Scientists Switch On an Embryonic Gene to Rescue a Lethal Blood Disorder

In a finding that could reshape the treatment of one of the world’s most severe inherited anemias, researchers at the University of Oxford have shown that reawakening a gene normally active only in the earliest weeks of human development can compensate for the missing adult hemoglobin that causes severe α-thalassemia. The study, led by Siyu Liu, Mira Kassouf and Douglas Higgs and published in Nature Genetics, identifies the precise DNA switches that silence the embryonic ζ-globin gene and demonstrates that editing those switches restores the gene to levels that rescue a lethal disease in mice and correct globin imbalance in human cells from patients.

α-thalassemia arises when the α-globin genes, which encode one of the two protein chains of adult hemoglobin, are deleted or defective. Healthy people carry four copies of the α-globin gene, two on each chromosome 16. Carriers with three or two functional copies are usually asymptomatic, and the carrier alleles are extraordinarily common in malaria-endemic regions, reaching frequencies of up to roughly 70 percent because they confer protection against the parasite. But when these alleles combine, the consequences escalate. People inheriting one Southeast Asian deletion allele and one α-globin deletion develop hemoglobin H disease, in which excess β-globin chains form useless β4 tetramers, and about 10 percent of these patients become transfusion dependent. The most devastating outcome occurs in fetuses that inherit no functional α-globin genes at all, a condition known as hemoglobin Bart’s hydrops fetalis syndrome, or BHFS, which normally causes death in utero during the second or third trimester.

Current options for the most severe patients are limited. Hematopoietic stem cell transplantation is the only curative treatment, but it requires a human leukocyte antigen-matched donor, and even when a donor is available, graft-versus-host disease and transplant-related mortality remain real risks. A small but growing number of BHFS survivors, salvaged by intrauterine transfusions and intensive neonatal care, face lifelong transfusion dependence and iron chelation. This therapeutic gap is what drew the Oxford team to an unusual idea: rather than trying to replace the missing α-globin gene, why not switch back on the embryonic gene that once did its job?

The logic rests on a quirk of human development. The α-globin locus on chromosome 16 contains, in sequence, the enhancers R1 through R4, the embryonic ζ-globin gene HBZ, and the two fetal and adult α-globin genes HBA2 and HBA1. During the first, primitive wave of red blood cell formation, ζ-globin is fully expressed and assembles into functional hemoglobin tetramers, including Hb Portland 1 (ζ2γ2) and Hb Portland 2 (ζ2β2), which can substitute for fetal and adult hemoglobins. But by six to seven weeks of gestation, as definitive hematopoiesis takes over, ζ-globin is completely silenced and remains off for the rest of life. Crucially, the most common disease-causing allele, the Southeast Asian deletion, removes both α-globin genes while leaving the enhancers and the ζ-globin gene intact, meaning that in many patients the dormant embryonic gene is still sitting there, waiting to be unlocked.

To find the off switch, the researchers first built an elegant mouse model called DZIA, in which the 4.4-kilobase segment containing the ζ-globin gene was replaced with a modified α-globin gene carrying twenty silent single-nucleotide polymorphisms that allowed its transcripts to be distinguished from endogenous ones. The result was striking: the substituted gene, placed in exactly the position of the embryonic gene, was fully transcribed in definitive erythroid cells, accounting for 40 to 50 percent of total α-like globin output, roughly equivalent to the combined output of two endogenous α-globin genes. This proved that the silencing machinery is encoded locally within that 4.4-kilobase region, not in distant flanking sequences or in the gene’s position within the chromosome’s three-dimensional architecture.

Next came a CRISPR-Cas9 saturation mutagenesis screen across the ζ-globin promoter. Using a reporter mouse in which ζ-globin expression is marked by a fluorescent protein, and refining the search in an immortalized mouse proerythroblast line, the team tiled 47 guide RNAs across 700 base pairs upstream of the transcription start site. The screen converged on two discrete repressive elements: predicted binding motifs for the transcription factors BCL11A and ZBTB7A, also known as LRF. Disrupting the BCL11A motif alone derepressed ζ-globin to about 4 percent of total α-like globin, the LRF motif alone to about 18 percent, and combined disruption, introduced with a cytosine base editor, pushed expression to roughly 26 percent, marginally more than the sum of the individual effects.

The team then built a mouse model carrying these promoter edits, termed ZPE, and crossed it with a new BHFS model that precisely replicates the human Southeast Asian deletion, including its flanking sequences. Untreated homozygous deletion embryos, which produce no α-globin at all, develop normally until embryonic day 15.5, when the last primitive, ζ-globin-expressing red cells disappear from circulation. Thereafter they become pale and growth restricted, develop edema, and by day 17.5 suffer severe hemolytic anemia with large numbers of immature, nucleated erythroblasts churning out nonfunctional β4 tetramers. Embryos carrying the promoter edits on top of the deletion showed dramatic improvement at day 17.5, with near-normal enucleated red blood cells expressing about 7.5 percent ζ-globin mRNA and producing detectable Hb Portland 2. The embryonic protein was, in effect, standing in for the missing adult one.

The human work followed a parallel path. In the HUDEP-2 definitive erythroid cell line, the researchers found that the architecture of the ζ-globin promoter differs subtly from the mouse: two consensus BCL11A motifs lie 148 base pairs apart, with the proximal one overlapping the essential CCAAT box, so editing that site is ineffective. The distal BCL11A site and the LRF site, however, proved to be the productive targets. Base editing both sites raised ζ-globin to about 15 percent of total α-like globin, comparable to levels achieved by knocking out both repressor genes entirely, while leaving erythroid differentiation untouched. Mechanistically, the authors propose a competition model: LRF binding, just 18 base pairs from the CCAAT box, blocks NF-Y from opening the chromatin, and together with BCL11A it prevents the master activator GATA1 from engaging the promoter. When both repressive motifs are mutated, GATA1 newly binds and the gene fires.

The most clinically resonant experiments used primary human CD34-positive stem cells from healthy donors, patients with hemoglobin H disease and a 21-week BHFS fetus. Adenine and cytosine base editors achieved efficiencies of 25 to 90 percent at the two sites, and edited cells differentiated normally. After normalizing for editing efficiency, the ratio of α-like to β-like globin in the hemoglobin H patient samples rose from 24.8 percent, a dangerously low level, to 48.4 percent, with ζ-globin contributing about half of the α-like output. In the BHFS fetal cells, the ratio climbed from 17.1 percent to 50.5 percent. The authors argue that restoring this ratio to around 0.5 would likely be sufficient to rescue the severe α-thalassemia phenotype, and HPLC analysis of edited BHFS cells showed a corresponding reduction in the pathological β4 tetramers that cause the disease.

Caution remains appropriate. Hb Portland 2 has weaker subunit interactions and higher oxygen affinity than adult hemoglobin, and its physiological behavior in adult life must be characterized before clinical application. Editing efficiencies in primary stem cells vary, off-target effects and long-term engraftment have yet to be rigorously assessed, and the authors note that residual low-level ζ-globin expression seen in human carriers appears to involve human-specific mechanisms absent from the mouse. Still, the study establishes a clear road map: a compact, locally acting cis-regulatory module enforces embryonic gene silencing, and precise base editing can dismantle it. Beyond α-thalassemia, the work reinforces an emerging paradigm, already validated by fetal hemoglobin induction in sickle cell disease and β-thalassemia, that developmentally silenced paralogs can be unlocked to compensate for adult gene defects, offering a potentially generalizable route to autologous genome-editing therapies for severe inherited disorders that today demand lifelong transfusion or a matched donor.

Subject of Research: Reactivation of the embryonic ζ-globin gene by promoter base editing as a therapeutic strategy for severe α-thalassemia

Article Title: Reactivation of the embryonic ζ-globin gene ameliorates severe forms of α-thalassemia

Article References: Liu, S., Ejaz, A., King, A. J., Olijnik, A.-A., Sloane-Stanley, J., Scott, C., Amid, A., Biggs, D., Davies, B., Eglinton, J., Martyn, G., Crossley, M., Babbs, C., Kassouf, M. T., & Higgs, D. R. (2026). Reactivation of the embryonic ζ-globin gene ameliorates severe forms of α-thalassemia. Nature Genetics, 58(10), 2619-2629. https://doi.org/10.1038/s41588-026-02770-0

Image Credits: AI Generated

DOI: 10.1038/s41588-026-02770-0

Keywords: α-thalassemia, ζ-globin, gene editing, base editing, BCL11A, ZBTB7A, hemoglobin, erythropoiesis, hydrops fetalis, CRISPR, globin gene regulation, hemoglobinopathies