Messenger RNAs (mRNAs) represent a fast-emerging class of biotherapeutics. They hold considerable promise, offering new opportunities for targeted treatment and flexible manufacturing, as demonstrated by the rapid development of mRNA-based COVID-19 vaccines [1,2]. However, the field is still in its infancy, and the clinical potential of mRNA extends beyond vaccines. To date, there have been more than 348 clinical studies [3] involving mRNA as an active ingredient or an enabler for cell and gene therapies such as CRISPR-Cas9 derived therapeutics. Lonza has built considerable expertise in the production of mRNA, as both drug substance (DS) and drug product (DP) [4,5].
Why use mRNA in vaccines or therapeutics?
Messenger RNAs are single-stranded nucleic acids transcribed from DNA. When used as a preventative vaccine or a therapeutic drug, mRNA is typically delivered to the cell’s cytoplasm, where it directs production of protein-based antigens. By mimicking the actions of natural mRNAs, therapeutical mRNAs use cells as natural “bioreactors” to produce clinically relevant proteins, thus avoiding challenges associated with some protein-based therapeutics [6]. The safety profile of mRNA therapeutics appears to be favorable, particularly as mRNA translation is transient, occurs in the cytoplasm rather than in the nucleus, and mRNA therapeutics are degraded by natural processes once their function is complete [7].
From a manufacturing perspective, mRNA vaccines and therapies offer advantages in terms of speed and scale when compared to protein-based therapies. As they are produced without using cell culture, they can be manufactured about five or six times faster (Fig. 1) than a typical biologic, using up to 200x times lower process volumes (Fig. 2) for a similar quantity of doses.
Such lower process volumes also result from the lower weight of drug substance per dose, i.e. typically 20 – 100 μg for an mRNA drug substance vaccine dose as compared to a wide range of 10 – 1000 mg for one dose of drug substance derived from Mammalian expression. Therefore, the commercial production processes of monoclonal antibody-based biologics require 1,000 – 20,000 L bioreactors, whereas the production of mRNA therapeutics can be achieved using 5 –100 L bioreactors (Fig. 2). Additionally, the cost of goods (CoGs) for manufacturing mRNA can be lower than for manufacturing protein-based therapeutics, which makes them attractive for numerous therapeutic applications.
Figure 1
Illustrative presentation of process steps for manufacturing LNP encapsulated mRNA bulk (upper part) and Mammalian Protein Drug Substance (lower part). Batch times differ by a factor of 5–6
Figure 2
Illustrative presentation of upstream reactor volumes for manufacturing encapsulated mRNA Drug Substance (left) and Mammalian Protein Drug Substance (right). Reactor volumes differ by a factor of about 200.*
Challenges in mRNA manufacturing
mRNAs are unstable molecules, being sensitive to shear stress and degradation. As a result, specific upstream and downstream processes are required to ensure integrity and stability during manufacturing. Work on stable manufacturing platforms is ongoing, as evidenced by the plethora of different processes and strategies that have been developed by biopharmaceutical companies to achieve their specific needs. An effective CDMO will be able to understand the process, and adapt it to meet cGMP standards.
Another challenge in the mRNA field is the implementation of analytical tools that allow the process development to be monitored, as well as the definition of an analytical release panel that meets the requirements of the various Health Authorities. In addition, the scarcity of cGMP-compliant equipment is acknowledged by both manufacturers and suppliers. Apart from the notable exception of affinity resins (oligo-dT), few resins or membranes are readily available for the cGMP manufacturing of mRNA products.To address the challenges in mRNA manufacturing, it is essential to make informed decisions in order to avoid late-stage changes to the process that would, inevitably, impact a project’s timeline.
What options are there for DNA template production?
The DNA template is a critical raw material for mRNA manufacturing, and the starting point for the in vitro transcription of mRNA. Currently, the vast majority of mRNA projects use plasmid DNA (pDNA) derived from microbial fermentation as a template. Manufacturing pDNA requires well-established workflows and expertise, while the process can usually be replicated quite easily at a manufacturing scale. Nevertheless, using pDNA has several drawbacks, including the time-consuming purification methods, and the necessary segregation of other manufacturing activities from mRNA manufacturing. Additionally, few manufacturers have the capacity and expertise to manufacture pDNA at the required quality levels for clinical or commercial manufacturing purposes [8, 9].
One alternative to pDNA is to use DNA derived from a polymerase chain reaction (PCR). This has advantages in terms of manufacturing speed and - as the components can be of animal- free origin - the analytical requirements for release can be simplified. However, using PCR-generated DNA poses challenges for large- scale production, and the PCR- generated template may have sequence restrictions at the 3’ end, derived from the use of primers. Adjustments can add costs and complexity to the process, and licensing costs and IP issues need to be considered. Concerns have also been raised about the fidelity of the DNA template derived from PCR [10, 11]. Finally, the use of PCR-derived DNA may require additional studies for regulatory approval.
Another alternative technology for mRNA production is the use of doggybone DNA (dbDNA). Such templates are currently offered by a single company, using an in vitro process with Φ29 DNA Polymerase and primers that initiate rolling circle amplification, producing long, doggybone-shaped double-stranded DNA. dbDNA has the advantage of speed — it is quick to produce, with a similar fidelity to pDNA. Touchlight, the originator of this technology, claims it has been used to deliver 5 g of cGMP dbDNA to an undisclosed major pharma company within three weeks [12]. As the process components are of animal-free origin, the downstream purification process is straightforward, with few enzymatic digestion steps. However, a thorough analysis of potential IP issues and supply chain challenges would be required.
* The scales for the manufacturing of Lipid Nanoparticles, as the currently leading non-viral delivery system for genetic drugs, are not discussed herein.
Technologies for mRNA in vitro transcription
mRNA is produced by incubating DNA template with an RNA polymerase – usually the T7-RNA-Polymerase – and nucleotides (NTPs) in a cell-free in vitro transcription (IVT) process performed in stirred-tank bioreactors as a principal process step. Capping and poly-A tail addition, which can be performed in a co-transcriptional or post-transcriptional manner, give in vivo stability and efficacy, while providing a sequence that allows highly specific affinity purification resins such as oligo-dT to be implemented. However, a strategy using post-transcriptional capping and poly-A addition requires additional enzymatic and thus downstream steps. This adds complexity, reduces the yields, and increases processing times (Fig 3, 4). Using modified nucleotides in addition to the enzymes and additives also has an impact on the mRNA quality, yields and impurities profile. Manufacturing mRNA to cGMP standards must use well-defined raw materials to minimize the number of contaminants that might be introduced into the reaction. This also reduces the numbers of purification steps required to provide high- quality mRNA, often resulting in higher yields.
Which polymerase?
There is a bewildering array of polymerase (T7, T3 and SP6 are commonly used for in vitro transcription) to choose from, including standard to thermostable and engineered versions. Standard wild type-derived T7 polymerases are the most commonly used and are available from multiple suppliers, but when choosing a polymerase for manufacturing scale, a biopharma company must consider whether their supplier will be able to provide sufficient volumes at the right quality, and at a price that is not going to impact your CoGs, supply chain or the safety of your final product. By placing these caveats on suppliers, the list of potential suppliers can be narrowed down to a small selection of suitable companies.
Wild type T7 polymerases are well studied, and the resulting impurity profile, limitations, and advantages are well known. Thermostable T7 polymerases at reaction temperatures up to 50°C have the benefits of producing fewer dsRNA byproducts, and more efficient co-transcriptional capping. This can reduce the number of purification steps, but, because of the energy cost of the high reaction temperatures, this can make your process expensive to run at scale. Engineered T7 polymerases can be used at 37°C, yet also produce fewer dsRNA byproducts, and yield more efficient co-transcriptional capping. This reduces the likelihood that an ‘Out of Specification’ batch will be produced.
When considering which polymerase to use, Lonza has experience of producing mRNA at different scales, and can advise on the cost benefits of using different technologies at scale. We can also perform comparability studies to determine which polymerase is optimal for a specific mRNA production process. For example, using a thermostable or engineered T7 polymerase can reduce the number of downstream purification steps. At a small scale, more steps may not add significant cost, but at manufacturing scale can become a major expense that impacts your CoGs. A CDMO can also provide guidance on licensing costs and supply chain issues, as thermostable and engineered T7 polymerases are available from only a handful of suppliers and not always in GMP Grade. Additionally, a CDMO with good regulatory experience can advise if it is appropriate to switch from using a low-cost polymerase to a more tailored, yet expensive, one, and whether it is necessary to perform comparability studies for each polymerase before manufacturing engineering batches or clinical material.
Figure 3
Co-transcriptional capping uses an mRNA cap analog in the transcription reaction. When using CleanCap® AG (left), a Cap-1 structure is accomplished with high efficiency (>90% ) and in the correct orientation.
Post-transcriptional (enzymatic) capping High yield can be achieved with both processes, and the determining factor in selection will be an assessment of the process time, total CoGs and security of supply chain, plus other considerations related to process implementation.
Ranking mRNA production materials
To produce mRNA to cGMP standards, a range of raw materials are required for buffer production, nucleotides, capping reagents, and enzymes for IVT, as well as enzymes that aid increased yield and downstream processes. Lonza continually evaluates suppliers and assesses their products according to criteria such as quality, price, volume required, security of supply chain, and the cGMP compliance of their manufacturing. Table 1 gives an illustrative example of four suppliers of these materials, excluding enzymes for IVT. As Lonza procures significant quantities for multiple projects, we routinely engage with suppliers in discussions about the pricing of their products. Lonza also has extensive experience in establishing supply chains, including preferred suppliers, thus granting customers freedom of choice and flexibility for successful projects.
Capping technology
mRNA capping is critical to the production of biologically active and stable mRNA. Two main capping technologies are currently used commercially for manufacturing cGMP grade mRNA: enzymatic [14] and co-transcriptional capping [15]. These are illustrated in Figure 3.
Enzymatic capping involves the use of an enzyme, usually the Vaccinia Capping Enzyme and mRNA Cap 2’-O-methyltransferase, to add the Cap-1 structure at the 5’ end of the mRNA. This method is well established, and has a high capping efficiency of up to 100%. The main drawback is the number of additional downstream steps required (Fig 3, 4).
The co-transcriptional method uses RNA polymerase to add a 5’ cap structure during the in vitro transcription. This means both, IVT and capping reaction take place simultaneously, thus reducing the number of reaction steps (Fig 3, 4). Capping efficiency varies, and wrong orientations have been reported. Some commercially available solutions, such as Anti Reverse Cap Analog N7-Methyl-3’-O-Methyl-Guanosine-5’-Triphosphate-5’-Guanosine (ARCA®) or CleanCap® appear to be more reliable in giving the correct orientation, and yields ranging from 70 – 95% have been reported. Co-transcriptional capping has been used in commercial mRNA production, most notably in the Pfizer/ BioNTech Comirnaty (BNT162b2) mRNA-based COVID-19 vaccine [16]. However, as the technology is proprietary and only available from one supplier, there are economic challenges that need to be considered.
When considering which capping technology to use, a CDMO such as Lonza with experience of testing different methods at a range of scales can advise you on the cost benefits of using different technologies. They can also provide guidance on when it makes commercial sense to switch from one technology to another. For example, the cost of reagents can often outweigh the cost of additional workflow steps when bulk pricing discounts are not available and larger reaction volumes are involved.
Figure 4
Process definition is impacted by the choices made for the IVT step. Implementation of tangential flow filtration (TFF) with one or more orthogonal chromatography steps can ensure the development of a cGMP-compliant process with an acceptable product impurity profile. Actual processes may differ.
Process design considerations
The design of a mRNA process depends largely on the choices of raw materials, capping and poly-A addition for the in vitro transcription. As discussed earlier, raw materials such as the template, whether natural and/or modified nucleotides are used, the T7 RNA Polymerase (wild-type vs engineered, and which supplier), additives and aiding enzymes, will have a significant impact in the mRNA in terms of quality, yields and the impurity profile (dsRNA, RNA integrity, poly-A tail length, residual raw materials etc.). All these factors will impact the definition of the downstream process that needs to be implemented to ensure the Target Product Profile is met. The definition of a process will also have to be aligned in terms of costs, process time and IP considerations, among others.
Conclusions
The production of mRNA for use in vaccines and therapeutics can be difficult and costly if small and medium-sized biopharma choose to keep it in- house. There are many choices to be made in the workflow for both the DNA and the mRNA. Without guidance from a CDMO, who can provide input on process implementation, cost benefit, IP, supply chain and regulatory issues, mRNA manufacturing can be challenging from both technical and economic standpoints. Partnering with a CDMO that has end-to-end mRNA manufacturing knowledge can add capacity and expertise to your projects, and can deliver gram-scale cGMP mRNA suitable for clinical and commercial programs.
Lonza has a global network with well-designed supply chains, and contracts with cGMP material suppliers that have been negotiated at competitive rates. Lonza assesses and chooses the most appropriate suppliers of equipment, production systems and materials, as well as providing guidance on which technology options to use to ensure your mRNA is manufactured to match the cost, efficacy, and safety specifications of your vaccine or therapeutic.
References
1 Stanton, D. Moderna contracts Lonza to help scale up COVID-19 mRNA vaccine candidate. Bioprocess Insider, 2020. https://bioprocessintl.com/bioprocess-insider/facilities-capacity/moderna-contracts-lonza-to-help-scale-up-covid-19-mrna-vaccine-candidate/
2 Thanh Le, T. et al. The COVID-19 vaccine development landscape. Nat Rev Drug Discovery. 2020; 19(5):305–306. https://doi.org/10.1038/d41573-020-00073-5
3 ClinicalTrials.gov Database Search for Clinical Trials with mRNA therapeutics | Recruiting, Active, not recruiting. https://www.clinicaltrials.gov/ct2/results?term=mRNA+therapeutics&Search=Apply&recrs=a&recrs=d&age_v=&gndr=&type=&rslt=Accessed 30.11.2022.
4 Mullin, R. Lonza to build mRNA and small-molecule plants in Visp. C&EN. 2021; 99(16). https://cen.acs.org/business/outsourcing/Lonza-build-mRNA-small-molecule/99/i16
5 Press release 2021: Lonza and Moderna Announce Further Collaboration for Drug Substance Manufacturing of COVID-19 Vaccine Moderna in the Netherlands. https://www.lonza.com/news/2021-06-02-07-02
6 Sahin, U., Karikó, K., and Türeci, Ö. mRNA-based therapeutics — developing a new class of drugs. Nat Rev Drug Discov. 2014; 13:759–780. https://doi.org/10.1038/nrd4278
7 Anand, P., Stahel, V.P. The safety of COVID-19 mRNA vaccines: a review. Patient Saf Surg. 2021; 15:20. https://doi.org/10.1186/s13037-021-00291-9
8 Pearson, S. Wacker Biotech Bullish on Microbial Biomanufacturing. GEN, 2019. https://www.genengnews.com/insights/wacker-biotech-bullish-on-microbial-biomanufacturing/
9 Lonza Invests in Microbial Development Lab Expansion. Pharmaceutical Technology. 2021; 16(12). https://www.pharmtech.com/view/lonza-invests-in-microbial-development-lab-expansion
10 Potapov, V., Ong, J.L. Examining Sources of Error in PCR by Single-Molecule Sequencing. PLoS One. 2017; 12(7): e0181128. https://doi.org/10.1371/journal.pone.0169774
11 McInerney, P., Adams, P., Hadi, M.Z. Error Rate Comparison during Polymerase Chain Reaction by DNA Polymerase. Mol Biol Int. 2014; Article ID 287430. https://doi.org/10.1155/2014/287430
12 Stanton, D. Doggybone DNA ‘ideal tech’ for pandemic vaccine response, says Touchlight. BioProcess Insider, 2021. https://bioprocessintl.com/bioprocess-insider/therapeutic-class/doggybone-dna-ideal-tech-for-pandemic-vaccine-response-says-touchlight/
13 Karda, R., Counsell, J.R., Karbowniczek, K. et al. Production of lentiviral vectors using novel, enzymatically produced, linear DNA. Gene Ther. 2019; 26:86–92. https://doi.org/10.1038/s41434-018-0056-1
14 Fuchs, A.L., Neu, A., Sprangers, R. A general method for rapid and cost-efficient large-scale production of 5’ capped RNA. RNA. 2016; 22(9):1454–1466. https://doi.org/10.1261/rna.056614.116
15 Jackson, N.A.C., Kester, K.E., Casimiro, D. et al. The promise of mRNA vaccines: a biotech and industrial perspective. npj Vaccines. 2020; 5:11. https://doi.org/10.1038/s41541-020-0159-8
16 Sahin, U., Muik, A., Derhovanessian, E. et al. COVID-19 vaccine BNT162b1 elicits human antibody and TH1 T cell responses. Nature. 2020; 586(7830):594–599. https://doi.org/10.1038/s41586-020-2814-7
