Self‑amplifying RNA (saRNA) vs mRNA: Key differences, trade-offs, and clinical outlook

What is self-amplifying RNA, and how does it differ from conventional mRNA?

Conventional mRNA and self-amplifying RNA (saRNA) are both transient RNA expression platforms that use the host-cell translation machinery to produce an encoded protein. In vaccines, this protein is typically an antigen, whereas in other therapeutic applications it may be an antibody or immunomodulatory protein.

The key difference is that saRNA contains additional sequences that enable intracellular RNA amplification. Most saRNA constructs investigated to date are based on positive-sense RNA alphavirus replicons, with commonly used backbones derived from Venezuelan equine encephalitis virus (VEEV), Sindbis virus (SINV), and Semliki Forest virus (SFV). In these engineered constructs, the viral structural genes are removed and replaced by the sequence encoding the protein of interest, while the non-structural protein genes required for RNA replication are retained. These genes encode the components of an RNA-dependent RNA polymerase (RdRP) complex, or replicase, that enables amplification of the RNA following successful delivery to the cytoplasm. [1,2,3]

Conventional mRNA follows a more direct expression pathway: once delivered to the cytoplasm, the transcript is translated by ribosomes until it is degraded. Protein expression therefore depends on the amount of intact, translationally competent mRNA reaching the cytoplasm, as well as its stability and translation efficiency. saRNA adds an intracellular amplification step, generating additional RNA templates from which the protein of interest can be expressed. This increases protein expression per unit of administered RNA and creates the potential for lower RNA doses compared with conventional mRNA. [1,2]

How does saRNA work inside cells?

For both conventional mRNA and saRNA, successful protein expression requires intact RNA to reach the cytoplasm. When formulated in ionizable lipid nanoparticles (LNPs), cellular uptake commonly occurs through endocytosis. Acidification within the endosome changes the ionization state of the ionizable lipids, promoting interactions with endosomal membrane lipids and facilitating the release of a fraction of the RNA payload into the cytoplasm.

As illustrated in Figure 1, conventional mRNA and saRNA commonly contain a 5′ cap and a 3′ poly(A) tail, but their architectures differ substantially. Conventional mRNA typically contains a protein-coding open reading frame (ORF) flanked by 5′ and 3′ untranslated regions (UTRs) and is translated directly from the delivered, non-replicating transcript.

Alphavirus-derived saRNA contains additional sequence elements that enable intracellular RNA replication. It typically contains two ORFs: the first encodes the four non-structural proteins (nsP1–4) that form the viral replicase, while the second encodes the protein of interest. The construct also contains viral-derived sequence elements required for replication, including 5′ and 3′ conserved sequence elements (CSEs) and a subgenomic promoter (sgPr) that controls expression of the protein of interest. [1,2]

The intracellular amplification cycle can be summarized in four main steps:

1. Replicase production. The first ORF of saRNA is translated to produce the non-structural proteins nsP1–4, which assemble into an RNA-dependent RNA polymerase (RdRP) complex.
2. Negative-sense RNA synthesis. The replicase uses the delivered positive-sense saRNA as a template to synthesize a complementary negative-sense RNA strand.
3. RNA amplification. The negative-sense strand serves as a template for the synthesis of new full-length positive-sense RNA copies, increasing the number of RNA templates inside the cell.
4. Subgenomic RNA and protein expression. The replicase also recognizes the subgenomic promoter on the negative-sense template and generates shorter positive-sense subgenomic RNA molecules containing the ORF encoding the protein of interest. These sgRNAs are subsequently translated by the host-cell machinery to produce the encoded antigen or therapeutic protein. [1,2]

Through this replication cycle, saRNA increases the number of intracellular RNA templates available for protein production beyond the initially delivered RNA molecules.

RNA replication also generates double-stranded RNA (dsRNA) intermediates that can be recognized by innate immune sensors. This response may contribute to immunogenicity in vaccine applications, whereas excessive innate immune activation can interfere with RNA translation and productive protein expression. saRNA performance therefore depends on the balance between RNA amplification, protein expression, and innate immune sensing. [1,2]

saRNA vs mRNA: what are the main differences?

Conventional mRNA and saRNA differ fundamentally in their RNA architecture and intracellular mechanism. As summarized in Table 1, these differences influence protein expression, expression kinetics, dose requirements, and innate immune sensing. The two platforms also differ in broader practical considerations, including manufacturing and clinical maturity.

A central distinction is the balance between intracellular amplification and molecular complexity. By generating additional RNA templates inside the cell, saRNA can support substantial protein expression from lower administered RNA doses, providing the basis for the dose-sparing effects observed in preclinical and clinical vaccine studies. At the same time, the larger and more complex RNA architecture introduces additional considerations for RNA formulation.

Table 1 provides a concise comparison, while the following sections examine these differences and their supporting evidence in greater detail. These characteristics should be interpreted as platform-level tendencies rather than fixed properties: the magnitude of dose sparing, expression kinetics, innate immune activation, stability, and delivery performance depends on the RNA construct, formulation, dose, administration route, tissue, and intended application.

Table 1. Comparison of conventional mRNA and saRNA.

Does saRNA reduce the required RNA dose compared to mRNA?

A key potential advantage of saRNA is the ability to achieve a desired biological response with a lower administered RNA dose. Because saRNA generates additional RNA templates after reaching the cytoplasm, protein expression is not limited to translation from the initially delivered RNA molecules. [1,4] Preclinical head-to-head studies have reported substantial dose-sparing effects, with approximately 6- to 100-fold lower saRNA doses than conventional mRNA in some experimental vaccine models. [1,4,6] For example, an influenza study in mice achieved comparable protection with 1.25 µg saRNA and 80 µg conventional mRNA, corresponding to a 64-fold difference in administered RNA dose. [5]

Importantly, this dose-sparing potential has also been demonstrated clinically. ARCT-154 is an LNP-formulated saRNA vaccine developed by Arcturus Therapeutics for the prevention of COVID-19. In a randomized phase 3 booster study conducted in Japan, adults who had previously received three doses of authorized mRNA COVID-19 vaccines received either 5 µg ARCT-154 or 30 µg BNT162b2 (Comirnaty, Pfizer/BioNTech). Despite the lower RNA dose, ARCT-154 met the prespecified non-inferiority criteria for the primary immunogenicity endpoints against the ancestral SARS-CoV-2 strain, including neutralizing antibody geometric mean titers and seroresponse rates. Both vaccines were generally well tolerated. [7] ARCT-154 subsequently received regulatory approval in Japan in 2023, under the product name Kostaive, providing an important example of the clinical translation of the saRNA platform beyond experimental studies.

Together, these studies support the dose-sparing potential of saRNA, although the magnitude varies across constructs and applications. The required RNA dose is influenced by the RNA construct and encoded antigen, as well as the formulation, delivery system, administration route, and experimental model.

saRNA and innate immunity: advantage, limitation, or both?

During intracellular replication, saRNA generates double-stranded RNA (dsRNA) intermediates that can be recognized by pattern-recognition receptors, promoting type I interferon and other inflammatory signaling pathways. This innate stimulation may contribute to vaccine immunogenicity, but excessive activation can also inhibit translation, promote RNA degradation, and limit productive RNA expression. [1,2,4]

The desired balance therefore depends on the application. Some innate stimulation may be beneficial for vaccination, whereas applications requiring sustained or high levels of protein production may favor tighter control of inflammatory signaling. Beyond innate immune sensing, alphavirus-derived replicase proteins can also interact with host cell stress pathways: depending on the replicon backbone, non-structural proteins have been associated with ER stress, altered stress granule and autophagy pathways, and apoptosis. [2,8]

The overall cellular response is shaped by both the saRNA construct and its delivery system, including replicon design, RNA quality and purification, formulation, dose, and administration route. Removing dsRNA impurities generated during in vitro transcription can reduce unintended innate sensing, although dsRNA intermediates are also generated intrinsically during saRNA replication. Innate immune activation can also be modulated through RNA and replicon design. For example, Akahata et al. showed that incorporation of 5-methylcytidine (5mC) into a VEEV-derived saRNA reduced RIG-I-mediated sensing and type I interferon induction while maintaining antigen expression and adaptive immune responses. [9] In another approach, Gong et al. demonstrated that modification of the saRNA replicon reduced replication-associated innate immune responses and translation inhibition while increasing payload protein expression. [10]

Innate immune activation should therefore be considered a context-dependent property of the complete saRNA product, requiring a balance between immune stimulation, efficient RNA amplification, and protein expression.

Which delivery systems are used for saRNA?

Like conventional mRNA, saRNA is a negatively charged and nuclease-sensitive molecule that typically requires an effective delivery strategy for in vivo applications. Lipid nanoparticles (LNPs) are among the most clinically advanced non-viral systems for RNA delivery and have been used for multiple saRNA vaccine candidates. LNP composition and formulation conditions can influence RNA encapsulation, cellular uptake, biodistribution, endosomal escape, protein expression, and reactogenicity. [4,11] For practical examples of how changes in LNP composition or formulation process parameters can affect nanoparticle properties and performance, see our studies on mRNA-LNP formulation screening and in vivo evaluation of standard LNP compositions.

Beyond LNPs, a range of lipid-based and polymeric systems have been investigated for saRNA delivery (Figure 2).

Polymeric delivery systems provide an alternative to LNPs. One example is pABOL [poly(cystamine bisacrylamide-co-4-amino-1-butanol)], a bioreducible polymer investigated for saRNA delivery. In a head-to-head preclinical comparison, pABOL-formulated saRNA produced higher protein expression than the tested LNP formulations, whereas LNP-formulated saRNA generated stronger vaccine immunogenicity. The findings show that the delivery system and administration route can substantially change protein expression and immune responses. [4] Cationic nanoemulsions have also been used to deliver saRNA in preclinical vaccine studies. [12]

Nanostructured lipid carriers (NLCs) have also been investigated for saRNA delivery, particularly with the aim of improving storage stability. In a preclinical COVID-19 vaccine study, a lyophilized saRNA/NLC formulation remained stable for at least 6 months at room temperature and at least 10 months under refrigerated conditions, while retaining its ability to induce humoral and cellular immune responses. [13]

How do saRNA and mRNA expression kinetics compare?

Conventional mRNA can produce protein rapidly after successful cytosolic delivery, after which expression declines as the transcript is progressively degraded. saRNA can also support early protein expression, but intracellular RNA amplification can generate later expression peaks and prolong overall expression, resulting in a different kinetic profile from conventional mRNA.

Casmil et al. directly compared conventional linear mRNA (linRNA) and VEEV-derived saRNA encoding the same luciferase reporter after intramuscular administration in mice. With LNP delivery, non-amplifying mRNAs reached peak expression approximately 4 h after injection and returned to baseline between approximately days 4–7 and 11–16, depending on the construct. saRNA also showed early expression but developed a second local maximum around day 7 and remained above baseline until approximately days 36–44. Moreover, saRNA produced approximately 3.2–11-fold greater cumulative expression than the optimized conventional linRNA construct, depending on dose. [8]

However, saRNA expression kinetics depend not only on the RNA construct, but also on the delivery vehicle, formulation, administration route, and target tissue. Bathula et al. compared LNPs with pABOL, an ionizable polymeric delivery system, and found substantial differences in saRNA expression according to both carrier and administration route. For example, 5 µg of LNP-formulated saRNA administered intramuscularly peaked around day 8 and remained detectable beyond day 29. [11] Casmil et al. similarly found that VEEV-saRNA returned to baseline around day 30 with LNPs compared with day 13 with pABOL, despite pABOL producing greater cumulative expression. [8]

Overall, these studies indicate that saRNA can support more sustained protein expression than conventional mRNA. The onset, magnitude, peak, and duration of expression depend on the interplay between RNA design, dose, formulation, delivery system, administration route, and target tissue.

What does the clinical evidence show for saRNA vaccines?

Clinical evidence has moved saRNA beyond a purely preclinical platform, with candidates progressing from early-phase studies to randomized phase 3 trials and regulatory authorization. ARCT-154 provides the clearest clinical example, combining a direct comparison with a conventional mRNA vaccine with subsequent regulatory approval.

ARCT-154 is an LNP-formulated saRNA COVID-19 vaccine administered intramuscularly. In a randomized phase 3 booster study in previously vaccinated adults, 5 µg ARCT-154 was compared with 30 µg BNT162b2 (Comirnaty). ARCT-154 met the prespecified non-inferiority criteria for the primary immunogenicity endpoints against the ancestral SARS-CoV-2 strain, providing clinical evidence of the dose-sparing potential of saRNA relative to a conventional mRNA comparator. [7] ARCT-154 was subsequently approved in Japan as Kostaive in November 2023 and received European Commission marketing authorization in February 2025, providing an important example of the clinical translation of saRNA.

GEMCOVAC-OM is an Omicron BA.1-specific saRNA vaccine formulated in a cationic lipid nano-emulsion rather than a conventional LNP. A 10 µg dose was administered intradermally using the needle-free Tropis system. In phase 3, GEMCOVAC-OM was compared with the intramuscular adenoviral-vector vaccine ChAdOx1 nCoV-19 (COVISHIELD) and met the prespecified non-inferiority criteria for neutralizing antibody responses; the study also reported superiority for BA.1 neutralizing antibody titers and seroconversion at day 29. [14] GEMCOVAC-OM received Emergency Use Authorization in India in June 2023.

VLPCOV-01 is an LNP-formulated saRNA vaccine encoding a membrane-anchored SARS-CoV-2 receptor-binding domain (RBD). In a randomized phase 1 booster study in adults previously vaccinated with BNT162b2, participants received 0.3, 1.0, or 3.0 µg VLPCOV-01, 30 µg BNT162b2, or placebo. VLPCOV-01 induced antibody responses even at 0.3 µg, with responses maintained for up to 26 weeks in non-elderly participants, as well as neutralizing responses against multiple variants. These findings provide early clinical evidence for effective immune responses at low saRNA doses, which could increase the number of vaccine doses produced from a given quantity of RNA and help support large-scale vaccine supply required during a pandemic. [15]

Across these trials, reactogenicity was predominantly mild to moderate and transient, and no vaccine-related serious adverse events were reported. [7,14,15] ARCT-154 showed a reactogenicity profile similar to BNT162b2. [7]

Together, these studies show that saRNA has progressed from an experimental platform to clinically validated and authorized vaccine products. However, the clinical evidence base remains smaller than that of conventional mRNA.

Table 2. Summary of selected clinical studies of saRNA COVID-19 vaccines.

What are the safety considerations for saRNA?

Like conventional mRNA, saRNA functions in the cytoplasm and does not require nuclear entry for protein expression, avoiding the genomic integration concerns. In saRNA constructs, the viral genes encoding structural proteins are replaced by the gene of interest, preventing the assembly into virus. [1]

However, self-amplification introduces additional safety considerations. saRNA encodes the viral-derived RNA-dependent RNA polymerase (RdRP) complex, and the immunological consequences of expressing these non-structural proteins require consideration. Replication also generates dsRNA intermediates that can activate innate immune pathways, making the magnitude and duration of inflammatory responses relevant to both safety and tolerability. [1]

Clinical experience with saRNA remains more limited than with conventional mRNA, although the growing number of clinical studies and regulatory authorizations provides an expanding safety dataset. Safety assessment should therefore remain product-specific, considering the saRNA construct, quality, purity, dose, delivery system, and intended application.

Why does saRNA introduce additional formulation and delivery considerations?

The longer size of saRNA constructs can introduce practical challenges for nanoparticle formulation. Longer RNA molecules can be more susceptible to fragmentation and loss of integrity during handling and processing. [1] The optimal formulation conditions for a conventional mRNA payload should therefore not be assumed to transfer directly to saRNA.

For saRNA-LNP formulations, formulation conditions need to be optimized for the larger RNA payload. These include parameters such as lipid and RNA concentrations and N/P ratio, as well as process parameters. For example, in microfluidic or other controlled-flow mixing methods, the flow rate ratio (FRR) and total flow rate (TFR) can influence LNP formation and resulting nanoparticle attributes. For a broader discussion of RNA-LNP formulation and manufacturing parameters, see our guide to lipid nanoparticle manufacturing methods. For practical formulation workflows, our LNP Starter Kits provide benchmark lipid compositions and can be used with the TAMARA microfluidic formulation system.

Beyond formulation, the larger size of saRNA can also create challenges for efficient intracellular delivery. One strategy being investigated to address the delivery constraints associated with large saRNA constructs is trans-amplifying RNA (taRNA). Instead of encoding the replicase and protein of interest within a single saRNA molecule, the system is divided into two RNA transcripts: a conventional mRNA encoding the nsP1–4 replicase complex and a second transcript encoding the protein of interest and containing the sequence elements required for amplification. [1,14] This approach reduces the size of the individual RNA components that must be delivered, but requires co-delivery of two RNA transcripts, which might bring additional considerations for formulation and dosing.

How do mRNA and saRNA characteristics influence their applications?

Conventional mRNA and saRNA should be viewed as complementary RNA platforms rather than successive generations in which one necessarily replaces the other. Conventional mRNA benefits from a simpler construct architecture, extensive clinical experience, and mature manufacturing and regulatory knowledge. For vaccine applications, it may be advantageous when rapid, controllable antigen expression and modular antigen design are priorities.

saRNA is currently most clinically advanced in vaccine applications, where dose sparing and prolonged antigen expression can be desirable. In addition, the dsRNA and other replication intermediates generated during saRNA amplification can activate innate immune pathways and provide adjuvanting effects that may contribute to vaccine immunogenicity. [4,13] However, excessive innate immune activation can also inhibit RNA translation and amplification, making the magnitude of this response an important design consideration. [4]

The larger size of saRNA constructs may, however, become a limitation when multiple antigens need to be incorporated into the same construct, since the replicase genes already account for a substantial portion of the transcript. Alternative designs may therefore be considered for multivalent approaches. Whether saRNA offers a meaningful advantage for a particular vaccine will ultimately depend on the intended application and required expression profile.

Importantly, the potential applications of mRNA and saRNA extend beyond vaccination, with platforms being investigated for applications including cancer immunotherapy and therapeutic protein expression. [11] In different applications, the desired duration and magnitude of protein expression and the acceptable level of innate immune activation may differ substantially.

Conclusion: saRNA extends the RNA toolbox

saRNA adds intracellular RNA replication to the established concept of transient mRNA expression. This can produce substantial antigen expression from lower administered RNA doses and has now been supported by both preclinical and clinical vaccine data. At the same time, saRNA introduces additional complexity in RNA design, innate immune sensing, formulation, analytical control, and delivery.

When comparing saRNA with conventional mRNA, the key consideration is how the characteristics of each platform align with the requirements of a specific application, including the desired expression kinetics and immune profile. As clinical experience with saRNA expands, these application-specific trade-offs can be defined with greater precision, helping to clarify where each RNA platform may be most appropriate.

References

[1] Bloom K, van den Berg F, Arbuthnot P. Self-amplifying RNA vaccines for infectious diseases. Gene Therapy. 2021;28:117–129.

[2] Silva-Pilipich N, Beloki U, Salaberry L, Smerdou C. Self-Amplifying RNA: A Second Revolution of mRNA Vaccines against COVID-19. Vaccines. 2024;12:318.

[3] Lundstrom K. Self-Replicating RNA Viruses for RNA Therapeutics. Molecules. 2018;23:3310.

[4] Blakney AK et al. Polymeric and lipid nanoparticles for delivery of self-amplifying RNA vaccines. Journal of Controlled Release. 2021;338:201–210.

[5] Vogel AB, Lambert L, Kinnear E, Busse D, Erbar S, Reuter KC, et al. Self-amplifying RNA vaccines give equivalent protection against influenza to mRNA vaccines but at much lower doses. Molecular Therapy. 2018;26:446–455.

[6] de Alwis R et al. A single dose of self-transcribing and replicating RNA-based SARS-CoV-2 vaccine produces protective adaptive immunity in mice. Molecular Therapy.

[7] Oda Y et al. Immunogenicity and safety of a booster dose of a self-amplifying RNA COVID-19 vaccine (ARCT-154) versus BNT162b2 mRNA COVID-19 vaccine: randomized phase 3 non-inferiority trial. The Lancet Infectious Diseases. 2024;24:351–360.

[8] Casmil IC, Friesen JJ, Bathula NV, Strumpel A, Ho CH, Guez I, et al. Divergent Delivery and Expression Kinetics of Lipid and Polymeric Nanoparticles across mRNA Modalities. Advanced Science. 2025;12:e08907. doi:10.1002/advs.202508907.

[9] Akahata W, Komori M, Morey AL, et al. Incorporation of 5-methylcytidine alleviates RIG-I-mediated innate immune responses to a self-amplifying RNA vaccine. Science Translational Medicine. 2025;17(824):eadz2276. doi:10.1126/scitranslmed.adz2276.

[10] Gong Y, Yong D, Liu G, Xu J, Ding J, Jia W. A Novel Self-Amplifying mRNA with Decreased Cytotoxicity and Enhanced Protein Expression by Macrodomain Mutations. Advanced Science. 2024;11(43):e2402936. doi:10.1002/advs.202402936.

[11] Bathula NV et al. Delivery vehicle and route of administration influences self-amplifying RNA biodistribution, expression kinetics, and reactogenicity. Journal of Controlled Release. 2024;374:28–38.

[12] Brito LA et al. A cationic nanoemulsion for the delivery of next-generation RNA vaccines. Molecular Therapy. 2014;22:2118–2129.

[13] Voigt EA et al. A self-amplifying RNA vaccine against COVID-19 with long-term room-temperature stability. npj Vaccines. 2022;7:136.

[14] An Omicron-specific, self-amplifying mRNA booster vaccine for COVID-19: a phase 2/3 randomized trial. Nature Medicine. 2024;30:1363–1372.

[15] Akahata W et al. Safety and immunogenicity of SARS-CoV-2 self-amplifying RNA vaccine expressing an anchored RBD: a randomized, observer-blind phase 1 study. Cell Reports Medicine. 2023.

[16] Beissert T, Perkovic M, Vogel A, et al. A Trans-amplifying RNA Vaccine Strategy for Induction of Potent Protective Immunity. Molecular Therapy. 2020;28:119–128.