Oligonucleotide correction of a rare CFTR intron 12 splicing variant from a cystic fibrosis patient

Extract

We describe the use of novel antisense oligonucleotides for correcting a cystic fibrosis (CF) splicing mutation not amenable to CF transmembrane conductance regulator (CFTR) modulator therapies.

Short abstract

SSOs corrected rare CFTR splicing mutations unresponsive to modulators, restoring functional CFTR in patient airway epithelial cells – even under epithelial inflammation – highlighting SSOs as a promising precision therapy for cystic fibrosis https://bit.ly/3U156kG

To the Editor:

We describe the use of novel antisense oligonucleotides for correcting a cystic fibrosis (CF) splicing mutation not amenable to CF transmembrane conductance regulator (CFTR) modulator therapies.

Over 2000 CFTR mutations have been identified with <30% CF alleles eliciting abnormal mRNA splicing. Pathogenic splice variants originate from mutations in sequences recognised by the splicing machinery on the pre-mRNA to splice-out the introns [1]. Cell type, age, and inflammation elicit spatial-temporal RNA splicing patterns for each gene [1]. Unsurprisingly, CFTR splicing errors produce variability in CF disease severity and among different tissues of the same patient [2]. For example, people with CF (pwCF) with the most frequent CFTR splicing mutation, 3849+10kbC>T (c.3718-2477C>T) can show near complete loss of productive mRNA and protein, resulting in severe CF manifestations [3], or sufficient CFTR protein levels to respond to CFTR modulators [4]. Similarly, a series of splicing variants localised close together in intron 12 of CFTR can produce a severe CF phenotype and resilience to modulator therapies, i.e. c.1679+1643G>T (c.1680-877G>T or 1811+1643G>T) [5], c.1679+1634A>G (c.1680-886A>G; 1811+1.6kbA>G) [6], and 1811+1.6kbT>A, and are all more frequently identified in pwCF from the South of Spain and France, and Latin America (1–5% CF alleles) [5, 6].

The now 17-year-old male of Latin American ancestry was diagnosed with CF at age 3 years with chronic cough, Pseudomonas aeruginosa infection, malabsorption, and a positive sweat test. He developed chronic methicillin-resistant Staphylococcus aureus infection at age 6 years and CF-related diabetes 5 years later. He has moderate-to-severe lung disease (baseline forced expiratory volume in 1 s ∼50% predicted) requiring 4–5 courses of i.v. antibiotics per year, and pancreatic insufficiency with moderate malnutrition despite nighttime G-tube feeds. Current radiography (figure 1a) illustrates marked hyperinflation, bilateral bronchiectasis, and bronchial wall thickening. CFTR genotyping performed at the Johns Hopkins Genomics DNA Diagnostic Laboratory through the Cystic Fibrosis Foundation Mutation Analysis Program in 2012 identified the c.1679+1643G>T splicing defect in one allele and the G542X nonsense mutation in the second allele. The G542X is an early termination codon variant eliciting an unstable mRNA and/or a truncated non-functional CFTR protein that cannot be rescued with modulators. The c.1679+1643G>T variant creates a cryptic splice site in intron 12, leading to the inclusion of a 53 base pair-intronic sequence as a pseudo-exon in the mature mRNA, disrupting normal CFTR mRNA and protein function [5] (figure 1b). Patient-derived human nasal epithelial cells (HNECs) were used to study CFTR mRNA expression, CFTR response to modulators, and the efficacy of oligonucleotides to correct the splicing defect and restore CFTR activity [3] in the presence of epithelial inflammation.

FIGURE 1.

Splice switching oligonucleotide (SSO) correction of the cystic fibrosis transmembrane conductance regulator (CFTR) c.1679+1643G>T splice variant. a, b) Clinical presentation of a cystic fibrosis (CF) patient with CFTR c.1679+1643G>T/G542X variants. a) Posteroanterior and lateral chest radiograph images demonstrating bronchiectasis with associated nodular opacities; b) schematics illustrating the effects of the intronic splicing defect c.1679+1643G>T and correction of the CFTR pre-mRNA with a SSO; the position of the PCR primers (F and R) are also shown (not to scale). The intronic splicing mutation (X) elicits aberrant splicing of intron 12 with the insertion of a 53 base-pair pseudo-exon (ψ, orange box) into the processed mRNA (top); this mutation produces not only the insertion of a non-coding sequence into the transcript but also a frameshift, which reveal early termination codons, a signal for the “nonsense mediated mRNA decay pathway”, an effective mechanism for limiting the translation of abnormal proteins. Abnormal CFTR transcript translates to deficient/defective CFTR protein and function, eliciting CF disease. A specific SSO (i.e. SSO-1 or SSO-2) can hinder the recognition of the mutant splicing site by the spliceosome, allowing for the expression of a wild-type (WT) spliced transcript (bottom), which translates into a WT functional CFTR protein and correction of the CF phenotype. c, d) SSO-1 correction of c.1679+1643G>T in human nasal epithelial cells (HNEC) from a patient. HNEC c.1679+1643G>T/G542X obtained during sinus surgery from the patient described in panels a and b were expanded, plated on permeable supports (250 000 cells·cm⁻²) at passage 3, and differentiated at air–liquid interface (ALI) conditions for 3–4 weeks following our published protocols [3, 7]. SSO-1 and a control SSO (SSOc), which has a scrambled nucleic base sequence, were synthesised with peptide-phosphorodiamidate (PPMO) chemistry as we previously described [3] (Gene Tools). SSO-1/PPMO and SSOc/PPMO were added directly in the media without any enhancing delivery vehicle or compounds onto well-differentiated HNEC on day 1, via the basolateral side (SSO-1=0.5 and 1 nmol, 12 h, basolateral 0.9 mL·cm⁻²) or apical side (SSO-1=0.5 nmol, 3 h, apical 0.09 mL·cm⁻²). On day 2, some cultures were treated with elexacaftor/tezacaftor/ivacaftor (ETI) (3 μM VX-445, 10 μM VX-661, 1 μM VX-770, 48 h, as we previously described [3, 7]). 72 h after SSO treatment, HNEC were subjected to CFTR RNA analysis via reverse transcriptase (RT)-PCR with primers recognising the exonic sequences flanking the mutant intron 12 (exons 12 and 14, shown in panel b), and CFTR function via Ussing chamber analysis following published protocols [3, 7]. c) Representative RT-PCR gel images are shown; CFTR upper and lower bands correspond to the intron 12 mutant (mut) and WT splice variants, respectively. The loss of the mutant upper band and the parallel increase in intensity of the WT lower band indicate effective dose-dependent splicing correction with SSO-1 treatment (maximal activity at 1 nmol, ∼1 µM). Neither ETI or the control SSOc facilitate splicing correction. The observed low level of WT lower band in samples not treated with SSO-1 could be the result of spontaneous normal splicing of c.1679+1643G>T and/or the detection of the G542X allele. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was the internal control (n=4–6). d) CFTR function analyzed as short circuit currents (I_sc) 72 h after SSO treatment; Ussing chamber analyses recorded the responses to sequential additions of amiloride (100 μM), forskolin (10 μM), VX-770 (1 μM), and CFTR inhibitor 172 (inh172) (10 μM) as shown in representative I_sc traces. Note: the CFTR function traces show that SSO-1-rescued CFTR I_sc was slightly enhanced by acute addition of potentiator VX-770/icenticaftor (1 μM), but not by chronic ETI addition (see panels e and f). The bar graph represents the post-amiloride, forskolin-induced I_sc peak data (mean±sd, n=3–4, p<0.01) in CF patient HNEC c.1679+1643G>T/G542X treated with SSO-1/PPMO or SSOc/PPMO at the indicated concentrations (0.5 or 1 nmol, 3 h, either basolateral 0.9 mL·cm⁻² or apical 0.09 mL·cm⁻²). Efficient dose-dependent rescue of CFTR I_sc was observed with SSO-1 but not with ETI, vehicle or SSOc/PPMO in the CF patient's HNEC. Both basolateral and apical SSO-1/PPMO delivery were efficacious in correcting the CFTR defect. For reference, the dotted line on the bar graph represents forskolin responses achieved in HNEC F508del/F508del treated with VX-770+VX-809 (components of Orkambi, comparable experimental conditions, previously published [7]). Transepithelial resistances did not significantly change between treatment groups (1712±107 versus 1718±72 Ω·cm² for vehicle and SSO-1, respectively). e, f) SSO correction in inflamed HNEC. CF-like epithelial inflammation was induced either with interleukin (IL)-1β or supernatant of mucopurulent material from airways of excised CF lungs (SMM) as we previously described [8, 9]. Well-differentiated HNEC c.1679+1643G>T/G542X were treated with IL-1β (10 ng·mL⁻¹, 24 h, basolateral 0.9 mL·cm⁻²) or SMM (50% dilution, 96 h, apical 0.09 mL·cm⁻²) on day 0. SSO-1/PPMO was administered at submaximal concentration (0.5 nmol, 12 h, basolateral 0.9 ml·cm⁻²) on day 1. In some studies, ETI (48 h) were administered on day 2. CFTR RNA and channel function analyses were performed 72 h after SSO treatment as above. The bar graphs represent: e) the RT-PCR ratiometric analysis calculated as the ratio of the intensity of lower band/(lower + upper bands)×100, calculated as previously described [3]; and f) the post-amiloride, forskolin-induced I_sc peak; mean±sd, n=4. Both, IL-1β and SMM-induced epithelial inflammation enhanced SSO-1 efficacy on restoring CFTR mRNA splicing and channel activity. g, h) A rapid test/proof-of-concept study identified novel SSO-2 for correction of c.1679+1643G>T and a second intron 12 splice variant. The goal of this study was to prove that one oligonucleotide can efficiently correct the main target, c.1679+1643G>T and other intron 12 splicing variant(s); e.g. c.1679+1634A>G. A rapid assay based on reverse transfection was optimised to test novel oligonucleotide sequences utilising simpler chemistry (2'-O-methoxyethyl phosphorothioate, Integrated DNA Technologies) and small quantities of patient cells. Novel SSO-2 designed to target c.1679+1643G>T was synthesised with 2'-O-methoxyethyl phosphorothioate chemistry (Integrated DNA Technologies) and tested in patient cells with either intron 12 splicing mutation. g) HNEC c.1679+1643G>T/G542X and h) human bronchial epithelial cells (HBEC) c.1679+1634A>G/F508del (de-identified HBEC from a CF donor obtained through the Cystic Fibrosis Foundation) were treated in suspension with a single dose of SSO-1 or SSO-2 (0.5 nmol for 24 h) utilising Lipofectamine RNAiMAX; vehicle and a non-targeted SSO (SSOc) were utilised as controls. SSO-treated cells were then plated on permeable supports (without SSO) for differentiation at ALI following our published protocols [3, 7] to perform RNA analysis (top; RT-PCR gel image at 1 and 6 days post-SSO treatment) and Ussing chamber measurements (bottom; I_sc traces and the bar graph representation of the post-amiloride, forskolin-induced I_sc peak at 15 days post-SSO treatment; mean±sd, n=4). Note the intense lower band (WT) in the RT-PCR gel of HBEC c.1679+1634A>G/F508del (SSOc and vehicle lanes) due to normal splicing of intron 12 of the F508del allele RNA; however, the loss of the upper, mutant band and the parallel enrichment of the lower, WT band are clearly observed with SSO-2 treatment. SSO-2 was effective in correcting the abnormal CFTR mRNA splicing and channel activity in both CF splice variants: our main target c.1679+1643G>T and a secondary target, the neighbour splicing mutation c.1679+1634A>G. Patient cells and data were obtained and utilised in accordance with the University of North Carolina Institutional Review Board, State, and Federal regulations. **: p<0.01.

CFTR RNA splicing and channel function [3] were studied in the patient's well-differentiated HNECs [7]. Reverse transcriptase (RT)-PCR analysis showed a prominent larger band resulting from the splice mutation-elicited inclusion of the pseudo-exon (figure 1c, vehicle). A smaller, lower intensity band with a size corresponding to wild-type CFTR likely resulted from spontaneous normal splicing [1, 3] and/or normal splicing of intron 12 in the G542X allele. Sequencing of the mutant band confirmed the splicing defect and inclusion of the 53 base-pair pseudo-exon [5]. Electrophysiological analysis [7, 10] showed no post-amiloride, forskolin-stimulated short-circuit currents (I_SC), indicating lack of CFTR activity from either CFTR allele (figure 1d, vehicle). Elexacaftor/tezacaftor/ivacaftor did not rescue mRNA defects and thus, failed to restore CFTR c.1679+1643G>T function (figure 1c–f).

We tested a strategy based on splice switching oligonucleotides (SSOs) to correct one of the CF alleles. Therapeutic SSOs, which have emerged as novel therapeutics for genetic diseases [11, 12], are short, single-stranded synthetic nucleic acids that modulate mRNA splicing by base-pairing to the pre-mRNA and steric blocking. SSOs are specific, well-tolerated, do not require viral or nonviral vehicles for delivery, can be administered systemically or via inhalation, and are ideal to correct CFTR splicing mutations or bypass some exonic mutations in precision medicine therapeutic applications [3, 11, 12]. Indeed, a SSO for the CFTR splicing defect 3849+10kbG>T is currently completing phase 2 clinical trials [13]. Several SSOs have been approved by the US Food and Drug Administration for other rare genetic disorders, including oligonucleotides tailored to one individual [11, 12].

The main objective of this study was to demonstrate the correction of the c.1679+1643G>T splicing variant with a novel SSO. SSO-1 was synthesised with peptide-phosphorodiamidate (PPMO) chemistry [3]; PPMOs combine the long-lasting corrective activity of phosphorodiamidates-morpholinos with the peptide-driven efficient cellular delivery in the lung and airway epithelia versus other SSO chemistries [3, 11]. A single treatment of SSO-1/PPMO to the patient's HNEC corrected the mRNA splicing defect and restored forskolin-stimulated I_sc in a dose-dependent manner with maximal activity at ∼1 µM; SSO-1 was efficacious added either to the basolateral or apical culture sides (figure 1c, d). SSO-1 was highly effective alone and did not require elexacaftor/tezacaftor/ivacaftor treatment to reach maximal CFTR c.1679+1643G>T splicing correction and activity (figure 1c–f). Maximal SSO-1-elicited CFTR I_sc, which reflected correction of only one CFTR allele was ∼70% of wild-type HNEC (I_sc=22.04±1.76 μA·cm⁻² [7]), and larger than Orkambi-rescued F508del/F508del I_sc (lumacaftor+VX-770=10–20% wild-type CFTR I_sc, tested in comparable experimental conditions [7], figure 1d), suggesting potential clinical benefit for SSO-1.

Epithelial inflammation modulates mRNA splicing, and specifically regulates CFTR biosynthesis, activity, and its responses to modulators [1, 14, 15]. To understand the effect of inflammation on SSO efficacy, we utilised two well-characterised CF airway inflammatory stimuli, i.e. interleukin-1β (IL-1β) and supernatant of mucopurulent material from airways of excised CF lungs (SMM) [8, 9]. Figure 1e, f illustrate IL-1β or SMM moderately increasing the levels of the wild-type splice variant mRNA (22–25%) in response to SSO-1, coupled with increased forskolin-stimulated I_sc (IL-1β (∼85% increase); SMM (∼95% increase)). We observed similar inflammation effects on the efficacy of SSOs targeting W1282X and 3849+10kbC>T CFTR mutations [16]. The data suggest that inflammation affects SSO-rescued CFTR at the mRNA and post-translational levels, and indicate the therapeutic potential of SSOs towards this patient's mutation enhanced by airway inflammation.

We sought to identify one SSO that could target not only c.1679+1643G>T but also other splice defects within intron 12. The c.1679+1634A>G splicing mutation [6] is located nine base pairs upstream from the c.1679+1643G>T defect [5]. We obtained human bronchial epithelial cells from a pwCF carrying the c.1679+1634A>G and the F508del mutations in separate alleles. First, we tested SSO-1; however, SSO-1 did not restore normal splicing or CFTR function in the c.1679+1634A>G variant (figure 1h), likely due to ineffective binding of SSO-1 to the c.1679+1634A>G mRNA. A newly designed oligonucleotide sequence (SSO-2) was tested utilising a rapid assay optimised for small quantities of patient cells and common chemistry oligonucleotides for proof-of-concept studies. SSO-2 was administered as a single treatment to suspensions of human bronchial epithelial cells or HNEC before cell differentiation. SSO-2 efficiently restored wild-type CFTR mRNA splicing and forskolin-elicited I_sc for both the main c.1679+1643G>T and secondary c.1679+1634A>G targets (figure 1g, h).

In conclusion, CFTR splice variants unresponsive to modulators can be efficaciously corrected with SSOs targeting a specific CFTR mutation or exon/intron regions. Designing a SSO that targets more than one CF mutation (e.g. within an intron) would increase the number of pwCF benefiting from a single SSO drug, improving its chances for clinical development. SSOs with PPMO chemistry are soluble and can be administered systemically [3], or by inhalation for improved airway targeting and reduced systemic toxicities. Our findings open avenues for the development of SSOs/PPMOs as novel CF therapeutics and indicate that inflammation-enhanced SSO efficacy must be a component of theratyping studies.