A pandemic concentrates the mind wonderfully. Ever since the COVID-19 pandemic struck, government, academia, business, and the general public have demonstrated an intense interest in mRNA technologies, not least of which are mRNA technologies that promise to speed the manufacture of mRNA-based vaccines. A representative manifestation of this interest was the 2024 mRNA Technology Conference, which was organized by the American Institute of Chemical Engineers.
At this event, which was held last April, stand-out presentations covered topics such as a rapidly deployable continuous flow RNA manufacturing platform; the continuous manufacturing of synthetic ribonucleotide triphosphates; adsorptive membranes for fast, efficient mRNA purification; and fast in-process analytics to enable robust mRNA manufacturing.
To hear the latest on these topics, GEN recently followed up with the presenters. We can confirm that although the intensity of the public’s interest in COVID-19 may have faded, the minds of the presenters remain concentrated on their work and how it can secure and improve the manufacturing processes for mRNA biologics. Several topics—such as the development of more compact, agile, and efficient ways to manufacture mRNA and its components, as well as the development of analytical techniques to ensure vaccine quality—are bound to be of perennial interest.
RNA on demand
Currently, most RNA biologics are manufactured in either big batch reactions or by discrete reactions linked together in a continuous way, with each batch requiring a significant amount of process development. Reaction kinetics change as nucleotide triphosphates (NTPs) and other reagents become depleted. Larger, conventional systems are designed and built to produce larger volumes of product, while at the same time, it is difficult for smaller liquid handling systems to be scaled up to meet higher demands.
These conventional systems, which are designed for tasks such as the delivery of higher volumes of a single species of mRNA to a substantial vaccination campaign, cannot easily be distributed to produce RNA at point of care or mobilized for parallel clinical trials for personalized therapy.
The new NTxscribe platform, which was released last month, is an end-to-end manufacturing process for the synthesis of mRNA, functional modification additions, and impurity removal, all in a single-use, closed, sterile process. Duval asserts that it “essentially allows you to put a template in one side and out the other side will come purified RNA.”
The benchtop NTxscribe can be set up in hours and requires minimal process development. According to Duval, it can make 25–50 mg of RNA in an hour and is capable of scaling, both by running the instrument longer (“it linearly scales so you continue to make RNA at the same rate”) and by running multiple units in parallel (“about 25 of these machines could make about 300 million doses of vaccine in about a month”). He adds, “When you’re finished with one synthesis, you could turn over and produce a completely different mRNA molecule on the same system with a different single-use cartridge.”
Hedge against the supply chain
Precursor nucleosides are required to manufacture RNA vaccines, and these may not always be available in the quantity needed, especially during times of supply chain disruption such as a global pandemic or heightened political tensions. To help secure the domestic supply chain and maintain a reliable source of these critical raw materials, Snapdragon Chemistry, a Cambrex company, partnered with the Biomedical Advanced Research and Development Authority to develop a platform that uses low-cost, fungible feedstocks to manufacture synthetic NTPs.
The researchers not only had to make the actual NPT molecules, but also needed to make solutions of several other precursor molecules that were not commercially available, that is, solutions of specific phosphate compounds that were not shelf-stable because the compounds would hydrolyze and degrade. These precursors “were really critical to be able to actually do that transformation, to convert it from the base into the NTP,” recalls Kevin Nagy, PhD, senior director of engineering and manufacturing at Snapdragon.
The combination of chemistry, engineering, and automation was critical to success. “Nothing we operated was turnkey off-the-shelf,” Nagy notes. “We basically built the precursor manufacturing equipment. We built the flow chemistry equipment fit for purpose … drawing from many other fields that may not be obvious to someone working in this space.” The continuous column chromatography system was inspired by process technology of the grain processing industry for de-ashing, for example.
Synthesis of the precursor was driven and enabled by process analytical technology and automation, Nagy says. “We had multiple columns running in parallel, in different phases. And then we were also doing a quench of effluent coming out of those columns with a neutralization reaction (to precisely control the pH).” The work was completed in about seven months, resulting in a skid-mounted platform that can provide raw materials for more than 250 million doses of COVID-19 or other mRNA-based vaccines.
Affinity, not diffusion
The COVID-19 pandemic engendered a rush to create massive numbers of doses of mRNA vaccine. To purify the mRNA, industry often relies on standard methods designed for relatively small biologics such as monoclonal antibodies. Yet an immunoglobulin G molecule is perhaps 10 nm in diameter, whereas mRNA is about 100 nm. “And that means that messenger RNA has about an order of magnitude lower diffusion coefficient than IgG,” says Thomas Neuman, a third-year doctoral student in chemical engineering at Rensselaer Polytechnic Institute.
He and his colleagues set out to establish an mRNA purification platform similar to the protein-A columns for antibodies. The group used an oligo-dT membrane to achieve a more fit-for-purpose solution to capture RNA.
When you consider a resin-based separator, which relies on diffusion, the molecule must travel through the chromatography matrix and then diffuse into the bead, “and that ends up having to take a really long time if you want the mRNA to interact with the resin,” Neuman explains. A membrane, he continues, relies on flow-through pores, “so you’re actually using pressure-driven flow to force the mRNA into the membrane, and you only need a small diffusional distance to interact with the membrane surface. And that plays a large role in improving matrix accessibility and reducing the residence time of the process, which really increases the productivity.”
The group examined different parameters, including oligo-dT length and spacing, as well as ligand density, using static testing and, ultimately, dynamic characterization to optimize the membranes in terms of cost, productivity, and sharpness of elution peak.
Neuman emphasizes that membranes are versatile and scalable. Membranes can be used in different configurations, including flat sheet, spiral wound, cassette, and hollow fiber devices; these devices can be deployed as modules; and multiple modules can be used to facilitate scale up. “Hollow fibers in particular are very attractive because they have a very good packing density,” he notes. “You can essentially just add more fibers without making major changes to your process.” These membranes are capable of operating over a relatively large flow rate with no change in efficiency.
The whole package
RNA is only one of the components that make up an mRNA vaccine. Because mRNA is not stable and is vulnerable to myriad RNAses throughout the body, it needs to be protected—generally by enveloping it in lipids. “You mix the mRNA with four different lipids which then form a particle,” termed a lipid nanoparticle (LNP), explains Aleš Štrancar, PhD, managing director, Sartorius BIA Separations.
In a perfect world, an LNP would consist of four lipids basically making a circle around a single mRNA molecule. Yet in practice, an LNP may be empty, or it may incorporate more than one mRNA. LNPs may form aggregates, complexes, and all kinds of structures. Consequently, LNP samples harbor considerable heterogeneity.
Štrancar maintains that employing orthogonal, analytical methods to determine encapsulation efficiency would be advantageous in two respects. First, it would improve quality control for the vaccines produced batch to batch, and from manufacturing site to manufacturing site. Second, it would provide better insight into LNP formation—as well as facilitate the selection of less problematic (for example, less immunogenic) lipids.
Štrancar points out that one way to embrace orthogonality is to use Sartorius’s PATfix Switcher method in addition to the established methods already recognized by regulatory bodies. According to Štrancar, not only can an orthogonal approach provide insights into a given product’s quality and reproducibility, but it can also provide more data to researchers, process developers, and quality control personnel so that they may compare the quality and reproducibility of different products.
