Oligonucleotide-based therapies for cystic fibrosis

Abstract

In the clinically successful era of CFTR modulators and Theratyping, 10–20% of individuals with cystic fibrosis (CF) may develop disease due to CFTR mutations that remain undruggable. These individuals produce low levels of CFTR mRNA and/or not enough protein to be rescued with modulator drugs. Alternative therapeutic approaches to correct the CFTR defect at the mRNA level using nucleic acid technologies are currently feasible; e.g., oligonucleotides platforms, which are being rapidly developed to correct genetic disorders. Drug-like properties, great specificity, and predictable off-target effects by design make oligonucleotides a valuable approach with fewer clinical and ethical challenges than genomic editing strategies. Together with personalized and precision medicine approaches, oligonucleotides are ideal therapeutics to target CF-causing mutations that affect only a few individuals resilient to modulator therapies.

There are >2000 mutations identified in the cystic fibrosis transmembrane conductance regulator (CFTR) gene but not all of these mutations may be associated with cystic fibrosis (CF) disease. Today, CF disease caused by some of the most frequent mutations can be treated with CFTR modulator drugs that correct the defect at the protein level. However, many mutations, some rare, remain undruggable because they cause low levels of CFTR mRNA and/or protein [1]. Before the COVID-19 vaccine implementation [2], messenger RNA (mRNA) therapies, although readily available seemed too complicated to draw attention and clinical benefit. Emerging mRNA technologies can be useful to design therapeutics for CF undruggable mutations, particularly those that are rare and very rare in the population. This review is a brief discussion of current therapeutic approaches aiming to correct CF-causing mutations in the CFTR mRNA with emphasis on therapeutic antisense oligonucleotides.

From gene to protein

Genetic mutations can be the result of single or multiple nucleotide changes in the coding or non-coding region sequences of a gene [3]. They can be silent (not associated with phenotype) or interfere with “gene to protein” expression and cause pathology. DNA is transcribed into a pre-mRNA in the cell nucleus. This early transcript contains exons and introns and is processed (splicing) by the spliceosome to eliminate the introns and render a mature mRNA with only the coding sequence [3]. Splicing requires specific sequences in the pre-mRNA to be recognized by the spliceosome and splicing regulatory factors to produce a specific splicing pattern for each gene. Splicing mutations can change the recognition sequences (e.g., splicing sites) by deleting or creating new ones. Thirty-60% of human genetic diseases may be caused by mutations affecting splicing [4]; e.g., the CFTR splicing mutation c.3718-2477C > T a. k.a. 3849 + 10 kbC > T (legacy name) [5], which alters CFTR normal splicing, affecting the production of functional protein and causing CF [5,6] (Figure 1a).

Figure 1. Correction of the CF 3849 + 10kb C > T splicing mutation using splice switching oligonucleotides in patient-derived airway epithelial cells (HBEC).

a) Splicing schematics describing aberrant and oligonucleotide-corrected splicing pattern of mutant CFTR mRNA; exon and intron are identified by legacy nomenclature. b, c) Splice switching oligonucleotide correction in HBEC from a CF patient homozygous for the 3849 + 10kbC > T mutation. Three independent oligonucleotides with LNA chemistry and mismatched oligonucleotide (MM) were administered at 50 nM via electroporation in patient cells. Cells were cultured and differentiated under ALI conditions for 14 days; after which, HBEC were analyzed at the mRNA level by RT-PCR (b) and for CFTR activity by Ussing chambers (C; n = 3) as described in Ref. [13]. The specific LNA oligonucleotide but not MM showed notable mRNA splicing correction and increased forskolin (Fs)-stimulated CFTR channel activity (post Amiloride) in the range of normal HBEC values. This patient’s cells do not respond to Orkambi^® or Trikafta^® treatments [13].

After splicing, 3′-capping and 5′-tailing, the mature mRNA moves from the cell nucleus to the ribosomes to be translated into a protein. A sequence of three consecutive nucleotides in the mRNA form a codon, which is translated into a particular amino acid with the aid of transfer RNAs (tRNAs) [3]. The most frequent, severe CF mutation in the USA, is the deletion of three consecutive nucleotides encoding for phenylalanine 508 in the CFTR protein (c.1521_1523del a. k.a. p. Phe508del), which eliminates one amino acid and causes CFTR protein misfolding and loss of function [1,7]. Nonsense or stop mutations can create a new premature stop codon (e.g., TAG, TGA, or TAA), causing a protein to terminate its translation earlier (note: splicing mutations often reveal cryptic stop codons). The most frequent outcome of early stop mutations is the rapid degradation of the RNA via the Nonsense-Mediated mRNA Decay (NMD) pathway, a surveillance mechanism that degrades mRNAs containing nonsense mutations resulting in negligible functional protein [3]. Examples include the CFTR c.1624G > T or G542X (legacy name) and c.3846G > A or W1282X (legacy name) mutations [1]. Stop mutations and severe splicing mutations in the CFTR gene cause little or no functional protein, resulting in severe CF disease and resilience to modulator therapies [8–14].

Therapeutic modalities that target CFTR mRNA

Several therapeutic strategies could correct CF mutations at the mRNA level to produce an RNA that can be translated into a functional CFTR protein: (1) Replacing the aberrant mRNA with a productive copy of CFTR mRNA; (2) Engineered tRNAs to translate an aberrant premature stop codon into an amino acid; (3) Therapeutic oligonucleotides to modulate the splicing or translation of aberrant CFTR mRNA into a productive messenger. Small molecules that affect translation reading-through or reduce the NMD pathway activity are also potential strategies that operate at the mRNA level and they are discussed elsewhere [8,9,15].

1. Therapeutic mRNA: Using a synthetic CFTR mRNA as a CF therapeutic approach presents great advantages and challenges. The main advantage is that only one therapeutic entity, i.e., one mRNA e one therapeutic design can correct different CFTR mutations serving many CF patients. The main challenge is delivery: how, what, where, when [8,9,16,17]. Administration of mRNA requires a vehicle or vector; the RNA needs to be modified to resist enzymatic degradation with minimal toxicity [2,17], the RNA needs to be delivered to the right cell types at the right levels and avoid potentially cytotoxic ectopic expression; and repetitive administration is required for consistent CFTR production and activity. Several groups are pursuing this approach; e.g., Translate Bio is using CFTR mRNA MRT5005 administered in lipid nanoparticles via inhalation. Initial clinical data was encouraging but the clinical trial was paused because of lack of activity (RESTORE-CF ClinicalTrials.gov ID NCT03375047) [18]. A similar strategy is being utilized by ReCode Therapeutics with promising pre-clinical data [19,20].

2. Engineered tRNAs: These synthetic tRNAs “read” the codon of a premature stop codon into an amino acid to complete the translation of CFTR protein [21]. The advantages of such a strategy are that engineered tRNAs could be used to correct different CF-causing nonsense mutations and even different diseases produced by nonsense mutations. The cons are delivery of large RNA molecules to target the right cells and tissues, requiring RNA modification, viral or non-viral vehicles, and repeated doses to achieve an acceptable therapeutic correction level. In addition, there are potential off-target effects by effecting the reading-through of native stop codons. Lueck and collaborators have demonstrated the application of this promising new class of RNA therapeutics in cell models with reduced off-target effects on native stop codons [21–23]. ReCode Therapeutics is also developing this strategy for clinical application in CF [19].

3. Therapeutic oligonucleotides: These are short, “drug-like” polymeric nucleic acids with the capacity to modulate gene expression and treat disease [13,17,24]. The advantages are a simple design with predictable high specificity/low off-target effects, good safety profile, and simple manufacturing and administration, i.e., oligonucleotides do not strictly require special vehicle or vectors [13,24]. The cons are still delivery of pharmacological concentrations to cellular targets (see below), repeated doses are needed but oligonucleotides can be long lasting potentially requiring only a few treatments a year [13,25,26], and the mutation-specific nature of many applications, resulting in small numbers of patients being served.

The first proof-of-concept study using a synthetic deoxyoligonucleotide to modulate gene expression was published in 1978 by Stepheson & Zamecnik [27]. It took several decades for advances in chemical modifications to the nucleotide structure, chemical synthesis, and automation to produce a breakthrough transformation of oligonucleotides into drug-like entities for potential clinical use. Thus, ~40 years later, the approval of Spinraza/nusinersen for the treatment of smooth muscle atrophy changed the field of oligonucleotides. Nusinersen developed by Ionis Pharmaceuticals and Biogen, is an 18 nucleotide-long, chemically modified oligonucleotide, which is administered via intrathecal injection in infants and adults; it increases the levels of SMM2 in patients with a deficiency in SMM1 [28]. Several successful oligonucleotides followed for the treatment of diseases like Duchenne muscle dystrophy [17,28,29]. Eluforsen, developed by ProQR Therapeutics, is a 33 nt, single-stranded chemically modified RNA oligonucleotide partly complementary to the p. Phe508del-CFTR RNA that corrected the mutation in pre-clinical studies [30] and in clinical trials via inhalation (ClinicalTrials.gov ID NCT02564354) [31]. Although safe, it was discontinued because of low CFTR corrective activity in patients who could be treated with modulators [9].

There are at least ten therapeutic oligonucleotides approved for clinical use in different disorders [17,32], and although none has reached clinical application to treat CF, oligonucleotide-based strategies are currently being developed to target mutations resilient to CFTR modulator therapies [11–13,15,33,34].

Therapeutic antisense oligonucleotides (ASOs)

There are different modalities of therapeutic oligonucleotide; e.g., antisense oligonucleotides (ASOs). ASOs are synthetic, short (16−35 nucleotides), single-stranded nucleic acids, with diverse chemistries, which bind to their cognate mRNA transcripts in a Watson-Crick fashion and modulate gene expression by various mechanisms [17,29]. Different CF therapeutics have been developed using this strategy, and are discussed below. Other types of oligonucleotides include RNAi and miRNA used as precision duplex silencers. A CF therapeutic was developed by Arrowhead Pharmaceuticals using an iRNA-targeting ENaC, the epithelial sodium channel that is often upregulated in CFepithelia [35–37]. ARO-ENaC is an iRNA oligonucleotide paired with a ligand that recognizes αvβ6 integrin for targeted cellular entry [37]. It was tested via inhalation in clinical trials (ClinicalTrials.gov ID NCT04375514) but the study was interrupted due to undesirable effects.

According to the mechanism of action, ASOs are divided in two types:

1. RNase H competent: The endogenous RNAse H enzyme catalyzes the degradation of RNA–DNA heteroduplexes formed by a DNA-based oligonucleotide binding to their cognate mRNA transcripts, and thus, silencing targeted gene expression. This mechanism has been exploited in potential CF-therapeutics. For example, Crosby et al. targeted ENaC. Thus, using an ENaC-specific ASO delivered via inhalation, CF-like lung phenotypes were ameliorated in a mouse model [33]. Based on these studies, Cofirasersen/IONIS ENAC 2.5Rx/ION827359 was developed by Ionis Pharmaceuticals in collaboration with the Cystic Fibrosis Foundation. ASO ION 827359 and iRNA ARO-ENaC are examples of clever designs to correct CF symptoms in a CFTR mutation-agnostic manner using oligonucleotides [35,36]. ION827359 was tested recently in clinical trials (ClinicalTrials.gov ID NCT03647228) with initial promising data but was suspended because of undesired effects.

2. Steric block (RNase H independent): The ASO binds to the mRNA and masks specific sequences within a target transcript and thus, interferes with RNA interactions with other RNAs or proteins, e.g., the spliceosome. The most widely used application of steric block ASOs is in the modulation of alternative splicing: to include or exclude exons by masking sequences in the pre-mRNA to the splicing machinery. The oligonucleotides that utilize this mechanism are also called Splice Switching Oligonucleotides because they switch the splicing pattern of a transcript (Figure 1). This is a useful mechanism to correct splicing defects as a result of mutations that affect exon/intron recognition, but also some mutations that cause frameshift, and/or premature stop codons [11–13,17,29,34]. The same technology can also be used for splice corruption to disrupt the translation of a gene and reduce the expression of a protein associated with a disease [13,17].

Splicing mutations are prevalent in many genetic diseases such as CF [1,8]. RNA splicing is a dynamic process that produces tissue-specific splicing patterns for each gene/transcript and is affected by physiological and pathological changes [3]. It is not surprising that CFTR splicing mutations can result in variable levels of productive CFTR protein and channel activity in patient- and tissue-specific manners [6,12,38–42]; the same CFTR splicing mutation can result in varying disease phenotypes and severity among CF patients, producing enough protein in some individuals to benefit from therapeutic modulators [43–45], while this may not be the case in other patients [13,46]. CFTR splicing mutations could be clinically corrected by ASOs/Splice Switching Oligonucleotides in patients that do not respond to CFTR modulator therapies (Figure 1). This strategy was introduced by Kole et al., in 1999, to successfully correct in cell models, the most frequent CF-causing splicing mutation, 3849 + 10 kb C > T (legacy name) [47], and it has been developed further [12,13,46,48,49] with recently published, promising pre-clinical data in airway epithelial cultures derived from CF patients and animal models by Oren & Kerem et al. [12] and Kreda & Juliano et al. [13].

Chemistry of drug-like oligonucleotides

Naturally occurring DNA and RNA polymers have poor pharmacologic properties, including low biodistribution and cellular uptake, and rapid hydrolysis by ubiquitously expressed enzymes. The recent successes of therapeutic oligonucleotides have stemmed in part from the development of chemical modifications to the nucleotide structure to give them better drug-like properties. Nucleotide chemical modifications provide one or more enhanced pharmacologic properties and also allow administration as drug-like naked oligonucleotides without the need of a specialized vehicle or vector, presenting a clear advantage compared to other nucleic acid-based therapies [13,17,29,32,50].

Nucleotides have three components: (1) a base (A, T, C, G, U), which is bound to (2) a sugar (ribose or deoxyri-bose) that is phosphorylated to create (3) the phosphodiester backbone linkages that string oligonucleotides together. Chemical modifications target each component. For example, modification of the bases such as N1-methylpseudouridine, increased COVID-19 mRNA vaccine effectiveness [2]. The frequently used phos-phorothioate (PS) modification involves replacing one non-bridging oxygen atom in the phosphate with a sulfur atom. The PS backbone provides enhanced stability, binding to serum proteins, and biodistribution [51]. The sugar may be modified at specific positions around the ring; for example, the 2′-OH of the ribose is most often replaced by 2′-O-(2-methoxyethyl) (2′MOE), 2′-fluo-roribose (2′FRNA), or 2′-O-methyl (2′OMe) groups, or the 2′ and 4′ carbons are bridged by a methylene (LNA or locked nucleic acid; used in Figure 1b and c) or constrained ethyl (cEt) linkage [50]. Alternatively, the sugar is replaced altogether with a sugar ring mimic, such as a morpholino a. k.a. PMO (phosphorodiamidate morpholino oligomer). Morpholinos are very stable, and have high biodistribution and low toxicity [32,52,53]. Morpholinos can be conjugated to a short peptide (PPMOs) to improve transport into cells, allowing for improved biodistribution and activity [13,24,50]. PMOs and PPMOs have been used successfully in bench [13,26,46,50,54–59] and clinical applications, e.g., Sarepta Therapeutics developed FDA-approved PMOs for Duchenne’s dystrophy Amondys 45/Casimersen; Exondys 51/Eteplirsen, Vyondys 53/Golodirsen and PPMO SRP-5051 in clinical trials (ClinicalTrials.gov ID NCT03375255). Our lab has developed a PPMO-based strategy to target CF-causing splicing mutations [13].

Oligonucleotide delivery to CF-relevant tissues

Delivery of nucleic acids to their cellular targets has been one of the main roadblocks in the therapeutic development of oligonucleotides-based strategies. Many of the chemical modifications described above in addition to those reviewed elsewhere [e.g. Refs. [17,26,29,60–63]] greatly improve tissue and cellular delivery of non-viral nucleic acids. However, even these heavily chemically modified oligonucleotides have reduced activity as the result of non-productive intracellular entrapment in endomembrane compartments. Cells have multiple pathways of endocytosis, pinocytosis, and fluid-phase uptake [24,64,65]. All forms of oligonucleotides, whether as free molecules, molecular conjugates, or associated with nanoparticles, utilize to varying degrees, these pathways which all converge in endosomes [66,67]. Within these compartments, oligonucleotides are pharmacologically inert since they cannot access their molecular targets in the cytosol or nucleus [68,69], and >90% of intracellular non-viral nucleic acid molecules may end degraded or exocytosed [13,24,70–73]. In the absence of a delivery moiety, oligonucleotides slowly leak into the cytosol and can then reach the nucleus. The currently approved oligonucleotide drugs all work this way. However, the slow leakage is rather inefficient; thus, enormous efforts have been made to enhance oligonucleotide delivery [29,68,74]. There are several major approaches to augmenting oligonucleotide delivery. The first is to use carriers such as lipid nanoparticles [75], but they can have limited systemic biodistribution [29,68,76]. Another approach is the use of conjugates of an oligonucleotide to a ligand to promote cell uptake [29,77,78], e.g., the carbohydrate conjugate of siRNA to target hepatocytes via the asialoglycoprotein receptor [17,29,79].

The Kreda laboratory in collaboration with the Juliano laboratory has developed a two-prong approach to increase oligonucleotide delivery to extra-hepatic tissues (Fig. 2). First, the oligonucleotide has PPMO chemistry: a morpholino conjugated to a cell-penetrating peptide, which greatly increases extra-hepatic biodistribution and cell uptake [13,50,54–56]; second, small-molecule “OECs” with endosomal permeability properties, identified by Juliano et al. and Initos Pharmaceuticals to enhance the delivery of non-viral nucleic acids from endosomal vesicles [13,24,71–73,80]. Although PPMOs have some innate endosome escape capability (Figure 2b), our work indicates that this can be substantially enhanced via the use of OECs [13]. Our novel delivery approach using PPMO + OECs to alter the intracellular trafficking of oligonucleotides increases their activity in the lung and other CF-relevant tissues [13,14,24]. Thus, one treatment of a PPMO + OEC designed to correct the CF-causing 3849 + 10 kb C > T splicing mutation reestablished CFTR channel activity to normal values in airway epithelial cells derived from a CF patient [13]. The PPMO was efficacious when administered either to the serosal or mucosal epithelial surfaces (>90% and ~80% CFTR mRNA correction, respectively), while Trikafta^®-like treatment was ineffective [13]. Our approach was ~40 times more efficacious in correcting the CF-causing splicing mutation with one PPMO + OEC treatment than using PMO or PMO aided by a transfection agent administered continuously for ~2 weeks [46]. Moreover, PPMO + OEC corrected a splicing mutation in vivo in the lung and other CF-relevant tissues of a mouse model for at least 3 weeks after one systemic treatment with no obvious toxicity [13,14]. Of note, conditions of concentrated mucus mimicking the CF milieu in the airway lumen did not affect the delivery of PPMOs or OECs [13,14]; similar data was described for other oligonucleotide chemistries in CF-like lung epithelial [81]. Our pre-clinical data suggest that the PPMO-based strategy may have substantial potential as a therapy for severe splicing mutations in CF.

Figure 2. Improving oligonucleotide delivery in patient-derived airway epithelial cells (HBEC).

a) OEC: HBEC were incubated basolaterally with a fluorescently-labeled oligonucleotide with PS-2′OMe chemistry (PS-FL, 0.5 mM, 16hs) in combination with OEC added either sequentially (7 μM, 2hs; top left) or concurrently (7 μM, 7hs; bottom left); or in the absence of OEC (top right). After the experiment, cells were labeled with fluorescent wheat germen agglutinin (green) to visualize cilia, and DAPI to visualize nuclei; confocal microscopy xz scanning as described in Ref. [13]. In these experimental conditions, the FL-PS oligonucleotide traffics efficiently into the nucleus only when OEC was present; note the supra-nuclear puncta, most likely oligonucleotide accumulated in endomembrane and endosomal compartments. b) PPMO: HBEC incubated with fluorescent-PPMO (PPMO-FL, 0.5 μM, 1 h) and imaged as described in Ref. [13]. The oligomer localizes efficiently and rapidly (<1 h) into the nuclei in the absence of OEC; image was obtained 6 months after one PPMO-FL treatment.

Conclusions

Although oligonucleotides can be used to affect gene expression via various mechanisms, most of them target one mutation at a time. Such an approach can benefit only patients with a specific mutation, bringing these therapies into the realm of personalized and precision medicine to correct rare and ultra-rare CF-causing mutations. Other genetic diseases resulting from multiple types of mutations present the same therapeutic dilemma. Traditional small-molecule drug development is often not suitable. It requires large, iterative screening efforts followed by extensive medicinal chemistry optimization, drug-specific toxicology tests, manufacturing and administration systems, and complex clinical trials for drug approval. In contrast, nucleic acids are chemically similar to one another but can be designed with great specificity for a particular mutation with minimal predictable off-target effects, greatly reducing the drug-screening process. Oligonucleotides can be grouped by chemistry and/or mechanism of action, simplifying the manufacturing process, administration strategy, and toxicology testing. The whole drug-developing path of designing a potential therapeutic oligonucleotide for a rare disease and producing and moving it through the regulatory process could take months rather than years [17,26,29]. This model for drug development was recently used by Yu and collaborators to produce Milasen, an oligonucleotide to treat Batten disease in one child [82] (https://answers.childrenshospital.org/milasen-batten-disease).

This is an exciting time for CF therapeutics, but there is yet the last long mile to finish the race and overcome every CF-causing mutation. The field of therapeutic oligonucleotides is growing rapidly. The advantages of fast drug development and simple administration in concert with precision and personalized medicine approaches make oligonucleotides an ideal therapeutic platform for undruggable rare CFTR mutations. Moreover, strategies to adapt and optimize the clinical trial design to quickly assess efficacy and safety, including the use of patient-derived specimens, animal models for safety data, and N-of-1 and “basket” trials, are being discussed by disease organizations, and academic, pharmaceutical and regulatory institutions (FDA-2021-D-0320, April 26 2021). Oligonucleotide-based treatments for CF may be at hand soon.