“Goldilocks Modifications” for mRNA Therapeutics Won the Nobel Prize

The use of chemically modified nucleotides propels oligonucleotide therapies to clinical maturity [1]. The use of modified nucleosides has also revolutionized mRNA therapeutics. Nucleoside-modified mRNA COVID-19 vaccines were developed in <1 year and have saved millions of lives. The 2023 Nobel Prize in Physiology or Medicine was awarded to Katalin Karikó and Drew Weissman, who have made seminal contributions to the nucleoside-modified mRNA technology.

In their two groundbreaking articles [2,3], Drs. Karikó and Weissman demonstrated that replacing all uridine nucleosides in an exogenous mRNA with pseudouridine (ψ) can significantly reduce the mRNA immunogenicity and increase the protein expression. At that time, although it was known that ψ is the most common post-transcriptional RNA modification, the role of ψ in RNA innate immune sensing and mRNA translation had been largely unknown. Drs. Karikó and Weissman noticed that mammalian tRNA, rich in ψ modification, was much less immunogenic than mRNA when transfected into cells.

After screening several modified nucleosides commonly found in tRNA, they discovered that ψ was the key modification for reducing RNA immunogenicity [2]. A derivative of ψ, N1-methylpseudouridine (m¹ψ), further enhances translation efficiency and is used in mRNA vaccines against COVID-19 [4,5]. The chemical difference between uridine and these modified nucleosides is subtle––ψ is isosteric to uridine, and m¹ψ is isosteric to 5-methyluridine. Why can such a subtle change drastically improve the efficacy and safety profile of mRNA vaccines?

“Goldilocks Modifications”

Early studies of mRNA therapeutics showed that unmodified mRNAs trigger a strong unwanted immune response when delivered to cells [6]. Therefore, the desired modified nucleoside for mRNA therapeutics would be in a “Goldilocks zone” in the chemical space––similar enough to canonical nucleosides for the mRNA to be translated efficiently, yet distinct enough to avoid triggering unwanted immune responses. In addition, these modified nucleosides should be compatible with a high-yield in vitro transcription system for large-scale manufacturing, and ideally, their incorporation will enhance the stability of mRNAs. Pseudouridine and m¹ψ are “Goldilocks modifications” because they meet all four criteria.

Let us start with translation. The high structural similarity between uridine and ψ/m¹ψ allows mRNAs containing these modified nucleosides to be decoded with high fidelity and speed. Experimentally, it has been demonstrated that m¹ψ modification has minimal impact on the yield and accuracy of translation [7]; stop codons containing ψ can also be efficiently recognized by release factors to ensure proper translation termination [8]. The structural similarity between uridine and ψ/m¹ψ also allows efficient manufacturing of nucleoside-modified mRNAs––ψ- and m¹ψ-triphosphates can be incorporated efficiently and faithfully by the widely used T7 RNA polymerase [2,9], and the yields are comparable with unmodified RNAs.

How does ψ/m¹ψ suppress the unwanted immune response? Studies on RNA innate immune sensing, including Drs. Karikó and Weissman's pioneering works, revealed that when unmodified mRNA is delivered into cells as exogenous RNAs through endocytosis, it can trigger a strong immune response. The exogenous RNA is first degraded by endosomal endonucleases, and these degradation products, including uridine-rich oligonucleotides and uridines [10,11], activate endosomal innate immune sensors Toll-like receptor 7 (TLR7) and TLR8 [2,10–15], which are part of the cellular antiviral mechanism.

The specificity of these receptors explains why replacing uridine is crucial for reducing the mRNA immunogenicity. Through a mechanism not fully established yet, ψ/m¹ψ completely blocks the activation of TLR7 and TLR8 [2,12,16].

A second reason that ψ/m¹ψ can mitigate unwanted immune responses to therapeutic mRNAs is their ability to diminish the formation of double-stranded RNA (dsRNA) side products during in vitro transcription. The T7 RNA polymerase can generate antisense RNA complementary to the primary transcript through either an RNA-dependent RNA polymerase activity [17] or a promoter-independent transcription of the DNA template [18].

The resulting dsRNA side products can be recognized by multiple innate immune sensors, including MDA5 [19] and oligoadenylate synthetase, OAS [20]. Pseudouridine and m¹ψ have been shown to reduce the dsRNA formed during in vitro transcription [18] and might also weaken the binding affinity of these innate immune sensors to dsRNAs.

Pseudouridine and m¹ψ also provide additional benefits for mRNA therapeutics by increasing the functional half-life of mRNAs. Pseudouridine modification confers resistance to some RNases, including RNase A and RNase L [21,22]. A more recent article also showed that ψ/m¹ψ enhances the thermal stability of mRNA secondary structures, which can increase mRNA functional half-life by enforcing spacing between translating ribosomes, thus reducing ribosomal collisions that can lead to the No-go decay of the mRNA [23].

Outlook

Drs. Karikó and Weissman's pioneering work on nucleoside-modified mRNAs resolved a key challenge in mRNA therapeutics: how to reduce the unwanted innate immune responses while maintaining the high translation efficiency of the mRNA. The nucleoside-modified mRNA technology is powering the development of a wide range of mRNA vaccines and mRNA immunotherapies (for a list of ongoing clinical trials with mRNA vaccines, please see Table 1 in ref. [24]).

Novel chemical modifications that can further extend the mRNA half-life will enable an even wider range of therapeutic applications beyond vaccines. Such modifications can include both novel modified nucleosides as well as chemically stabilized cap analogs. Notably, β-S-ARCA, a phosphorothioate cap analog, can increase the stability and translation efficiency of mRNAs in human dendritic cells [25], and it has entered clinical trials [26].

Finding new chemical modifications for better nucleic acid therapeutics is an optimization problem in the chemical space under (usually) multiple constraints. Yet, for the past three decades, the ingenuity of nucleic acid chemists has led to breakthroughs in one field after another. The relatively young mRNA therapeutics field still has many important challenges to solve. I can only expect more great discoveries in this modality.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

I acknowledge financial support from NIH R21-AI172204, UMass Center for Clinical and Translational Science Pilot Project Program, and start-up funds from UMass Chan Medical School.