Introduction
Synthetic mRNA produced by in vitro transcription (IVT) is the active pharmaceutical ingredient of approved vaccines and of many drugs under development. IVT mRNA typically contains several hundred bases of non-coding untranslated regions (UTRs) that are involved in the stabilization and translation of the mRNA. Remarkably, new research has discovered that these UTRs can be virtually eliminated without compromising translation of the encoded protein of interest.
This blog summarizes how Swiss researchers (Mamaghani et al.) achieved this drastic structural minimalization in the form of “minRNA” having only 4 nucleotides (nts) in the UTR. In this way, overall production of IVT minRNA for research use and pharmaceuticals can be significantly simplified. Trilnk’s CleanCap® analog and modified nucleotide 5’-triphosphates were used for this research, which has already had more than 1,100 views since its publication on July 24, 2024.
Background
Following the early 1980s, when IVT mRNAs initially employed 5’ and 3’ UTRs found in beta globin genes, academics and companies screened for optimized (opt) 5’ and 3’ UTRs using different technologies in order to identify sequences that afford better functionality for the coding sequence (CDS) of interest (Figure 1, top). For example, optRNA in the Pfizer-BioNTech Comirnaty® vaccine has a 52-nt 5’ UTR from human alpha globin and a 295-nt 3’ UTR from two genes: mtRNR1 and AES, while optRNA in the Moderna SpikeVax® vaccine has a 57-nt 5’ UTR of undisclosed origin and a 111-nt 3’ UTR from human alpha globin.
Mamaghani et al. reasoned that, while the secondary and tertiary structures in UTRs in opt mRNA can provide higher stability/translation, these can be problematic sequences to introduce in a DNA template and to transcribe. As outlined in the following sections, systematic studies were therefore carried out to investigate whether easier to produce minRNA (Figure 1, bottom) would maintain its function compared to optRNA.
FIGURE 1. Schematic structures of 7-methyl guanosine (G) 5’ capped IVT mRNA with either (top) optimized, or opt, 5’ and 3’ UTRs or (bottom) a translation initiator of short 5’ UTR (TISU) element, with no 3’ UTR, each with a poly(A) tail. Image from Mamaghani et al. and used under CC BY 4.0 license.
Minimal 5’ UTRs are compatible with the translation of IVT mRNA
In 2008, Elfakess and Dikstein at the Weizmann Institute of Science in Israel reported that approximately 4% of mammalian mRNAs possess a very short 4-nt 5’ UTR sequence, CAAG, termed TISU (translation initiator of short 5’UTR), which is absent in the usual 9-nt Kozak consensus sequence. Using gaussia luciferase (Gluc) and firefly luciferase (Fluc) as luminescent reporter enzymes, Mamaghani et al. synthesized base-unmodified minRNAs (Figure 1, bottom) comprised of CleanCap® trinucleotide m7GpppAG followed by TISU, the AUG start codon, CDS, UAA stop codon, and poly(A) tail—without inclusion of a 3’ UTR.
As shown in Figure 2, transfection of HEK-293 cells with each of these TISU-based reporters led to significantly better performance compared to controls with either no mRNA or mRNA with no 5’ UTR. While the TISU construct was not significantly better than the Kozak-based RNA for Gluc, TISU was significantly better than Kozak for Fluc.
FIGURE 2. Functionality of the TISU 5’ UTR in unmodified RNA. Luciferase activity in HEK-293 cells 24 h after transfection of unmodified mRNAs containing no 5’ UTR, a Kozak 5’ UTR, or a TISU 5’-UTR and coding for gaussia luciferase (white bars) or firefly luciferase (grey bars), compared to no mRNA transfection. Experiments were performed in triplicate and data are presented as the mean ± SD; ns = not significant, ** p < 0.01, **** p < 0.0001. Image from Mamaghani et al. and used under CC BY 4.0 license.
Comparison of minRNA to Classical OptRNA In Vitro
Because of the widespread use of nucleoside-modified IVT mRNA as vaccines and therapeutics, Mamaghani et al. next compared the performance of TISU minRNA to a representative optRNA wherein each type of design was comprised of variously modified nucleosides. Based on their previous work, the optRNAs incorporated an elongation factor (eIF4G)-recruiting aptamer as a 5’ UTR and a beta globin tandem repeat as a 3’ UTR. The minRNA and optRNA encoding Fluc were generated with either unmodified uridine (U) or complete substitution of U by pseudouridine (Ψ), N1-methylpseudouridine (m1Ψ), or 5-methoxyuridine (moU).
As shown in Figure 3, in HEK-293 cells moU optRNA coding Fluc showed somewhat more luminescence-based protein expression than minRNA. However, there were no consistent statistically significant differences between minRNA and optRNA for the other three constructs. Comparable performance between Fluc minRNA and optRNA was also found in human peripheral blood mononuclear cells (PBMCs), although in this case each type of m1Ψ- and moU-modified construct was substantially more active than the corresponding Ψ-modified and unmodified U.
FIGURE 3. Comparison of protein expression for minRNA (black bars) and optRNA (gray bars) encoding Fluc. Constructs were either unmodified uridine (U) or complete substitution of U by pseudouridine (Ψ), N1-methylpseudouridine (m1Ψ), or 5-methoxyuridine (moU). Luminescence-based activity 24 h after transfection in HEK-293 cells (left panel) and PBMCs (right panel). Experiments were performed in triplicate and data are presented as the mean ± SD; ns = not significant, * p < 0.05. Image from Mamaghani et al. and used under CC BY 4.0 license.
Comparison of minRNA to classical optRNA for therapy or vaccination
Mamaghani et al. next compared the functionality of minRNA and optmRNA in several different model systems for therapy or vaccination. The first therapeutic approach studied was built on their previous demonstration that chimeric antigen receptor (CAR) T cells engineered with mRNA encoding a natural killer group 2D (NKG2D) could kill glioma cells (Meister et al.). In the present case, they co-encoded the NKG2D-CAR and the reporter protein termed RQR8 (Philip et al.) as either a minRNA or an optRNA having the SpikeVax® 5’ and 3’ UTRs (see Background), both constructs being m1Ψ-modified.
The percentage of expression of RQR8 on the surface of electroporated T cells was similar when using these minRNA or optRNAs, although the signal intensity was slightly higher with minRNA. Target cell killing was likewise similar for CAR T cells engineered with these minRNA or optRNAs.
The next model therapeutic was based on reported promising results for the cytokine interlukin-2 (IL-2) in treatment of cancer and autoimmunity. Mamaghani et al. generated IL-2-encoding minRNA and optRNA (5’ and 3’ UTRs from SpikeVax®, see Background), both constructs being m1Ψ-modified. These were transfected into HEK-293 cells and human primary fibroblasts for IL-2 measurement by ELISA after four days of culture. MinRNA was significantly (* p < 0.05) better than optRNA in HEK-293 cells and comparable to optRNA in primary cells.
Mamaghani et al. next tested Fluc expression in vivo following injection of mice with minRNA or optRNA (an eIF4G aptamer as a 5’ UTR and a beta globin tandem repeat as a 3’ UTR), both constructs being m1Ψ-modified and formulated in lipid nanoparticles. In summary, light emission recorded by in vivo imaging indicated similar functionality of the two constructs.
Additional in vivo studies involved vaccination comparing m1Ψ-modified minRNAs encoding either the model antigen ovalbumin (OVA) or SARS-CoV-2 spike and optRNAs encoding either OVA or SARS-CoV-2 spike (the eIF4G aptamer 5’ UTR and the mtRNR1-AES 3’ UTR of Comirnaty®, see Background). In transfected cells, the two RNAs coding for spike gave equivalent expression of the viral protein on the cell surface. In vivo, after three subcutaneous injections of the LNP-formulated RNAs, ELISA antibody responses against spike were statistically equivalent between the min and opt designs, and undetectable for the OVA negative control (Figure 4).
FIGURE 4. ELISA quantification of serum antibodies against SARS-CoV-2 spike following vaccination with LNPs containing either optRNA or min RNA constructs and corresponding OVA negative controls in mice injected subcutaneously. N = 5 mice per group. Data are presented as the mean ± SD; ns = not significant. Image from Mamaghani et al. and used under CC BY 4.0 license.
Concluding comments
The above results demonstrate that minRNAs with only a 4-nt TISU 5’ UTR (CAAG) and without a 3’ UTR are functionally equivalent to commonly used optRNAs having optimized 5’ and 3’ UTRs. This holds true for unmodified U or the modified nucleosides Ψ, m1Ψ, or moU.
Mamaghani et al. state that their data will have “a major impact on the future of mRNA-based vaccines and therapies, as [their] design makes it easier to produce the DNA template required for mRNA production.” They add that, when ordering synthetic genes, structured or repeated UTRs can cause problems and lead to delays and additional costs.
In addition, they say, the reduced structural complexity of minRNA reduces potential unexpected effects of cell type-dependent functionality of 5’ UTRs (Castillo-Hair et al.) or 3’ UTRs (Mayr). The minRNA design also mitigates the risk of an atypical start in the 5’ UTR (Trulley et al.) or an override of the stop codon leading to translation of the 3’ UTR (Mangkalaphiban et al.).
Mamaghani et al. conclude by suggesting that minRNA design will lead to easier, faster, and cheaper production, enabling the acceleration of RNA vectors for the protein-based therapies and vaccines of the future.
Your comments are welcomed, as usual.
Please feel free to share this blog with your colleagues or on social media.
