Skip to content
RNA Therapeutics

Circular RNA: The Next Frontier in RNA Medicine

How circular RNAs are made, why they're promising as therapeutics and vaccines, and the delivery and stability advantages driving a wave of clinical programs.

PR Nadia Farooq, MSc 6 min read circRNA Circular RNA IRES

Circular RNAs (circRNAs) were long considered curiosities — noisy splicing byproducts of little functional relevance. That view collapsed after 2013, when high-throughput RNA-Seq showed circRNAs are abundant, cell-type-specific, and often functional. In the last three years, engineered circRNA has emerged as a serious therapeutic modality — several startups (Orna, Laronde/now Sail Bio, Circio) and big-pharma programs are pushing it into the clinic.

What makes circRNA different

  • No 5’ cap, no free 3’ end. Covalently closed ends make it invisible to exonucleases like XRN1 and the exosome.
  • Cap-independent translation. Ribosomes are recruited via an internal ribosome entry site (IRES) — either a viral IRES (EMCV, CVB3) or a designed synthetic element.
  • Half-life advantage. In cultured cells, engineered circRNA can persist ~4-10× longer than a matched linear mRNA (~48-72 h vs 8-24 h in typical experiments).
  • Reduced innate immunity. Because there’s no 5’ triphosphate and no free 3’ end, RIG-I / MDA5 sensing is lower — though this depends heavily on chemistry and purification.

How you make it

Two dominant approaches:

  1. Group I intron self-splicing (e.g. the Anabaena or T4 phage systems). The vector encodes a linear precursor with permuted exon fragments; self-splicing produces a covalently closed circle in vitro. Orna’s technology follows this route.
  2. In-vitro RNA ligation with T4 RNA ligase or splint-mediated ligation. Reliable at small scale but harder to scale.

Purification of circRNA away from linear byproducts is critical — residual linear species drive both immunogenicity and off-target translation.

Why it matters for therapeutics

mRNA replacement therapy. For an enzyme replacement indication that requires steady-state protein for weeks (e.g. propionic acidemia), circRNA can potentially convert monthly dosing into quarterly or better.

Vaccines. Longer expression from a single injection may enable stronger, more durable antibody responses at lower doses. Preclinical circRNA vaccines against SARS-CoV-2 variants have shown neutralizing titers comparable to linear mRNA vaccines with a fraction of the dose.

Cancer immunotherapy. Sustained antigen expression from circRNA may extend the productive antigen-exposure window for T-cell priming.

Protein-of-interest expression platforms. For CAR-T generation in vivo, or for transient reprogramming, longer expression from a single delivery event is highly desirable.

Delivery

The delivery playbook is nearly identical to linear mRNA: lipid nanoparticles with an ionizable lipid, a helper phospholipid, cholesterol, and a PEG-lipid. The larger effective molecular weight of circular molecules doesn’t materially change LNP formulation.

For discussion of the shared LNP chemistry, see our mRNA vaccine article.

Design challenges

  • IRES choice determines translation efficiency and cell-type tropism. Viral IRESes are potent but can trigger PKR-mediated innate immunity; synthetic IRESes are being iterated with the help of ML models.
  • Circularization efficiency in-vitro can be poor for long inserts (> 2 kb), which currently limits payload size.
  • Purification. Removing linear precursor and small RNase-generated fragments is essential to avoid runaway immune activation.
  • Structural stability during shipping and dosing. Frozen storage is still required for most current formulations.

The bioinformatics angle

  • circRNA discovery from RNA-Seq: CIRCexplorer2, CIRI2, find_circ, DCC. Use at least two tools and take the intersection — false positives are common.
  • circRNA-miRNA sponge prediction: TargetScan-style seed matching within known circRNA sequences; validated for a subset of endogenous circRNAs.
  • IRES prediction: emerging deep-learning models trained on known viral and cellular IRESes.

Clinical landscape (as of early 2025)

  • Multiple preclinical vaccine programs from Circio, Orna, Sail Bio.
  • Early-phase academic trials for engineered circRNA in oncology settings.
  • No approved circRNA product yet — but the modality is broadly regarded as the most likely next RNA therapeutics wave after mRNA vaccines.

Bottom line

Circular RNA is not a solved technology — the design space around IRES elements, circularization chemistry, and purification is still moving fast. But the stability advantage is real, the delivery infrastructure is shared with mRNA, and the biology is credible. Over the next five years, expect the first circRNA product approvals and a clearer picture of where circRNA outperforms linear mRNA and where it doesn’t.

Related reading: mRNA vaccine technology, siRNA vs ASO comparison, and RNAi mechanism.

FAQ

Q. How is circular RNA different from linear mRNA?

A. circRNA has no 5' cap and no free 3' end. Its ends are covalently joined, making it resistant to exonucleases. Translation initiates internally via an IRES or engineered internal element rather than via cap-dependent scanning. The result: substantially longer intracellular half-life and, in many cases, more sustained protein expression per dose.

Related in RNA Therapeutics