RNA-based therapies in liver metabolic diseases

Abstract

RNA-based therapeutics have rapidly emerged over the past decade, offering a new class of medicines that differ significantly from conventional drugs. These therapies can be programmed to target or restore defective genes, allowing for more personalised treatments and reducing side effects. Notably, RNA therapies have made significant progress in the treatment of genetic liver diseases, exemplified by small interfering RNA treatments for hereditary transthyretin amyloidosis, which use liver-targeting strategies such as GalNAc conjugation to improve efficacy and safety. RNA-based gene-editing technologies, such as base editor and prime editor clustered regularly interspaced short palindromic repeats systems, also show promise with their ability to minimise genomic rearrangements and cancer risk. While RNA therapies offer high precision, challenges remain in optimising delivery methods and ensuring long-term safety and efficacy. Lipid nanoparticle-mRNA therapeutics, particularly for protein replacement in rare diseases, have gained support from preclinical successes. Compared with viral gene therapies, mRNA therapies present a safer profile with reduced risks of genomic integration and oncogene activation. However, clinical trials, especially for rare diseases, face limitations such as small sample sizes and short observation periods. Further preclinical studies, including non-human primates, will be essential for refining trial designs. Despite their potential, the high costs of RNA therapies pose a challenge that will require cost–utility models to guide pricing and accessibility. Here, we discuss the fundamental aspects of RNA-based therapeutics and showcase the most relevant preclinical and clinical developments in genetic liver metabolic diseases.

Key messages.

• RNA-based drugs are emerging as one of the most promising therapeutic strategies, able to reach molecular targets that, in principle, are undruggable. Their broad therapeutic activities range from the silencing of harmful genes via small interfering RNAs, antisense oligonucleotides and clustered regularly interspaced short palindromic repeats (CRISPR)-Cas mediated DNA editing, to the replacement of deficient proteins via mRNA therapeutics.

• The design and production of RNA-based drugs can be quite straightforward, are flexible and have great potential for streamlined manufacturing automation.

• Important issues in RNA therapies include organ targeting, the preservation of RNA integrity and the avoidance of immune responses. The liver is a privileged organ for RNA-based therapeutics, as the two most successful drug delivery strategies, that is, encapsulation in lipid nanoparticles and conjugation with N-acetylgalactosamine (GalNAc), have strong hepatocellular tropism.

• Among RNA therapeutics, one may find some of the most successful examples of personalised therapies. Prominent methods include RNA interference-based approaches to treat genetic metabolic diseases.

• CRISPR technology holds significant potential for treating metabolic diseases by enabling precise gene editing to correct genetic mutations, thereby addressing the root cause of these disorders. Base and prime editors, advanced CRISPR-based technologies, enable precise DNA edits without double-strand breaks.

• mRNA-based therapies have the potential to transform the treatment of severe genetic hepatic metabolic diseases with systemic implications. Interim results from a phase 1/2 mRNA clinical trial for propionic acidaemia and a novel, clinically relevant non-human primate model for acute intermittent porphyria confirm safety and translatability of repeated systemic mRNA administration without triggering immune responses.

Introduction

The exceptionally fast development of effective mRNA vaccines to tackle the COVID-19 pandemic did not happen by chance. Decades of fundamental research and countless discoveries on the structure and biological roles of RNA paved the way for the use of RNA-based molecules to modulate biological pathways and treat or prevent disease.1,3 In addition to coding for proteins, RNA molecules play a broad range of dynamic functions including catalytic, scaffolding and gene expression regulatory activities. These functions are mediated to a great extent through the ability of RNA to bind other RNA and DNA molecules with high affinity and specificity via base pairing. Harnessing the biological properties of RNA can lead to the generation of new classes of drugs endowed with unique properties in comparison with conventional therapeutic agents. These advantages include the capacity of RNA-based drugs to engage with targets that in principle are undruggable such as non-coding RNAs involved in the pathogenesis of numerous diseases, or with messenger RNAs (mRNAs) encoding proteins that cannot be effectively targeted by small molecules or antibodies.¹ This versatility is complemented by the capacity of mRNAs to drive protein synthesis when appropriately delivered, thus enabling not only the development of vaccines but also promising applications in protein replacement therapies.^{4 5} From a conceptual perspective, and according to the different mechanisms of action, RNA therapies can be divided into three main categories: (1) compounds that interfere with cellular RNAs; (2) RNAs used as guides for therapeutic genome editing and (3) mRNAs that mediate protein expression5,8 (figure 1).

Figure 1. Current RNA therapeutics for metabolic monogenic disorders. ASOs and siRNAs are complementary oligonucleotides that target specific mRNAs through Watson-Crick base pairing, while exogenous mRNAs are delivered to cells to produce the desired proteins. CRISPR-Cas9, a gene-editing tool, disrupts disease-causing genes at specific locations via an sgRNA guide with a sequence that is complementary to the target gene. CRISPR leverages the Cas9 protein as ‘molecular scissors’ to cut DNA, activating a cellular repair process that can either disrupt the faulty gene, silence it, or permanently correct deleterious mutations to restore proper protein expression. There are multiple Cas proteins with distinct functions. While Cas9 remains the most widely used for genome editing, Cas12, Cas13, and Cas14 expand CRISPR’s applications to RNA editing, disease detection, and more precise DNA modifications, respectively. ASO, antisense oligonucleotide; CRISPR-Cas9, clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein 9; mRNA, messenger RNA; sgRNA, single guide RNA; siRNA, short interfering RNA.

Compared with small molecules and other biologicals such as recombinant proteins and antibodies, the design and production of RNA-based drugs, once the nucleotide sequences have been established, can be quite straightforward. The physicochemical characteristics of RNAs are well known, and their manufacture relies on the same base materials; therefore, production can be standardised in versatile platforms that are constantly upgraded for simpler and more agile manufacturing.^{3 9} To these advantages of RNA therapies, we can add their virtual lack of genotoxicity, a potential risk of DNA therapies which use viral vectors that can integrate in the genome.¹ However, despite of the great potential of RNA therapeutics, there are obstacles that need to be faced. On one hand is their anionic charge and poor stability, particularly that of large mRNAs susceptible to RNase-mediated degradation in blood and tissues, which compromises their cellular permeability and bioavailability. Another limitation is RNA immunogenicity, which is especially challenging when repeated administrations are needed in chronic treatments. These drawbacks are actively being addressed by different strategies, including the development of packaging systems using lipid nanoparticles (LNPs) that protect the RNA cargos from degradation while enabling intracellular delivery and the modification of RNA chemistry to reduce immunogenicity and enhance stability.^{5 10 11}

One unresolved theme in RNA drug development is the control for cell specific delivery, particularly when targeting solid organs. Interestingly, in this context, the liver is a privileged tissue, being a highly perfused organ endowed with a fenestrated vascular endothelium. Indeed, the majority of LNP formulations are avidly absorbed by hepatocytes,⁵ and the most successful strategy for targeting small RNAs, that is, conjugation with N-acetylgalactosamine (GalNAc), results in a preferential liver tropism.¹⁰ Here, we first provide an overview of the molecular design, delivery strategies and mechanisms of action of the three most prominent RNA therapies mentioned above (figure 1). Aptamers and micro-RNA-based therapies fall outside the scope of this text, and readers are referred to recent reviews.12,15 Finally, we also summarise preclinical and clinical studies that tested the safety and efficacy of these new classes of drugs in liver metabolic diseases.

The RNA-based therapeutic toolbox

RNA targeting therapies

Antisense oligonucleotides (ASOs) were the first RNA-targeting strategy devised almost 50 years ago.² These are single-stranded synthetic oligonucleotides of approximately 6–9 kDa (~18–30 nucleotides) with extensive chemical modifications to improve their stability and resistance to nucleases.⁶ ASOs can alter mRNA expression through three different mechanisms of action: (1) binding mRNAs through base pairing and triggering RNase H-mediated degradation of double-stranded DNA-RNA heteroduplexes; (2) the interference with the splicing machinery interacting with pre-mRNAs to modulate protein expression and (3) the impairment of mRNA translation.^{1 2 6} The stability, specificity and potency of ASOs have been improved over time with innovative chemistries.¹⁶ Current RNase-H dependent ASOs include the so-called ‘gapmers’, characterised by a chimeric structure in which a DNA core of 10 nucleotides is flanked by 5’ and 3’ RNA-like residues. This structure increases nuclease resistance and enhances target hybridisation while preserving RNase H activation.^{10 17}

Small interfering RNAs (siRNAs) are double-stranded RNAs of approximately 13 kDa (a duplex of 21 nucleotides with 19 complementary bases and terminal 2 nucleotide 3’ overhangs).⁶ The guide or antisense strand is complementary to the target mRNA, whereas the other strand is referred to as the sense or passenger strand. siRNAs use the endogenous RNAi pathway, in which the guide strand forms a complex with the argonaut 2 (AGO2) protein to generate an RNA-induced silencing complex (RISC). The catalytic AGO2 protein within the RISC complex mediates the cleavage of the target mRNA, reducing the expression of protein-coding genes^{17 18} (figure 2A).

Figure 2. Strategies for targeting siRNA and mRNA therapies to hepatocytes. (A) Delivery and mechanism of targeted silencing by siRNAs. Targeted delivery of siRNA to hepatocytes is achieved through chemical conjugation with a trimer of GalNAc, which avidly binds to the ASGPR that is predominantly expressed on liver hepatocytes. On cellular entry, the siRNA is processed by DICER, which discards the passenger strand and incorporates the guide strand into the RISC complex. The guide strand then directs the RISC to complementary mRNA sequences, where AGO2, an endonuclease within the RISC, cleaves the target mRNA, effectively silencing the target gene and preventing protein synthesis. (B) Delivery and mechanism of action of therapeutic mRNAs. LNPs are currently the most effective nanocarrier system for liver-targeted therapies following systemic administration. LNPs include amino ionisable cationic lipids, cholesterol, polyethylene glycol-lipid conjugates, helper phospholipids (phosphatidylcholines) and active mRNA molecules. Owing to the negative charge of mRNA, encapsulation is often achieved via electrostatic interactions with cationic lipids. An exchange of proteins from other lipid particles results in ApoE being present in LNPs, which facilitates hepatic LNPs uptake via LDLRs, highly abundant on the surface of hepatocytes. Additionally, delivery systems that incorporate GalNAc-lipid modified LNPs allow for LDLR-independent hepatocyte targeting via the ASGPR. On cellular uptake, LNPs become protonated in acidic environments, facilitating endolysosomal escape and mRNA release into the cytosol to be translated into the therapeutic protein. Proteins may remain within hepatocytes at their specific subcellular locations or may be secreted into the circulation to reach other target organs. ApoE, Apolipoprotein E; AGO2, argonaut 2; ASGPR, asialoglycoprotein receptor; GalNAc, N-acetylgalactosamine; LDLR, low-density lipoprotein receptor; LNPs, lipid nanoparticles; RISC, RNA-induced silencing complex; siRNAs: small interfering RNAs.

The immunogenicity of ASOs and siRNA therapeutics is an issue that needs to be considered. Both single and double-stranded RNA can be detected by pattern recognition receptors in the extracellular compartment (Toll-like receptors 3, 7 and 8 in endosomes) and by cytoplasmic receptors such as protein kinase R (PKR), RIG-I-like receptors and oligoadenylate synthases, triggering inflammatory reactions and RNA degradation.^{19 20} Chemical modifications at the 2’ position of the sugar moiety have been introduced in both ASOs and siRNAs to increase binding affinity while lowering off-target effects and immunogenicity.^{16 18}

Effective delivery of oligonucleotide therapeutics to target tissues is a major challenge, as ASOs, and particularly siRNAs, are relatively large negatively charged molecules. Various strategies, including delivery vehicles and targeting ligands, have been developed to improve their stability in circulation and their uptake by target cells while reducing kidney clearance.^{10 17 18} LNPs are among the most successful delivery vehicles for ASOs, siRNAs and, as will be discussed below, also for gene-editing and mRNA-based therapeutics (figure 2). LNPs include four different components: amino ionisable cationic lipids, cholesterol, polyethylene glycol (PEG)-lipid-conjugates and helper lipids (phosphatidylcholines) (figure 2B). Encapsulation in LNPs protects RNAs from nucleases in the circulation and in endosomes; moreover, LNPs harbouring ionisable cationic lipids are neutrally charged at physiological pH, which decreases their immunogenicity in the blood.¹⁶ In addition, the inclusion of PEG-lipid-conjugates, positioned at the external interface, further contributes to reducing LNPs immunogenicity and their uptake by macrophages.^{5 21} Interestingly, LNPs also associate with apolipoprotein E (APOE) in the circulation, and this interaction facilitates low-density lipoprotein receptor (LDLR)-mediated endocytosis by hepatocytes.^{10 17} On internalisation, once in the endosome, ionisable lipids become positively charged due to the low pH, which promotes the fusion of LNPs with endosomal membranes and the release of the RNA cargo into the cytoplasm.¹⁶

The most validated strategy for ASO-targeted and siRNA-targeted delivery is the conjugation with GalNAc. GalNAc is a carbohydrate moiety trivalent ligand that binds with high affinity to the asialoglycoprotein receptor (ASGPR) (figure 2A). Importantly, the ASGPR is a highly expressed and fast-recycling receptor expressed on the surface of hepatocytes, therefore, GalNAc-conjugated ASOs or siRNAs (which are GalNAc-conjugated in their passenger strand) are efficiently taken up to be released in the cytosol after endosomal trafficking.^{10 17 18 22} However, it has been estimated that only 1%–2% of GalNac-conjugated ASOs, and about 0.3% of GalNAc-conjugated siRNAs, escape from endosomes in vivo.^{23 24} This is also a very slow process that markedly influences the pharmacodynamic responses to these RNA therapies.^{23 24} The mechanisms that determine endosomal escape are not well understood, and several hypotheses have been proposed. These include the spontaneous formation of short-lived breaches in the endosomal lipid bilayer, the temporary disruption of the membranes during fusion events between endosomes or with lysosomes and RNA retro-transport from endosomal compartments to the Golgi, from where breaching across the lipid bilayer may be easier.²³ Endosomal escape is, therefore, a rate-limiting step for the activity of GalNAc-conjugated RNA therapeutics and active research is conducted to improve it while avoiding cellular toxicity.^{23 25}

Notably, GalNAc siRNA conjugates can be administered subcutaneously and still reach the liver very efficiently, which differs from LNP formulations that, for hepatic delivery, need to be given intravenously.^{5 17} Interestingly, strategies combining both vehicles, GalNac and LNPs, have been successfully developed to maximise the hepatic targeting of RNA-based medicines26,28 (figure 2). Nevertheless, despite advances in delivery, there is still room for improvement in RNA targeting therapies. Innovations in chemical modifications, new formulations and the discovery of ligands beyond GalNAc are expected to enhance the targeted delivery of RNAi therapies. While advances in delivery systems can help reduce immune-associated toxicities, they may not suffice on their own. Incorporating novel RNA modifications to mitigate immune activation is also essential.²⁹

Genome editing tools

Genome editing is an emerging therapeutic strategy aimed at correcting pathogenic mutations in genomic DNA. This approach is based on clustered regularly interspaced short palindromic repeats (CRISPR)/Cas nucleases, a component of the bacterial immune system that fights viral infections by cutting foreign DNA.³⁰ CRISPR/Cas nucleases can recognise their targets via a single-guide RNA (sgRNA) molecule loaded in the Cas protein. This sgRNA can be designed to guide Cas nucleases to specific regions in the genome, where they generate double-strand breaks (DSBs)³¹ (figure 3). This is a critical step, as off-target effects can result in unintended genome alterations. Improved predictive algorithms for selecting optimal target sequences are being elucidated, although complete accuracy cannot yet be guaranteed.³² DSBs are subsequently repaired by either homology-directed repair, or, more frequently, non-homologous end joining (NHEJ) or microhomology-mediated end joining.⁷ However, these repair systems pose important challenges for Cas nuclease-based therapies, as they can give rise to small insertions and deletions, particularly those in the NHEJ pathway,³³ and DSBs can also lead to large deletions and chromosomal translocations, among other undesired effects.⁷ These are important limitations that have led to the development of more precise genome editing strategies, such as the use of CRISPR base editors (BE), which enable the correction of single-nucleotide mutations without inducing DSBs or requiring a repair template³⁴ (figure 3). Although BE may correct most of the pathogenic mutations, they are restricted to cytidine and adenosine conversions and cannot mediate targeted deletions or insertions. Another technique known as prime editors (PEs) has recently emerged to address these limitations.³⁵ PE enables any single-nucleotide conversion and small insertions and deletions without producing DSBs (figure 3). Both BE and PE require guide RNAs; in the case of PE, it is known as prime editing guide RNA (pegRNA), which, in addition to directing the enzyme to the target locus, serves as a template for the edit of interest. Detailed descriptions of the molecular mechanisms of action of BE and PE can be found in recent dedicated reviews.^{36 37}

Figure 3. Gene editing strategies using CRISPR-Cas systems, BEs and PEs. CRISPR editing with Cas9 guided by a sgRNA induces double-strand breaks at target DNA sites. DNA repair then occurs through the HDR or NHEJ pathways, enabling therapeutic applications. BE and PE represent advanced CRISPR-based technologies that enable precise DNA edits without inducing double-strand breaks. BEs facilitate single-base conversions; for example, adenine deaminase in ABE converts adenine to inosine, which is recognised as guanine during DNA replication, whereas cytidine deaminase results in a C-to-T base conversion via CBE. PEs enable a broader range of edits, including base substitutions, insertions and deletions, although their larger, more complex structure can present delivery challenges in certain cellular contexts. ABE, adenine base editor; BE, base editors; Cas9, CRISPR-associated protein 9 nuclease; CBE, cytosine base editor; HDR, homology-directed repair; NHEJ: non-homologous DNA end joining; PE: prime editors; sgRNA: single-guide RNA.

The components of the gene editing machinery can be delivered into cells as DNAs or mRNAs encoding their expression or as ribonucleoproteins (RNPs). For the delivery of CRISPR/Cas and sgRNAs in DNA from, viral vectors such as adenoviruses and adeno-associated viruses (AAVs) have been used. While liver-tissue tropism can be optimised with certain AAV serotypes, continuous expression of gene editors increases the risk of off-target editing, and as observed in gene therapy studies, there are concerns related to dose-limiting toxicity and insertional mutagenesis.⁷ Adenoviruses are devoid of genome integration-related risks but are known to trigger a potent immune response including T-cell-mediated cytotoxicity to hepatocytes expressing the gene editing machinery.⁷ The LNPs described above are also used for efficient delivery of Cas mRNA and sgRNAs to the liver, and as occurs with mRNAs, the sgRNAs also need to be chemically modified to increase their intracellular stability.^{7 38 39} Virus-like particles (VLPs), which are formed from viral protein assemblies devoid of infective capacity, are emerging as a promising strategy to deliver mRNA or RNP cargos,⁷ and engineered versions of these VLPs have shown high efficiency for liver targeting.⁴⁰ One important advantage of mRNA-based and RNP-based delivery strategies is the transient expression of the editing machinery, which can suffice to edit the genome permanently with a lower risk of off-target effects and without genome integration.^{2 7 38 39}

mRNA-based therapies

One of the main functions of RNA in nature is to serve as an intermediary for transmitting the genetic information contained in DNA to proteins, which are the primary end effectors. The RNA molecules that convey genetic information are known as mRNAs. In most cases, these mRNAs are linear molecules. To produce therapeutic proteins in the target tissue in vitro-generated mRNA sequences encoding the therapeutic protein can be used. In addition to the coding region, the mRNA molecule is composed of the following structures: (1) a 5' cap, (2) 5' and 3' untranslated regions (UTR) flanking the coding sequence and (3) a 3' polyA tail (figure 4). To develop mRNA medicines, the mRNA molecules are generated through in vitro transcription. A DNA sequence complementary to the desired mRNA sequence is cloned and inserted into a plasmid containing a T7 RNA polymerase promoter. The plasmid can be amplified and produced in high quantities, then linearised with a restriction enzyme and incubated with T7 RNA polymerase and nucleotides. The final product must be purified to eliminate fragmented and double-stranded mRNA impurities.⁴¹ However, it should be mentioned that despite significant advances in the mRNA manufacturing process, its scalability to produce the required doses for long-term treatments is still a challenge.⁴¹

Figure 4. Structure of optimised mRNAs for therapeutic applications. (A) Structure of therapeutic mRNAs, including the five functional elements: 5′ cap, 5′ UTR, ORF, 3′ UTR and poly(A) tail. The 5′ cap protects mRNA from exonucleases and enables ribosome recognition. Optimised cap structures enhance translation and also reduce immune activation by recognition via RLRs. In the ORF, chemical modifications of nucleobases such as pseudouridines (Ψ), N1-methyl-pseudouridine (N1Ψ) or methylated nucleobases such as m1A, m6A, and m5C, and GC-rich sequences can be introduced to reduce immunogenicity, prevent unwanted secondary structures and increase protein expression. The 5’ UTR can be modified to increase translation by placing Kozak sequences from genes such as globin or Hsp70, and the modification of the 3’ UTR by including synthetic sequences from the albumin or Hsp70 genes can improve mRNA stability. UTRs can also be customised to enhance tissue-specific translation. The length of the poly(A) tail is also optimised to improve the synthetic capability of a given mRNA. (B) saRNAs use the self-replication basis of RNA alphaviruses to amplify RNA transcripts in the cytosol. They include two ORFs, one for the viral nsp proteins (1-4), which enable RNA replication and amplification within the cell increasing protein expression from small doses, and the ORF of the protein of interest. Another alternative to mRNAs is circRNA, which is devoid of a 5’ cap and a 3’ poly(A) tail, has a self-splicing intron (group I intron) to promote circularisation and an engineered IRES to drive protein synthesis. Compared with linear mRNAs, circRNAs exhibit reduced immunogenicity and enhanced stability leading to prolonged duration of translational activity compared to linear mRNAs. Ψ, pseudouridine; circRNA, circular mRNA; IRES, internal ribosome entry site; m5C, 5-methylcytosine; m6A, N6-methyladenosine; N1Ψ, N1-methyl-pseudouridine; NSP, non-structural protein; ORF, open reading frame; RLRs, retinoic acid-inducible gene I (RIG-I)-like receptors; saRNA, self-amplifying mRNA; 3' UTR, 3' untranslated region.

To generate mRNA medicines from a labile molecule such as mRNA, each stage of this process and each part of the molecule must be carefully optimised to create a product that is as stable as possible, with the greatest capacity to produce the therapeutic protein and the least activation of the immune system.^{4 11} An essential optimisation is the substitution of some of the nucleotides by modified nucleotides such as pseudouridine (Ψ) or 1-methyl-pseudouridine (m1Ψ) to prevent the recognition of these RNA molecules by the innate immune system through the Toll-like receptors TLR7/8⁵ (figure 4A). Importantly, it was recently noted that the presence of m1Ψ modifications can affect mRNA translation fidelity, resulting in the generation of protein products that differ in the amino acid sequence.⁴² This situation may compromise the protein product’s performance or trigger toxic or immune reactions and should be taken into account.

To improve the protein expression, mRNA-encoded viral-derived replicases can be incorporated into the mRNAs to generate self-amplifying mRNAs (saRNAs). Once produced in the target cell, the viral replicase can generate multiple copies of the mRNA encoding the therapeutic protein (figure 4B). Therefore, higher protein yield and longer expression can be achieved.⁴³ An alternative RNA topology to generate protein-coding RNA is circular RNA (circRNA). These molecules are devoid of free ends accessible to exonucleases and, therefore, are resistant to several mechanisms of RNA degradation. The production of circRNAs can be achieved by different approaches, including ribozymatic methods involving self-splicing introns,⁴⁴ and the resulting molecule is characterised by long-term stability in the target cells (figure 4B). To initiate translation, a cap-independent mechanism must be used such as internal ribosome entry sites (IRESs). However, cap-independent translation generally has low efficiency,⁴⁵ and IRES sequences need to be engineered and/or chemically modified (inclusion of m6A) to enhance their activity.¹¹

The successful delivery of mRNA to the cytoplasm of target cells for ribosome binding and protein translation necessitates an ideal vector that meets several crucial criteria, particularly for enzyme replacement therapies that need chronic dosing. First and foremost, it must demonstrate safety at doses surpassing those required for efficacy. Additionally, the vector should consistently maintain its activity across various batches, be redosable without compromising efficacy or safety, remain imperceptible to the immune system and be biodegradable. Scalability in manufacturing is also essential,⁴¹ along with an acceptable profile of on-target delivery versus off-target delivery.¹⁰ Interestingly, off-target delivery of mRNA cargo can be mitigated by the inclusion of tissue-specific miRNA binding sites on the UTRs.⁴⁶ Nevertheless, these properties have been attained to a significant extent with the development of LNPs, particularly for liver-directed therapies, as mentioned above, and are being constantly improved through chemical engineering to increase mRNA delivery.^{47 48} Once mRNAs are released into the cytoplasm and translated, the therapeutic proteins will undergo their physiological post-translational modifications and reach their natural location according to their protein localisation signals^{49 50} (figure 2B).

In addition to LNPs, which have emerged as the most widely used carriers for mRNA delivery, alternative delivery systems warrant consideration due to their potential to offer unique advantages.⁵ Cationic polymers, for instance, provide flexibility in structural design and can facilitate endosomal escape, a critical step in mRNA delivery.2351,54 DNA-based nanostructures present a programmable and biocompatible platform, allowing precise control over size and cargo loading.^{55 56} Similarly, exosomes, as naturally occurring vesicles, offer inherent biocompatibility and the ability to traverse biological barriers effectively.^{57 58} Exploring these alternative delivery systems alongside LNPs may expand the toolbox for mRNA medicine development. Nonetheless, irrespective of the delivery system, improving the thermostability and physical stability of mRNA formulations is one critical aspect that needs to be improved for the translation of mRNA therapy to the market.⁴¹

RNA-based therapies: the trip from the toolbox to the clinic

RNA targeting therapies

ASOs have shown very robust therapeutic effects on liver diseases in preclinical as well as clinical applications. Fomivirsen was the first ASO approved by the Food and Drug Administration (FDA) back in 1998 for the treatment of cytomegalovirus retinitis.^{2 6} However, since then, the ASO platform that has reached the clinical arena with the most promising outcomes in the last few years has been the gapmers developed by Ionis Pharmaceuticals, with up to 26 clinical trials ongoing.¹⁶ There are four FDA-approved drugs based on this technology for diseases originating from proteins produced in the liver: mipomersen, volanesorsen, inotersen and eplontersen. Mipomersen^{59 60} suppresses ApoB-100 expression to treat homozygous familial hypercholesterolaemia, and volanesorsen⁶¹ targets apolipoprotein C-III mRNA for degradation reducing triglyceride levels in patients with familial chylomicronaemia syndrome. On the other hand, inotersen is an ASO that targets the mutant (but also the wild type) transthyretin (TTR) mRNA which is responsible for hereditary transthyretin-mediated amyloidosis (hATTR disease).⁶² TTR is mostly produced in the liver, and the secretion and accumulation of mutant TTR in different tissues is responsible for most of the systemic pathological manifestations of hATTR. Inhibition of liver-produced mutant TTR by inotersen improved neurological function and patients’ quality of life; however, due to significant side effects its availability has been restricted.⁶² Eplontersen is another ASO that also targets TTR mRNA and was recently approved by the FDA; it has shown promising clinical efficacy. Eplontersen is GalNAc-conjugated, and therefore, its enhanced hepatocellular uptake permits a lower dose and frequency of administration compared with inotersen^{60 62} (table 1).

Table 1. FDA-approved ASO and siRNA therapeutics in clinical studies for treating metabolic disorders.

Disease	Drug	Type of RNA (formulation)	Target gene	Delivery	Company	Clinical trial	Regulatory approval
Transthyretin amyloidosis	Patisiran (Onpattro) Vutrisiran (Amvuttra) Inotersen (Tegsedi) Eplontersen (Wainua)	siRNA (LNPs) siRNA (GalNAc) ASO ASO (GalNAc)	TTR TTR TTR TTR	Intravenous Subcutaneous Subcutaneous Subcutaneous	Alnylam Pharmaceuticals Alnylam Pharmaceuticals Ionis Pharmaceuticals Ionis Pharmaceuticals	Active: Phase 3 for cardiomiopathy ATTR (NCT03997383) Completed: Phase 3 for polyneuropathy ATTR (NCT01960348) Active: Phase 3 for polyneuropathy (NCT03759379) and cardiomyopathy (NCT04153149) ATTR Completed: Phase 3 for polyneuropathy (NCT04136184) Active: Phase 3 for cardiomiopathy ATTR (NCT04136171) Completed: Phase three for polyneuropathy ATTR (NCT04136184)	FDA and EMA 2018 approval (for polyneuropathy) FDA and EMA 2022 approval (for polyneuropathy) FDA and EMA 2018 approval (for polyneuropathy) FDA 2023 approval (for polyneuropathy)
Primary hyperoxaluria type 1	Lumasiran (Oxlumo) Nedosiran (Rivfloza)	siRNA (GalNAc) siRNA (GalNAc)	GO LDHA	Subcutaneous Subcutaneous	Alnylam Pharmaceuticals Dicerna Pharmaceuticals	Completed: Phase 3 (NCT03681184) Active: Phase 3 (NCT04042402) and Phase 1 for PH type 3 (NCT04555486)	FDA and EMA 2020 approval FDA and EMA 2023 approval
Acute hepatic porphyrias	Givosiran (Givlaari)	siRNA (GalNAc)	ALAS1	Subcutaneous	Alnylam Pharmaceuticals	Completed: Phase 3 (NCT03338816)	FDA 2019 and 87 EMA 2020 approval
Familial chylomicronaemia syndrome	Volanesorsen (Waylivra)	ASO (GalNAc)	APOC3	Subcutaneous	Ionis Pharmaceuticals/Akcea Therapeutics	Completed: Phase 3 (NCT02658175)	EMA 2019 approval
Hypercholesterolaemia	Inclisiran (Leqvio)	siRNA (GalNAc)	PCSK9	Subcutaneous	Novartis Pharmaceuticals	Completed: Phase 3 (NCT04929249)	EMA 2020 and FDA 2021 approval
Familial hypercholesterolaemia	Mipomersen (Kynamro)	ASO	APOB	Subcutaneous	Ionis Pharmaceuticals	Completed: Phase 3 (NCT01598948, NCT00694109, NCT00607373, NCT00794664, NCT00770146, NCT01475825, NCT00706849)	FDA 2013 approval
Alpha-1 antitrypsin deficiency	Fazirsiran (TAK-999/ARO-AAT)	siRNA (GalNAc)	Mutant SERPINA1	Subcutaneous	Takeda	Active: Phase 3 (NCT03945292)
Atherosclerotic cardiovascular disease	SGB-3403	siRNA (GalNAc)	PCSK9	Subcutaneous	Suzhou Sanegene Bio	Active: Phase 1 (NCT06239714)
Dyslipidaemia	AZD8233	ASO	PCSK9	Subcutaneous	AstraZeneca	Completed: Phase 2 (NCT04641299) Completed: Phase 2 (NCT04823611) Completed: Phase 2 (NCT04964557)
AA amyloidosis	nL-SAA1-01	ASO	SAA1	Subcutaneous	Mayo Clinic	Active: Phase 1 (NCT06397001)
Familial hypercholesterolaemia	Vupanorsen (IONIS ANGPTL3-LRx)	ASO (LICA)	ANGPTL3	Subcutaneous	Pfizer	Completed: Phase 1 (NCT04459767) Completed: Phase 2 (NCT04516291)

ALAS1, delta-aminolevulinate synthase 1; ANGPTL3, angiopoietin-like 3; APOC3, apolipoprotein C3; ASO, antisense oligonucleotides; ATTR, transthyretin amyloidosis; FDA, Food and Drug Administration; GO, glycolate oxidase; LDHA, lactate dehydrogenase A; LICA, LIgand Conjugated Antisense; SAA1, serum amyloid A1; TTR, transthyretin.

Major advances in this field of siRNA therapies have been achieved, leading to the clinical success of some of these drugs (table 1). In fact, liver disease has been the spearhead of siRNA therapies with the approval of patisiran, givosiran, lumasiran, nedosiran, inclisiran and vutrisiran led mainly by Alnylam Pharmaceuticals. All these drugs, with the exception of patisiran, are targeted to the liver via GalNAc conjugation to interact with the ASGRP on hepatocytes.⁶³ Patisiran is formulated in LNPs and was the first FDA-approved RNAi drug⁶⁴ and, like vutrisiran,⁶⁵ targets TTR mRNA to treat hATTR. Patisiran, vutrisiran and fazirsiran have demonstrated biochemical and clinical efficacy in human studies in patients with hATTR.^{62 66} Givosiran has been approved for the treatment of acute intermittent porphyria (AIP) and targets the hepatic expression of 5’-aminolevulinate synthase 1 (ALAS1), the enzyme that catalyses the rate-limiting step in the heme biosynthetic pathway,⁶⁷ reducing the accumulation of the neurotoxic heme intermediates delta-aminolevulinic acid (ALA) and porphobilinogen.⁶⁸ Lumasiran aims to block hepatic oxalate production by targeting the expression of hydroxyacid oxidase 1 (HAO1)⁶⁹ in patients with primary hyperoxaluria (PH), a group of three rare disorders of hepatic glyoxylate metabolism characterised by oxalate overproduction.⁷⁰ Nedosiran (Dicerna Pharmaceuticals), which targets lactate dehydrogenase A (LDHA), the last gene involved in oxalate production, is being tested in clinical trials for the three PHs with promising results.^{71 72} Inclisiran inhibits the expression of protein convertase subtilisin/kexin type 9 (PCSK9) to increase the levels of LDLR in hepatocytes and reduce LDL-cholesterol levels in patients at high risk of atherosclerotic cardiovascular disease.⁷³ This drug, which needs to be administered only twice a year, has shown extraordinary efficacy in reducing LDL-cholesterol in numerous clinical trials.⁷⁴

Genome editing tools

Liver-targeted CRISPR-Cas9 technology has also been experimentally tested as a potential therapy for a broad range of metabolic conditions.^{38 75 76} Interestingly, recent reports demonstrated the efficacy of a selected adenine BE mRNA/gRNA administered via LNPs in a humanised mouse model harbouring the most common pathogenic variants of phenylalanine hydroxylase, the cause of phenylketonuria.^{77 78} Clinical studies are also ongoing to test CRISPR-Cas9 technology in systemic diseases in which the liver plays a central role (table 2). A clinical trial sponsored by Intellia Therapeutics and Regeneron Pharmaceuticals explores the treatment of hATTR amyloidosis. NTLA-2001 consists of an LNP-encapsulated Cas9 encoding mRNA and a TTR-targeted sgRNA, and its administration efficiently reduces serum TTR levels with mild adverse effects.⁷⁹ More recently, VERVE-101, a CRISPR BE targeted to disrupt hepatic PCSK9 expression on LNP-mediated delivery, was shown to reduce LDL cholesterol levels in patients with heterozygous familial hypercholesterolaemia. However, cardiac and hepatic toxicities were observed, and the trial was halted.^{80 81}

Table 2. Clinical trials of CRISPR-Cas9-based genome-editing therapies for inherited metabolic disorders.

Disease	Inheritance	Classification	Phenotype (affected metabolism)	Subcellular location	Target gene to inactivate	Company	Drug name (delivery vehicle)	Clinical trial
Hereditary transthyretin amyloidosis (ATTR)	Autosomal dominant	Patients with polyneuropathy (ATTRv-PN) and patients with related cardiomyopathy (ATTR-CM)	Deposit of unsoluble protein fibrils in the extracellular matrix (Transthyretin)	Plasma (secreted protein)	TTR	Intellia therapeutics	NTLA-2001 (LNP, IV)	NCT04601051 (Phase III) 2020-002034-32 (EudraCT Number)
Hereditary Angioedema (HAE)	Autosomal dominant or recesive	HAE1 and HAE2 are disorders caused by mutation in the SERPING1 gene	Recurring and unpredictable inflammatory attacks in various organs and tissues of the body	Plasma (secreted protein)	KLKB1 (Kallikrein B1)	Intellia therapeutics	NTLA-2002	NCT05120830 (Phase II)
Heterozygous familial hypercholesterolaemia	Autosomal dominant	Type 2 familial dyslipidaemia	Hypercholesterolaemia (high LDL-c)	Cell surface and to the endosomes/lysosomes in the presence of LDLR	PCSK9	Verve therapeutics	VERVE-101 (LNP, IV)	NCT05398029 (Phase I)
Heterozygous familial hypercholesterolaemia or premature coronary artery disease	Autosomal dominant	Type 2 familial dyslipidaemia	Hypercholesterolaemia (high LDL-c)	Cell surface and to the endosomes/lysosomes in the presence of LDLR	PCSK9	Verve therapeutics	VERVE-102 (GalNAc-LNP, IV)	NCT06164730 (Phase I)
Homozygous familial hypercholesterolaemia	Autosomal recessive	Patients with refractory hyperlipidaemia	Hypercholesterolaemia (High LDL-c)	Plasma (secreted protein)	ANGPTL3	Verve therapeutics	VERVE-201 (LNP, IV)	NCT06451770 (Phase I)
Homozygous familial hypercholesterolaemia	Autosomal recessive	Cardiovascular disease	Hypercholesterolaemia (High LDL-c)	Plasma (secreted protein)	ANGPTL3	CRISPR Therapeutics AG	CTX310	Not yet available (IND)
Lipoprotein(a) deficiency	Autosomal dominant	Plasma lipoprotein assembly, remodelling, and clearance and cholesterol metabolism	Coronary artery anomaly (lipoprotein(a) quantitative trait locus)	Plasma (LDL-like particle)	LPA	CRISPR Therapeutics AG	CTX320	Not yet available (IND)
Essential hypertension	>200 genes associated	Refractory hypertension	Increased risks of cerebral, cardiac and renal complications and pre-eclampsia (angiotensinogen)	Plasma (secreted protein)	AGT	CRISPR Therapeutics AG	CTX340	Not yet available (IND-enabling)
Acute hepatic porphyria	Autosomal dominant	Acute hepatic porphyrias	Abdominal-psychoneuroological acute attacks (heme synthesis)	Hepatocyte mitochondria	ALAS1	CRISPR Therapeutics AG	CTX450	Not yet available (IND-enabling)

AGT, angiotensinogen; ALAS1, 5'-aminolevulinate synthase 1; ANGPTL3, angiopoietin like 3; CRISPR, clustered regularly interspaced short palindromic repeats; KLKB1, kallikrein B1; LDL-c, low-density lipoprotein cholesterol; LPA, lipoprotein(A); PCSK9, Proprotein Convertase Subtilisin/Kexin Type 9; TTR, transthyretin.

mRNA-based therapies

mRNA therapy is transforming modern medicine by enabling the expression of new therapeutic proteins, replacing defective ones or introducing novel functions. As previously mentioned, mRNA therapies offer significant advantages, including rapid design and production compared with protein-based therapies.⁸² Additionally, RNA displays a safer profile than DNA therapies, as it does not integrate into the genome. Over the past few years, a wealth of preclinical evidence has been provided on the safety and efficacy of liver-targeted LNP-mRNA therapies for rare metabolic diseases and their systemic complications4983,113 (table 3). Early studies were performed in mouse models of the paediatric ultrarare disease methylmalonic acidaemia (MMA), which is caused by deficits in the mitochondrial enzyme methylmalonyl-CoA mutase (MUT). Mice treated with LNP-hMUT mRNA presented clear metabolic benefits and improved survival with no signs of liver toxicity or immune or inflammatory reactions due to repeated LNP-hMUT mRNA administration.^{94 105} A new version of hMUT mRNA with improved protein expression and reduced immunogenicity further validated this therapeutic approach.¹⁰⁸ These pioneering works were followed by similar studies demonstrating the efficacy of LNP-mRNA therapies for ornithine transcarbamylase (OTC) deficiency, a urea cycle disorder associated with severe brain damage^{91 92} and for propionic acidaemia (PA), caused by defects in the mitochondrial enzyme propionyl-CoA carboxylase (PCC) alpha (PCCA) and/or beta (PCCB) subunits, leading to impaired amino acid and fatty acid metabolism, severe neurological symptoms and, eventually, multiorgan failure.⁸⁹ Interestingly, this PA study used LNPs loaded with mRNAs coding for both the PCCA and PCCB subunits, demonstrating long-term therapeutic potential and the feasibility of chronically administering multiple mRNAs to produce large enzyme complexes applicable to other genetic disorders.⁸⁹ The treatment of glycogen storage diseases (GSDs) has also been experimentally addressed by the administration of LNP-mRNA. This is the case for GSD1a, in which mutations in glucose 6-phosphatase lead to life-threatening metabolic alterations, liver and kidney injury, and eventually to the development of liver tumours. mRNA therapy corrected the severe metabolic alterations found in GSD1a models and reduced the risk of cancer development.^{85 86}

Table 3. Current status of the development of liver-targeted new mRNA therapeutic strategies for rare metabolic diseases.

Disease	Inheritance	Onset of symptoms	Affected metabolism (enzyme)	Subcellular location	Therapeutic mRNA (reference)	Company	Preclinical data	Clinical trial
Progressive familial intrahepatic cholestasis type 3	Autosomal recessive	All ages	Phospholipd (ABCB4)	Membrane	hABCB4 mRNA	Moderna	Wei et al 202164
Methylmalonic acidaemia	Autosomal recessive	Infancy	Propionate (MUT)	Mitochondria	hMUT mRNA (mRNA-3705)	Moderna	Coughlan et al 202467 and Baek et al 202470	Phase1/2 (NCT04899310) Active
Crigler‐Najjar syndrome type 1	Autosomal recessive	Infancy	Bilirubin (UGT1A1)	ER	hUGT1A1 mRNA hUGT1A1 mRNA (mRNA-3351)	Alexion Pharmaceuticals Moderna	Apgar et al 201868 Greig et al 202369
Phenylketonuria	Autosomal recessive	Newborn	Phenylalanine (PAH)	Cytoplasm	PAH mRNA (mRNA-3210)	Moderna	Baek et al 202470	Phase 1/2 (NCT06147856) Withdrawn
Acute intermittent porphyria	Autosomal dominant	15–40 years	Hemo (PBGD)	Cytoplasm	hPBGD mRNA	Moderna	Jiang et al 201839 and Córdoba et al 202491
Variegate porphyria	Autosomal dominant	15–40 years	Hemo (PPOX)	Cytoplasm	hPBGD mRNA	Moderna	Jericó et al 202192
Classic galactosaemia	Autosomal recessive	Infancy	Galactose (GALT)	Cytoplasm	hGALT mRNA	Moderna	Balakrishnan et al 202471
Glycogen storage disease type 1a	Autosomal recessive	Infancy	Glucose (G6Pase)	ER membrane	hG6Pase mRNA (mRNA-3745) hG6Pase mRNA	Moderna Alexion Pharmaceuticals	Cao et al 202175 Roseman et a 201874	Phase 1/2 (NCT05095727) Active
Glycogen storage disease type 3	Autosomal recessive	Infancy	Glycogen (AGL)	Cytoplasm	GDE mRNA (UX053)	Ultragenix/Arcturus Therapeutics		Phase 1/2 (NCT04990388) Terminated
Alpha-1 antitrypsin deficiency	Autosomal recessive	25–50 years	Proteins (AAT)	ER	hAAT mRNA	Alexion Pharmaceuticals Moderna	Connolly et al 201877 Karadagi et al 202076
Propionic acidaemia	Autosomal recessive	Infancy	Propionate/Urea cycle (PPC)	Mitochondria	PCCA/PCCB mRNA (mRNA-3927)	Moderna	Baek et al 202470 & Attarwala et al 202379	Phase 1/2 (NCT 04159103; NCT 05130437) Active
Ornithine transcarbamylase deficiency	X-linked	Infancy	Urea cycle (OTC)	Mitochondria	OTC mRNA	Translate Bio (MRT5201) Arcturus Therapeutics (ARCT-810/LUNAR-OTC) Phase Rx Tsukuba Research Laboratories	Yu et al 202294 Prieve et al 201880 Yamazaki et al 202381	Phase 1/2 (NCT03767270) Withdrawn Phase 1 (NCT04442347) Completed; Phase 2 (NCT06488313) Active
Primary hyperoxaluria 1	Autosomal recessive	Childhood	Glyoxylate (AGT)	Peroxisome	AGT mRNA	Alexion Pharmaceuticals	Kukreja et al 201982
Fabry disease	X-linked	Childhood/ Teenagehood	Glycolipids (GLA)	Lysosome	Gal A mRNA	Moderna Translate Bio	Zhu et al 201985 DeRosa et al 201984
Maple syrup disease	Autosomal recessive	Newborn	Amino acids (BCKDHA, BCKDHB, DBT)	Mitochondria	hBCKDHA, hBCKDHB & hDBT mRNA	Moderna	Greig et al 202486
Liver arginase deficiency	Autosomal recessive	Childhood	Urea (ARG1)	Cytoplasm	hARG1 mRNA	Moderna	Khoja et al 202287
Citrullinaemia type II	Autosomal recessive	Adult	Aspartate/glutamate transporter (SLC25A13)	Mitochondria	hCitrin mRNA	Moderna	Cao et al 201989
Hepatorenal tyrosinaemia type I	Autosomal recessive	Newborn	Tyrosine (FAH)	Cytoplasm	FAH mRNA	University of Texas Curevac/MRM Health	Cheng et al 201873 Cacicedo et al 202290
Argininosuccinic aciduria	Autosomal recessive	Newborn	Urea (ASL)	Cytoplasm	ASL mRNA	Moderna	Gurung et al 202493

AIP was among the first rare metabolic disorders in which mRNA therapy was experimentally tested. The administration of LNPs loaded with porphobilinogen deaminase (hPBGD) mRNA in an AIP mouse model was very effective in correcting the biochemical alterations (increased levels of the haem precursor porphobilinogen and the neurotoxic ALA), as well as the clinical manifestations of the disease, including pain, motor disturbances and hypertension.⁴⁹ hPBGD mRNA administration was safe and effective after repeated administration in large animals, rabbits and non-human primates (NHPs).^{49 104} Remarkably, a recent study confirmed the therapeutic efficacy and safety of recurrent PBGD mRNA administration in a new model of AIP developed in NHPs that faithfully reproduces the clinical manifestations of the human disease.¹⁰³

Phase 1/2 clinical trials testing LNP-mRNAs are ongoing for a few rare metabolic disorders such as GSD1a and GSD3, OTC deficiency, MMA and PA (table 3). An interim analysis of a phase 1/2 dose-optimisation trial testing LNP-encapsulated PCCA/PCCB mRNAs in PA patients was recently released.¹¹⁴ This preliminary analysis indicated that repeated treatment (16 participants, more than 340 intravenous doses) was well tolerated, the levels of disease-related metabolites were reduced, and the frequency of metabolic decompensation events, that is, the exacerbation of PA-related symptoms, was also reduced.¹¹⁴ Although the small sample size precluded performing statistical analyses, these are encouraging observations. A follow-up study of the long-term safety and clinical activity of this type of mRNA therapy is ongoing (table 3).

Conclusions and future perspectives

Over the past decade, we have witnessed the vigorous emergence of RNA-based therapeutics, a new class of medicines that substantially differ from conventional drugs. Therapeutic RNAs can be accurately programmed to target any pathogenic gene or to restore the expression of a missing or defective gene. These properties have opened the possibility for the development of precise and personalised therapies, thus minimising the side effects of traditional drug treatments. Advancements in RNA therapies are particularly notorious in the field of genetic liver diseases and in systemic conditions with strong liver involvement, as illustrated by the transformative advent of siRNA therapies in the management of hATTR. Optimised chemistry and very successful liver-targeting strategies, such as GalNAc conjugation, have led to the use of siRNA drugs that can be administered subcutaneously and display very good safety profiles, low immune-related effects and high efficacy.¹⁸

Gene-editing approaches hold promise for diseases in which an eventual reversal of the therapy is not necessary, particularly the highly specific BE and PE CRISPR systems which minimise cancer risk owing to large genomic rearrangements and chromosomal lesions. Transient strategies, as discussed here, can reduce off-target gene editing and adaptive immune responses that eliminate transduced and edited cells. Nevertheless, better delivery methods need to be developed, and additional preclinical research is needed to ensure safety and long-lasting on-target effects. Ongoing clinical trials are expected to provide valuable information not only on efficacy but also on long-term safety.

While siRNAs and ASOs account for most of the approved drugs among RNA medicines, the potential of LNP-mRNA therapeutics for protein replacement strategies is increasingly supported by an overwhelming number of successful preclinical studies. Compared with virus-based gene therapy, mRNA therapy offers increased safety by minimising the risks of genomic integration, insertional mutagenesis and oncogene activation. LNP formulations also evade immune recognition and clearance in vivo, improving therapeutic efficacy and duration. The recently released interim analyses of a first-in-human phase 1/2 trial with PCCA/PCCB LNP-mRNAs in PA patients are very encouraging. These therapies have the potential to reduce mortality and morbidity and improve the quality of life of patients with severe rare metabolic diseases. However, further refinements are needed, including the development of more potent LNP-mRNAs that need to be less frequently dosed. However, clinical trials, particularly those involving rare metabolic diseases, face limitations such as small sample sizes per cohort, the absence of control groups and shorter observation periods for higher-dose cohorts than for lower-dose cohorts. To overcome these limitations preclinical studies testing LNP-mRNA efficacy, safety and pharmacokinetics in different animal species, including NHPs as reported for AIP and PA,^{90 103 115} can be instrumental to inform the design of human trials. On the other hand, despite of their promising therapeutic potential the access to mRNA therapies for rare diseases will likely be constrained by substantial financial costs to healthcare systems. To address this important limitation, the development of cost–utility models to estimate patient outcomes and management costs could be helpful. In this context, a recent such study found that the mRNA therapies for MMA and PA were indeed cost-effective.¹¹⁶

We are convinced that RNA-based drugs will deliver new therapeutic opportunities for a broad range of diseases, many of which are considered very difficult to treat. Nevertheless, we also believe that to achieve this goal, collaboration between academia and industry, together with guidance and support from regulatory agencies, is essential.

The financial sponsors had no role in the analysis or the development of conclusions.