Tributaries of the 2023 Nobel Prize in Physiology or Medicine, and lessons learned

Abstract

Almost without exception, scientific breakthroughs are not epistemological orphans. Historians of science have developed a body of scholarship on this, and the cases arising in our era continue to confirm the phenomenon. The work by Katalin Karikó and Drew Weissman that proved foundational for the subsequent development of mRNA vaccines for COVID-19 had its antecedent roots yet is also a striking example of both serendipity and their persistence. Their receipt of the 2023 Nobel Prize in Physiology or Medicine was greatly deserved and, as Alfred Nobel likely envisioned the broad impact to be for all the prizes, affirms to the public at large that there is such a thing as the scientific method, and that there are such things as facts. The importance of society recognizing this has always been critically important, perhaps never more so than now.

Discoveries in science (and probably in all human pursuits) almost always have antecedents. Newton wrote in a letter to Robert Hooke that he had seen further because he stood on the shoulders of giants, though he was certainly not the first to use this phrase (Merton 1965). That notwithstanding, he did get gravity and coinvented the calculus, the latter arguably a rare example of a breakthrough without known precedence. (The Persian mathematicians had done so much before in arithmetic, but that impressive body of work was not a direct precursor for the calculus.) Certain later discoveries have also been alleged to have been “totally unanticipated,” one frequently mentioned case being the detection of the cosmic microwave background radiation by Arno Penzias and Robert Wilson. But although this had not been explicitly predicted, the intriguing earlier finding by Andrew McKellar that interstellar space hovers at about 2.9°K was an important clue. In our own field, the discovery of RNA interference was anticipated from seminal work in plants. And as to the case for serendipity, often a handmaiden to breakthroughs, it was the property of bacteriophage T7 RNA polymerase to often turn around and start a short transcript from the other DNA strand, known to Andrew Fire, that led him and Craig Mello to each, in full cooperation, use fully double-stranded RNAs and “the rest is history” (Fire et al. 1998).

How science works and advances, and how important it is for the public to know something of it, had champions in the 19th century like Louis Pasteur, who regularly engaged non-scientist audiences (Dubos 1950). In our time, we recognize historians like Robert Merton (vide supra) and Thomas Kuhn. In a pathfinding study, Merton's student Harriet Zuckerman found that a considerable number of Nobel laureates worked not with laureates, but with scientists whom they thought likely to become ones (Zuckerman 1977), an analog of the strategy of the hockey player Wayne Gretzky who famously said, “I don't go where the puck is, I go where it is going to be.”

mRNA AS THERAPY

The first report of an mRNA introduced into an animal involved cationic liposome transfection into mouse muscle tissue of a T7 in vitro transcript coding for chloramphenicol acetyltransferase (Wolff et al. 1990). It contained 5′ and 3′ untranslated sequences from the mouse β-globin gene and also had a 5′ cap, shown earlier to be critically important for the stability and translation of exogenous mRNAs (Green et al. 1983; Kreig and Melton 1984). Much of the data in this paper were from parallel DNA transfections, but expression from the introduced mRNA was demonstrated. At the same time, my laboratory succeeded in transfecting U2 small nuclear RNA into human cells using calcium phosphate, as in DNA transfection, and showed that it behaved properly as to its accurate 3′ end processing, association with U2 snRNP particle proteins and 5′ cap hypermethylation (Kleinschmidt and Pederson 1990). But this was a short, non-messenger RNA and, moreover, it was introduced lacking its usual modifications, including 10 ribose-2′-O-methyl groups and 13 pseudouridines. We did not check to see if these modifications occurred in the cells, but we later showed that correct site-specific pseudouridine formation does occur on U2 snRNA in HeLa cell extracts (Patton et al. 1994), and thus it is reasonable to assume it did in the aforementioned study. Pseudouridine will arise again in the story.

KARIKÓ AND WEISSMAN

This article is not intended to describe the training and early careers of Katalin Karikó and Drew Weissman as these have been described in an excellent review (Shapiro and Losick 2021) and interview (Neuman 2021). In due course, and through very different pathways, they individually ended up at the University of Pennsylvania, Karikó a molecular biologist working on RNA surveillance and Weissman an infectious disease physician-scientist inter alia. Between 1987 and 1991 she published six important papers on the RNase L-2′,5′-oligoadenylate system. In that period and subsequently, both investigators were enormously productive in their respective research domains, with Karikó focusing primarily on the intracellular expression of therapeutic proteins and Weissman on molecular and immunological aspects of viral infectious disease. A chance encounter at a Xerox machine led them to join forces, combining her expertise in RNA and his in virology and immunology. (That copier encounter now is becoming as famous as the one at a deli in 1972 at which Herb Boyer and Stanley Cohen conceived of recombinant DNA.)

INNATE IMMUNITY AND RNA

By this time, it had become clear that synthetic mRNAs (meaning transcribed by DNA bacteriophage T7 or SP6 RNA polymerases) introduced into mammalian cells are sometimes unstable whereas transfer RNAs, isolated from cells, were much more stable. In due course, Karikó and Weissman realized that this difference might not be the sizes of the RNAs, but rather the fact that natural tRNAs are highly modified—indeed they were where RNA modification was first reported.

Within a year, they published a seminal paper showing that synthetic RNAs containing modified nucleosides suppress TLRs when transfected into mammalian cells (Karikó et al. 2005). Many have hailed this paper as “the” or at least “a” turning point. (It had been previously rejected by three journals, with one regarding the work to be “an incremental advancement”.)

We shall return to RNA modifications shortly but first there is an important heuristic point to be interjected here. The primary motivation of Karikó had always been to express therapeutic proteins in situ, with the notion that the therapeutic effect would be manifest in those very same cells. However, in 2000 Karikó and Weissman demonstrated that transfection of cultured dendritic cells with an mRNA encoding the HIV gag protein resulted in potent immune responses in cocultured CD4+ and CD8+ T-cells (Weissman et al. 2000). From this, they envisioned that perhaps this would work as well in the intact organism, albeit with the requirement that the translation product would need to be stable enough in the circulation to trigger the desired immune response. Thus, it is to be recognized that this “vision” of theirs arose two decades prior to the advent of COVID-19.

THE PSEUDOURIDINE MYSTIQUE

Pseudouridine is an isomer of uridine in which the glycosidic bond is to C5 of the pyrimidine instead of to the usual N1. This creates additional hydrogen bonding potential and stability, such versatility increasing further since in pseudouridine the glycosidic linkage can be oriented either syn or anti to the ribose (Griffey et al. 1985). Historically, pseudouridine's stabilizing feature with complementary RNA was envisioned as a possible way, as deoxypseudouridine, to enhance the action of antisense DNA oligos (Supplemental Fig. 1). Among the modified nucleosides Karikó and Weissman investigated, they elected pseudouridine for further investigation and found that pseudouridine-substituted mRNA is not only non-immunogenic but also fails to trigger the RNA-dependent protein kinase, thus enhancing translational efficiency as well (Karikó et al. 2005, 2008; Anderson et al. 2010). Another major advance was defining how these modified nucleobases operate in terms of RNA structural biology (Nallagatla and Bevilacqua 2008).

mRNA VACCINOLOGY MATURES

During the period 2015–2020, Karikó and Weissman applied the idea of mRNA-based antigen expression and immune response to the Zika virus, with spectacular results (Pardi et al. 2017), and Weissman's group pursued the same approach for influenza, in collaboration with the leading expert Peter Palese, this study also producing impressive results (Freyn et al. 2020). That the approach might work for respiratory viruses was afoot, since ones like MERS had become major challenges. Then, everything changed when in Wuhan, China a new respiratory illness emerged and, with blinding speed, Chinese scientists isolated the virus, sequenced its genome, and released the data to the world. It was assumed that conventional, protein-based vaccine strategies would be undertaken and of course they were. But the catalytic power of bacteriophage RNA polymerases (initiation and elongation rate, and high turnover number) strutted into this domain as a confident grandee. Candidate mRNA-based SARS-CoV-2 vaccines were soon developed by BioNTech led by Karikó, Uğur Şahin and Ӧzlem Tϋreci, and Moderna, cofounded by the Harvard stem cell biologist Derrick Rossi, who had envisioned and indeed deployed mRNA transfection for desired proteins in his outstanding work on hematopoietic stem cells and other progenitor cells.

THREE MORE STEPS

In the case of the double helix, it has been said that in late 1952 Rosalind Franklin was only “two steps away.” This was correct in that she did not know, nor could have known, the meaning of the space group's C2 symmetry (Crick did) nor the chemistry of base-pairing, which Watson had wrong until a visiting scientist corrected it (Pederson 2020). Karikó and Weissman had realized that introduced mRNA needed to circumvent the RNA surveillance pathways that cue on the absence of usual modifications. They had nailed that, and now, there were three more steps. One was more efficient lipid nanoparticle delivery of the mRNA, and here the key technology was found in a Canadian company, Acuitas Therapeutics (Vancouver). The second key step involved understanding how antigenic the expressed SARS-CoV-2 spike protein would be. It does not normally leave infected cells by itself but in a virion, and in any case, does not use an N-terminal signal peptide for export via the endoplasmic reticulum. And the emerging decision to focus on intramuscular administration, given its hypervascularization, also played into the strategy. But the most critical “second step” was how stable the spike protein would be in the human bloodstream, and whether there was foundational knowledge about this. It turns out that there was.

ENTER NIH

As many observers know, the U.S. government funded Moderna, whereas BioNTech was funded through a partnership with Pfizer, each investment essential to the development of the vaccines. But the U.S. government supplied something else, and it was critical.

A group at NIH led by Barney Graham had for some years been emerging as leaders in the structural biology of viral proteins. These proteins typically have one conformation in the virion and yet sometimes another one when they dock on a receptor. Everyone in the mRNA vaccine field recognized that the expressed antigenic protein would need to elicit a strong immune response and that this might be enhanced if the most antigenic protein structure could be created and stabilized. For example, in one case Graham's group, in collaboration with Merck, introduced two prolines into an HIV protein, knowing that this amino acid cannot participate in an α-helix. Subsequently, Graham and colleagues tackled the SARS-CoV-2 spike protein and revealed how it folds and introduced mutations to create a more stable folding. It is not widely known that Graham's group contributed so very importantly to the COVID-19 vaccines, as did other laboratories, and I therefore emphasize this here. With this knowledge in hand, both BioNTech and Moderna designed their mRNAs to encode this stabilized form of the spike protein.

And then there was yet another step that Karikó and Weissman took. They had been prescient to introduce pseudouridine, but subsequently it was found that N¹-methylpseudouridine substitution stabilizes mRNA even more and, in addition, enhances its translational efficiency (Andries et al. 2015; Svitkin et al. 2017). This and other refinements led to the COVID-19 vaccines brought by BioNTech and Moderna, saving the lives of hundreds of millions of people throughout the world.

LESSONS

There are two centerpieces in this story. One is how antecedent science always is at play. The second is how certain scientists have the vision and intrepid courage to pursue their dream, despite all odds. Many, many obstacles rose up in front of Karikó and Weissman, but they stayed on focus. The late cancer biologist Judah Folkman once made a distinction between “obstinance” and “persistence,” arguing that the former is often irrational, but the latter is not when accompanied by a sense that there is a reasonable probability of being right (Folkman 1988).

Bravo to Kati and Drew, who have so vividly and valiantly displayed this.

SUPPLEMENTAL MATERIAL

Supplemental material is available for this article.

COMPETING INTEREST STATEMENT

The author owns stock in Moderna Therapeutics, Inc.