Insight

mRNA vaccines feature rapid production and robust process versatility, conferring unparalleled advantages in responding to emergent public health epidemics. The successful deployment of SARS-CoV-2 mRNA vaccines has greatly propelled the development of numerous candidate mRNA vaccines targeting infectious diseases, cancers, genetic disorders and other conditions, demonstrating enormous application potential. Currently, extensive research is underway focusing on nucleotide modification and optimization of lipid nanoparticle (LNP) delivery systems, aiming to further enhance the immunogenicity, efficacy and stability of mRNA vaccines. This article outlines the fundamental manufacturing workflow of mRNA vaccines.

The core mechanism of mRNA vaccines is to deliver antigen-encoding mRNA into host cells via appropriate delivery vectors. Host cellular machinery is then utilized to translate and synthesize target antigen proteins, which trigger immune responses against viruses or pathogenic cells and ultimately achieve immunization or therapeutic effects.

Two major categories of mRNA vaccines have attracted widespread attention: conventional mRNA vaccines and self-amplifying mRNA (saRNA) vaccines. Distinct from conventional mRNA vaccines, saRNA carries an additional open reading frame (ORF) that encodes RNA-dependent RNA polymerase (RdRp). This enzyme drives the replication of antigen-encoding mRNA, enabling higher and more sustained antigen expression. Accordingly, saRNA vaccines can elicit potent immune responses at lower dosages. Nevertheless, the excessive length of saRNA sequences and potential innate immune activation remain key challenges for saRNA vaccine development. Both conventional mRNA and saRNA share core structural elements including the 5′ cap, untranslated regions (UTRs), ORF and poly(A) tail, all of which are indispensable for efficient translation and antigen expression. Rational optimization of these components can improve translation efficiency and reduce the effective dose of mRNA required for immune activation.

Eukaryotic mRNA possesses a 7-methylguanosine (m7G) cap at the 5′ terminus. This structure binds to eukaryotic initiation factor 4E (eIF4E) to initiate protein translation. The 5′ cap also maintains mRNA stability by protecting the transcript from exonuclease degradation and helps exogenous RNA evade recognition by the innate immune system. For in vitro transcribed (IVT) mRNA, 5′ capping is commonly achieved either by employing vaccinia virus-derived enzymes or co-transcription with cap analogs such as ARCA and CleanCap. Incomplete capping leaves uncapped transcripts vulnerable to recognition as viral RNA, which triggers RNA degradation and undesirable immune activation. Therefore, enzymatic dephosphorylation is routinely performed post-IVT to eliminate uncapped mRNA and mitigate adverse immune reactions.

5′-UTR and 3′-UTR regulate the translation efficiency of mRNA. The 5′-UTR facilitates ribosome loading and translation initiation, whereas long sequences with high GC content or stable secondary structures inhibit these processes. Aberrant start codons within the 5′-UTR may disrupt the translation of the target ORF. The 3′-UTR modulates mRNA stability, subcellular localization and translation activity. Optimization of UTRs is critical for improving overall vaccine performance. Common strategies include adopting 5′-UTRs derived from highly expressed human genes (e.g., α-globin and β-globin), screening optimal 3′-UTRs via random UTR libraries and SELEX technology, and designing high-efficiency 5′-UTRs through computational prediction of ribosome loading and translation efficiency.

Codon optimization of the ORF enhances translation efficiency and boosts protein production. However, excessively rapid translation may lead to incorrect protein folding and loss of immunogenicity. Related research directions include using rare codons to modulate translation rate, adjusting the GC content of the ORF to improve mRNA stability and translation efficiency, and leveraging artificial intelligence (AI) to optimize ORF sequences, predict structural stability and modification sites, and redesign secondary structures to eliminate degradation-prone regions.

The poly(A) tail is a hallmark of native eukaryotic mRNA, playing vital roles in maintaining transcript stability, regulating translation efficiency and extending the in vivo half-life of mRNA. It is recognized by poly(A)-binding protein (PABP), which further interacts with eIF4G. The length of the poly(A) tail is functionally critical. Studies have verified that a poly(A) tract consisting of 120 nucleotides delivers optimal protein expression for IVT mRNA vaccines.

Two mainstream approaches are applied to append the poly(A) tail to IVT mRNA:

The entire production process starts with identifying the optimal antigen sequence for the targeted pathogen or disease. The ORF is then codon-optimized for the host organism, followed by the construction of plasmid vectors. Plasmid DNA (pDNA) is subsequently produced and purified using Escherichia coli expression systems. The purified pDNA is linearized and used as the template for in vitro mRNA transcription, after which crude mRNA products undergo purification. The purified mRNA is encapsulated into lipid nanoparticles (LNPs) or other delivery vectors to form mRNA-LNP complexes. Finally, the bulk material undergoes formulation, concentration and lyophilization to yield the final mRNA vaccine product.

Once the optimized antigen ORF is finalized, the target sequence is inserted into a plasmid vector containing all essential mRNA structural elements. The integrity of the full-length mRNA construct is verified via DNA sequencing and restriction enzyme digestion. Recombinant plasmids are transformed into E. coli, and positive clones are selected for large-scale cultivation and amplification. GMP-grade pDNA templates are essential for high-quality in vitro mRNA synthesis. Strict quality control criteria must be established to monitor impurities and process inhibitors, which may compromise IVT efficiency and subsequent mRNA purification.

In vitro transcription synthesizes mRNA transcripts using DNA templates (pDNA or PCR products). Plasmid DNA is the preferred template for large-scale GMP production due to its cost-effectiveness and low mutation risk. First, restriction endonucleases are used to linearize pDNA at specific sites downstream of the poly(A) sequence to ensure precise transcription termination. RNA polymerases are then employed to transcribe linearized plasmids into mRNA. Rigorous quality control strategies are implemented throughout the process to maximize yield and minimize byproducts, particularly double-stranded RNA (dsRNA).

High-purity mRNA is a prerequisite for homogeneous, stable formulations, as well as high yield and purity of in vivo antigen expression. Post-IVT, crude mRNA must be purified to remove residual reaction components including enzymes, template DNA, truncated transcripts and free nucleotides, with special emphasis on dsRNA removal. Prior to purification, DNase I is added to digest template DNA, which reduces feed viscosity. Resulting small DNA fragments are further removed via ultrafiltration or chromatographic purification. Available purification technologies include lithium chloride (LiCl) precipitation, ion exchange chromatography, affinity chromatography, size-exclusion chromatography and tangential flow filtration (TFF).

Naked mRNA is susceptible to degradation by ribonucleases and cannot efficiently cross cell membranes. Advanced delivery systems facilitate cellular uptake of mRNA, prolong its in vivo half-life and mitigate unwanted immune activation. A variety of delivery platforms have been developed, including lipid-based carriers, peptide vectors, dendritic cell (DC)-targeted systems and polymeric nanoparticles. Among these, LNPs represent the most clinically validated delivery system, which have been successfully applied in FDA-approved COVID-19 mRNA vaccines and other investigational products such as Iribovax COVID-19 mRNA vaccine.

mRNA encapsulation into LNPs is typically performed via microfluidic mixing, where aqueous mRNA solution is combined with lipid-containing organic solution. This technology enables scalable manufacturing and homogeneous particle distribution of final mRNA-LNP formulations.

LNPs for mRNA vaccine delivery consist of four core components: ionizable lipids, neutral helper lipids (e.g., distearoylphosphatidylcholine, DSPC; dioleoylphosphatidylethanolamine, DOPE), polyethylene glycol (PEG) lipids and cholesterol. Ionizable lipids carry positive charges at low pH and become neutral under physiological pH, ensuring excellent biocompatibility and minimal non-specific interactions with cell membranes. PEG lipids stabilize nanoparticles by preventing particle aggregation and reducing binding to serum proteins. Cholesterol maintains the structural integrity of lipid bilayers. A key functional feature of LNPs is the capability to facilitate endosomal escape: after cellular internalization, the acidic environment within endosomes protonates ionizable lipids, which disrupt the endosomal membrane and release mRNA into the cytoplasm for antigen translation.

If the concentration of crude mRNA-LNP fails to meet formulation specifications, concentration via ultrafiltration is required. Both flat sheet cassettes and hollow fiber membranes are commonly used. The molecular weight cutoff (MWCO) of ultrafiltration membranes is selected based on the hydrodynamic radius of mRNA-LNP complexes, with a typical range of 50–300 kDa. Critical quality attributes monitored during ultrafiltration include particle size, polydispersity index (PDI), zeta potential, encapsulation efficiency and overall product recovery before and after concentration. Compared with flat sheet cassettes, hollow fiber modules feature open flow channels with low shear stress, which preserves particle size uniformity and minimizes product retention at the end of concentration. These advantages make hollow fiber ultrafiltration well-suited for large-scale mRNA vaccine production. To reduce microbial load prior to sterile filtration, gamma irradiation or pre-sterilized fluidic paths can be adopted.

Lyophilization is an effective approach to improve the thermal stability of mRNA vaccines. Optimized formulations combined with appropriate lyoprotectants maintain high product stability during freezing and long-term storage. After reconstitution, lyophilized LNPs should fully restore their original particle size and transfection activity.