how-to manual for baby improvers," meaning people who would seek to use embryo editing not to prevent disease but to select or enhance traits such as intelligence, height, or physical appearance. This path leads directly to discussions about eugenics, the discredited 20th-century ideology that sought to engineer human populations by selecting which people should reproduce.

Egli himself was direct about the risks: "We can provide the data for discussion, but then essentially there you stop and let others take over." He emphasized that the question of whether to move toward clinical use of embryo editing is not one for scientists alone to answer. It is a societal decision requiring broad public debate.

A separate and important objection raised by geneticist Urnov and others: preimplantation genetic testing already allows couples undergoing in vitro fertilization to test embryos for hundreds of known genetic mutations and select unaffected embryos for transfer. For most single-gene disorders, this approach achieves the medical goal without any editing and without creating heritable changes. Whether embryo editing is necessary for most cases, rather than merely technically interesting, is a legitimate scientific question.

The case where embryo editing might be uniquely useful is for couples where both partners carry two copies of a disease-causing mutation (homozygous carriers), leaving no unaffected embryos to select. This is a smaller subset of cases.

What Experts Say

Reactions from the scientific and bioethics community have been sharply divided.

Supportive perspectives: George Daley, a stem cell biologist at Harvard Medical School, stated that he is "generally supportive of the concept of embryo editing to prevent genetic disease," noting that the field needs exactly this kind of careful, incremental research. The absence of chromosomal damage in Egli's results represents a real advance over prior CRISPR work.

Critical perspectives: Ana Iltis, a bioethicist at Wake Forest University, cautioned that the absence of immediate chromosomal damage is merely "step one" in proving safety, and that many harms might only emerge after birth.

Fyodor Urnov questioned the medical rationale given that preimplantation genetic testing already provides a safe and effective alternative for most couples at risk of passing on genetic disease.

Egli's own position: The lead researcher is cautious and explicit: the technology is not ready for clinical use. He views this as research that should inform public discussion, not as a step toward immediate application. The study was conducted on embryos not intended for implantation.

Warning Signs

What Makes This Technology Risky

Several distinct risks must be separated:

Mosaicism: Uneven editing creates embryos where different cells carry different versions of the gene. The health consequences of such genetic patchwork in a developing human are unknown.

Off-target edits: Base editors can change DNA letters at unintended locations in the genome if the guide RNA that directs them is not sufficiently specific. The study found off-target activity was "dependent on the guide RNA," meaning guide RNA design is critical and an off-target edit at a cancer-related gene could have serious consequences.

Unknown long-term effects: Safety in embryonic development does not predict safety across an entire human lifespan. No edited human embryo has been brought to term in this study (and the study was not designed for that purpose). Long-term effects on health, fertility, and cancer risk cannot currently be assessed.

Heritable changes: Any edit that passes into the germline will be transmitted to future generations. An error is not correctable after the fact.

Misuse potential: Technology developed for disease prevention could be adapted for trait enhancement. The boundary between treating disease and engineering traits is not always sharp.

Timeline to Clinical Use

When Could This Reach the Clinic?

The short answer: not soon, and not without a very different regulatory and ethical consensus than currently exists.

Current status of the research

• Preprint published May 30, 2026 on bioRxiv; not yet peer-reviewed or published in a peer-reviewed journal
• Conducted in research embryos not intended for implantation
• Mosaicism problem is unsolved
• Long-term safety data do not exist
• No regulatory authority has approved germline editing in humans for clinical use

Steps needed before any clinical consideration: 1. Independent replication of results by other laboratories 2. Peer review and publication in a peer-reviewed journal 3. Development of methods to eliminate or reliably detect mosaicism 4. Comprehensive off-target analysis across the full genome 5. Regulatory and societal discussion in each jurisdiction 6. Clear definition of which clinical indications, if any, would justify germline editing given existing alternatives

In parallel, non-heritable base editing therapies (editing cells in a living patient rather than an embryo) have already reached patients. In 2025, a child with a serious genetic disorder was successfully treated using a customized base editing therapy, and base editing for sickle cell disease is in clinical trials. These somatic applications of base editing carry none of the germline ethics concerns and are advancing rapidly.

FAQ

What is the difference between base editing and CRISPR?

CRISPR-Cas9 works by cutting both strands of the DNA double helix at a specific location and relying on the cell's DNA repair machinery to make the desired change. In human embryos, this repair frequently goes wrong, causing large deletions or chromosomal loss. Base editing, developed by David Liu at Harvard in 2016, does not cut both DNA strands. Instead, it uses an enzyme fused to a Cas9-derived protein that chemically converts one DNA letter to another: adenine (A) to guanine (G) using an Adenine Base Editor, or cytosine (C) to thymine (T) using a Cytosine Base Editor. This avoids the double-strand break and its associated risks.

What is mosaicism, and why is it a problem?

Mosaicism occurs when the cells of a single embryo end up with different genetic sequences. In the Columbia study, base editing sometimes reached only some cells in an embryo, leaving others with the original unedited sequence. A person who develops from a mosaic embryo would have a genetic patchwork: some cells with the intended edit and others without it. Whether this creates health problems depends on which cells are edited and which are not, and the consequences are difficult to predict in advance. For a therapy intended to prevent disease, this is a fundamental obstacle: you cannot guarantee that the disease-relevant cells will carry the correction.

Could this lead to designer babies?

This is the central ethical concern. The technology itself does not distinguish between removing a disease-causing mutation and introducing a trait-enhancing change. Both are single-letter edits at defined locations. The difference is intent and target. Scientists including Urnov worry that once a technique for embryo editing exists and is demonstrated to work, commercial or social pressure could drive its use toward enhancement rather than disease prevention. This is the primary reason why virtually every national regulatory body has so far prohibited any clinical use of heritable human gene editing.

Is preimplantation genetic testing not enough?

For most couples at risk of passing on a genetic disease, preimplantation genetic testing (PGT) conducted during in vitro fertilization allows selection of embryos that do not carry the problematic mutation. This is a well-established, clinically proven approach that involves no editing and no heritable modification. Embryo editing would primarily add value in the rare situation where a couple has no unaffected embryos to select, as occurs when both partners carry two copies of the disease mutation. This is a much smaller category than the general concern over inherited disease.

What happened when Chinese scientist He Jiankui edited human embryos?

In 2018, He Jiankui announced he had used CRISPR to edit human embryos that were subsequently implanted, resulting in the birth of twin girls with edited CCR5 genes, intended to make them resistant to HIV infection. The announcement caused widespread international condemnation. The edit was considered unnecessary (HIV can be prevented by other means), inadequately safe, and conducted without proper informed consent. He was convicted of illegal medical practice in China and sentenced to three years in prison. The case became the defining example of why the scientific community emphasizes that embryo editing must not proceed to clinical use without broad societal consensus.

Summary

The Columbia University base editing study represents a genuine scientific advance: for the first time, genes in human embryos have been edited with single-letter precision without the chromosomal damage that accompanied previous CRISPR attempts. The targeted genes, PCSK9 linked to heart disease and HBG1 and HBG2 linked to fetal hemoglobin and blood disorders, are medically relevant and the edits mimic naturally protective variants.

The major unresolved problem is mosaicism. A technology that edits some cells but not others in an embryo cannot yet be considered safe or effective for clinical use. Long-term safety, off-target effects, and the absence of peer review are additional reasons for caution.

The ethical questions are as important as the technical ones. Heritable gene editing affects not only a future person but their descendants. The existence of preimplantation genetic testing as a well-established alternative means that embryo editing does not currently address an unmet medical need for most couples at risk. Whether it ever should remains a question for society, regulators, and ethics bodies, not only for scientists.

References

1. Jerabek S, Kim J, Sung J, et al. Efficient base editing and development in human embryos without chromosomal alterations. bioRxiv. 2026.05.30.728989. https://doi.org/10.64898/2026.05.30.728989 [Preprint; not yet peer-reviewed]

1. Nature News. First precise genome editing of human embryos triggers praise and alarm. June 2026. https://www.nature.com/articles/d41586-026-01827-8

1. Komor AC, Kim YB, Packer MS, Zuris JA, Liu DR. Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage. Nature. 2016;533(7603):420-424. https://doi.org/10.1038/nature17946

1. Gaudelli NM, Komor AC, Rees HA, et al. Programmable base editing of AT to GC in genomic DNA without DNA cleavage. Nature. 2017;551(7681):464-471. https://doi.org/10.1038/nature24644

1. Zuccaro MV, Xu J, Mitchell C, et al. Allele-specific chromosome removal after Cas9 cleavage in human embryos. Cell. 2020;183(6):1650-1664. https://doi.org/10.1016/j.cell.2020.10.025 [Egli 2020 CRISPR catastrophic outcomes study]

1. Rees HA, Liu DR. Base editing: precision chemistry on the genome and transcriptome of living cells. Nature Reviews Genetics. 2018;19(12):770-788. https://doi.org/10.1038/s41576-018-0059-1

1. Sherber N et al. Base editing of HBG1 and HBG2 promoters for sickle cell disease. New England Journal of Medicine. 2026. https://doi.org/10.1056/NEJMoa2504835

1. National Academies of Sciences, Engineering, and Medicine. Heritable Human Genome Editing. Washington DC: National Academies Press; 2020. https://doi.org/10.17226/25665

1. Nuffield Council on Bioethics. Genome editing and human reproduction: social and ethical issues. London; 2018. https://www.nuffieldbioethics.org/publications/genome-editing-and-human-reproduction

1. Columbia University Medical Center. Study Identifies Pitfall for Correcting Mutations in Human Embryos with CRISPR. 2020. https://www.cuimc.columbia.edu/node/23079

1. Fertility and Sterility. CRISPR/CAS9 base editing enables precise gene editing in preimplantation human embryos without chromosomal changes. 2024. https://www.fertstert.org/article/S0015-0282(24)00788-X/fulltext

Key Takeaways

• Base editing changes individual DNA letters without cutting both strands of DNA, avoiding the chromosomal damage caused by CRISPR-Cas9 in human embryos
• Egli's Columbia University team edited three genes in human embryos: PCSK9 (heart disease risk / LDL cholesterol) and HBG1 and HBG2 (fetal hemoglobin, sickle cell disease, thalassemia)
• The study was published as a preprint in May 2026 and has not yet undergone peer review
• Editing efficiency was approximately 70% across analyzed alleles, with no chromosomal abnormalities detected
• The central unresolved problem is mosaicism: many embryos had a mix of edited and unedited cells within the same embryo
• Mosaicism means the technology is not yet safe or reliable for clinical use in humans
• Bioethicist Ana Iltis cautions that harms from embryo gene editing may not appear until after birth
• Geneticist Fyodor Urnov warns that the research risks serving as a manual for embryo enhancement, not just disease prevention
• Preimplantation genetic testing already allows selection of disease-free embryos during IVF and addresses most clinical needs without editing
• No regulatory authority has approved germline editing in humans, and the scientific community consensus is that clinical use is not currently justified

Medical Disclaimer

This article is intended for general information only. It does not constitute medical, genetic, or legal advice, and is not a substitute for consultation with a qualified healthcare professional or genetic counselor.

The research described here is a preprint that has not undergone peer review. Findings should be interpreted with appropriate caution until independent replication and peer review are completed.

In a medical emergency: 101 in Israel (Magen David Adom), 112 in Europe, 911 in the United States.