The RNA revolution, advancing beyond traditional vaccines to new therapeutic modalities and constructs such as self-amplifying RNA (saRNA) and circular RNA (circRNA), continues to place increasing pressure on downstream purification. Diffusion-limited resins, the time-tested workhorse of protein purification, are fundamentally incompatible because mRNA is an enormous (>40 nm) molecule with low diffusivity. Our perspective, rooted in core chemical engineering principles, applies transport analysis and re-examines published performance data to demonstrate why even optimized perfusion chromatographic resin systems, exhibiting 0.1 % of total flow through the resin particles, cannot overcome the inherent diffusional barriers preventing efficient RNA purification. Alternatively, convection-based devices, notably membranes and monoliths, are well situated as their transport characteristics are not limited by the molecular transport properties of RNA. Ultimately, pressure-driven flow enables the potential for increased device capacity at orders of magnitude (103x) lower process time and smaller device footprint contributing to markedly improved productivity. Taken together, these findings suggest that a paradigm shift is required toward convective membrane systems to create a platform capable of delivering scalable, economic, and ultimately industrially attractive mRNA purification. https://doi.org/10.1016/j.biotechadv.2025.108774
Microfiltration (MF) membranes are key components for bioprocessing applications such as sterile filtration, clarification and protein/nucleic acid purification. The morphology of MF polymer membranes synthesized by phase inversion (PI) is characterized by a lognormal pore size distribution and regions in pore space with fast flow channeling which negatively affects selectivity and surface area utilization due to preferential solute transport. Here, we identify and quantify flow channeling and assess its effect on the selectivity and permeability of a commercial, 0.22 μm mean pore size, polyvinylidene fluoride (PVDF) membrane. The 3D membrane microstructure, acquired with focused ion beam-scanning electron microscopy, was used for in silico fluid flow, and particle filtration studies. In silico filtration results were in good agreement with experimentally observed particle rejection behavior. From flow field simulations, fast flow channels associated with larger diameter pores were observed and despite occupying only 2.5 % of pore space volume, these channels contributed to 15 % of total fluid flow. The longest channel length was nearly 7 times the mean pore throat diameter. Particle trajectory analysis showed that particle capture in channels which are characterized by fast flow is lower than that in other slower-flow regions, thereby resulting in poor separation. In silico filtration studies in which all the fast flow channels were eliminated resulted in selectivity increasing from 2.5 1.2 (original membrane) to 482 350 whereas the permeability decreased by 4 times. Simulations showed that eliminating a greater volume fraction of channels led to improved performance, exceeding the selectivity – permeability upper bound for commercial polymeric microfiltration membranes. This in silico framework enables the design of a new class of polymer membranes with reduced channeling and an improved selectivity–permeability trade-off. https://doi.org/10.1016/j.memsci.2025.124747
It is commonly recognized that hydrophilic surfaces reduce protein membrane adhesion during aqueous bioprocessing due to water's strong binding through electrostatic and hydrogen bonding capability. Here, we show that when (i) a protein displaces bound water at a polymer interface, hydrogen bonding and electrostatic interactions with the polymer membrane surface drive protein adhesion, and (ii) comparing two commonly used commercial hydrophilic polymer membranes with different polar surface modification chemistries, the one with higher hydrogen bonds (modified polyethersulfone (mPES)) exhibited three times higher adhesion force to a hydrophilic protein (streptavidin) than the one with less hydrogen bonds (modified polyvinylidene fluoride (mPVDF)). Stronger protein-membrane hydrogen bonding for mPES as corroborated by its higher electron donor surface energy component and higher hydrogen bonding propensity observed from surface energy measurements and by solvation shell spectroscopy, respectively, support our explanation of these results. Atomic force microscopy (AFM) colloid probe technique was used here to measure intermolecular forces/energy between streptavidin and two polymeric membrane surfaces. Non-contact forces at separations greater than 2 nm were modeled using the DLVO theory, while contact/adhesion forces, which include hydrogen bonding, were measured at separation ~0.16 nm. These findings highlight the importance of protein-polymer membrane hydrogen bonding interactions in selecting polymers for membrane downstream purification and other applications. https://doi.org/10.1016/j.jcis.2025.138530
The increasing clinical trials of single-stranded mRNA (ss-mRNA) therapeutics highlight the urgent need to develop efficient, scalable, and economic purification methods. Current diffusion-driven, resin-based purification techniques constrain productivity and rely on expensive oligo(dT) ligands for target ss-mRNA poly(A) tail hybridization. To overcome these challenges, we use interfacial molecular forces, such as charge and hydrogen bonds, between nucleic acid variants and a positively charged synthetic microporous membrane to purify ss-mRNA, a desirable therapeutic, from an undesirable impurity, immunogenic double-stranded RNA (dsRNA). Membranes achieved high binding capacities (1.28 mg/m2) and up to 100% ss-mRNA recovery at ~pH 9.0, with optimized surface density (4000 to 10,000 nmol/m2). Purification was operated at rapid flow rates (1.5 ml/min,1000 MV/min) with reusability (>10 trials) and negligible ligand leaching. The key discovery of this cost-effective ligand-less multimodal surface-modified approach is that the addition of the polyamine spermine, which selectively neutralizes dsRNA charge at amine-to-phosphate ratios >450, enhanced separation efficiency. https://doi.org/10.1126/sciadv.adv8656
In the rush to produce mRNA vaccines and ensure integrity and purity, capital-intensive purification steps were used, typically involving slow diffusion-driven, inefficient, and costly chromatographic resin processes. We present the first results that microporous affinity synthetic polymer membranes can outperform commercial monolithic and resin approaches for capture of mRNA, thereby promoting fast purification of mRNA vaccines and therapeutics. Our seminal finding is that capture of Fluc-mRNA by synthetic affinity membranes enables efficient intact mRNA recovery (∼100 % of bound mass), faster processing times and lower dispersion by leveraging convective flow with short diffusional path lengths. We also demonstrate higher ligand accessibility and device productivity. Accessible ligand per device volume was 87–94 %, ∼74 % and ∼53 % for the membranes, monoliths, and resin column, respectively. A calibrated mathematical framework predicted mRNA pulse breakthrough with dispersion that was measured in the process tubing and membrane module using a non-reactive tracer and modeled using ideal reactors. https://doi.org/10.1016/j.seppur.2024.129310