Trends and Commonalities of Approved and Late Clinical-Phase RNA Therapeutics

Abstract

Background/Objectives: After many years of research and the successful development of therapeutic products by a few industrial actors, the COVID-19 vaccines brought messenger RNAs, as well as other nucleic acid modalities, such as antisense oligonucleotides, small interfering RNA, and aptamers, into the spotlight, eliciting renewed interest from both academia and industry. However, owing to their structure, relative "fragility", and the (usually) intracellular site of action, the delivery of these therapeutics has frequently proven to be a key limitation, especially when considering endosomal escape, which still needs to be overcome. Methods: By compiling delivery-related data on approved and late clinical-phase ribonucleic acid therapeutics, this review aims to assess the delivery strategies that have proven to be successful or are emerging, as well as areas where more research is needed. Results: In very specific cases, some strategies appeared to be quite effective, such as the N-acetylgalactosamine moiety in the case of liver delivery. Surprisingly, it also appears that for some modalities, efforts in molecular design have led to more "drug-like" properties, enablingthe administration of naked nucleic acids, without any form of encapsulation. This appears to be especially true when local administration, i.e., by injection, is possible, as this provides de facto targeting and a high local concentration, which can compensate for the small proportion of nucleic acids that reach the cytoplasm. Conclusions: Nucleic acid-based therapeutics have come a long way in terms of their physicochemical properties. However, due to their inherent limitations, targeting appears to be crucial for their efficacy, even more so than for traditional pharmaceutical modalities.

Keywords: delivery; endosomal escape; local administration; nucleic acids; targeting; topical.

Figure 1

Drugging the RNA genome. Reproduced with permission from Warner et al., 2018 [5]. This figure illustrates the total human genome as composed of non-coding regions in grey, regions coding for non-coding RNAs in dark blue (estimated at 70% of the genome), and of genes coding for RNAs, which in turn code for proteins (estimated to be 1.5% of the genome). Of these protein-coding genes, the authors estimate that 15% are associated with diseases. In addition, it is estimated that the proteins that are currently “drugged” by an approved product correspond to 0.05% of the total genome. Hence, considering targets in the RNA genome offers a wider array of therapeutic options than the more traditionally explored proteome.

Figure 2

Examples of different RNA modalities. Reproduced with permission from Bejar et al., 2022 [7]. The most common modalities of therapeutic nucleic acids allow different forms of control over proteins that may be disease-related, more so than using traditional therapies that act at the protein level. For example, aptamers (bottom left) can bind to receptors in a “classical manner” to act as antagonists, while siRNAs (top right) and ASOs (top left) can lead to mRNA degradation and thus inhibit activity of the protein that would have been translated. ASOs can even act in the nucleus to provide alternative splicing of mRNAs that lead to modified proteins. Finally, mRNAs can increase the amounts of a given protein that are produced, leading to increased activity.

Figure 3

siRNA and miRNA pathways. Reproduced with permission from Hu et al., 2020 [32]. A: The figure shows the miRNA pathway, in which the complementarity of the RNA strand and its targeted mRNA (blue) is not complete. It interacts with a cytoplasmic protein complex known as RISC and the mRNA-RISC interaction can lead to the modulation of protein expression (indicated by X) and mRNA degradation. B: siRNAs (red) are duplex nucleic acids, of which one strand can interact with the RISC complex. The RISC complexed single strand can then pair with a target mRNA leading to its degradation and thus inhibiting translation (red cross). This decreases the amount of protein produced and, therefore, the activity of said protein.

Figure 4

Main ASO activity pathways and example of corresponding chemical modifications. Reproduced with permission from Tarn et al., 2021 [34]. In this figure, the authors show how the structure of an ASO can lead to a different mechanism of action. Indeed, for RNase H to cleave an mRNA, it needs to recognize a DNA/RNA duplex. Therefore, by selectively modifying the 2′ position of nucleotides at the extremities of an ASO, leaving a DNA “gap” in the center, cleavage by RNase H is possible. However, if all nucleotides of an ASO have their 2′ position modified, the DNA/RNA duplex cannot be recognized by RNase H. In this case, the mRNA is not degraded, but the ASO can have an activity by preventing other cellular factors from interacting with the mRNA by steric hindrance.

Figure 5

Comparison between ASGPR and other receptors mediated cellular uptake. Reproduced with permission from Dowdy, 2017 [45]. The ASGPR receptor is a very singular example of transmembrane receptor mediated uptake. Although the main barrier to cytosol activity remains the endosomal escape, binding to this receptor leads to increased cytosol concentration thanks to two main characteristics of the receptor. First, ASGPR is particularly densely expressed at the surface of hepatocytes, up to a hundred times more than other receptors, which means that more targeted siRNA can bind to the same number of hepatocytes compared to non-ASGPR expressing cells. Furthermore, the turnover of this receptor is higher than most, with a cycle estimated to be six times shorter than more common receptors. This means that each time a receptor is bound to a targeted siRNA leading to endocytosis in a hepatocyte, the delay before this receptor is available at the cell surface again to bind a new siRNA is shorter than other receptors. Combined, these two properties lead to more siRNAs reaching the endosomal escape bottleneck. Therefore, if the proportion of siRNAs that undergo endosomal escape remains the same, the number of siRNAs that go through the bottleneck is higher.

Figure 6

Example of siRNA and shRNA chemical modification platforms. Reproduced with permission from Hu et al., 2020 [32]. In response to the main hurdles faced in nucleic acid therapeutics, one key strategy has been to chemically modify the nucleic acid chains. These modifications are used to enhance stability towards nuclease but also modulate immunogenicity and increase cellular uptake. Most actors in this field have developed their own platforms, using different types of modifications, in different patterns. Interestingly, siRNAs developed by a single company, based on the same platform, show differences in design. This shows that the strategy used in terms of chemical modifications must be adapted on a case-by-case basis to lead to the best compromise between stability, immunogenicity, uptake, and activity.