Tackling intrinsic instability of mRNA vaccine by a rapid onsite microfluidic assembly (ROMA) technology

doi:10.21203/rs.3.rs-5047616/v1

Download PDF

Article

Tackling intrinsic instability of mRNA vaccine by a rapid onsite microfluidic assembly (ROMA) technology

https://doi.org/10.21203/rs.3.rs-5047616/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Despite achieving unprecedented success, current mRNA vaccines face significant challenges, including thermo-instability, degradation, and infrastructure-dependence, making customizable supply a distant goal. Here, we describe a Rapid Onsite Microfluidic Assembly (ROMA) technology capable of generating ready-to-inject mRNA vaccines with a real-time quality inspection as a solution. Diverging from traditional manufacturing mechanism of directly assembling mRNA and lipids into mRNA-LNPs, ROMA technology utilizes mRNA and pre-made empty LNPs to form mRNA-LNPs that exhibit equivalent physiochemical parameters and in vivo expressions compared to conventional ones. Our ROMA prototype offers personalized options for mRNA vaccines, including lipid nanoparticle (LNP) sizes, compositions, mRNA types, and dosages tailored to individual needs, at a throughput of 200 doses/hour (∼100 µg mRNA/dose) with scalable potential. Crucially, ROMA mRNA vaccine, immediately deployable without the need for storage, fundamentally avoids the intrinsic thermal instability and degradation risks associated with conventional ones. This transformative ROMA technology offers unparalleled user-end convenience, unlocking the translational potential for personalized mRNA vaccines and treatments, thereby significantly expanding the scope of mRNA-based therapeutics.

Physical sciences/Nanoscience and technology/Nanomedicine/Drug delivery

Physical sciences/Engineering/Chemical engineering

Physical sciences/Engineering/Biomedical engineering

Physical sciences/Materials science/Biomaterials/Biomaterials – vaccines

Physical sciences/Nanoscience and technology/Nanobiotechnology/Microfluidics

A distinctive double-edged characteristic of mRNA vaccine lies in the susceptibility of mRNA to degradation^1–5, ensuring its superior biological safety compared to conventional ones⁶, but determining that mRNA cannot be used independently. Therefore, mRNA must be protected within a carrier to achieve its optimal effectiveness⁷.

Traditional mRNA vaccine manufacturing mechanism typically employs lipid nanoparticles (LNPs) as carriers to encapsulate mRNA, aiming to achieve outstanding encapsulation results, biosafety, and vaccine potency^7,8. In this approach, mRNA and lipid molecules simultaneously and directly self-assemble to form complex nanostructures known as mRNA-LNPs⁹ (Fig. 1a). Due to the susceptibility of mRNA to degradation, raw mRNA-LNPs normally have to undergo a series of unintermittent and Good Manufacturing Practice (GMP) processes like tangential flow filtration, sterile filtration, and filling to become the final product¹⁰. This heightens the demands on infrastructure and thus determines that conventional mRNA vaccine cannot be manufactured in the point-of-care. In addition, as a non-equilibrium thermodynamical system, mRNA-LNPs inherently suffer from poor thermal stability and require storage either at extremely low temperature^11–13 or in costly water-free formulations, such as freeze-dried powder^14–16, patches^17–19, and microneedles^20–22. Even under ideal storage conditions, loss of mRNA integrity^23,24 and irreversible adduction reaction²⁵ between mRNA and lipid molecules have been found during long-term storage, undermining efficacy of mRNA vaccine gradually over time. Such intricate manufacturing and storage processes leave users no options to modify or replace any component of conventional mRNA vaccine after production. Taken together, the inherent instability of mRNA diminishes efficacy of vaccine product, results in infrastructure-dependence and cost elevation, and limits development of personalized and point-of-care mRNA therapy.

In this study, we develop a Rapid Onsite Microfluidic Assembly (ROMA) technology to fundamentally address the aforementioned challenges. In contrast to the mechanism of conventional mRNA vaccine manufacturing that mRNA and lipids directly self-assembling to mRNA-LNPs, ROMA technology employs a two-step self-assembly strategy (Fig. 1b). While the preparation of empty LNPs in the first step still requires the use of laboratory or GMP-grade equipment, the second step, loading mRNA into the empty LNPs to form mRNA vaccine, can be performed in the point-of-care using a simple microfluidic device within seconds. The feasibility of this ROMA strategy lies in the exceptional stability of mRNA and empty LNPs. Our experimental data, along with previous study²⁶, indicate that separately stored mRNA and empty LNPs exhibit significantly better stability than their combined mRNA-LNP complex under the same storage conditions (Fig. 1c). Their stability at 4°C and room temperature are even comparable to those of mRNA-LNP vaccine stored at -80°C (Supplementary Fig. S1). Therefore, mRNA and empty LNPs in the ROMA method can be considered as readily available and easily storable raw materials, allowing users to onsite and on-demand load mRNA into empty LNPs as mRNA vaccine.

ROMA mRNA vaccine shows unprecedented advantages in terms of production, storage, and usage (Supplementary Fig. S2). Using pre-configured empty LNPs as raw materials, the encapsulation of mRNA by ROMA now becomes an onsite, organic solvent-free, and self-pH-neutralizing procedure, which requires no additional processing prior to use. It reduces the mRNA-LNP manufacturing cycle from several days to a matter of minutes at the user-end. In addition, since ROMA mRNA vaccine can be made and used immediately without the need for storage, it minimizes the risk of adduction reaction between mRNA and lipids Therefore, the intrinsic thermal instability and degradation issues of mRNA-LNPs during storage are fundamentally eliminated. Intriguingly, ROMA technology is a versatile platform and applicable to various nucleic acids and LNPs of different formulations. Moreover, we develop a spectrophotometry-based ROMA detector, which enables a real-time access to the quality of mRNA vaccine product. We further find that ROMA mRNA-LNPs are highly similar to conventional ones in terms of physical structure, cellular expression, in vivo distribution, and vaccine efficacy.

Utilizing 3D printing technology, we build up a ROMA prototype and a ROMA detector. From the illustration and photos (Fig. 2a, b, and c), one can easily see that a typical ROMA prototype integrates various 3D-prinited and electrical components as annotated by number icons. The whole ROMA device is housed in a medical stainless-steel hull and highly miniaturized (< 50 cm) and portable (< 2 kg) and accepts either an external power source or batteries. As core of the ROMA prototype, the microfluidic chip features a 1200 mm-long serpentine channel with 2000 pairs of triangular obstacles (Fig. 2d and Supplementary Fig. S3), which can periodically alter the fluid direction to affiliate the encapsulation of mRNA into empty LNPs. A typical ROMA raw material package includes mRNA (in DI water, amber vial), pre-made empty LNPs (in acidic aqueous buffer, transparent vial), and neutralizing buffer (transparent vial) with equivalent volumes, and disposable microfluidic chip, medical syringes, needles, sterile filter, and tubing (Supplementary Fig. S4). In practice (Fig. 2e), the microfluidic chip, mRNA vial, empty LNP vial, and syringe containing neutralizing buffer are first installed onto respective working platforms. Upon initiating the machine, the microfluidic chip is supposed to be heated to a preset temperature (typically 70 ℃ to 80 ℃) in about 30 seconds (Supplementary Video 1). The UV disinfection LED initiates its programmed sequence thereafter (Supplementary Video 2). Users can easily set the stroke length (namely product volume) and production speed via a touchscreen, as well as freely switch between languages and automatic/manual modes (Supplementary Video 3). After clicking the start button, the lifting platforms are programmed to elevate the mRNA and empty LNP solutions up to the needles (Supplementary Video 4). Simultaneously, both solutions are aspirated at a same flow rate through the microfluidic chip where mRNA and empty LNP self-assemble into mRNA-LNP. Consequently, the mRNA-LNP mixture flows through a 0.22 µm sterile filter into the collection vial and gets neutralized, becoming a sterile, organic solvent-free, and ready-to-inject mRNA vaccine (Supplementary Video 5). To ensure the quality of the vaccine products, we also develop a ROMA detector based on the principle of spectrophotometry. With only about 0.2 mL of sample, it can quickly assess the quality of the vaccine within several seconds through red or green light signals (Supplementary Video 6, 7, and 8).

Strikingly, according to our calculation, it takes only 3.5 seconds for an mRNA molecule to enter and exit the chip while being encapsulated as mRNA-LNP, with a typical mRNA encapsulation flow rate of 0.3 mL/min of individual channel. By further increasing the mRNA concentration and flow rate, we can easily produce ready-to-use mRNA-LNPs that is approximately equivalent to 200 doses (∼100 µg mRNA/dose) of vaccine within one hour by repeatedly running the ROMA prototype (Methods). This throughput adequately meets the demand for large-scale vaccination, clinical treatment, and personalized vaccine production. In the future, we can further upgrade the production capacity of the current laboratory-scale ROMA prototype by implementing parallel chips or increasing microfluidic channel flux.

We discover that ideal ROMA encapsulation of mRNA into empty LNPs is a comprehensive result of microfluidic flow environment, pH, temperature, and ionizable lipid components (Fig. 3a). In brief, the high-temperature environment in the microfluidic channels leads to a rapid phase transition of empty LNPs^27–29, enhancing permeability and fluidity of lipids. Meanwhile, the acidic pH environment in the microfluidic channels induces ionization of ionizable lipids in the empty LNPs with strong positive charges³⁰. These conducive conditions enable the mRNA, which is negatively charged, to be encapsulated into the empty LNPs rapidly (Fig. 3a).

In the conventional production of mRNA vaccines, a crucial prerequisite for achieving a high encapsulation efficiency (E.E.) and monodisperse particle size of mRNA-LNPs is homogenous mixing of precursor mRNA and lipid solutions³¹. Interestingly, we find this prerequisite is just as important to ROMA encapsulation, however, not enough (Fig. 3a).

We first compare the encapsulation results from three most common mixing conditions: an Eppendorf (EP) tube, a conventional straight microfluidic mixing channel, and an EP tube with vortex stirring (Methods), which represent flow mixing features of static, unidirectional, and stochastically oscillatory, respectively (Fig. 3b). Disappointingly, it turns out that none of these three classical methods can prepare mRNA vaccines that simultaneously meet both high E.E. and monodispersity (Supplementary Fig. S5), which are not suitable for clinical use. Cryo-TEM images also clearly indicate that these conventional flow mixing methods fail to prevent, and in some cases even induce aggregation and fusion of mRNA-LNPs during encapsulation (Fig. 3b).

While previous research has provided evidences of pH^{32, 33}and high temperature^34–36-induced aggregation and fusion of mRNA-LNPs, our experiment demonstrates that empty LNPs do not aggregate or fuse even under acidic or heated conditions, at least within the timeframe of ROMA encapsulation process (Supplementary Fig. S6). This suggests that the aggregation and fusion of mRNA-LNPs are not solely due to interactions among LNPs themselves but driven by the interplay between mRNA and LNPs. We suppose that the most likely reason is the interaction between positively charged LNPs and negatively charged mRNA.

Hence, we devise a microfluidic channel to address this issue by employing spatially arranged obstacles to periodically alter fluid direction (Fig. 3b, Supplementary Fig. S7-S12). We thus define fluid frequency f as the number of times the fluid changes direction within the microfluidic channel per second. We observe that when f is between 200–7000 Hz, mRNA-LNPs show high E.E., small particle sizes, and narrow size distributions (Fig. 3c, Supplementary Fig. S13). However, E.E. gets soon declined and particle size increases if f is below 200 Hz. This is because that lower f brings a longer duration (Methods, Supplementary Fig. S14) when mRNA and empty LNPs remain in a unidirectional flow state in one cycle, resulting in aggregation and fusion of mRNA-LNPs observed in the conventional mixing conditions (Supplementary Fig. S15). Moreover, for the microfluidic chip with a fixed obstacle pattern, f is decided by flow rate so that lower f means slower flow rate and longer exposure time of mRNA in the high-temperature acidic environment, leading to declination of E.E. and aggregation of LNP (Supplementary Fig. S15) and degradation of mRNA (Supplementary Fig. S16). On the other hand, we notice that excessively high f (> 7000 Hz) can also increase LNP size. We hypothesize that this effect might be attributed to the elevated shear forces generated by higher f (Methods, Supplementary Fig. S14), disrupting the membrane structure of LNPs^37–39and subsequently inducing aggregation and fusion.

We find pH and E.E. exhibit a peak-shaped relationship with peaking at pH 4.5 (Fig. 3d). It is well-known that ionizable lipids only yield positive charges to attract negatively charged mRNA in an acidic environment⁴⁰, therefore, it is easily comprehensible that E.E. is suboptimal under relatively neutral pH (pH > 5.0). However, we observe that excessively acidic pH (pH < 4.0) can be harmful to encapsulation as well. The most plausible explanation is that overly strong electrostatic attraction of mRNA-LNPs at low pH overwhelms repulsive effects of surface PEG groups^{41, 42} and periodically oscillatory flows and thus results in aggregation and fusion of mRNA-LNPs, which is clearly reflected in the increase of particle size and PDI (Supplementary Fig. S17), as well as Cryo-TEM images (Fig. 3d). Therefore, achieving optimal ROMA mRNA encapsulation is contingent upon maintaining a modest acidic pH range.

As a material consisting of polymer mixtures, empty LNPs exhibit polymer-like temperature-induced phase transition behaviors, a typical characteristic of which is enhanced fluidity of lipids⁴³. Differential scanning calorimetry experiment suggests that our empty LNPs has a critical phase transition temperature (\(\:{T}_{p}\)) range roughly from 39 ℃ to 53 ℃ (Supplementary Fig. S18). Since higher temperature causes faster completion of the phase transition in empty LNPs and improves E.E. of mRNA (Fig. 3e), we believe that the increase in the fluidity of LNP favors its enhanced permeability to mRNA.

It is noteworthy that when the temperature rising from 25 ℃ to 75 ℃, the particle diameter of mRNA-LNPs gradually decreases, with a drastic drop of PDI from approximately 0.3 to 0.03 (Fig. 3e, Supplementary Fig. S19). This indicates that a faster completion of phase transition is not only beneficial for the encapsulation of mRNA but also helps reduce the aggregation and fusion of mRNA-LNPs. Although Cryo-TEM images indicate that structural integrity and uniformity of mRNA-LNPs are inert to high temperature (Fig. 3e), excessively high temperature (e.g., 85 ℃) can be harmful to the integrity of mRNA and then bioactivity of mRNA (Supplementary Fig. S20). This suggests that there may be a critical temperature for significant mRNA degradation (\(\:{T}_{d}\)) and highlights a temperature-dependent trade-off (\(\:{{T}_{p}<T<T}_{d}\)) between encapsulation and degradation speed of mRNA in ROMA encapsulation.

We prepare empty LNPs with different molar ratios of ionizable lipids (Supplementary Fig. S21) and observe a proportional increase in E.E. with the higher content of ionizable lipids (Fig. 3f). We particularly note that empty LNPs with 0% ionizable lipids can hardly encapsulate mRNA, even with optimal \(\:f\), pH, and \(\:T\). This emphasizes once again that the electrostatic interaction between ionizable lipids and mRNA is the primary driving force for the encapsulation of mRNA into empty LNPs. In addition, higher content of ionizable lipids also results in smaller particle size and PDI, making them more suitable for clinical applications (Fig. 3f, Supplementary Fig. S22). However, the question of whether ionizable lipids have a monotonic promoting effect on mRNA encapsulation is yet to be explored using empty LNP formulations with higher molar ratios of ionizable lipids in future studies.

As discussed above, ideal encapsulation (shown as green color in Fig. 3c, d, e, and f) is only achieved within a limited range of parameters (proper \(\:f\), pH, \(\:T\), and molar ratio of ionizable lipid). Outside of this comfort zone, non-ideal encapsulation (shown as red color in Fig. 3c, d, e, and f) occurs. However, is there any difference in the self-assembly mechanism between the two states? To answer this question, we measure changes in the number of LNPs before and after encapsulation (Methods). Surprisingly, we see a significant difference in the quantity of mRNA-LNPs between ideal and non-ideal encapsulation conditions. Under ideal encapsulation conditions, the quantity of product mRNA-LNPs is nearly identical to that of initial empty LNPs (Fig. 3g and Supplementary Fig. S23). This implies that mRNA-LNPs maintain their structural integrity well and the minor particle size increase is primarily due to mRNA encapsulation. In contrast, under non-ideal encapsulation conditions, the quantity of product mRNA-LNPs is significantly lower than that of initial empty LNPs, and there are substantial increases in particle size and PDI (Supplementary Fig. S24). This indicates severe fusion and aggregation occur between mRNA-LNPs and the Cryo-TEM results also confirm these facts (Fig. 3c, d, and e). Considering aggregation and fusion of mRNA-LNPs are induced by electrostatic attraction and thermodynamic instability, we propose the crucial factor distinguishing ideal and non-ideal encapsulation mechanism is whether the electrostatic and thermodynamic interactions between LNPs can be delicately balanced.

Real-time quality control is a prerequisite for point-of-care vaccine production. It can be easily concluded that three most important variables determine the result of ROMA encapsulation: \(\:f\), pH, and \(\:T\), which need to be closely monitored during the reaction. According to the design of ROMA prototype, \(\:f\) (decided by preset fluid flow rate and microfluidic chip geometry) and \(\:T\) are controlled and displayed on the touchscreen. Therefore, real-time monitoring and correction of \(\:f\) and \(\:T\) can be easily achieved with minimal risk and the key issue moves forward to the detection of pH. Despite that the small volume and rapid flow pattern make in situ monitoring of fluid pH challenging, interestingly, we find that the fluid pH is closely related to the size of the resulting mRNA-LNPs (Fig. 3h). We further discover that, when the concentration and volume of mRNA and empty LNPs are fixed, the size of resulting mRNA-LNP is qualitatively correlated to its optical density (OD). This trend is clearly reflected as the visual appearance of mRNA-LNPs changing from thick to thin white color with pH increasing in ROMA reaction (Fig. 3h). Therefore, we can establish a quantitative relationship between OD and pH, which inspires us to develop a portable and rapid ROMA detector to provide qualitative test results using green (pass) and red (fail) light signals (Supplementary Video 6, 7, and 8). This ROMA detector, as an immediate and non-invasive means in conjunction with the built-in inspection of \(\:f\) and \(\:T\), effectively ensures the quality of mRNA-LNP upon its production. To our knowledge, such real-time vaccine quality detection function is lacking in other GMP equipment.

It should be noted that, in this article, qualitative judgment criteria of the ROMA detector for OD values are based on a fixed combination of mRNA and LNP (Methods). It can be estimated that as the types of mRNA and LNP change, the OD values of the resulting mRNA-LNP complexes will also vary. In the next step, we plan to test different types of mRNA and LNP and use these results to enrich our database, thereby enhancing the ROMA detector's capability for onsite detection and its clinical translation potential.

We use small-angle X-ray scattering (SAXS) technique to compare ROMA and conventional mRNA- LNPs. Surprisingly, we find that despite ROMA and conventional methods following vastly different mechanisms, both ROMA and conventional mRNA-LNPs have similar sizes (Supplementary Fig. S25) and highly overlapped X-ray scattering patterns (Fig. 3i). Subtracting the X-ray scattering profiles of empty LNPs as background from each sample, we find that the resulting X-ray scattering profiles of the enveloped mRNA are also still closely resembling (Fig. 3i). This trend has been further confirmed by confocal microscopic imaging results (Fig. 3j) that the morphology, fluorescence intensity and pattern of ROMA and conventional LNPs containing the same mass ratio of mRNA are alike. This evidence strongly indicates that ROMA LNPs and traditional LNPs have highly similar physical structures and inner mRNA distributions, which makes ROMA LNPs a seamless alternative to traditional LNPs.

Our ROMA prototype can superbly accommodate various nucleic acids and lipids, LNP sizes, and mRNA doses, as well as feed ratios of mRNA, and raw material formulations ranging from liquid to lyophilized, long-term stored to freshly made (Fig. 4a).

We test representative nucleic acids including linear RNA, circular RNA, and plasmid DNA (Fig. 4b). We are excited to find all ROMA LNPs display comparable cellular expressions to conventional ones (Fig. 4b), with respect to satisfying encapsulation results (Supplementary Fig. S26 and Fig. S27). This suggests that the ROMA method can applied to numerous nuclei acids-related biomedical fields, including but not limited to therapeutic and prophylactic vaccines, gene editing, cell therapy and so on.

It is well known that cationic/ionizable lipid is a key factor influencing encapsulation and delivery results^{44, 45}. Therefore, we are motivated to test empty LNPs containing different representative cationic/ionizable lipids, such as ALC-0315, SM-102, MC-3, and DOTAP. We find that all these cationic/ionizable lipids could achieve fine E.E. of mRNA (Fig. 4c) and maintain good integrity of mRNA (Supplementary Fig. S28). Different cationic/ionizable lipids exhibit varying capacities for mRNA encapsulation, which may be related to their degrees of ionization (pKa values)^{40, 46}that further influence the extent of electrostatic attraction between empty LNPs and mRNA.

Importantly, the ROMA prototype consistently maintains excellent E.E. (Fig. 4d), particle size, PDI, and mRNA integrity (Supplementary Fig. S29) over a wide range of mRNA input quantities, even at very low N/P ratios, where conventional methods typically face challenges⁴⁷. Interestingly, we note a clear positive correlation between the LNP particle size and the mRNA feed ratio. This is most likely due to the increased volume of mRNA-LNPs as more mRNA enters the empty LNPs, which is also clearly reflected in the Cryo-TEM images (Supplementary Fig. S30).

It is noteworthy that empty LNPs of different sizes, after an ideal encapsulation (Supplementary Fig. S31), show a consistent proportion (62%∼66%) of particle diameter growth when the mRNA feed ratio is fixed (mRNA/Empty LNPs = 10%, mass ratio) (Fig. 4e). This suggests that, beyond electrostatic interactions, the mRNA encapsulation capacity of empty LNPs is also influenced by their inner spatial volumes. This also indicates that the size of LNPs after encapsulation is predictable, thereby opening up the possibility of customizing different mRNA-LNP sizes.

We find that different storage conditions and processes have negligible impact on the resulting mRNA-LNP vaccine. We test thirteen different combinations of raw materials as indicated by the Roman numbers (I) to (XIII), including 4°C/RT/RT-lyophilized, long-term stored and freshly made forms of mRNA and empty LNPs (Fig. 4f). These varied combinations, after quick onsite encapsulation, have equivalently outstanding E.E. and particle size (Fig. 4f), with respect to PDI, mRNA integrity, and cell transfection rate (Supplementary Fig. S32). Cryo-TEM images also confirm the high similarity of ROMA mRNA-LNPs composed of different raw materials (Supplementary Fig. S33).

Our rigorous testing regimen for the ROMA prototype involves continuous operation, with mRNA-LNPs collected at intervals of 1 minute, 30 minutes, and 60 minutes for comparative analysis. Our results show that across all time points, the E.E. consistently exceeds 95%, while maintaining a uniform size distribution ranging between 100–110 nm and a low PDI of less than 0.05. (Fig. 4g). We also note that, although mainstream commercial devices can also achieve comparable particle size and E.E., they statistically fail to keep a monodisperse particle size distribution as ROMA prototype does⁴⁸. This should be attributed to the fact that traditional manufacturing normally requires a cumbersome purification process after mRNA encapsulation, during which mRNA-LNPs inevitably undergo fluid shearing, pH, and buffer change. Such adverse factors have been widely recognized to result in particle rupture or fusion^{49, 50}, and eventually loss of particle uniformity.

We find that ROMA and conventional mRNA-LNPs, whether by intravascular (I.V.) or intramuscular (I.M.) injection, have similar distribution patterns and sustained expression durations (0–48 h) as presented by their bioluminescence (Fig. 5a). This preliminarily indicates that ROMA mRNA-LNPs maintain similar cellular and organ tropism as those formed by conventional methods, which is reasonable since the Cryo-TEM and SAXS results show that ROMA mRNA-LNPs have identical physical morphology and structures as conventional ones.

We further utilize a SARS-CoV-2 spike mRNA as a model to study the immune responses (Fig. 5b). We prepare ROMA vaccines with mRNA produced one year ago and mRNA freshly produced, respectively, and compare them with conventional vaccines with mRNA freshly produced (Fig. 5c). Interestingly, we find that ROMA vaccines, whether encapsulating old or new mRNA, their IgG antibody titers after the first and second vaccinations are statistically equivalent to that of conventional vaccines encapsulating new mRNA (Fig. 5d). As mentioned earlier, our ROMA method employs a strategy of storing mRNA and empty LNPs separately, which minimizes the adduction reaction between mRNA and lipid during long-term storage²⁵. Therefore, the integrity of mRNA (Supplementary Fig. S34) and the efficacy of mRNA vaccines are not affected by the long-term storage issue observed in conventional mRNA vaccines²⁵. Moreover, we observe that ROMA vaccines made of refrigerated (4°C stored) and RT-lyophilized empty LNPs both stimulate high IgG antibody titers after the first vaccinations (Fig. 5e). This once again emphasizes the excellent versatility of our ROMA method.

Noteworthy is that, Balb/c mice vaccinated with ROMA mRNA-LNPs, regardless of the LNP formulations and mRNA production dates, exhibit statistically less weight loss compared to conventional ones (Fig. 5f). We speculate that despite the high similarity in physicochemical structure and biodistribution between ROMA and conventional mRNA-LNPs, ROMA LNPs may induce less inflammation compared to conventional ones. However, this hypothesis requires further experimental validation in the future.

ROMA technology represents a first-in-class platform for customizable mRNA-LNP vaccine, with a real-time quality inspection upon production that can be universally extended to various nuclei acid and LNP candidates. This decentralized paradigm shift, transforming previously uninterrupted procedure into one that allows raw material preparation and vaccine production to be carried out in different places, for the first time makes onsite production of ready-to-inject mRNA vaccine possible.

For mRNA research, ROMA not only saves substantial time and materials but also empowers labs or individuals lacking mRNA-LNP production capabilities to enter the field. For clinical translation, ROMA technology is promising to accelerate the promotion of mRNA vaccine in infrastructure-lacking areas since it circumvents the dependence on intricate manufacturing and storage facilities of conventional ones. Furthermore, this advancement will revive numerous LNP and mRNA formulations characterized with highly efficient intracellular protein expression but limited thermal stability to be effectively utilized in clinical settings.

We thus envision that the ROMA technology can harmoniously meet urgent needs of basic and translational applications, significantly shorten turnover time of personalized vaccine and bedside therapy, and broadly benefit fields like personalized tumor vaccines, protein replacement therapy and gene therapy.

Acknowledgements

The authors wish to express their heartfelt appreciation to Prof. Robert Langer for his invaluable guidance and insightful suggestions on the experiment, as well as for his constructive comments on the manuscript. The authors would like to express their gratitude to the diligent and dedicated technical support team in Virogin Biotech for their support in sample characterizations and animal experiments. J.X. wants to extend special thanks to Prof. Zhenyu Yang, Mrs. Nan Luo, and Ms. Keqing Xu for their help with scientific illustration design. This work was supported by Virogin Biotech Funding and the Shanghai Rising Star Talent Program (Grant no. 23QB1401300, to J.X.).

Contributions

Z.G., W.J., and J.X. initiated the concept of ROMA prototype and supervised the whole project. Z.Z. and J.X. led the simulation, design, and fabrication of microfluidics. Z.Z., Q.W., Z.D., and J.X. performed encapsulation experiments. R.Y., Z.Y., Y.G., and F.H were responsible for synthesis of nuclei acids and in vitro experiments. J.D. and J.X. led in vivo experiments and communications with outsourced research services. W.Z. performed SAXS characterizations. Y.J. and J.X. led the manufacturing and assembly of the ROMA prototype. All authors wrote the manuscript.

Conflicts of interests

Virogin Biotech Co. Ltd has filed patent applications for ROMA encapsulation methods and prototype (CN202310058166.8; PCT/CN2023/142475).

Data availability

The datasets generated and analyzed during these studies are available from the corresponding author upon reasonable request.

Polack, F. P. et al. Safety and efficacy of the BNT162b2 mRNA Covid-19 vaccine. New England journal of medicine 383, 2603–2615 (2020).
Barbier, A. J., Jiang, A. Y., Zhang, P., Wooster, R. & Anderson, D. G. The clinical progress of mRNA vaccines and immunotherapies. Nat. Biotechnol 40, 840–854 (2022).
Cheng, Q. et al. Selective organ targeting (SORT) nanoparticles for tissue-specific mRNA delivery and CRISPR–Cas gene editing. Nat. nanotechnol 15, 313–320 (2020).
Anand, P. & Stahel, V. P. The safety of Covid-19 mRNA vaccines: a review. Patient Saf Surg 15, 1–9 (2021).
Wayment-Steele, H. K. et al. Deep learning models for predicting RNA degradation via dual crowdsourcing. Nat. Mach Intell 4, 1174–1184 (2022).
Pardi, N., Hogan, M. J. & Weissman, D. Recent advances in mRNA vaccine technology. Curr Opin Immunol 65, 14–20 (2020).
Han, X. et al. Adjuvant lipidoid-substituted lipid nanoparticles augment the immunogenicity of SARS-CoV-2 mRNA vaccines. Nat Nanotechnol 1–10 (2023).
Kon, E., Elia, U. & Peer, D. Principles for designing an optimal mRNA lipid nanoparticle vaccine. Curr Opin Biotechnol 73, 329–336 (2022).
Chen, S. et al. Nanotechnology-based mRNA vaccines. Nature Reviews Methods Primers 3, 63 (2023).
Chaudhary, N., Weissman, D. & Whitehead, K. A. mRNA vaccines for infectious diseases: principles, delivery and clinical translation. Nat Rev Drug Discov 20, 817–838 (2021).
Kim, B. et al. Optimization of storage conditions for lipid nanoparticle-formulated self-replicating RNA vaccines. Journal of Controlled Release 353, 241–253 (2023).
Hou, X., Zaks, T., Langer, R. & Dong, Y. Lipid nanoparticles for mRNA delivery. Nat Rev Mater 6, 1078–1094 (2021).
Hatziantoniou, S., Maltezou, H. C., Tsakris, A., Poland, G. A. & Anastassopoulou, C. Anaphylactic reactions to mRNA COVID-19 vaccines: A call for further study. Vaccine 39, 2605 (2021).
Lamoot, A. et al. Successful batch and continuous lyophilization of mRNA LNP formulations depend on cryoprotectants and ionizable lipids. Biomater Sci. (2023).
Ai, L. et al. Lyophilized mRNA-lipid nanoparticle vaccines with long-term stability and high antigenicity against SARS-CoV-2. Cell Discov 9, 9 (2023).
Muramatsu, H. et al. Lyophilization provides long-term stability for a lipid nanoparticle-formulated, nucleoside-modified mRNA vaccine. Molecular Therapy 30, 1941–1951 (2022).
Prins, M. L., Prins, C., de Vries, J. J., Visser, L. G. & Roukens, A. H. Establishing immunogenicity and safety of needle-free intradermal delivery by nanoporous ceramic skin patch of mRNA SARS-CoV-2 vaccine as a revaccination strategy in healthy volunteers. Virus Res 334, 199175 (2023).
Koh, K. J. et al. Formulation, characterization and evaluation of mRNA-loaded dissolvable polymeric microneedles (RNApatch). Sci Rep 8, 11842 (2018).
Tran, K. T. et al. A Single-Administration Microneedle Skin Patch for Multi-Burst Release of Vaccine against SARS-CoV-2. Adv Mater Technol. 8, 2200905 (2023).
Vander Straeten, A. et al. A microneedle vaccine printer for thermostable COVID-19 mRNA vaccines. Nat Biotechnol 1–8 (2023).
Yu, J. et al. Intradermal delivery of mRNA using cryomicroneedles. Acta biomater 148, 133–141 (2022).
McMillan, C. L. et al. Complete protection by a single-dose skin patch–delivered SARS-CoV-2 spike vaccine. Sci Adv 7, eabj8065 (2021).
Blenke, E. O. et al. The storage and in-use stability of mRNA vaccines and therapeutics: not a cold case. J Pharm Sci 112, 386–403 (2023).
Crommelin, D. J., Anchordoquy, T. J., Volkin, D. B., Jiskoot, W. & Mastrobattista, E. Addressing the cold reality of mRNA vaccine stability. J Phar Sci 110, 997–1001 (2021).
Packer, M., Gyawali, D., Yerabolu, R., Schariter, J. & White, P. A novel mechanism for the loss of mRNA activity in lipid nanoparticle delivery systems. Nat Commun 12, 6777 (2021).
Jones, L. J., Yue, S. T., Cheung, C.-Y., & Singer, V. L. RNA quantitation by fluorescence-based solution assay: RiboGreen reagent characterization. Anal Biochem 265, 368–374 (1998).
Cholakova, D., Glushkova, D., Tcholakova, S. & Denkov, N. Nanopore and nanoparticle formation with lipids undergoing polymorphic phase transitions. ACS Nano 14, 8594–8604 (2020).
De Campo, L. et al. Reversible phase transitions in emulsified nanostructured lipid systems. Langmuir 20, 5254–5261 (2004).
Von White, G., Chen, Y., Roder-Hanna, J., Bothun, G. D. & Kitchens, C. L. Structural and thermal analysis of lipid vesicles encapsulating hydrophobic gold nanoparticles. ACS Nano 6, 4678–4685 (2012).
Han, X. et al. An ionizable lipid toolbox for RNA delivery. Nat commun 12, 7233 (2021).
Evers, M. J. et al. State-of-the-art design and rapid-mixing production techniques of lipid nanoparticles for nucleic acid delivery. Small Methods 2, 1700375 (2018).
Yu, H. et al. Real-time pH-Dependent Self-Assembly of Ionisable Lipids from COVID-19 Vaccines and In Siut Nucleic Acid Complexation. Angewandte Chemie 135, e202304977 (2023).
Larson, N. R. et al. pH-dependent phase behavior and stability of cationic lipid–mRNA nanoparticles. J Pharm Sci 111, 690–698 (2022).
Markova, N. et al. Biophysical characterization of viral and lipid-based vectors for vaccines and therapeutics with light scattering and calorimetric techniques. Vaccines (Basel) 10, 49 (2021).
Ball, R. L., Bajaj, P. & Whitehead, K. A. Achieving long-term stability of lipid nanoparticles: examining the effect of pH, temperature, and lyophilization. Int j nanomedicine 305–315 (2017).
Ayat, N. R. et al. Formulation of biocompatible targeted ECO/siRNA nanoparticles with long-term stability for clinical translation of RNAi. Nucleic Acid Ther 29, 195–207 (2019).
Yamamoto, K. & Ando, J. Endothelial cell and model membranes respond to shear stress by rapidly decreasing the order of their lipid phases. J. Cell Sci 126, 1227–1234 (2013).
Petersen, E. N., Chung, H.-W., Nayebosadri, A. & Hansen, S. B. Kinetic disruption of lipid rafts is a mechanosensor for phospholipase D. Nat commun 7, 13873 (2016).
Marmottant, P., Biben, T. & Hilgenfeldt, S. Deformation and rupture of lipid vesicles in the strong shear flow generated by ultrasound-driven microbubbles. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 464, 1781–1800 (2008).
Patel, P., Ibrahim, N. M. & Cheng, K. The importance of apparent pKa in the development of nanoparticles encapsulating siRNA and mRNA. Trends pharmacol Sci 42, 448–460 (2021).
Guo, X. & Szoka, F. Steric stabilization of fusogenic liposomes by a low-pH sensitive PEG- diortho ester- lipid conjugate. Bioconjug Chem 12, 291–300 (2001).
Thevenot, J., Troutier, A.-L., David, L., Delair, T. & Ladavière, C. Steric stabilization of lipid/polymer particle assemblies by poly (ethylene glycol)-lipids. Biomacromolecules 8, 3651–3660 (2007).
Karal, M. A. S., Billah, M. M., Ahmed, M. & Ahamed, M. K. A review on the measurement of bending rigidity of lipid membranes. Soft Matter (2023).
Rachel S. Riley et al., Ionizable lipid nanoparticles for in utero mRNA delivery.Sci Adv 7, eaba1028 (2021).
Margaret M. Billingsley, Nathan Singh, Pranali Ravikumar, Rui Zhang, Carl H. June, and Michael J. Mitchell. Ionizable lipid nanoparticle-mediated mRNA delivery for human CAR T cell engineering. Nano Letters 20, 1578–1589 (2020).
Sun, D. & Lu, Z.-R. Structure and function of cationic and ionizable lipids for nucleic acid delivery. Pharm Res 40, 27–46 (2023).
Li, S. et al. Payload distribution and capacity of mRNA lipid nanoparticles. Nat Commun 13, 5561 (2022).
Maeki, M., Uno, S., Niwa, A., Okada, Y., & Tokeshi, M. Microfluidic technologies and devices for lipid nanoparticle-based RNA delivery. J Controlled Release 344, 80–96 (2022).
Geng, C., Zhou, K., Yan, Y., Li, C., Ni, B., Liu, J., Wang, Y., Zhang, X., Wang, D., & Lv, L. A preparation method for mRNA-LNPs with improved properties. J Controlled Release 364, 632–643 (2023).
Schoenmaker, L., Witzigmann, D., Kulkarni, J. A., Verbeke, R., Kersten, G., Jiskoot, W., & Crommelin, D. J. A. mRNA-lipid nanoparticle COVID-19 vaccines: Structure and stability. Int J Pharm 601, 120586 (2021).

Yes there is potential Competing Interest. Virogin Biotech Co. Ltd has filed patent applications for ROMA encapsulation methods and prototype (CN202310058166.8; PCT/CN2023/142475).

SupplementaryInformation.docx
Video1.mp4
User interface and temperature regulator
Video2.mp4
UV disinfection
Video3.mp4
Control panel of ROMA encapsulation
Video4.mp4
Flexible movement of sample lifting platforms
Video5.mp4
Collection of ROMA mRNA-LNPs
Video6.mp4
Characterization of a successful sample (ideal encapsulation) by ROMA detector
Video7.mp4
Characterization of a failed sample (low encapsulation) by ROMA detector
Video8.mp4
Characterization of a failed sample (LNP aggregation) by ROMA detector

Download PDF

Version 1

posted

You are reading this latest preprint version

Tackling intrinsic instability of mRNA vaccine by a rapid onsite microfluidic assembly (ROMA) technology

Status:

Version 1

Abstract

Figures

Main

Structure and workflow of the ROMA prototype

Mechanism of ROMA technology

Customizable mRNA vaccines from universality of ROMA prototype

Potent in vivo expressions of ROMA vaccines

Perspective and conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1