CircRNA with full nucleoside modification
A Fudan University circRNA platform combines nucleoside modification with rolling-circle translation, matching semaglutide in obese mice.
Why we wrote this. Solving the nucleoside-modification barrier for circRNA is the kind of early-stage signal the semaglutide cluster should capture before it becomes mainstream coverage.
In this article (5 sections)
A paper published in Nature Biomedical Engineering on 29 June 2026 describes an engineered circular RNA platform that solves a problem that has blocked the technology from clinical use: how to apply the same nucleoside modifications used in approved mRNA vaccines to circular RNA, without destroying the molecule's ability to translate protein[1]. The work comes from Fudan University's Frontier Innovation Center and Clinical Center for Biotherapy, with Wang Xinyue, Pan Qian, Yin Jie, and Kou Haomeng as equal contributors.
Why nucleoside modification matters
The mRNA vaccines developed against SARS-CoV-2 are chemically modified: uridine residues are replaced with N1-methylpseudouridine (m1psi), which reduces the innate immune response the body would otherwise mount against synthetic RNA. Without that modification, injected RNA triggers toll-like receptor signalling, pro-inflammatory cytokine release, and degradation before cells can read the sequence.
Circular RNA (circRNA) has attracted attention as an alternative delivery format because its closed-loop structure makes it inherently more resistant to exonucleases than linear mRNA. However, previous circRNA designs relied on internal ribosome entry sites (IRES) or other cap-independent translation elements that are disrupted when nucleosides are chemically altered. The result was a choice: accept the immunogenicity of unmodified circRNA, or modify the molecule and lose most of its translational output[1].
What the Fudan team found
The researchers screened cap-independent translation enhancer (CITE) sequences from plant and insect RNA viruses, looking for elements that tolerate nucleoside modification while still initiating rolling-circle translation. They identified a sequence from black beetle virus, which they term BBV-CITE, that retains activity in fully m1psi-modified circular RNA.
Rolling-circle translation is the mechanism that makes circRNA productive: because the molecule has no free ends for ribosomes to fall off, a single ribosome can complete multiple passes around the loop, generating many copies of the encoded protein from a single RNA molecule. Preserving that property in a modified circRNA is the core advance the paper reports.
The team also developed a scarless production method via in vitro transcription that assembles circular RNA without leaving the extra sequence tags that earlier ligation or splicing methods required. That matters for manufacturing: cleaner RNA means fewer sequences that could provoke an immune response or interfere with the translation element.
Animal data: the GLP-1 result
Two therapeutic applications are reported. For oncology, the nucleoside-modified circRNA vaccine produced greater tumour growth inhibition in mouse models than an equivalent m1psi-modified mRNA vaccine, while generating lower pro-inflammatory cytokine expression. For metabolic disease, a semaglutide-class application: the team encoded a GLP-1 peptide variant in their modified circRNA and tested it in obese mice[1]. Blood glucose reduction matched the commercial semaglutide comparator over the observation period.
A parallel 2026 paper in the Journal of Advanced Research, working with a different scarless circRNA platform (SLPIE), reported similar matching of semaglutide efficacy over seven days in diabetic mice using circRNA encoding a GLP-1 receptor agonist variant.
These are both preclinical results in rodent models. Neither paper reports human data, pharmacokinetic profiles in primates, or safety characterisation beyond the cytokine readouts. The gap between a mouse efficacy match and a human clinical dose-response is large.
What this means for peptide therapeutics
The approved GLP-1 receptor agonists are peptide drugs manufactured through chemical synthesis or microbial fermentation, then formulated for subcutaneous injection or oral absorption. The research described here points toward a different supply chain: encode a peptide sequence in RNA, inject modified RNA that instructs the body's own cells to produce the peptide, and let rolling-circle translation sustain production for days rather than hours.
If that approach reaches clinical use, it could change the dosing cadence, the manufacturing cost structure, and potentially the immune profile of peptide-based therapeutics. The modified circRNA platform addresses one of the main historical objections to RNA-based peptide delivery, which was that unmodified RNA was too immunogenic for chronic therapeutic use.
The research is at an early stage. The main barriers between this platform and a clinical product include primate and human pharmacokinetic data, a defined safety package across repeated dosing, regulatory classification of the delivery modality, and the logistics of stable formulation and cold-chain distribution comparable to current mRNA vaccine infrastructure.
What this is not
This paper is not a clinical trial, a human safety study, or a regulatory submission. It is a proof-of-concept study in cell culture and mouse models, published in a high-impact journal but not yet replicated or extended to human subjects. The GLP-1 comparator result is a single mouse-model readout, not a head-to-head clinical comparison with approved semaglutide products. None of the circRNA platforms described here have entered Phase 1 trials as of this writing.
Frequently asked
What is circular RNA and how does it differ from mRNA?
Circular RNA (circRNA) is a closed-loop RNA molecule with no free ends. Unlike linear mRNA, it cannot be degraded by exonucleases that chew from the end of a strand, making it intrinsically more stable. It uses cap-independent translation elements rather than the 5' cap that linear mRNA requires. Rolling-circle translation allows a ribosome to traverse the loop multiple times, generating more protein per RNA molecule than a single-pass linear template would. The trade-off historically was that circRNA required specialised sequence elements that were disrupted by the nucleoside modifications used to reduce immune activation in mRNA therapeutics.
Why does nucleoside modification matter for RNA therapeutics?
Unmodified synthetic RNA triggers innate immune responses through toll-like receptors and other pattern-recognition systems, leading to pro-inflammatory cytokine release and accelerated RNA degradation. Replacing uridine with modified bases such as N1-methylpseudouridine (m1psi) dampens this response, which is why the approved COVID-19 mRNA vaccines use it. The Fudan University paper reports a cap-independent translation element from black beetle virus that retains its activity in fully nucleoside-modified circular RNA, resolving the previous incompatibility.
Does this mean a circular RNA version of semaglutide is coming soon?
Not imminently. The GLP-1 result in the Fudan paper is from obese mouse models, not human trials. The gap between a rodent proof-of-concept and an approved drug involves primate pharmacokinetics, multi-dose safety studies, regulatory classification, formulation stability, and manufacturing scale-up. No circRNA-based GLP-1 therapeutic has entered a Phase 1 clinical trial as of this writing. The research establishes that the technical barrier around nucleoside modification and rolling-circle translation can be solved at the bench level; clinical development is a separate and much longer process.
How does this relate to existing GLP-1 drugs like semaglutide?
Existing GLP-1 receptor agonists such as semaglutide are chemically synthesised peptides formulated for injection or oral delivery. The circRNA approach would instead instruct the body's own cells to produce a GLP-1 variant from an RNA template, using the cellular translation machinery rather than a pharmaceutical manufacturing process. If it advances clinically, the difference would primarily show up in dosing duration (rolling-circle translation could sustain production longer than a fixed dose of injected peptide), manufacturing approach, and potentially the immune and tolerability profile. Current approved drugs are not being displaced by this research; it is an early-stage alternative delivery modality.
Sources
- [1]Wang X et al. Engineered circular RNA compatible with complete nucleoside modification and rolling circle translation through a Cap-independent translation mechanism. Nature Biomedical Engineering (2026 Jun 29). PMID 42373803.Tier 1 · primary↩
- [2]Synthesis of scarless circular RNAs expressing long-acting GLP-1RAs for type 2 diabetes therapy. Journal of Advanced Research (2026). PMID 42150718.Tier 1 · primary↩
No revisions yet. First published .