Preliminary results of standard analysis methods for Covid vaccines

Summary #

The breakthrough for the production of Biontech/Pfizer and Moderna’s mRNA vaccines was the development of a cationic lipid shell. Pure mRNA would be easily degraded by cleaving enzymes, so-called ribonucleases, and would also have difficulty crossing the cell membrane. A double layer of lipid molecules forms small globules and the mRNA is packed inside. Because of their small size, they are referred to as lipid nanoparticles. To stabilize this in turn, a PEG (polyethylene glycol) coating is required. Regardless of the genetic manipulation of the body’s own cells to express S-protein antigens, the question of the „quality“ of these nanolipids and their PEG protective shells arises. Can these materials themselves be harmful to health, and are different „qualities“ possibly the cause of production-lot-dependent side-effect profiles?

The mRNA vaccines against COVID-19 approved by the EMA contain parts of the genetic information of SARS-CoV-2 in the form of messenger ribonucleic acid. During vaccination, this mRNA is injected into the muscle and thus reaches body cells. The lipid nanoparticles are then incorporated into the foreign body cell via endocytosis, i.e. through an invagination in the area of ​​the cell membrane, and can then release the transported mRNA into the cytosol. In these cells, viral S protein parts are synthesized in the ribosomes according to the encoded blueprint of the mRNA and then expressed on the membrane, which is intended to trigger a specific immune response. The mRNA itself is broken down again by the body after some time, with periods of several weeks also being mentioned.

To characterize the vaccines we used (A) bright field microscopy BFM, (B) dark field microscopy DFM, (C) scanning electron microscopy SEM, (D) energy dispersive X-ray spectroscopy EDS, (E) total reflecting X-ray fluorescence spectroscopy (TXRF) and (F) time-of-flight mass spectroscopy TOF-MS.

The formation of crystalline structures could be confirmed in BFM and DFM. A comparison with solid cholesterol samples confirmed the assumption that the decomposition products of the nanolipid particles resemble the „liquid crystal structures“ of cholesterol. Cholesterol is built into the nanolipid layer and thus facilitates endocytosis, among other things. Surface tension and static electricity influence the formation of microscopic crystals with geometries of varying complexity.

In the SEM, after the drying process had taken place, very large rectangular platelets were found in Moderna samples, which turned out to be NaCl crystals after an EDX analysis. However, there is reason to assume that these are larger table salt crystallizations on much smaller cholesterol crystals. Furthermore, isolated particulate impurities could be found in the SEM, for example metal particles from stainless steel (Fe-Ni-Cr). However, it can only be roughly estimated whether these are significant quantities in the harmful range. Additional measurements with TXRF from larger sample areas did not show any spectral signals in this respect, which means that the proportion of such metallic contamination can probably be estimated in the low ppm range. However, TXRF showed conspicuous peaks of Ca, but also small amounts of Al and Mg. All three elements are not specified in the manufacturer’s data sheets.

The most important result for the time being is a connection between the polymerization behavior of the PEG and the vaccine side effect profile at Pfizer/Biontech. MALDI TOF-MS revealed a batch-dependent degree of polymerization of PEG. If one then plots the maximum of the distribution of measured PEG molecular weights against the reported number of side effects, a more homogeneous short-chain coating indicates an increase in the number of vaccination side effects. A homogeneous coating could increase the decay half-life and thus promote onward transport to other tissue and organ parts. Conversely, the amount of cholesterol crystal formation as a breakdown product may be related to a more unstable and longer chain PEG coating. Of course, cholesterol crystals, especially in microcapillaries, are also a potential hazard.

Further investigations with infrared spectroscopy FTIR and evaluations of studies with imaging Raman methods of in vitro cell substrates with vaccines revealed no need for or indications of contamination with graphene or graphene oxide GO.

Introduction #

The global COVID-19 pandemic has given the development of mRNA vaccines an unprecedented boost. This can be explained by the fact that RNA vaccines, in contrast to conventional vaccines, can be developed quickly, produced in large quantities and can theoretically be adapted to different pathogens. The development but also the control of RNA-based drugs relies on advanced analytical methods such as scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), total reflecting X-ray spectroscopy (TXRF), infrared spectroscopy (FTIR) and mass spectrometry (MS) and requires a deep understanding for fragmentation mechanisms to reliably identify inorganic, organic and RNA structures.

The stability of RNA is very low and strongly dependent on the surrounding environment. The fact that DNA is stable and RNA is unstable can be attributed to the different sugars in the backbone. Ribose in RNA contains a hydroxyl group at the C2′ position, which destabilizes the phosphodiester bond. This 2′-OH group can attack the phosphate group at C3′ of the nucleotide intramolecularly. This leads to autohydrolysis [1] of the RNA even in the absence of degrading enzymes. This OH group is not present in deoxyribose in the backbone of the DNA. Therefore, DNA does not autohydrolyze and DNA fragments can remain stable for hundreds or even thousands of years.

Fig.1 The stability of RNA is impaired by autohydrolysis. a) Schematic of the autohydrolysis of an RNA strand with the 2′,3′-cyclic phosphate intermediate. Intramolecular nucleophilic attack of the 2′-OH group initiates cleavage of the phosphodiester. The absence of the 2′-OH group in DNA prevents cleavage there. b) Structures of the cyclic nucleoside monophosphate isomers. c) Nomenclature and structures of the RNA nucleobases (N).

The clear challenge for mRNA therapeutics is their susceptibility to nucleases, illustrated by a serum half-life of <5 min [2]. Although chemical modifications of siRNA are very successful in improving stability and reducing immunogenicity [3], so far they have not been successful for mRNA due to the sensitivity of the translation machinery to these modifications [4]. Another challenge for mRNA is the lack of cellular uptake of naked mRNA in most cell types [5], with the exception of immature dendritic cells [6]. These two challenges are addressed by incorporating a nucleoside-modified or sequence-modified mRNA into a delivery system that both protects the mRNA from enzymatic attack and facilitates cellular uptake. For example, incorporation into lipid nanoparticles protects the mRNA from enzymatic attack and increases cellular uptake and expression up to 1000-fold compared to naked mRNA when administered in animal models [7,8].

The fragmentation of ions can be studied using mass spectrometry. Thus, structural information about lipids and nucleic acids can be obtained. c fragments result from cleavage of the bond between the 5′-O and the phosphorus atom. The c-fragments of RNA dinucleotides have the same mass-to-charge ratio (m/z) as the base-catalyzed RNA fragments after autohydrolysis. The exact composition of Pfizer-BioNTech LNP [9] and Moderna LNP [10] has been published.

Fig. 2 mRNA-lipid nanoparticle structure with 1-10 copyable mRNAs. The lipid polyethylene glycol (PEG) forms the surface of the lipid nanoparticle (LNP) together with DSPC, which forms a bilayer. Cholesterol and the ionizable lipid in charged and uncharged forms can be distributed across the LNP

Cationic polymers have been used on a relatively large scale for nucleic acid delivery for many years. In the simplest case, cationic polymers are mixed in excess with nucleic acid to form electrostatically bound cationic polyplexes. Although many polymers have been developed, they are not as advanced as lipid nanoparticles for nucleic acid delivery, and the number of animal studies in which they have been successfully applied to vaccines is very limited. Nano-lipids are then coated with hydrophilic PEG to stabilize them in aqueous media and limit protein and cell interactions when administered in vivo.

Cationic polymers have been used on a relatively large scale for nucleic acid delivery for many years. In the simplest case, cationic polymers are mixed in excess with nucleic acid to form electrostatically bound cationic polyplexes. Although many polymers have been developed, they are not as advanced as lipid nanoparticles for nucleic acid delivery, and the number of animal studies in which they have been successfully applied to vaccines is very limited. Nano-lipids are then coated with hydrophilic PEG to stabilize them in aqueous media and limit protein and cell interactions when administered in vivo.

Decay of mRNA vaccines #

The Moderna COVID-19 vaccine must be stored at -25°C to -15°C but is also stable for up to 30 days between 2°C and 8°C and up to 12 hours between 8°C and 25°C [11]. The Pfizer/BioNTech COVID-19 vaccine will be stored at -80°C to -60°C and thawed at 2°C to 8°C for up to 5 days prior to dilution with saline prior to injection [12] . The dry ice temperatures required for Pfizer vaccine are more difficult to achieve during distribution and storage than the normal freezing temperature required for Moderna vaccine. The reasons for these temperature differences are not obvious as both vaccines contain similarly high concentrations of sucrose as a cryoprotectant. Moderna’s mRNA LNPs are frozen in two buffers, Tris and acetate [10], while Pfizer/BioNTech’s vaccine uses only a phosphate buffer [9]. Phosphate buffers are known to be suboptimal for freezing as they tend to precipitate and cause abrupt pH changes at the onset of ice crystallization [13,14].

The mechanism of RNA autohydrolysis in vitro has been extensively studied [1]. The reaction is initiated by nucleophilic attack of the 2′-OH group on the 3′-phosphate and leads via a phosphorane to a cyclic 2′,3′-phosphate, the key intermediate in RNA autohydrolysis (Figure 1a). The intermediate is then hydrolyzed in aqueous solution to 2′- and 3′-phosphate. Under acidic or basic conditions, the autohydrolysis of RNA can be accelerated millionfold compared to spontaneous hydrolysis at neutral pH [15].

If there are no gaps in the PEG coating, it can be assumed that the entire mRNA will also remain intact for at least 5 days at 2 °C to 8 °C. Higher temperatures can reduce this time to a few hours. In the meantime, however, shelf lives have been extended and cooling requirements further reduced [16] and the 5 days increased to 30. The manufacturer should therefore have further improved the stabilization of LNPs. It was therefore of interest to investigate whether the stabilizing PEG coating could possibly show qualitative production-related differences. For example, light-optical microscopic studies of mRNA vaccines reveal very unusual crystal-like impurities whose concentration increases with storage time. These are platelet-shaped geometric structures that bear a strong resemblance to rhomboid cholesterol crystals [17].

Fig. 3 Ch solid and liquid crystals observed by polarized light microscopy: (d) tube-like crystal fractures (e) typical ChM crystals with angles of 79.2° and 100.8° and often notched corners

Optical microscopy #

There are now many microscopic studies of mRNA vaccines (concentrated and diluted) in BF (bright field) and DF (dark field) with similar results. Researchers report finding strange and unidentifiable geometric impurities alongside the LNPs in these vaccine samples. Observations with a DF microscope revealed for vaccines from Moderna

Fig. 4 mRNA vaccine from the manufacturer Moderna stored at -18° C. and magnified 1000 times in the DF microscope

Crystal formation increases after storage at room temperature. The milky turbidity of Pfizer/Biontech Comirnaty is likely to have a lot to do with crystallization. This was also confirmed by simple slit lamp examinations of originally closed vaccine doses. This becomes clearer in the dark field microscope and is also documented with video recordings.

Fig. 5 mRNA vaccine from Pfizer/Biontech stored at room temperature and magnified 1000x in a DF microscope

A comparison with an aqueous solution of solid cholesterol (brand: Fisher Chemical C/5360/48) in the DF microscope clearly shows similar crystal structures as can be observed in the vaccines

Fig. 6 Cholesterol crystals in aqueous solution

Investigations of dried seed samples on silicon carriers in the reflected light microscope and in the SEM (scanning electron microscope) showed further crystallization phenomena on ordered surfaces such as Si or ChM through vaccine-specific buffer and salt solutions

Fig. 7 Crystal formation of mRNA-1273 (Moderna) in the reflected light microscope (top) and in the scanning electron microscope (bottom)

Scanning Electron Microscopy #

A more detailed SEM examination of mRNA-1273 in the edge areas also shows the crystallization of larger platelet-shaped structures. Electron beam scans of these crystals show a clear NaCl signal in the EDS.

Fig.8 NaCl crystal platelets in dry mRNA-1273 in the scanning electron microscope and their EDS spectrum

Figures 4, 5, and 6 show appearances of solid Ch crystals and optical textures of liquid crystals. Arc-like crystals in Fig. 4 were short curved rods. Sometimes longer filaments have been found, which can also be longer than 10 µm. Long arcs can also be coiled into irregular spirals. Tubular crystals were also observed (Fig. 9) breaking at their ends. Plate-like ChM crystals mostly occur with angles of 79.2° and 100.8° and often a notched corner. After some time, small liquid crystals began to accumulate (Fig. 5) and formed 1-5 µm particles, which can be called aggregated liquid crystals.

Fig. 9 Tubular Ch crystals in SEM

When the drying process begins, the salts in the buffer solution will also crystallize on existing Ch crystals and envelop them. Figure 10 shows a rectangular NaCl crystal showing an inner substructure with rounded corners indicative of Ch crystals

Fig. 10 SEM NaCl crystal with Ch embedded

The following questions arose from the optical observations given here as examples:

1. Which material changes limit the lifespan of the mRNA vaccines and can these be represented with simple measurement methods
2. Can material changes also be detected in certain batches due to the manufacturing process and how do they affect the safety profile

It was explained above why nano-lipids are coated with hydrophilic PEG for stabilization. Reference was also made to the combination with DSPC and cholesterol. A mass spectrometric method can be used to characterize the PEG coating. For this we had a Bruker MALDI time-of-flight mass spectrometer at the University of Natural Resources and Life Sciences in Vienna.

Mass Spectrometry #

From an analytical point of view, PEG samples are suitable for conventional spectroscopic methods such as matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) [18, 19].

The PEG samples were prepared according to a standard protocol for MS measurement. MALDI-TOF mass spectra were recorded using the reflectron mode on an Ultraflex TOF-TOF mass spectrometer (Bruker Daltonik GmbH, Bremen, Germany). The serial instrument was equipped with a 337 nm, 50 Hz N2 laser. Depending on the range examined, the peak resolution achieved was between m/z 250 – 950 or 1000 – 4000. If the measurement is limited to the mass range defined by the calibration substance (m/z 3000-4000), more than 95% of the peaks were Measured errors below 5 ppm.

In all Comirnaty batches examined, the material was taken from the original vaccine doses. The task was to identify the „flyable“ components of the solid fraction and to examine them for qualitative differences. For example, the measurement allows conclusions to be drawn about the number of ethylene oxide units (n) and the structure of the end group R of ALC-0159. You also get information about the quality of the ALC-0135.

Fig 11 TOF-MS fragmentation of RNA results in the same cyclic phosphate intermediate formed during RNA autohydrolysis.

Fragmentation of RNA leads to 2′,3′-cyclic phosphates. The resulting c fragments have the same m/z ratio as the cyclic phosphate intermediate formed during RNA autohydrolysis [18]. Measurements from different batches showed relatively strong differences in the fragmentation at m/z=328. Figure 12 shows spectra of different batches with different c-fragmentation. High relative levels of c fragments (EP2166) possibly indicate high levels of intact mRNA. This observation also corresponds to the following observations for PEG.

Figure 12 MS of four batches of Comirnaty. m/z=328 differences indicate different mRNA loadings of the LNPs

The splitting at m/z=766 and m/z=764 of ALC-0315 probably has to do with the substitution of two hydrogens by a C=C double bond. It remains to be clarified whether this has to be explained with the MALDI methodology or the manufacturing quality. Also the ALC-0315 intensity seems to match well with the following PEG results.

The mass ranges m/z 1000 – 4000 show the PEG polymer distributions of the batches examined. In Fig. 12 these differences become clear and call for explanation. First, truncated polymer chains could be generated either during MALDI analysis through in-source fragmentations or during sample preparation. More likely, however, are incomplete reactions leading to the presence of by-products (unreacted starting materials, possibly mono-substituted PEG, or unexpected end-group modification). These products could be stable in the remaining steps of the synthesis or continue to react, yielding increasingly complex mixtures with ions detected over a wide mass range.

Figure 13 PEG range m/z 2000-3500 shows lot dependent polymerization results

Let’s then summarize the measured PEG polymer distributions in terms of the number of ethylene glycol members

Fig. 14 Overview of PEG polymerisation

We interpret a higher number of polyethylene elements as possible incomplete polymerization of the LNP surface.

Fig. 15 Incomplete PEG polymerization of the LNP surface

In the next step, we asked ourselves whether such PEG inhomogeneities could have an impact on side effects (ADR) after the injection. This is a valid question given the strong dependence of LNPs and mRNA on the PEG stabilization mechanism. The batch-related vaccine side effects can be found under [20] in „How Bad is My Batch“. Our measurements revealed a statistically measurable relationship between the maximum of the polymer distribution of the PEG and the ADRs from this database.

Fig. 16 Relationship between PEG polymerization and vaccination side effects

Thus, we first hypothesize that a homogeneous polymerization of the LNP surface can stabilize the vaccine for a long time, preventing it from degradation before the LNPs are transported to unwanted body organs. Conversely, incomplete polymerisation and coating are insufficient protection against rapid degradation of the LNPs and destroy their function as protective mRNA transporters. With the decomposition of the LNPs, cholesterols are also released and can form into crystals. Of course, the question also arises as to whether the cholesterols contained in the LNPs are fully incorporated into the cell membrane during endocystosis, or whether they then also crystallize out in the extracellular fluid. More research is needed here, especially accurate animal studies to learn more about the effects of PEG and cholesterol.

References #

1. K. Greis, C. Kirschbaum, M. I. Taccone, M. Götze, S. Gewinner, W. Schöllkopf, G. Meijer, G. von Helden, K. Pagel, Angew. Chem. Int. Ed. 2022, 61, e202115481;
2. Michael D. Buschmann, Manuel J. Carrasco, Suman Alishetty, Mikell Paige, Mohamad Gabriel Alameh, Drew Weissman,Vaccines (Basel) 2021 Jan; 9(1): 65. Published online 2021 Jan 19. doi: 10.3390/vaccines9010065, PMCID: PMC7836001
3. Behlke MA., Oligonucleotides. 2008 Dec;18(4):305-19. doi: 10.1089/oli.2008.0164. PMID: 19025401
4. Choi J., Indrisiunaite G., DeMirci H., Ieong K.-W., Wang J., Petrov A., Prabhakar A., Rechavi G., Dominissini D., He C., et al.,  Nat. Struct. Mol. Biol. 2018;25:208–216. doi: 10.1038/s41594-018-0030-z
5. Lorenz C., Fotin-Mleczek M., Roth G., Becker C., Dam T.C., Verdurmen W.P.R., Brock R., Probst J., Schlake T.,  RNA Biol. 2011;8:627–636. doi: 10.4161/rna.8.4.15394.
6. Diken M., Kreiter S., Selmi A., Britten C.M., Huber C., Türeci Ö., Sahin U., Gene Ther. 2011;18:702–708. doi: 10.1038/gt.2011.17.
7. Oberli M.A., Reichmuth A.M., Dorkin J.R., Mitchell M.J., Fenton O.S., Jaklenec A., Anderson D.G., Langer R., Blankschtein D., Nano Lett. 2017;17:1326–1335. doi: 10.1021/acs.nanolett.6b03329
8. Pardi N., Hogan M.J., Naradikian M.S., Parkhouse K., Cain D.W., Jones L., Moody M.A., Verkerke H.P., Myles A., Willis E., et al., J. Exp. Med. 2018;215:1571–1588. doi: 10.1084/jem.20171450
9. COVID-19 Vaccine Emergency Use Authorizations (EUAs), EUA letter of authorization, https://www.fda.gov/media/144412/download
10. Clinical Trial of a COVID-19 Vaccine in Healthy Adults, https://www.modernatx.com/sites/default/files/mRNA-1273-P301-Protocol.pdf
11. Fact Sheet for Healthcare Providers Administering Vaccine, https://www.fda.gov/media/144637/download
12. Pfizer-BioNTech Fact Sheets and FAQs, https://www.fda.gov/media/144413/download
13. Gómez G, Pikal MJ, Rodríguez-Hornedo N., Pharm Res. 2001 Jan;18(1):90-7. doi: 10.1023/a:1011082911917. PMID: 11336359.
14. Tang X, Pikal MJ., Pharm Res. 2004 Feb;21(2):191-200. doi: 10.1023/b:pham.0000016234.73023.75. PMID: 15032301
15. Emilsson GM, Nakamura S, Roth A, Breaker RR., RNA. 2003 Aug;9(8):907-18. doi: 10.1261/rna.5680603. PMID: 12869701; PMCID: PMC1370456
16. EMA press office, https://www.ema.europa.eu/en/news/more-flexible-storage-conditions-biontechpfizers-covid-19-vaccine
17. D Q Wang, M C Carey, Journal of Lipid Research, 1996, vol 37(3), 606-630, ISSN 0022-2275
18. Christine Enjalbal et al, MALDI-TOF Journal of the American Society for Mass Spectrometry, 2005, vol 16(5), 670-678, ISSN 1044-0305
19. Na, D.H., Youn, Y.S. and Lee, K.C. (2003), Rapid Commun. Mass Spectrom., 17: 2241-2244. https://doi.org/10.1002/rcm.1175
20. Batch codes and associated deaths, disabilities and illnesses for Covid 19 Vaccines, https://howbad.info/