SARS-CoV-2 viral evolution

It’s funny that this pandemic can prove all anti-evolutionists wrong. 

Nothing like seeing Charles Darwin natural selection right in front of your eyes at the speed of light. Just look at the SARS-CoV-2 viral evolution…

In his book “On the origin of species” written in 1859, Charles Darwin defined natural selection as the “principle by which each slight variation of a trait, if useful, is preserved”. What does this mean (?), it means the individuals best adapted to their environments are more likely to survive and reproduce. 

What we are seeing now, is that the mutations of SARS-CoV-2 that better promote spreading, are the ones that are becoming more common among the population, even when derived in different locations. The virus is changing and evolving to spread more rapidly, because natural selection will optimize the level of virulence that maximizes pathogen fitness – expressed as the basic reproductive number (R0)1,2

On average, comparative data from previous studies tell us that, low-virulence infections have a greater chance of successfully establishing transmission cycles in humans than virus with higher mortality3. As such, as before, the virus actually just wants to spread and not kill.

But, since the environment also affects transmissibility, there are more factors on the equation “when will this madness end” than we would have wished for.

For example, in the evolutionary trade-off between virulence and transmissibility, because intra-host virus replication is needed to allow inter-host transmission, it is almost impossible for natural selection to optimize all traits simultaneously1 and give us some peace. 

For example, in the case of the Myxoma virus (MYXV) in rabbits, this evolutionary trade-off leads to an ‘intermediate’ virulence being more advantageous to the virus than a higher virulence1,4. This happens because the rabbit host dies before inter-host transmission, in the case of higher virulence; and, with lower virulence the virus goes absolutely nowhere, because it does not increase virus transmission rates. A similar trade-off model has been proposed to explain the evolution of HIV virulence1,5.

Unfortunately, experimental studies in some viruses have shown that high virulence can promote certain advantages, as in the case of malaria, where a higher virulence was shown to provide the Plasmodium parasites with a competitive advantage within hosts1,6. Or, in the case of the rabbit haemorrhagic disease virus (RHDV), where there is evidence that virulence has increased through time, probably because virus transmission often occurs through flies that feed on animal carcasses, making host death selectively favourable1,7.

Let’s thank the Gods that SARS-CoV-2 is NOT transmissible through flies.

So, current evolutionary theory tells us that it is possible to anticipate the direction of virulence evolution, if the key relationship between virulence and transmissibility, and hence viral fitness, is understood1. Crucial to this is the analysis of the intersection between genomics and evolutionary studies, what is called phylogenomics.

This field of science provides a way to understand virulence evolution, and creates a number of hypotheses that can be tested using appropriate experimental cell assays and bioinformatic tools8,9

The collaboration of public health and research teams worldwide has now allowed the publication of 620,338 SARS-CoV-2 genomes in GISAID (http://www.gisaid.org/) (as of February 25, 2020)9. At the same time, a dynamic nomenclature system for SARS-CoV-2 has been described to facilitate real-time epidemiology revealing links between global outbreaks that share similar viral genomes10. At the root of the phylogeny are two lineages, A and B; where, A is likely ancestral, as it shares two distinguishing variants with the closest known bat viruses. Further linage designations link new variants to geographically distinct populations, B.1 in the Italian outbreak, then other parts of Europe and the world; and, B.1.1 being the major European lineage which was spread throughout the world. However, many of the major lineages are now present in most countries, and recapitulate the global diversity of SARS-CoV-2, indicating that most local epidemics were seeded by a large number of independent introductions of the virus.

The current evolutionary tree of SARS-CoV-2 shows multiple introductions of different variants across the globe, with introductions from distant locations seeding local epidemics, where infections sometimes went unrecognized for several weeks and allowed wider spread11. The tree topology actually indicates that SARS-CoV-2 viruses have not diverged significantly since the beginning of the pandemic11. These results show that, so far, SARS-CoV-2 has evolved through a non-deterministic, noisy process; and, that random genetic drift has played the dominant role in disseminating unique mutations throughout the world11.

There remains an urgent need for a SARS-CoV-2 vaccine as a primary countermeasure to contain and mitigate the spread; and, the virus’s surface S (Spike) protein continues to be an attractive vaccine target,because it plays a key role in mediating virus entry into the cells, and is known to be immunogenic.

Of course, the virus was only recently identified in the human population with a short time frame relative to the adaptive processes that can take years to occur.

But, the most recent findings show us that the SARS-CoV-2 viruses that are currently circulating, constitute a homogeneous viral population, to which the current vaccines available will be sufficient to mitigate the spread. 

Soon, SARS-CoV-2 will become just another viral acquaintance during the winter, like a common cold

Tree

References:

1          Geoghegan, J. L. & Holmes, E. C. The phylogenomics of evolving virus virulence. Nature Reviews Genetics 19, 756-769, doi:10.1038/s41576-018-0055-5 (2018).

2          Bull, J. J. & Lauring, A. S. Theory and empiricism in virulence evolution. PLoS Pathog 10, e1004387, doi:10.1371/journal.ppat.1004387 (2014).

3          Geoghegan, J. L., Senior, A. M., Di Giallonardo, F. & Holmes, E. C. Virological factors that increase the transmissibility of emerging human viruses. Proc Natl Acad Sci U S A 113, 4170-4175, doi:10.1073/pnas.1521582113 (2016).

4          Kerr, P. J. et al. Next step in the ongoing arms race between myxoma virus and wild rabbits in Australia is a novel disease phenotype. Proceedings of the National Academy of Sciences 114, 9397-9402, doi:10.1073/pnas.1710336114 (2017).

5          Fraser, C., Hollingsworth, T. D., Chapman, R., de Wolf, F. & Hanage, W. P. Variation in HIV-1 set-point viral load: epidemiological analysis and an evolutionary hypothesis. Proc Natl Acad Sci U S A 104, 17441-17446, doi:10.1073/pnas.0708559104 (2007).

6          de Roode, J. C. et al. Virulence and competitive ability in genetically diverse malaria infections. Proc Natl Acad Sci U S A 102, 7624-7628, doi:10.1073/pnas.0500078102 (2005).

7          Di Giallonardo, F. & Holmes, E. C. Viral biocontrol: grand experiments in disease emergence and evolution. Trends Microbiol 23, 83-90, doi:10.1016/j.tim.2014.10.004 (2015).

8          Stern, A. et al. The Evolutionary Pathway to Virulence of an RNA Virus. Cell 169, 35-46.e19, doi:10.1016/j.cell.2017.03.013 (2017).

9          Sjaarda, C. P. et al. Phylogenomics reveals viral sources, transmission, and potential superinfection in early-stage COVID-19 patients in Ontario, Canada. Scientific Reports 11, 3697, doi:10.1038/s41598-021-83355-1 (2021).

10        da Silva Filipe, A. et al. Genomic epidemiology reveals multiple introductions of SARS-CoV-2 from mainland Europe into Scotland. Nat Microbiol 6, 112-122, doi:10.1038/s41564-020-00838-z (2021).

11        Dearlove, B. et al. A SARS-CoV-2 vaccine candidate would likely match all currently circulating strains. bioRxiv, 2020.2004.2027.064774, doi:10.1101/2020.04.27.064774 (2020).

Mutations

RNA Viruses, such as the SARS-CoV-2HIV and Influenza, tend to pick up mutations rather quickly, once they are copied inside the cells of our bodies. This happens because the little cell workers that copy the RNA (called enzymes) are sloppy workers and prone to make errors. 

The SARS-CoV-2 virus genetic code has just 30,000 nucleotides blocks of RNA, or letters that can spell at least 29 genes1; and, the most common mutations are single-nucleotide changes between viruses from different people, that have little effect on the overall performance of the virus infection rate, but that allow researchers to track the spread by linking closely-related viruses2

In fact, sequencing data actually suggests that coronaviruses change more slowly than most other RNA viruses, because they have “proofreading” mechanisms that correct potentially fatal copying mistakes, accumulating only two single-letter mutations per month in its genome — a rate of change about half that of Influenza and one-quarter that of HIV2.

The problem is, that scientists can spot mutations faster than they can make sense of them, or what problems they may cause. Many mutations will have no consequence for the virus’s ability to spread or cause disease, because they do not alter the shape of a protein; whereas those mutations that do change proteins are more likely to harm the virus than improve it2. For example, in the beginning of the pandemic in Singapore, the ∆382 deletion made the infection of the virus milder3. On the contrary, the G614 variant showed a fitness advantage of infection, where individuals had a higher concentration of pseudotyped virions, suggestive of higher upper respiratory tract viral loads, but not with increased disease severity4

If we think about it, the virus just wants to spread, but not actually kill – because, if it kills, it can no longer spread.

More recently, a new SARS-CoV-2 virus variant showed up in the UK, where multiple spike protein mutations are seen (deletion 69-70, deletion 144, N501Y, A570D, D614G, P681H, T716I, S982A, D1118H)5. Changes are also seen in other genomic regions, inclusively in the receptor binding domain (N501Y, “Nelly”). 

The unusually high number of spike protein mutations suggests that the variant has not emerged through gradual accumulation of mutations6. Instead, one possible explanation for the emergence of the variant, is prolonged SARS-CoV-2 infection in a single patient, potentially with reduced immunocompetence, similar to what has previously been described7,8. Such prolonged infection in immunocompromised patients can lead to accumulation of immune-escape mutations at a higher rate. 

This explanation is also suggested to the current SARS-CoV-2 virus variant found in Brazil (P.1)9, and in South Africa (501Y.V2)10

Unfortunately, this shows that the emergence of successful variants with similar properties is not so rare6; and, just like the need to develop new influenza vaccines every season, that also might be the case for SARS-CoV-2. The persistence of the pandemic, may enable accumulation of immunologically relevant mutations in the population, even as vaccines are developed. 

But, because science tries to be one step ahead, this is exactly what manufacturers are already working on, specifically the ones using mRNA technology that can quickly adapt their platforms to the current mutations seen.

Mutation luck
Mutation of luck

References:

1          Naqvi, A. A. T. et al. Insights into SARS-CoV-2 genome, structure, evolution, pathogenesis and therapies: Structural genomics approach. Biochim Biophys Acta Mol Basis Dis 1866, 165878-165878, doi:10.1016/j.bbadis.2020.165878 (2020).

2          Callaway, E. The coronavirus is mutating — does it matter? Nature 585, 174-177, doi:https://doi.org/10.1038/d41586-020-02544-6 (2020).

3          Young, B. E. et al. Effects of a major deletion in the SARS-CoV-2 genome on the severity of infection and the inflammatory response: an observational cohort study. The Lancet 396, 603-611, doi:10.1016/S0140-6736(20)31757-8 (2020).

4          Korber, B. et al. Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G Increases Infectivity of the COVID-19 Virus. Cell 182, 812-827.e819, doi:10.1016/j.cell.2020.06.043 (2020).

5          Andrew Rambaut, N. L., Oliver Pybus, Wendy Barclay4, Jeff Barrett5, Alesandro Carabelli6, et al. Preliminary genomic characterisation of an emergent SARS-CoV-2 lineage in the UK defined by a novel set of spike mutations: COVID-19 genomics UK consortium. https://virological.org/t/preliminary-genomic-characterisation-of-an-emergent-sars-cov-2-lineage-in-the-uk-defined-by-a-novel-set-of-spike-mutations/563 December 2020 (2020).

6          ECDC. Rapid increase of a SARS-CoV-2 variant with multiple spike protein mutations observed in the United Kingdom. https://www.ecdc.europa.eu/sites/default/files/documents/SARS-CoV-2-variant-multiple-spike-protein-mutations-United-Kingdom.pdf December 2020 (2020).

7          Choi, B. et al. Persistence and Evolution of SARS-CoV-2 in an Immunocompromised Host. N Engl J Med 383, 2291-2293, doi:10.1056/NEJMc2031364 (2020).

8          McCarthy, K. R. et al. Natural deletions in the SARS-CoV-2 spike glycoprotein drive antibody escape. bioRxiv, 2020.2011.2019.389916, doi:10.1101/2020.11.19.389916 (2020).

9          Kupferschmidt, K. New mutations raise specter of ‘immune escape’. Science 371, 329-330, doi:10.1126/science.371.6527.329 (2021).

10        Karim, S. S. A. The 2nd Covid-19 wave in South Africa. https://www.scribd.com/document/488618010/Full-Presentation-by-SSAK-18-Dec December 2020 (2020).

Moderna mRNA COVID-19 vaccine: how did they do it?

Moderna uses an efficient method to make large quantities of individual mRNAs with high purity, based on a technique developed in 1984 by D. Melton and colleagues at Harvard University and published in Nucleic Acids Research. In here bacteriophage polymerases were used to transcribe plasmid DNA and make purified RNAs. D. Melton went further and showed that he could transcribe a synthetic mRNA injected into frog oocytes, and get expression of b-globin (Krieg & Melton, Nucleic Acids Research, 1984). With this report, the protocol to make pure mRNAs was settled.

Today, at Moderna, they moved on from this idea to a GMP drug-product inside of a vial in under two months. In the case of the COVID-19 vaccine, they went from a sequence of the virus to having the material ready for a Phase I clinical trial in 45 days. 

This automated process starts with back-translation using Moderna’s proprietary computer algorithm, that has been defined over the years to pick the best mRNA sequence for the target in question. 

Then, the therapeutic mRNA is engineered to avoid the innate immune sensors that defend against RNA viruses and might destroy the vaccine. These defences come in to main types: Toll-like receptors (TLR) in the endoplasmic reticulum, and RIG-I/MDA-5-like receptors in the cytoplasm. 

To avoid TLRs, Moderna places modified nucleotides. This technique is based on a paper by Diebold and colleagues published in Science (2004), where the researchers took macrophages from mice and incubated them with different polymers (polyA, polyC, polyG, polyI, polyU), then checking for an IFN-alpha response. What they have seen was that the only polymer that created the IFN-alpha response was the one that contain Uridine (polyU), which seems to be the only nucleotide that is being recognized by the TLR-driven innate immune response. As such, at Moderna, mRNAs are engineered to replace all U’s with N1-methyl-pseudiuridine without affecting base-pairing, but avoiding TLR recognition. N1-methyl-pseudiuridine is a naturally occurring nucleotide that our bodies recognize as native.

Another problem though, is avoiding the creation of double-stranded RNA (dsRNA) to escape the RIG-I/MDA-5-like receptors. The state-of-the-art used in Moderna to avoid this trap, is HPLC purification to reduce dsRNA content. This is a technique based on Karikó & Weismann (2011, Nucleic Acids Research), that results in very little dsRNA in the end-product, and an avoidance of the innate immune response. The developed technique at Moderna has recently been published in Science Advances (June 2020), where Moderna scientists describe a highly sensitive assay to detect dsRNA and purify therapeutic mRNAs.

(It seems Moderna scientists are now going even further, and have reengineer T7 RNA polymerase not to produce dsRNA at all – although this has not been published yet)

The third step to make a good mRNA-drug, and a successful vaccine, is to find an efficient delivery method.  With that in mind, Moderna researchers looked back at work done by Wolff and colleagues in the 90s (Science, 1990, vol.247, 1465-1468), where they were able to directly inject RNA in the mouse skeletal mouse and have protein expression, with no special delivery system involved. 

In 2014, researchers at Moderna were able to go further, and measured with a Luciferase system, the in vivo expression of biologically active proteins in a dose-dependent manner, by injecting mice with Moderna’s naked mRNA. Nowadays, Moderna targets different tissues via multiple Routes of Administration with specific delivery vehicles (ROAs: direct injection in the heart or tumours, subcutaneous, intravenous, intramuscular). 

Most ROAs require mRNA encapsulation with Lipid nanoparticles (LNPs, 80-100nm). Each encased particle contains 2-6 molecules of mRNA, phospholipid cholesterol, a PEG lipid that keeps the lipid nanoparticles stable avoiding aggregation, and an ionizable lipid that at low pH interacts with the mRNA. These mRNA-LNPs are akin to endogenous lipid transport complexes, slightly bigger than Very-Low Density Lipoproteins (VLDLs), and seen as “friendly bubbly characters” by our bodies.

At Moderna, they empower a rational-structure LNP design by analysing and engineering each individual step of the component mixture, to enhance chemical/physical stability of mRNA-LNPs to allow an optimal biodistribution, cellular uptake, endosomal escape and protein expression. In comparison to the competitor LNPs in the field, which have a 1-2% delivery rate, Moderna’s LNPs can deliver 30% mRNA into cells. To date, Moderna has published two peer-revied articles describing these efforts and their applications in different ROAs (Sabis … Benenato, Mol Ther, 2018Hassett … Britto, Mol Ther Nuc Acids, 2019). 

Finally, there needs to be the right knowledge to engineer the best mRNA sequence for a particular purpose in order to make a good mRNA-drug or vaccine. 

So, there needs to be translation initiation fidelity, so that it always starts in the right place and there’s faithful decoding; and, also, the ribosome needs to stop in the right place. 

The mRNA needs to have a functional half-life, to function before it gets destroyed. 

Also, the therapeutics or vaccine needs to hit the correct cell type, which is achieved by putting “off-logic gates”, like microRNAs target sites in the 3’prime UTR, that cause the mRNA to be degraded in case it hits an undesired cell type.

At Moderna, they have all steps in place for a rapid development of mRNA as medicines:

  1. Efficient methods to make large quantities of individual mRNAs at high purity;
  2. Ability to avoid the innate immune sensors that defend against RNA viruses;
  3. Efficient delivery methods,
  4. Knowledge of how to engineer the best mRNA sequence for a particular purpose.

Besides proprietary algorithm to define the mRNA sequence and unique LNP design; one of the big differences between BioNTech/Pfizer vaccine and Moderna, is that Moderna does not use self-amplifying RNA in their vaccines. According to Dr Melissa Moore, Chief Scientific Officer at Moderna Therapeutics Inc., self-amplifying RNA first starts by building dsRNA which pushes-up the immune response against the vaccine, as such Moderna hasn’t gone down that road to avoid such road-blocks.

The important take-out message is not which vaccine we take.

With a mRNA vaccine, the cell is doing the protein itself, and the antigen-presenting cells are exhibiting it on their surface, which causes an activation of B and T cells.

This protects us and others from developing this devastating disease.

COVID-19 mRNA vaccine, the BNT162b2 candidate

The principle of mRNA therapeutics is to introduce therapeutic messenger (m) RNAs encoding the genetic information for a protein, into a cell of interest. The mRNA structure is designed to increase half-life, translation and protein functionality. 

The mRNA synthesis is made by in vitro transcription using a linear DNA template, which provides more than 500 copies of mRNA per template. They promote a transient expression of the encoded protein/antigen, and are degraded into nucleotides, without the formation of toxic metabolites. Since mRNA is highly sensitive, it gets degraded after a short-time, with no risk of genomic integration. 

This one process can be used to manufacture essentially any mRNA sequence. 

The raw reaction mixture has mRNA and all types of impurities, both process-related (T7 RNA polymerase, remaining building blocks, hydrolyzed DNA…), and product-related impurities (break-off transcripts, side products…), that need to be further purified with high affinity chromatographic methods.

Once purified RNA is obtained, the therapeutic is ready to deliver to the patient.

BioNTech together with their partner Pfizer and FosunPharma, managed to do all this in a record time of 84 days to develop a COVID-19 vaccine. 

Their scientists designed multiple antigen variants using the SARS-CoV-2 Spike (S) protein.

Next, they ran preclinical and toxicology evaluation tests in vitro and in vivo (e.g., in vitro expression data, antibody titers in animal models, pseudo-virus neutralization (pVN) assay in animal models), so they could discover the most active antigen and bring it into the clinic. Also, using their Good Manufacturing Practice (GMP) facilities they produced small scale batches for fast entry into the clinic to demonstrate safety, tolerability (i.e., low reactogenicity), and immunogenicity. 

Afterwards, they initiated clinical trials, where not one but four vaccine candidates were simultaneously studied, in order identify the safest and most effective candidate for further development. In parallel, there was already preparation of larger manufacturing batches of the candidate with the best output in the clinical studies; and, the identification of collaboration partners to develop and provide the vaccine worldwide (Pfizer, FosunPharma).

BioNTech and Pfizer studied three mRNA types during all Phase I/II studies, to evaluate safety and efficacy. The basis for such studies was that different types of mRNA vaccines lead to different responses. 

As such, vaccine prototype (1) was a Uridine mRNA (uRNA) with strong intrinsic adjuvant effect, strong antibody response, that stimulates the production of more CD8+T cells then CD4+T cells.

Vaccine prototype (2) was a Nucleoside-modified mRNA (modRNA), with a moderate intrinsic adjuvant effect, very strong antibody response, lower cytokine induction, that stimulates more CD4+T cells then CD8+T cells.

In the case of vaccine prototype (3), this was a Self-amplifying mRNA (saRNA) with long duration, and a higher likelihood for good efficacy antibody response with lower dosage. 

Early and constant interaction with the regulatory agencies (e.g.EMA) has allowed that BNT162b2 vaccine candidate raced to the finish line, being now the leading candidate from BioNTech & Pfizer; and, currently in review for conditional marketing authorization approval.

It is outstanding that if a new variant of the virus is required (e.g., due to a mutated virus sequence), a new GMP batch of a possible vaccine could be available within half of the time (around 42 days).

If there is a positive message out of these dark days, is that Science is there to save us from our doom. 

Let’s embrace Science.

mRNA vaccine