Using modified mRNA to help heart muscle cells grow and study or treat inherited heart diseases

Heart failure (HF) is characterised by a reduced capacity of the heart to fill and/or pump blood [1] and is a leading cause of death worldwide, affecting an estimated 64 million people [2]. Patient numbers are projected to increase in the coming years, placing a huge burden on the healthcare system with treatment costs predicted to be $70 billion by 2030 in the US alone [3]. Aside from the financial impact, HF significantly diminishes the quality of life of those afflicted, making it clear that improved treatment options are required. Currently, first-line treatment for HF patients is pharmacotherapy, which focuses on symptom control by reducing blood pressure (e.g. ACE inhibitors) and blood volume (e.g. loop diuretics). However, these interventions do not address the underlying cause of disease and ultimately do not prevent disease progression [4]. As our understanding of the molecular mechanisms governing HF has improved vastly in recent decades, research has provided promising novel targets for therapeutic intervention. To manipulate these novel targets, multiple gene therapy methods have been employed including plasmids, proteins and viruses (lenti-, adeno- and adeno-associated). However, these approaches can cause random genome integration, ineffective translation due to pre-existing neutralising antibodies and/or uncontrolled gene expression limiting their application [5].

A therapeutic modality that overcomes these drawbacks is modified mRNA (modRNA). The use of modRNA has been recently highlighted due to its involvement in the swift development of vaccines to combat the COVID-19 pandemic [6,7]. ModRNA translation is transient, allowing for short-term, robust expression of one or multiple proteins. This synthetic mRNA has an enhanced gene packaging capacity and induces minimal immunogenic response [8,9]. Aside from viral vaccine development, modRNAs have been designed to treat varying pathologies including metabolic disorders, cystic fibrosis and multiple cancers [10]. Therapeutic development has been extended to cardiovascular diseases with VEGF-A modRNA therapy recently completing a phase 2a clinical trial [11]. This review will focus on the emerging field of modRNA and its use in therapeutics for driving cardiomyocyte (CM) proliferation as well as treating and modelling genetic diseases.

Traditional use of pharmacotherapy to treat HF has seen drug cocktails developed to regulate fluid retention, reduce cholesterol levels and manage contraction, which has halved mortality in moderate HF patients [12]. These approaches have continued to evolve with the use of SGLT2 inhibitors [13] and cardiac myosin activators [14], further improving HF treatment strategies. However, pharmacological intervention does not correct the underlying cause of the disease and is often accompanied by side effects. With increasing knowledge of the molecular drivers of HF, the use of gene therapy has revolutionised HF therapeutic options (reviewed in [15]).

Gene therapy involves the delivery of genetic material (Table 1) to cells resulting in gene repression, overexpression or editing depending on the approach taken and pathology in question. The use of DNA-based vectors has dominated the gene therapy field and can be distributed using non-viral or viral vectors. Using a non-viral vector, plasmid DNA has been utilised due to its expression period of one to four weeks and low immunogenicity, which has seen numerous clinical trials undertaken (reviewed in [16]). However, plasmids exhibit low transfection efficiency [17], whilst a lack of GMP-grade manufacturing facilities exists [18], restricting their clinical application.

The development of viral vectors has provided a way to circumvent transfection issues whilst providing longer expression of the transgene. Retroviruses undergo stable genome integration which enables transgene expression for many months [19]. However, they are unable to transfect non-dividing cells [20], limiting their use with CMs. Lentiviruses are a complex retrovirus which can transduce postmitotic and quiescent cells [21] and provide expression for many months [22]. However, like its family members, they have an increased risk of insertional mutagenesis [23] affecting their safety profile. Adenoviruses have been the most used viral vector for clinical trials thus far and rarely undergo genome integration, demonstrating rapid expression of transgenes up to 35 kb in size [24]. Whilst being popular in the preclinical and clinical pipeline, adenoviruses are highly immunogenic and offer expression for only weeks to months (reviewed in [25]). The viral vector with the most FDA-approved therapies [26] is adeno-associated viruses (AAVs) which have been used prominently in preclinical cardiac studies (reviewed in [27]). AAVs demonstrate low immunogenicity and transgene expression lasting up to 15 years [28], making them an attractive vector for genetic disorders. However, AAVs possess a small gene packaging capacity and neutralising antibodies are prevalent in the community (reviewed in [29]). Additionally, the prolonged expression has been shown to be detrimental in a cardiac regeneration study with arrhythmias causing death in a porcine model [30]. To overcome these drawbacks, recent work has seen the development of drug-inducible AAV systems whereby administration of splice-modulating drugs enables transgene expression providing temporal control [31] and has already been applied to the cardiac regeneration field [32].

Although DNA delivery using viral vectors has been the standard for cardiac gene therapy, the issues around insertional mutagenesis, pre-existence of neutralising antibodies and the relatively small transgene capacity still ultimately restrict their application. Compared with DNA expression strategies, modRNAs are unable to integrate into the genome and have no gene packaging capacity [33]. Further improving the safety profile, the introduction of modified caps and nucleosides diminishes the immunogenic response [9], and they are degraded within a short timeframe compared with DNA, making them an increasingly appealing gene therapy modality.

The first published use of exogenous mRNA for gene overexpression was recorded in 1989 [34] and was soon followed with in vivo delivery of chloramphenicol acetyltransferase mRNA to skeletal muscle in 1990 [35]. Following these and other proof of principle experiments, the majority of research using exogenous mRNA for therapeutic approaches was applied to vaccine development for viral infections [36] and cancer immunotherapy [37]. However, the clinical application of exogenous mRNA was limited by its instability and immunogenic properties. When a cell is subjected to foreign DNA or RNA, an inflammatory response arises due to the innate immune system whereby retinoic acid-inducible gene-I-like receptors [38] and toll-like receptors [39] recognise the foreign nucleic acid driving interferon [40] and tumour necrosis factor (TNF) production [41]. For instance, when monocyte-derived dendritic cells were transfected with different types of RNA, unmodified mRNA induced similar levels of TNF production as bacterial RNA and mammalian mitochondrial RNA [8]. This activation potential displayed an inverse correlation with the degree of nucleoside modification, with bacterial and mitochondrial RNA exhibiting few modifications. When exogenous mRNA was synthesised with modified nucleosides, the inflammatory response was suppressed in a proportional manner to the number of modified nucleosides present. It was this seminal work by Katalin Karikó and Drew Weissman that enabled the development of the first FDA-approved mRNA-based biologic vaccine in 2020 to combat COVID-19 [42] and ultimately earned them the 2023 Nobel Prize in Physiology or Medicine. Reducing the immunogenicity elicited by mRNA allowed the use of exogenous modRNA to extend into other therapeutic strategies including neutralising antibodies [43–46], genome editing [47–49] and protein replacement [50–53].

The structural composition of modRNA shares the classical features of any mature mRNA (Figure 1). The 5′ 7-methyl-guanosine cap is essential for mRNA functionality by regulating nuclear export [54], stability [55] and translation initiation [56]. For modRNA, the cap can be introduced co-transcriptionally or post-transcriptionally. With the former, the polymerase incorporates a cap analogue during in vitro transcription (IVT) leading to capping efficiencies between 50-95% depending on analogue type. Early cap analogue iterations, such as anti-reverse cap analogue, produce a cap 0 structure which leads to immune activity without an extra methylation reaction [57]. The development of CleanCap, which provides a cap 1 structure containing this methylation group, minimises immune activation, increases capping efficiency and yield [58]. Conversely, capping post-IVT can be achieved using Vaccinia virus which results in 100% capping efficiency [59] but at the expense of further handling and purification.

Flanking the coding sequence are the untranslated regions (UTRs) which regulate post-transcriptional processes including subcellular localisation [60], mRNA stability [61] and translation efficiency [62]. As the 5′ UTR is crucial for translation initiation, multiple studies focused on 5′ UTR design have highlighted the importance of sequences lacking multiple start and stop codons [63] as well as avoiding secondary structure formation to enable effective mRNA-ribosome binding [64]. Conversely, secondary structures in the 3′ UTR correlate with elevated protein expression [64]. For 3′ UTR design, the traditional approach has been the use of abundantly expressed genes such as α- and β-globin to provide a stable transcript [65]. However, recent work has shown a modRNA with a four-base 5′ UTR and no 3′ UTR produced functionally comparable protein with a β-globin 3′ UTR illustrating further optimisation is possible [66].

The poly(A) tail is required for mRNA stability [67] and formation of a closed-loop structure in the translation-initiation complex [68]. Similar to the 5′ cap, the poly (A) tail can be introduced co-transcriptionally based on the DNA template or post-IVT using a poly(A) polymerase [69]. The consensus has seen modRNA poly(A) length to be 100–120 adenosines in length to enable increased mRNA stability and translational efficiency via efficient poly(A) protein binding [70,71]. However, recent work has illustrated that a modRNA with 80 adenosines produces protein at levels equivalent to 120 adenosines [72], indicating that binding of at least two poly(A) binding proteins is sufficient to protect from deadenylation [73].

The component that enables the widespread application of this modality is the coding sequence which consists of the open reading frame (ORF) of one or multiple proteins depending on the gene therapy application. Aside from the ability to incorporate any ORF, there are several modifications to the coding sequence which have major implications on mRNA stability and translation. This includes the choice of modified nucleosides for any of the four bases which can diminish the cellular immune response [8,74] and enhance protein expression [9,72]. Additionally, recent work has highlighted the importance that codon optimisation and mRNA structure play in translation and stability, respectively. Development of multiple online tools [75–77] has streamlined this process, enabling the design of optimally translated mRNA to achieve elevated protein expression. These wide-ranging considerations in the design of modRNA make it an extremely versatile system where its application has been observed for both cardiac regenerative and genetic disease research.

With cardiovascular diseases the leading cause of death worldwide [78], modRNA has begun to be utilised for therapeutic development in this space. After numerous studies focussed on delivery optimisation (reviewed in [79]), myocardial infarction (MI) has been the primary disease model for cardiovascular modRNA therapeutic development. After an MI, several pathological processes occur, including CM death, immune cell infiltration and myofibroblast activation compounded by insufficient neovascularisation (reviewed in [80]). Multiple studies using modRNA have demonstrated its ability to address these perturbations through reducing CM apoptosis [81,82], inflammation [83] and cardiac fibrosis [84–86], whilst also promoting angiogenesis [87]. However, the most advanced therapeutic target for cardiovascular modRNA medicine is VEGFA which drives neovascularisation [11,88–91] and improves CM engraftment [92]. Whilst replacement of lost CMs through fibroblast reprogramming has been demonstrated using modRNA [87,93], this review will focus on modRNA therapeutics targeting CM proliferation.

In the first week of postnatal life, CM maturation occurs, coinciding with reduced capacity for CM proliferation, resulting in poor regenerative potential within adult mammals. This short regenerative window was first demonstrated in mice [94] and has since been observed in rats [95], rabbits [96], sheep [97], pigs [98,99] and anecdotally in newborn humans [100,101]. When regeneration is possible, pre-existing CMs undergo proliferation to replace the damaged tissue [94], making strategies targeting CM cell cycle re-entry an attractive therapeutic approach. As uncontrolled CM cell cycle re-entry can lead to arrhythmia and death [30], the transient, pulse-like nature of modRNA makes it ideal for cardiac regenerative gene therapy. Currently, there are six published articles or preprints using modRNA to drive CM proliferation which we will preview here (Table 2, Figure 2). These studies highlight the potential of using modRNA as a vector for gene therapy to drive CM proliferation both in vitro and in vivo. In addition to achieving improved cardiac function through increased CM proliferation, the transient nature of modRNA was demonstrated with no prolonged CM proliferation (four weeks post-MI) quantified in several of the studies reviewed here. This finding is important as it indicates that modRNA therapies will likely prevent persistent CM dedifferentiation which, if uncontrolled, has led to arrhythmogenesis and/or decreased cardiac output in zebrafish [118], mice [119–121] and pigs [30].

These findings are encouraging and demonstrate how the transient expression of modRNA is suitable for cardiac regenerative purposes. Whilst encouraging, the timing of therapeutic intervention is crucial and is a major difference between preclinical and clinical models. To this point, most in vivo regenerative studies in mice have seen therapeutic administration occur simultaneously with ligation of the left anterior descending (LAD) coronary artery and no subsequent reperfusion. In contrast, most patients suffering a MI receive immediate heart reperfusion to restore blood flow, resulting in recovery times of at least 8 weeks [122], during which significant ventricular remodelling and fibrosis occurs [123]. The consequences of these perturbations on modRNA efficacy have not been fully investigated to see if CM proliferation is still possible and, if so, sufficient to improve function. This lack of a clinically relevant chronic model also has implications for modRNA delivery. modRNA dosage has been tested previously, however only without MI [124] or in an acute post-MI setting [125]. To enhance modRNA delivery, lipid nanoparticles have been incorporated into the modRNA pipeline, but again testing has only occurred without MI [126] or in the acute environment [127]. These considerations are vital for cardiac modRNA therapeutic development as the use of clinically inferior models may impede clinical trial progression.

Although MI is the primary cause of HF [128], the condition has numerous aetiologies including coronary heart disease, cardiomyopathy, congenital heart disease and amyloidosis [129]. From numerous epidemiological studies [130–134], it is apparent that a major genetic component is intertwined in the development of these pathologies. For modelling and treating genetic diseases, the use of Cas9 has transformed medicine through genome editing which has transferred to the cardiovascular field (reviewed in [135]). However, only recently has modRNA been utilised to deliver the Cas9 mRNA with in vivo gene editing applied to treating transthyretin amyloidosis [48] and hereditary angioedema [136]. Whilst primarily known for its transient expression, modRNA has been utilised for protein replacement therapy of genetic disorders with clinical trials undertaken to treat methylmalonic acidemia [50], propionic acidaemia [51] and cystic fibrosis [52]. Although an emerging area, we highlight the published studies using modRNA for modelling and treating genetic conditions affecting CMs.

Hurdles are still limiting the clinical translation of modRNA; however, different avenues are being interrogated to improve the applicability of the modality. Whilst this review has focussed on conventional mRNA, the use of circular RNA (circRNA) and self-amplifying RNA (saRNA) is emerging in the RNA therapeutics field (Figure 3A). circRNA is a type of closed RNA lacking 5′ and 3′ ends that occurs naturally in cells and exhibits tissue-, development- and disease-specific expression [165]. As circRNAs lack a 5′ cap, most were classified as non-coding RNA until the discovery that circRNA use cap-independent translation for protein expression [166]. This has enabled synthetic circRNAs to be developed for therapeutic purposes to deliver non-coding [167,168] and protein-coding therapeutics with the latter currently used primarily in vaccine development [169]. Encouragingly, unmodified circRNA are also less immunogenic than unmodified conventional mRNA [170], increasing their appeal as a potential therapeutic vector. SaRNA utilises the non-structural proteins of single-stranded RNA viruses to produce a self-replicating vector for transgene expression [171]. Like conventional mRNA, saRNA is linear; however, due to its self-replication, it has been found that an influenza saRNA vaccine requires 64-fold less material to produce the same antigenic response [172]. Due to the presence of the viral proteins required for replication, saRNA exhibits enhanced immunogenicity, although this can be suppressed with the addition of anti-immunogenic genes [173]. However, inclusion of these genes increases an already large construct which can affect production. Together, circRNA and saRNA provide enhanced protein expression over longer durations, which reduces dosage whilst providing an overexpression vector better suited for treating chronic diseases. However, they have not yet been applied as a therapeutic modality in the cardiac medicine field.

Whilst great strides have been made to improve tissue targeting, more advanced in vivo delivery systems (Figure 3B) are required to enhance the translatability of modRNA. To prevent degradation by nucleases and enable cellular uptake and ultimately translation, most RNA therapeutics have relied on the use of lipid nanoparticles (LNPs) [174]. LNPs are highly homogeneous and scalable, but they exhibit low biocompatibility, high toxicity [175] and a strong propensity to accumulate in the liver [176], increasing the risk of side effects. Although unable to prevent liver accumulation, direct cardiac injection of LNP-encapsulated mRNA enables markedly reduced dosage [127,177] potentially reducing off-target effects. An alternative to LNPs is the use of cells as a delivery vector which sees the mRNA introduced to cells ex vivo before utilising the natural paracrine functions for mRNA delivery upon cellular injection [178]. This approach is far more biocompatible and extends longevity in circulation but is accompanied by limitations surrounding donor haplotype compatibility and potential tumourigenesis. Another biological approach is the use of extracellular vesicles – encompassing both exosomes and microvesicles – which are naturally produced thereby improving biocompatibility and utilising endogenous cellular uptake mechanisms (reviewed in [179]). Multiple examples exist where exosomes have been used for cardiac regeneration (reviewed in [180]) although they have not been used as a delivery vector to the heart. However, exosomes are highly heterogenous and of biological origin, limiting their scalability and cargo loading, which has been relatively inefficient, impeding their clinical translation. A newer approach has seen the development of biological and synthetic hybrids, known as biomimetics, which combine a synthetic core with a cellular membrane coating (reviewed in [181]). This combination enables efficient cargo loading into the core whilst the native cellular membrane reduces immunogenicity, providing high biocompatibility and increased stability [182]. However, the cargo release mechanism is currently unclear, and a lack of manufacturing criteria exists. Collectively, mRNA delivery system development is a focal point for its clinical translation, yet it is likely that the system used will be determined by the pathology and the need for transient or persistent mRNA translation.

Although packaging systems are also being developed to exhibit tropism [183], sequence-specific modRNA regulation is an evolving avenue for cell type-specific translation. This has been primarily employed in the cardiac regeneration field [104,107,114,116] through utilisation of a synthetic RNA circuit where the archaeal ribosomal protein, L7Ae, suppresses modRNA translation [184]. By incorporating a cell-specific miRNA binding site in the L7Ae-expressing modRNA, its degradation will occur in a cell-type specific manner, allowing translation of the therapeutic modRNA (Figure 4A). Alternative miRNA-dependent translational regulation of exogenous mRNA has been demonstrated with miRNA-binding site incorporation into the ribosome binding site [185] (Figure 4B) controlling mRNA translation; however, this was performed using circRNA. Another approach saw a miRNA-binding site incorporated in the poly(A) tail (Figure 4C) which upon miRNA binding, reveals the tail enabling translation and has been used to purify CMs from iPSC-CM differentiated culture [186]. Aside from miRNAs, there is a promising technology harnessing the double stranded RNA editing enzyme, ADAR, which edits mismatched adenosines to inosines – read as guanosine by the translational machinery [187]. Targeting cell-specific transcripts, modRNAs [188–191] are designed with a sequence complementary to the endogenous transcript except for an in-frame stop codon (UAG) which will be edited to tryptophan (UGG) enabling downstream translation in only the target cell (Figure 4D). Whilst in their early stages of development, these mechanisms are encouraging as they utilise cell state-specific transcript expression providing a versatile yet targeted approach to translational regulation.

The use of modRNA for modelling and therapeutic intervention in cardiac pathologies is a constantly evolving and exciting space. The strengths of modRNA compared with other gene therapy vectors lie in its lack of gene packaging restrictions, reduced immunological response and transient but robust protein expression. The studies reviewed here demonstrated the importance of the last point for both pro-proliferative therapies and genome editing as the pulse-like expression enables safer outcomes by preventing prolonged CM de-differentiation and off-target effects, respectively. Whilst the future involvement of modRNA in cardiac research and medicine is anticipated, improved delivery systems are required to improve clinical translatability and the use of other mRNA systems is likely necessary to tackle chronic diseases. Addressing these challenges will allow modRNA to become a leading gene delivery vector for modelling and therapeutic purposes enabling the development of future treatments to combat HF.