Mark ở vũ trụ khác · Another Universe Genes and Rare Diseases #10 · Medical ethics · Series finale

The gene-edited babies of Shenzhen

?

He Jiankui · Shenzhen · 2018 · WHO governance framework, 2021

The storythe first children born from gene-edited embryos
Who did itresearcher He Jiankui and two collaborators
Outcomecondemned worldwide, 3 years in prison (2019)
LabelFalse, ethically and scientifically

Source: AP and STAT 26 Nov 2018; Shenzhen court ruling 30 Dec 2019 via Xinhua and AP (secondary sources); WHO release 12 July 2021.

For reading and understanding, not a substitute for medical advice.

Recap in 1 minute

Essence

November 2018A researcher in Shenzhen announced twin girls born from embryos whose CCR5 gene had been edited.
Not neededThe father had HIV, but there was already a way to avoid passing it to the children without editing genes.
Not as claimedThe manuscript showed the babies did not carry the natural HIV-resistance variant, but new variants nobody understands.
Prison and new rulesA Shenzhen court gave 3 years (2019). WHO issued a governance framework (2021); China banned clinical germline editing (2024).
Edit an embryo, edit a family lineA change in an embryo enters every cell and passes to descendants. That needs broad agreement, not secrecy.

Source: AP, STAT, NPR Nov 2018; MIT Technology Review Dec 2019 (manuscript excerpts); Xinhua via AP 30 Dec 2019 (secondary); WHO 12 July 2021; China's Ministry of Science and Technology 8 July 2024 via Xinhua; MedlinePlus Genetics.

Part 1

The question

Piece #9 showed CRISPR scissors treating one person at a time. Used on an embryo, though, the change reaches the children and grandchildren too. Who gets to decide that, and what happens when one person decides for the whole world?

The question and wording are Mark's, based on the WHO governance framework (2021) and the International Commission report (2020).

2015–2018In secretHuman embryo experiments kept quiet, then announced by video.
→
2018–2019Condemnation, trialScientists object; a court finds three people guilty.
→
2020–2026RulesWHO, the academies and China set limits. The debate goes on.
Part 2

The story

Who, where, when. The team never published a peer-reviewed paper, so almost every detail comes from the press and the court ruling.

2015 · The first experiment on human embryos, in the labHuang Junjiu's team (Guangzhou) tried CRISPR on human embryos that could not develop, targeting the gene behind beta thalassaemia. Result: poor repair, "mosaic" embryos (edited and unedited cells mixed), off-target cuts. Their conclusion: far from ready for people.
26 Nov 2018 · A YouTube videoHe Jiankui, a researcher in Shenzhen, announced twin girls (pseudonyms Lulu and Nana) born a few weeks earlier from embryos whose CCR5 gene, the door HIV uses to enter cells, had been edited. The father had HIV.
28 Nov 2018 · The Hong Kong summitHe presented at the Second International Summit on Human Genome Editing. Scientists asked: where is the medical need, who did the ethics review, where are the data? The answers did not convince.
Late Nov 2018 · HaltedChinese authorities halted the team's research and opened an investigation.
Dec 2019 · The manuscript surfacesMIT Technology Review published excerpts from the manuscript major journals had rejected. The babies did not carry the natural HIV-resistance variant, but new variants with unknown effects.
30 Dec 2019 · The verdictA Shenzhen court sentenced He Jiankui to 3 years in prison and a 3 million yuan fine for illegal medical practice; two collaborators were also convicted. The court said the team had forged ethics review documents, and confirmed a third baby.
Apr 2022 · ReleasedHe Jiankui finished his sentence. The three children's identities remain protected.

Source: Liang et al. (2015), Protein & Cell 6(5): 363–372; AP, STAT, NPR 26 Nov 2018; Hong Kong summit via NPR and GEN; research halt per Xinhua (via press); MIT Technology Review Dec 2019 (manuscript excerpts); ruling of the Nanshan District court, Shenzhen, 30 Dec 2019 via Xinhua, AP, Bloomberg; MIT Technology Review 4 Apr 2022 (release). Except the 2015 paper, all are secondary sources.

Why did scientists call the experiment "unnecessary"?

To stop a father with HIV from passing it on, medicine already had "sperm washing" before IVF. David Liu and Jennifer Doudna, two of the people behind gene-editing tools, both made this point: there was no unmet medical need.

In other words, healthy babies bore the risk of an untested experiment for a benefit that could be had more safely.

Source: remarks by David Liu and Jennifer Doudna, via GEN and the South China Morning Post, Nov 2018 (secondary sources).

Part 3

How the gene works

1. Editing body cells (like the medicine in piece #9) affects one person. Editing a one-cell embryo puts the change into every cell, including the future eggs or sperm, so it passes to descendants.

2. The CCR5 gene makes a "door" on white blood cells that most HIV strains use to get in. Some people naturally carry the Δ32 variant (32 letters missing), which keeps the door shut.

3. Scissors in an embryo cannot be fully controlled: each cell may repair differently, making a "mosaic" embryo, and new variants can appear instead of Δ32.

Two kinds of gene editing over three generations. Left, editing body cells: only the treated person carries the change, descendants do not. Right, editing an embryo: that person, the children and the grandchildren all carry the red change. editing body cells editing an embryo (germline) treated personthe change stops with one person baby from edited embryochildren, grandchildren all carry it, nobody asked childrengrandchildren
Why embryo editing is different. A mistake in an embryo is a risk not to one person but to people not yet born, who cannot consent.Source: MedlinePlus Genetics (body cells and germline cells); Council of Europe Oviedo Convention (1997), Article 13. The family tree is an illustration.
Three locks. Left: normal CCR5, the door opens, HIV gets in. Middle: the natural Δ32 variant, the door stays shut, studied for years. Right: new variants made by the 2018 experiment, odd shapes, effects unknown. normal CCR5door open, HIV gets in natural Δ3232 letters missing, well studied ? new variants (2018)effects unknown
Right gene, wrong result. He said he had "copied" the natural HIV-resistance variant; the manuscript showed otherwise.Source: role of CCR5 and Δ32 per AP and NPR, Nov 2018; manuscript excerpts and Fyodor Urnov's comment, MIT Technology Review Dec 2019 (secondary sources). The figure is an illustration.
Part 4

Fact check: what is true and what is rumour

14 things often said about the case. Because the team has no peer-reviewed paper, many rows rest on the press and the court ruling, marked as secondary.

SOLIDIn Nov 2018, He Jiankui announced twin girls born from edited embryos

Announcement video 26 Nov 2018 (AP, STAT); the 2019 ruling confirmed the births.

SOLIDThe aim was to disable CCR5 against HIV; the father had HIV

As He told AP and NPR (secondary).

SOLIDThere was already a way to avoid passing HIV from father to child without gene editing

"Sperm washing" before IVF; raised by David Liu and Jennifer Doudna at the time (GEN, SCMP; secondary).

SOLIDA Shenzhen court gave 3 years in prison and a 3 million yuan fine

On 30 Dec 2019, for illegal medical practice; Zhang Renli 2 years, Qin Jinzhou an 18-month suspended sentence (Xinhua via AP, Bloomberg; secondary).

SOLIDThere was a third gene-edited baby

The ruling said the team was involved in three births to two mothers (Xinhua via AP).

SOLIDThe team forged ethics review documents

Per the court ruling, via Xinhua and AP (secondary).

SOLIDWHO issued a global governance framework and recommendations in 2021

WHO release 12 July 2021: a governance framework with illustrative scenarios, plus recommendations on research registries, medical travel and education. The expert committee was set up in Dec 2018, right after the case.

SOLIDInternational science bodies concluded heritable editing is not acceptable for now

International Commission of the US academies and the UK Royal Society (3 Sep 2020): not ready; Third International Summit, London (8 Mar 2023): "remains unacceptable at this time".

CAUTION"At least one baby is mosaic"

Scientists who saw the documents told AP so; the full data were never published for independent checking.

CAUTION"The children are perfectly healthy, nothing to worry about"

No independent medical data have been published. Their identities are protected, rightly, which also means nobody outside knows for sure.

FALSE"He recreated the natural HIV-resistance variant"

Manuscript excerpts: the right gene, CCR5, but new variants, not Δ32. Fyodor Urnov called the claim a misrepresentation of the data (MIT Technology Review Dec 2019; secondary).

FALSE"He was the first to use CRISPR on human embryos"

In 2015, Huang Junjiu's team had tried it on non-viable embryos in the lab (Liang et al., Protein & Cell 2015). He was the first to announce children born from edited embryos.

FALSE"Science proved that CCR5-Δ32 carriers die early"

A 2019 Nature Medicine paper said so, but its authors retracted it over an error in the UK Biobank data; a 2020 reanalysis found no effect on lifespan.

FALSE"After this, nobody wants to edit human embryos any more"

In 2025 at least two US start-ups (Preventive, Manhattan Genomics) raised money to research embryo editing; many scientists object (MIT Technology Review 31 Oct 2025; secondary).

Source: as given in each row; full list in Part 9.

Part 5

Where we are today

After 2018 the world put up fences. But the door is not fully shut.

3 Sep 2020 · International CommissionSet up by the US National Academy of Medicine, National Academy of Sciences and the UK Royal Society. Conclusion: edited embryos must not be used for pregnancy, as no technology can yet make precise changes without unwanted ones. If ever allowed, start only with serious single-gene diseases.
12 July 2021 · WHOThe first global governance framework and recommendations on human genome editing, including a registry of human genome editing studies and the issue of medical travel. It repeated: now is not the time for heritable editing.
8 Mar 2023 · London summitThe Third International Summit: heritable editing "remains unacceptable at this time"; it also noted that somatic editing has cured diseases and called for affordable access.
8 July 2024 · ChinaThe Ministry of Science and Technology's ethics guidelines: any clinical research on germline editing "is currently irresponsible and not allowed"; using edited embryos for pregnancy is banned.
Oct 2025 · Start-upsA US company has about 30 million USD to research embryo editing and says it will not rush. Many scientists call the direction dangerous and a distraction from the real benefits of gene editing for patients (secondary).
In EuropeThe Oviedo Convention (1997) bans interventions aimed at changing the genome of descendants; 30 states have ratified it.

Source: release of the US National Academies and the UK Royal Society, 3 Sep 2020; WHO 12 July 2021; statement of the Organising Committee of the Third International Summit, Royal Society 8 Mar 2023; guidelines of China's Ministry of Science and Technology, 8 July 2024, via Xinhua and gov.cn; Parliamentary Assembly of the Council of Europe (PACE) 2023 report on Article 13 of the Oviedo Convention; MIT Technology Review 31 Oct 2025 (secondary). Checked 16 Oct 2026.

Part 6

Verdict

Scores are Mark's assessment based on the sources in Part 9, not data from a study. They rate the 2018 experiment, not the person.

Real medical need
0/5
Safety, scientific groundwork
1/5
Transparency, consent
0/5
The scientific community's response
4/5
What went right afterwards
  • Scientists spoke up at once, in public
  • WHO and the academies set out frameworks
  • The three children's identities were protected
What was wrong with the experiment
  • No unmet medical need
  • Secrecy and forged ethics documents
  • New, poorly understood variants that can be inherited

Label: FALSE, ethically and scientifically. Fast does not mean right; and a good tool (piece #9) can still be misused.

More: why does Mark rate the "response" 4/5 rather than 5/5?

Scientists reacted quickly and set out clear guidance. But the experiment went all the way to births without anyone outside the team stopping it; at the Hong Kong summit itself, David Baltimore called that a failure of self-regulation, due to a lack of transparency.

And by 2025, embryo editing was being discussed again by private companies. The fences exist, but they need guarding.

The "More" section is Mark's commentary; "failure of self-regulation" is David Baltimore's phrase at the Hong Kong summit, via NPR (secondary).

Part 7

Lessons: what to ask before believing, and before taking part

Three children had no choice. Adults can ask.

For reading and understanding, not a substitute for medical advice.

1 · Is there a real need?Is there already a safer way? If so, why try a new one?
2 · Who reviewed it?Did an independent ethics committee approve the study? Is it publicly registered?
3 · Fully explained?Risks, alternatives, the right to withdraw: were they written and explained before signing?
4 · Peer-reviewed paper?A "breakthrough" with only a video or a press release is not yet evidence.
Beware offers to "edit your baby's genes"Editing embryos for pregnancy is banned in many countries and judged unacceptable by WHO and the academies. If someone offers it, check with a genetic doctor.
Want to know your inherited risk?The question to ask is: "Does my family need genetic counselling or carrier testing?" (as with thalassaemia in piece #5).
Self-check: rereading the 2018 case with the four questions

1. Real need? No: sperm washing existed.

2. Who reviewed it? The court found the ethics documents forged.

3. Fully explained? Cannot be verified; and the babies could not consent.

4. Peer-reviewed paper? None; only a video and a rejected manuscript.

Four questions, four "no"s. Next time you read a "breakthrough", count how many "yes"es it gets.

The four questions are Mark's suggestions, based on principles in the WHO governance framework (2021) and the facts in Part 2.

Part 8

Read next, and thank you

Genes and Rare Diseases #9: CRISPR, scissors borrowed from bacteriaThe same tool used the right way: the first medicine in 2023 and baby KJ in 2025. Read
WHO (2021), Human genome editing: a framework for governanceThe original document, with readable illustrative scenarios.
Thank you for reading all 10 pieces of "Genes and Rare Diseases". From Photo 51 to the babies of Shenzhen, each piece was one person, one family, one gene. If you remember only one thing, Mark hopes it is this: genes are probabilities, not destiny, and before every "breakthrough", ask one more question.
  1. Photo 51
  2. There is no "intelligence gene"
  3. Huntington's
  4. A drop of blood from the heel
  5. Thalassaemia
  6. Sickle cell
  7. Progeria
  8. Feeling no pain
  9. CRISPR
  10. The gene-edited babies (this piece)
Drop by Mark Ở Vũ Trụ Khác to read other topics, or to learn with Mark.
Part 9

Sources

Primary sources are documents from WHO, the academies, international summits and government rules. He Jiankui's team has no peer-reviewed paper, so case details come from the press and the court ruling (via Xinhua), marked as secondary. The working environment blocks WHO and NIH pages; points were checked against excerpts in search results.

  1. WHO (12 July 2021), "WHO issues new recommendations on human genome editing for the advancement of public health" and two companion reports (governance framework; recommendations). who.int
  2. National Academy of Medicine, National Academy of Sciences, Royal Society (2020), Heritable Human Genome Editing, release 3 Sep 2020. royalsociety.org
  3. Organising Committee of the Third International Summit on Human Genome Editing (8 Mar 2023), statement. royalsociety.org
  4. Council of Europe (1997), Oviedo Convention on Human Rights and Biomedicine, Article 13; PACE report 2023.
  5. Liang P, Huang J et al. (2015), "CRISPR/Cas9-mediated gene editing in human tripronuclear zygotes", Protein & Cell 6(5): 363–372. PMC
  6. China's Ministry of Science and Technology (8 July 2024), ethical guidelines for human genome editing research, via Xinhua. gov.cn
  7. MedlinePlus Genetics, "What are the ethical issues surrounding gene therapy?". medlineplus.gov
  8. Wei X, Nielsen R (2019), Nature Medicine 25 (retracted Oct 2019); reanalysis: "No statistical evidence for an effect of CCR5-∆32 on lifespan in the UK Biobank cohort", Nature Medicine 2020.
  9. Secondary sources: AP, STAT, NPR 26 Nov 2018 (announcement, CCR5, HIV, mosaicism); GEN and South China Morning Post Nov 2018 (David Liu, Jennifer Doudna, David Baltimore); MIT Technology Review Dec 2019 (manuscript excerpts), 4 Apr 2022 (release), 31 Oct 2025 (start-ups); Xinhua via AP, Bloomberg, Sixth Tone 30 Dec 2019 (ruling); CRISPR Medicine News 2025.