Upregulation of a CFTR mRNA isoform has therapeutic potential for the treatment of 3′ CFTR PTC variants

Abstract

Nonsense or premature termination codon (PTC) variants of the CFTR gene are pathogenic and found in ∼10% of North American people with cystic fibrosis. In addition to encoding incomplete proteins, PTC variants induce nonsense-mediated mRNA decay (NMD), leading to ∼80%–90% reduction in full-length mRNA. This reduction is a contributor to PTC mutation-related pathology. E22 trunc is a naturally occurring truncated CFTR mRNA that terminates before the W1282X PTC variant and is resistant to NMD. To induce its expression, antisense oligonucleotides (ASOs) were tiled across intron 22 splice donor (SD) and splice acceptor (SA) sites. Top SD/SA ASO pairs were assessed for their impact on e22 trunc mRNA, e22 trunc protein, and CFTR-mediated chloride (Cl⁻) transport in immortalized and primary human bronchial epithelial (hBE) cell cultures. We demonstrate that e22 trunc mRNA generates a truncated CFTR protein whose Cl⁻ transport function can be enhanced with elexacaftor/tezacaftor/ivacaftor (ETI) treatment. ASO and ETI treatment in combination restore ∼20% and 25% of wild-type CFTR Cl⁻ transport function in immortalized epithelial and primary hBE cells homozygous for CFTR W1282X, respectively. This study provides a foundation for advancing ASO-mediated upregulation of e22 trunc mRNA and protein as a therapeutic approach for cystic fibrosis caused by 3′-terminal CFTR PTC mutations.

Graphical abstract

Antisense oligonucleotides targeting CFTR splicing upregulate a naturally occurring, NMD-insensitive mRNA isoform (e22 trunc), enabling partial functional rescue of CFTR in W1282X models. Combined with modulators, this strategy restores up to 25% wild-type activity, offering a novel therapeutic approach for cystic fibrosis caused by 3′ CFTR nonsense mutations.

Introduction

Cystic fibrosis (CF) is a severe autosomal recessive disease caused by mutations in the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR) protein, which mediates Cl⁻ ion transport in different tissues. The most common pathological features are pancreatic insufficiency and bronchiectasis with chronic airway infections resulting in respiratory failure and premature death.¹ Previously, only symptomatic treatments were available to people with CF (pwCF), such as pancreatic enzymes to aid digestion, airway clearance techniques to reduce mucus accumulation, and antibiotics for pulmonary infections.² Over the last decade, a new class of drugs, CFTR modulators, that target the underlying molecular defect(s) to improve channel gating and/or CFTR trafficking to the plasma membrane have been employed therapeutically. As of September 25, 2024, 1,167 variants have been reported by the Clinical and Functional Translation of CFTR (CFTR2) project: https://cftr2.org; 1,085 of these variants are classified as CF-causing. Based on genotype, 93%–94% of pwCF in North America could benefit from the modulator therapies^3,4; however, the remaining 6%–7% have CFTR variants, including nonsense mutations, for which current modulator treatments are not approved.

Nonsense variants are difficult to treat therapeutically. The mRNA harboring such variants is subject to nonsense-mediated mRNA decay (NMD), reducing the level of available transcript (down to 10%–20% of wild-type [WT] levels for many CFTR premature termination codon [PTC] variants). Truncated proteins translated at low levels from the remaining mRNA are often nonfunctional, rapidly degraded by the proteasome and, specifically in the case of CFTR, do not respond to modulators. An exception is the CFTR W1282X variant, whose protein retains residual function when expressed at the cell surface in sufficient quantities.^5,6

From CFTR mRNA isoform analysis, we have identified a low-abundant, naturally occurring alternative isoform that terminates in intron 22 (upstream of W1282X), which we term “e22 trunc mRNA.” This mRNA species arises from skipping the splice donor (SD) site at the 3′ end of exon 22 followed by the usage of an alternative polyadenylation (ApA) site located inside intron 22 (whereas the major full-length CFTR mRNA transcript is a product of canonical exon 22/23 splicing). Translation of e22 trunc mRNA results in a protein that is truncated at CFTR amino acid 1,239 (the end of exon 22), followed by 9 novel amino acids encoded from the retained portion of intron 22. We have observed e22 trunc mRNA in various primary and immortalized airway epithelial cells and primary intestinal cells at a relative abundance of ∼5%–10% compared to full-length CFTR mRNA (Table S1).

Since e22 trunc mRNA isoform has an in-frame stop codon in the retained part of intron 22 and terminates at an ApA in the same intron (i.e., before the W1282X position) with no additional splicing downstream of the stop, it does not exhibit characteristics of a PTC. This led us to predict that, in contrast to the fully spliced, full-length mRNA isoform harboring the W1282X variant (FL-W1282X), e22 trunc is not subject to NMD. We further hypothesized that e22 trunc protein may have partial function and could provide therapeutic benefit if expressed at sufficient levels.

Here, we show that ASO-mediated blockage of CFTR exon 22/23 splicing enhances intron 22 ApA usage and increases expression of this NMD-insensitive e22 trunc mRNA. Furthermore, the resultant truncated protein product is trafficked to the plasma membrane and exhibits anion channel activity that can be ∼10-fold enhanced by treatment with a CFTR potentiator. This approach may provide clinical benefit for pwCF harboring certain 3′ PTC variants, e.g., W1282X, Q1313X, and E1371X.

Results

Intron 22 ApA usage results in a naturally occurring, low-abundant CFTR mRNA isoform that ends after exon 22, is insensitive to NMD, and encodes a truncated protein

A partial, exon 22-truncated mRNA isoform of CFTR was initially identified by Burch,⁷ and subsequently, additional isoforms truncated after exon 22 were reported in the Ensembl database: ENST00000648260.1 and ENST00000649406.1. We also detected a truncated mRNA isoform in 16HBE14o- cells utilizing 3′ rapid amplification of cDNA ends (RACE) with a 5′ primer targeting exon 8, followed by long-read sequencing. Consistent with the transcripts reported by Burch and in the Ensembl database, exon 22-truncated CFTR transcripts extended approximately 140 bp into intron 22. We observed processed truncated mRNA that terminated with 10–30 post-transcriptionally added adenosine residues that align with putative consensus hexanucleotide ApA and dinucleotide cleavage sites in intron 22, consistent with utilization of those sites during transcription (Figure S1). We designated these transcripts as “e22 trunc.” Subsequently, long-read cDNA sequencing of 3′ RACE CFTR amplicons generated with a 5′ primer targeting exon 1 was employed to confirm the presence of “complete” (containing all of the exons 1–22) e22 trunc mRNA in 16HBE14o- cells (Figure 1; Table S2). To the best of our knowledge, this study presents the first report of an exon 22-truncated CFTR mRNA containing all 22 upstream exons.

Figure 1.

CFTR genomic locus, e22 trunc mRNA, and protein

(A) CFTR genomic locus with intron 22 ApA sites. Introns not shown to scale. (B) e22 trunc mRNA representation showing alternative 3′ UTR and poly A tail. (C) e22 trunc protein representation showing additional nine alternative amino acids and missing NBD2.

Translation of e22 trunc mRNA produces a truncated CFTR protein (1239 + 9 aa) that retains transmembrane domains 1 and 2 (TMD1 and TMD2), nucleotide-binding domain 1 (NBD1), R domain, and partial NBD2, followed by nine additional amino acid residues (VRFEHCLLC) encoded in the retained part of intron 22. In the mRNA sequence, those nine intron-derived codons are followed by a stop codon, turning the rest of the retained ∼140 bp intronic sequence (up to the ApA site) into 3′UTR. With no additional splicing events occurring downstream of that stop, we hypothesized that the e22 trunc isoform (which terminates before the location of W1282X) would be NMD-insensitive in contrast to full-length CFTR W1282X mRNA. We quantified FL-WT, FL-W1282X, and e22 trunc mRNA expression levels in primary human bronchial epithelial (hBE) cells at air-liquid interface (ALI), intestinal organoids (IOs), and immortalized airway cells with WT or W1282X^+/+ CFTR genotypes (Table S3). As expected, FL-W1282X mRNA levels were reduced by approximately 80%–90% (85.6% ± 5.7%; range, 80.1%–92%) relative to the full-length mRNA levels observed in WT cells, while e22 trunc mRNA exhibited similar, albeit low expression, levels (8.4% ± 2.5% relative to FL-WT; range, 6.1%–9.3%) in both WT and W1282X cells.

Therapeutic rationale for upregulation of a CFTR mRNA isoform for the treatment of 3′ CFTR PTC variants

Given that e22 trunc mRNA is expected to be insensitive to NMD, we aimed to quantify the half-life (t_1/2) of FL-WT, FL-W1282X, and e22 trunc transcripts (Figure 2A) in an actinomycin D time course experiment. As shown in Figure 2B, the t_1/2 of e22 trunc mRNA was reduced compared to FL-WT mRNA (3.9 ± 0.2 h vs. 12.6 ± 1.8 h) but was greater than FL-W1282X mRNA (<<2 h). It has been established that the CFTR_trunc-1281 protein, which lacks the majority of NBD2, is transported to the apical membrane and maintains partial channel function.^6,8 To investigate whether the e22 trunc protein also retains function, e22 trunc cDNA was expressed in Fischer rat thyroid (FRT) cells, and CFTR-mediated Cl⁻ transport was evaluated.

Figure 2.

Therapeutic rationale for upregulation of a nonsense-mediated decay (NMD)-insensitive CFTR mRNA isoform for the treatment of 3′ CFTR PTC variants

(A) Diagram of FL-WT, FL-W1282X, and e22 trunc mRNA isoforms. (B) Kinetics of the mRNA isoform abundances measured after treatment with actinomycin D; fitted exponential decay curves (dashed) with 95% confidence interval of the prediction (gray-filled bands), p values of the nonlinear least-squares fits and the fitted half-life times (t_1/2) with 95% confidence intervals are also shown. The NMD-driven decay of FL-W1282X transcript is too fast for a meaningful determination of its half-life on the timescale used in the experiment. (C) CFTR Cl⁻ transport (conductance, G_t) time course measured in FRT cells overexpressing WT CFTR cDNA (vehicle only) or F508del, C832X, and e22 trunc cDNAs, in vehicle (−) or treated for 24 h with 3/3 μM ELX/TEZ (+); 1 μM IVA acutely added during TECC-24 assay, as indicated. FRT parental, 96 h post-transfection, treatment (apical and basolateral) 24 h prior to assay. (D) Summary transport activity metric extracted from time course data shown in (C): defined as ΔInhibitor (total change in conductivity [mS/cm²] in response to addition of CFTR inhibitors, calculated relative to the plateau values reached after forskolin [Fsk] + IVA treatments). (E) Transepithelial electrical resistance, a widely accepted measurement reporting on the integrity of tight junctions in cell culture models of epithelial monolayers, was unaffected by WT, F508del, C832X, or e22 trunc transfections or treatment with 3/3 μM ELX/TEZ. (F) The rationale for exon 22/23 splice blocking to promote e22 trunc mRNA isoform as a therapeutic strategy for 3′ terminal CFTR PTCs downstream of exon 22. We hypothesize that inhibition of exon 22/23 splicing leads to retention of intron 22, including its ApA sites, thereby making them accessible for ApA usage. Utilization of intron 22 ApA sites generates a truncated mature CFTR mRNA lacking exon junction complexes downstream of the stop codon, allowing it to evade NMD. Consequently, this mechanism results in elevated levels of truncated CFTR transcripts. pA, polyadenylation site.

As shown in Figures 2C and 2D, cells transfected with e22 trunc cDNA exhibited CFTR-mediated Cl⁻ conductance of 1.41 ± 0.10 mS/cm² upon treatment with elexacaftor (ELX)/tezacaftor (TEZ)/ivacaftor (IVA) (ETI) (24-h incubation with 3/3 μM ELX/TEZ followed by in-assay addition of CFTR potentiator IVA). This was significantly greater than for cells expressing full-length cDNA harboring PTC variants, W1282X or C832X (conductance of 0.16 ± 0.03 and 0.05 ± 0.01 mS/cm², respectively; p < 0.001 in 2-tailed Student’s t test). The conductance was fully inhibitable with CFTR inhibitors CFTRinh-172 and GlyH-101, confirming that the observed Cl⁻ conductance was mediated by the truncated CFTR protein translated from the e22 trunc cDNA. Transfections of FL-WT (positive control) and F508del (ETI-responsive CF variant) cDNA were performed for comparison. In all transfection conditions, the baseline transepithelial electric resistance exceeded the quality control threshold of 500 Ω/cm², suggesting formation of tight junctions and maintenance of cell layer integrity (Figure 2E).

Upon confirming that e22 trunc mRNA demonstrates improved t_1/2 compared to FL-W1282X and that e22 trunc protein retains partial function, we sought to determine if the levels of e22 trunc mRNA and of the resulting protein could be increased in the endogenous genomic context for the therapeutic benefit of pwCF harboring the W1282X variant.

We hypothesize that ApA sites within intron 22 compete with canonical exon 22-exon 23 splicing; therefore, disruption of splicing between the intron 22 SD and its acceptor site as well as all downstream acceptor sites (introns 23–27) is expected to promote premature transcription termination within intron 22. We further hypothesized that this mechanism would result in maximal upregulation of the truncated e22 mRNA isoform and its corresponding protein product (Figure 2F).

To test this hypothesis and force the maximum achievable levels of e22 trunc isoform, we generated by gene editing a 16HBE14o- cell line, in which we deleted part of the CFTR genomic DNA starting in intron 22 just downstream of the 11 putative ApA sites and through ∼159 bp past the end of the native 3′ UTR (16HBEge-ΔE23-3′ UTR cell line, Figure 3A). PCR screening was used to identify nine clones with the correct Δexon 23-3′ UTR deletion junction. E22 trunc mRNA expression levels of the clones ranged from 0.5× to 1.2× of parental full-length CFTR mRNA levels, suggesting that in the absence of competing splicing events, the ApA sites in intron 22 can be used fully and efficiently. Three 16HBEge-ΔE23-3′ UTR clonal cell lines (2-H07, 3-B09, and 3-D01) with e22 trunc levels similar to parental WT-FL mRNA levels were selected for further characterization of CFTR mRNA, protein expression, and Cl⁻ transport function. The ∼12-fold increase in e22 trunc mRNA expression in the three selected clones compared to the parental e22 trunc levels was significant at p < 0.001 in a 2-tailed Student’s t test (Figure 3B). Western blot analysis revealed the presence of the shifted molecular weight bands consistent with partially (Band B) and fully (Band C) glycosylated e22-truncated CFTR protein (Figures 3C and 3D). As illustrated in Figure 3E, Fsk-stimulated CFTR Cl⁻ transport function in the 16HBEge-ΔE23-3′ UTR clonal cell lines was modest; however, treatment with CFTR correctors and acute addition of a CFTR potentiator resulted in an ∼2-fold enhancement in channel function. The results for dose escalation of IVA (0.1–10 μM) in the 2-H07 clone are shown in Figure 3F, where CFTR Cl⁻ transport activity is represented as AUC/min Fsk+IVA, the area under the portion of the trace calculated starting at the IVA addition time point (Figure 3E) normalized to the measurement time interval between IVA and CFTR inhibitor additions. Increased CFTR activity was observed at all doses relative to untreated ± ELX/TEZ (3/3 μM) (p < 0.005, 2-tailed Student’s t test). The maximum response of ∼42% WT function was observed at ELX/TEZ (3/3 μM) + IVA (3 μM) in the 2-H07 clone. These results suggest that although e22 trunc proteins have gating and, potentially, trafficking defects, CFTR modulators can significantly improve their channel function to therapeutically meaningful levels, well beyond what is currently achievable, starting with the native, full-length mRNA harboring PTC variants such as W1282X.

Figure 3.

Inhibition of exon 22 splicing forces intron 22 ApA usage and elevates e22 trunc mRNA and protein levels that can be corrected with CFTR modulators to therapeutically relevant levels

(A) Edited CFTR gene (ΔE23-3′ UTR) to “force” e22 trunc mRNA expression in 16HBE14o- parental cells. (B) ddPCR absolute mRNA copies of FL-WT exon 25/26 (black) and e22 trunc (gray) from 16HBE14o- parental cells and 16HBEge-ΔE23-3′ UTR- clonal lines 2-H07, 3-B09, and 3-D01. (C) Western blot analysis of 16HBE14o- parental line ± ELX/TEZ and clones 2-H07, 3-B09, and 3-D01 ± ELX/TEZ. Antibody UNC 596 detects CFTR; ACTB serves as a loading control. (D) Western blot of PNGaseF-treated (de-glycosylated) FL-WT, FL-W1282X, and e22 trunc cDNA overexpression from HEK293 cells and 16HBEgeΔE23-3′ UTR-2-H07 and 16HBE14o- parental cells. Samples were diluted to produce equivalent band intensity. (E) Electrophysiological traces of 16HBEgeΔE23-3′ UTR gene-edited cell line clones 2-H07, 3-B09, and 3-D01 from TECC-24 assay after DMSO (vehicle) treatment (three upper panels) or 48 h treatment with ELX/TEZ (3/3 μM) (three bottom panels). All samples treated in-assay with IVA (1 μM). (F) TECC-24 I_eq assay results of 16HBE14o- (gray bar) and 16HBEge W1282X (light gray bars) and IVA dose escalation in 16HBEgeΔE23-3′ UTR-2-H07 (black bars) ± ELX/TEZ 3/3 μM. AUC FSK+IVA is area under the curve calculated starting at the IVA addition time point in electrophysiological traces as shown in (E).

ASO blockade of exon 22/23 splicing promotes intron 22 ApA usage and increases e22 trunc mRNA and protein levels, which can be functionally rescued in the 16HBEge W1282X cell line model of CF

Subsequently, we aimed to extend our research to a therapeutically viable approach. We postulated that antisense oligonucleotide (ASO) blockade of intron 22 SD and splice acceptor (SA) would inhibit intron 22 removal and promote utilization of intron 22 ApA sites during pre-mRNA processing. We sought to identify optimal SD- and SA-blocking ASOs that facilitated the expression of e22 trunc mRNA and designed 21 overlapping ASOs in 1-nucleotide “steps” across the intron 22 SD (n = 11) and SA sites (n = 10) (Figure S2). The SD and SA ASOs were evaluated for their capacity to increase e22 trunc mRNA in 16HBE14o- cells. The most effective ASOs, SA-08 and SD-18, yielded e22 trunc mRNA levels measured at 36.9% ± 1.3% (4-fold increase over untreated, p < 0.001, 2-tailed Student’s t test) (Figures S3A and S3B) and 25.4% ± 1.8% (2.8-fold over untreated, p < 0.001, 2-tailed Student’s t test) (Figures S4A and S4B) of the FL-WT mRNA level in untreated 16HBE14o- cells, respectively, and were utilized for subsequent experiments.

We evaluated the effects of SD-18 alone or in combination with SA-08 on e22 trunc and FL-W1282X mRNA expression in 16HBEge W1282X cells (Figure 4A). Neither FL-W1282X nor e22 trunc mRNA exhibited significantly different log2 fold-change (log-FC) between (−) ASO_Scrambled (100 μM) control and vehicle control (p > 0.05, 2-tailed Student’s t test). All SD-18 and SD-18/SA-08 combination treatments increased e22 trunc mRNA levels relative to vehicle control: SD-18 (10 μM), 5.2-fold; SD-18 (100 μM), 5.9-fold; SD-18/SA-08 (10/10 μM), 7.4-fold; and SD-18/SA-08 (100/10 μM), 8.4-fold (p values< 0.05, 2-tailed Student’s t test). Unexpectedly, SD-18/SA-08 (10/10 μM) resulted in a larger fold increase in e22 trunc mRNA than SD-18 (100 μM) treatment. At the same time, the expression levels of FL-W1282X under those treatment conditions did not exhibit statistically significant log-FCs vs. the vehicle (p > 0.1, 2-tailed Student’s t test).

Figure 4.

ASO blockade of exon 22/23 splicing promotes intron 22 ApA usage and increases e22 trunc mRNA and functionally rescuable protein levels in the 16HBEge W1282X cell line model of CF

(A) e22 trunc (dark gray) and FL-W1282X (light gray) mRNA quantitation via ddPCR as described above from 16HBEge W1282X cells post-TECC-24 assay with DMSO, (−) ASO, SD-18, or SD-18/SA-08 48-h treatments. All samples were treated in-assay with IVA (3 μM). The data are averaged between two experimental replicates, each with three technical repeats (see materials and methods) and shown as fold-change of each species, under each treatment relative to e22 trunc mRNA in untreated control. The error bars show the standard error of the weighted average. (B) Western blot analysis from 16HBEgeW1282X cells post-TECC-24 assay as described above. Full-length CFTR, ΔEx23 (CFTR with exon 23 deleted), FL-W1282X (truncated at W1281), e22 trunc, and blank (non-transfected) constructs overexpressed in HEK293 cells as sizing controls (lanes 1–5). 16HBEge-W1282X incubated with (−) ASO_Scrambled (10 μM), SD-18 (10 μM), or SD-18 (10 μM), and SA-08 (2 μM) treated for 48 h ± ELX/TEZ (3/3 μM) (lanes 6–11). CFTR antibody UNC 596 was used to detect CFTR, and beta actin was used as loading control. All lysates were incubated with PNGaseF. Note the presence of the translation product of W1282X mRNA (upper band) and e22 trunc (lower band) in the (−) ASO controls and the presence of only e22 trunc with ASO treatments. (C) TECC-24 assay from CFF-16HBEge W1282X cells treated with vehicle, (−) ASO_Scrambled, SD-18, SA-08, or SD-18 and SA-08 (black bars) for 48 h and also ± ELX/TEZ (3/3 μM) treated for 48 h. All samples received in-assay addition of IVA (3 μM). CFTR function is represented as AUC/min FSK+IVA (see text). Error bars show standard error. (D) Correlation between the level of e22 trunc mRNA (measured by ddPCR) and CFTR function (measured in TECC assay and represented as AUC/min FSK+IVA). The data points are from the same controls or treatments as shown in (A and C). The AUC/min values are those measured in the TECC assay and shown in (C). The e22 trunc mRNA copy counts used in this plot are from one biological replicate of ddPCR experiment, in which + ELX/TEZ measurements were also available (not shown in the averaged data in A). The mRNA copies are normalized to the count of FL-1282X, the same way as in (A).

In an alternative approach, we ran non-parametric Jonckheere-Terpstra (JT) tests on log-FCs of the studied mRNA isoforms at different SD-18 levels vs. their respective levels in vehicle. JT evaluates the null hypothesis of no trend in the data vs. the alternative of ordered group means (μ_control<μ_dose=10<μ_dose=100). It is an omnibus test that utilizes all treatment groups in a single calculation and thus avoids multiple testing issues. The p values for e22 trunc and FL-W1282X were p < 0.01 and p = 0.5, respectively, indicating a statistically significant trend of increasing RNA abundance with increasing ASO dosage for the former species, but no such trend for the latter.

Finally, we also applied a weighted linear model for log-FC (see materials and methods) to assess potential interaction between SA-08 and SD-18. The main effects of SD-18 and SA-08 on the fold change of e22 trunc were both significant at p < 0.05, while the interaction was insignificant (p > 0.9); therefore, the available data do not suggest any synergy beyond a simple additive effect of SA-08. Both the main effects and the interaction were insignificant in the model fitted for log-FC of the FL-W1282X (p > 0.5), in line with no significant effect of our ASOs on FL isoform observed with other tests.

To rule out the possibility that ASO treatments induced unspliced intron 22-retained transcripts, one of the repeats of this experiment was performed with an intron 22/exon 23 junction probe included in the droplet digital PCR (ddPCR) assay (Figure S5), resulting in very low expression levels detected with the latter probe in vehicle or (−) ASO_Scrambled treatment controls; those levels remained unchanged under treatments with SD-18 alone or in combination with SA-08 (p > 0.05, 2-tailed Student’s t test).

To assess the effect of ASO treatments on e22 trunc protein, CFF-16HBEgeW1282X cells were treated with (−) ASO_Scrambled (10 μM), SD-18 (10 μM), or SD-18/SA-08 (10/2 μM) ASOs for 48 h ± ELX/TEZ (3/3 μM) treatment. Western blots of CFTR protein from PNGaseF-deglycosylated cell lysates were compared to full-length CFTR (∼170 kDa, 1,480 amino acids [aa]), CFTR Δexon23 (ΔEx23) (∼157 kDa, 1,428 aa), CFTR_trunc-1281 (∼141 kDa, 1,281 aa), and e22 trunc (∼138 kDa, 1,258 aa) HEK293 cDNA overexpression constructs (lanes 1–4) (Figure 4B). Faint bands aligning with e22 trunc and CFTR_trunc-1281 overexpression controls were present in (−) ASO_Scrambled control cells independent of ELX/TEZ (3/3 μM) treatment (lanes 6 and 7). ASO treatments with SD-18 (10 μM) (lanes 8 and 9) or SD-18/SA-08 (10/2 μM) (lanes 10 and 11), ± ELX/TEZ (3/3 μM) incubation, resulted in a single high-intensity band that aligned with the ∼138-kDa e22 trunc protein from the HEK293 overexpression control (lane 4). ELX/TEZ (3/3 μM, 48 h) treatment had minimal effect on band intensity and had no effect on the size of the protein products.

The ASO-induced rescue of CFTR-mediated chloride transport in CFF-16HBEge W1282X cells was directly measured in the TECC (Trans-Epithelial Clamp Circuit )-24 assay (n = 3), the results are shown in Figure 4C (CFTR function is represented as AUC/min FSK+IVA, the area under the TECC trace curve after FSK+IVA additions, normalized to the time interval between IVA and inhibitor additions). The AUC/min values for samples treated with (−) ASO_Scrambled (100 μM) were 2.3 ± 0.2 and 1.6 ± 0.03 for ±ELX/TEZ, respectively, and were not meaningfully different from the values measured for vehicle treatment (3.3 ± 0.1 and 0.5 ± 0.1, respectively). All ASO treatments ± ELX/TEZ resulted in not only statistically significant, but much larger increases in CFTR-mediated Cl⁻ current relative to ASO control (p < 0.001, 2-tailed Student’s t test). Specifically, the average AUC/min values for SD-18 (10 μM), SD-18 (100 μM), SD-18/SA-08 (10/10 μM), and SD-18/SA-08 (100/10 μM) treatments were 5.5 ± 0.4, 4.6 ± 0.2, 9.3 ± 0.1, and 12.1 ± 0.6 without ELX/TEZ and 10.3 ± 0.7, 11.5 ± 0.3, 18.1 ± 0.8, and 21.9 ± 0.4 with ELX/TEZ (3/3 μM), respectively Of note, SD-18/SA-08 combination treatment even at 10/10 μM concentration resulted in a larger functional response than treatments with SD-18 alone up to 100 μM, which is consistent with mRNA response data (Figure 4A). To that end, we assessed correlations between CFTR function (measured in TECC assay and expressed as AUC/min FSK+IVA) and the amount of e22 trunc mRNA (measured in ddPCR assay and expressed as fold change vs. FL-W1282X in vehicle), with or without ELX/TEZ correctors, across all studied ASO treatments. The results are plotted in Figure 4D. On the log-log scale, the correlations are 0.95 (no correctors) and 0.97 (+ELX/TEZ), and they are highly significant (p = 0.003 and p = 0.002, respectively). The 90% confidence intervals of the slopes of the two least-squares linear fits shown in the plot are (0.24, 0.48) (no correctors) and (0.20, 0.37) (+ELX/TEZ); strong overlap between the intervals indicates that there is no evidence for the slopes being different even at the p = 0.1 significance level.

It is of particular importance that the maximum functional response observed in our experiments with SD-18/SA-08 (100/10 μM) + ELX/TEZ (3/3 μM) combination treatment was equivalent to ∼20% WT levels observed in 16HBE14o- parental cells.

In primary human bronchial epithelial cells homozygous for CFTR W1282X, ASO blockade of exon 22/23 splicing promotes intron 22 ApA usage, resulting in increased expression of e22 trunc mRNA and rescue of CFTR function

To extend our analysis into a more translationally relevant model, we employed the “gold-standard” CF model, primary hBE cells in ALI cultures. Unlike the previous immortalized CFF-16HBEge W1282X cultures, hBEs cells at ALI form an airway-like epithelium that creates a barrier to free uptake. Therefore, to achieve sufficient upregulation of e22 trunc mRNA, we found it necessary to subject cultures to long ASO exposures starting at the undifferentiated state and refreshing ASOs with each media change throughout 21 days of differentiation. Undifferentiated CFTR W1282X^+/+ hBE cells were treated with SD-18 (10 μM), SD-18 (100 μM), SD-18/SA-08 (10/10 μM), SD-18/SA-08 (100/10 mM), and (−) ASO_TNMD (100 μM) ASOs for 48 h and then differentiated for 21 days with media/ASO changes every 2 days. ASO treatment was discontinued 48 h before functional assays, and cells were harvested for protein and mRNA quantification.

Primary CFTR W1282X^+/+ hBE cells treated with (−) ASO_TNMD (100 μM) had 450 ± 26 and 745 ± 34 baseline absolute copies/20 ng total RNA of e22 trunc and FL-W1282X mRNA, respectively (Figure 5A). All SD-18, SA-08, or combination treatments significantly increased e22 trunc mRNA levels (p < 0.001, 2-tailed Student’s t test). SD-18 (10 μM) increased e22 trunc 3.7-fold and FL-W1282X mRNA remained unchanged. SD-18 (100 μM) not only increased e22 trunc mRNA 7.4-fold but also unexpectedly increased FL-W1282X mRNA. SD-18/SA-08 (10/10 μM) treatment increased e22 trunc mRNA 4.8-fold, and again, FL-W1282X mRNA remained unchanged. Last, consistent with the 16HBEge W1282X data, SD-18/SA-08 (100/10 μM) treatment resulted in the highest induction of e22 trunc, 8.2-fold, and did not significantly increase FL-W1282X mRNA levels (p > 0.05, 2-tailed Student’s t test).

Figure 5.

ASO blockade of exon 22/23 splicing promotes intron 22 ApA usage and increases e22 trunc mRNA and functionally rescuable protein levels in primary hBE cells homozygous for CFTR W1282X

(A) FL-W1282X (light gray) and e22 mRNA trunc quantitation (dark gray) via ddPCR from fully differentiated CFTR W1282X^+/+ hBE cells at ALI after TECC-24 assay. Cells were treated with (−) ASO_TNMD, SD-18, SA-08, or SD-18/SA-08 for 21 days and incubated ± ELX/TEZ (3/3 μM) for 48 h prior to the assay. All samples treated in-assay with IVA (3 μM). Data are shown as fold-change of each species, under each treatment relative to e22 trunc mRNA in (−)ASO_TNMD control. (B) Western blot analysis of fully ALI-differentiated CFTR W1282X^+/+ hBE cells post-TECC-24 assay as described above. Fully ALI-differentiated CFTR W1282X^+/+ hBE cells (−) ASO_TNMD (100 μM), SD-18 (10 μM or 100 μM), or SD-18/SA-08 (10/10 μM or 10/100 μM) pretreated for 48 h ± ELX/TEZ (3/3 μM). CFTR detection with antibody UNC596; beta-actin was used as loading control. (C) TECC-24 assay from fully ALI-differentiated CFTR W1282X^+/+ hBE cells treated with (−) ASO_TNMD, SD-18, SA-08, or SD-18/SA-08 for 21 days followed by ± ELX/TEZ (3/3 μM) treatment for 48 h prior to the assay. All samples treated in assay with IVA (3 μM). CFTR function is represented as AUC/min FSK+IVA (see text). Error bars show standard error.

Next, western blot analysis was used to measure the impact of ASOs on e22 trunc protein expression ± ELX/TEZ (3/3 μM) (Figures 5B and S6). Protein lysates were harvested from assay filters used in the TECC-24 functional assessments as described above. e22 trunc Band C intensity and ACTB bands were quantified, and Band C intensities were normalized to ACTB and referenced to (−) ASO_TNMD (100 μM) ± ELX/TEZ (3/3 μM). All ASO blockade treatments increased Band C intensity relative to (−) ASO_TNMD without ELX/TEZ. Treatment with ELX/TEZ further increased Band C intensity relative to vehicle control for all ASO treatments. As observed with e22 trunc mRNA copy number, the order of Band C intensity was SD-18/SA-08 (100/10 μM)> SD-18 (100 μM)> SD-18/SA-08 (10/10 μM)> SD-18 (10 μM), with relative intensities of 5.1, 4.9, 3.3, and 2.5 and 3.2, 2.4, 2.1, and 1.3 for with and without ELX/TEZ (3/3 μM), respectively.

To evaluate the functional impact of ASO treatments, TECC-24 assays were completed with fully differentiated CFTR W1282X^+/+ hBE cells at ALI (Figure 5C). (−) ASO_TNMD (10 or 100 μM) (n = 2) with vehicle resulted in negligible CFTR-mediated Cl⁻ current of 0.28 ± 0.06 and 0.45 ± 0.58 AUC/min Fsk+IVA (μA/cm²) that was increased marginally with ELX/TEZ (3/3 μM) to 1.1 ± 0.37 and 1.1 (n = 2) AUC/min Fsk+IVA (μA/cm²). Surprisingly, unlike the gene-edited model and the exon 22/23 splice-blocking ASO treatments in CFF-16HBEge W1282X, ASO treatments in hBE cells at ALI in the absence of ELX/TEZ (3/3 μM) did not increase CFTR-mediated Cl⁻ currents (p > 0.05, 2-tailed Student’s t test), even though increased mRNA and protein levels were observed. After 21-day ASO treatment at ALI followed by 48-h incubation with ELX/TEZ (3/3 μM), CFTR-mediated Cl⁻ currents were significantly increased (p < 0.001, 2-tailed Student’s t test) to 4.9 ± 0.8, 8.6 ± 0.7, 8.4 ± 2.6, and 9.3 ± 0.6 μA/cm² with SD-18 (10 μM), SD-18 (100 μM), SD-18/SA-08 (10/10 μM), and SD-18/SA-08 (100/10 μM) relative to (−) ASO_TNMD (100 μM), respectively. SD-18/SA-08 (100/10 μM) with ELX/TEZ (3/3 μM) resulted in ∼25% WT function as measured in non-CF hBE control cells.

Discussion

CFTR PTC variants are a challenging class of pathogenic variants to address, and there are currently no therapies available. Readthrough agents, such as aminoglycosides, that allow ribosomes to “overread” the PTC have been studied extensively since they, in principle, could provide a “universal therapeutic” for any disease caused by nonsense variants.^{9,10,11,12,13,14,15} However, efficacy and potency are low, especially concerning the level of functional restoration required for CF, the therapeutic index is narrow, and nephrotoxicity and ototoxicity are associated with chronic use.^16,17 Additionally, the common notion that PTCs have a ∼10-fold higher propensity to be readthrough compared to natural termination codons^18,19,20,21 may be true at very low levels of readthrough, but remains to be shown at a therapeutically meaningful range. Furthermore, protein products resulting from successful readthrough often contain a non-native amino acid that might negatively impact protein function and stability and thereby further limit the efficacy of this approach.²² Last, the problem of low mRNA template levels (10%–20% of FL-WT mRNA) due to NMD remains a significant challenge that any readthrough therapeutic must overcome.⁹

Here, we describe a novel approach based on the upregulation of a truncated CFTR mRNA isoform that may benefit pwCF who harbor PTC variants that reside in the 3′ terminus of CFTR, e.g., c.3845G>A|c.3846G>A [p.Trp1282X; legacy: W1282X], c.3937C>T [p.Gln1313X; legacy: Q1313X], and c.4111G>T, [p.Glu1371X; legacy: E1371X]). We do not expect this method to suffer the liabilities of global readthrough approaches.

We first observed the presence of a naturally occurring e22 trunc mRNA resulting from intron 22 ApA site usage while conducting experiments investigating the potential for secondary mRNA splicing consequences resulting from common CFTR PTC variants. Briefly, we utilized Human Splicing Finder: http://www.umd.be/HSF3/ to predict the likelihood of five CFTR PTC variants (c.366T>A [p.Tyr122X; legacy: Y122X], c.1624G>T [p.Gly542X; legacy: G542X], c.1657C>T [p.Arg553X; legacy: R553X], p.Arg1162X; legacy: R1162X], and c.3845G>A|c.3846G>A [p.Trp1282X; legacy: W1282X]) to generate alternative splicing signals. Subsequently, we characterized the distribution of CFTR transcript isoforms through 3′ RACE and long-read sequencing on primary cells and CFF-16HBEge isogenic lines carrying these variants. Importantly, it was in these experiments that we first observed a naturally occurring e22 trunc mRNA and predicted that it would be resistant to NMD for PTCs downstream of exon 22.

Interestingly, an in silico analysis of CFTR resulted in 293 putative ApA sites identified in all introns that matched the ApA motif (A(A/T)TAAA followed by CA 10–30 nucleotides downstream).²³ We questioned the observed usage of intron 22 ApA sites, while other putative intronic ApA sites remained relatively unused. One possible explanation is that constitutive ApA repression via telescripting relies on U1 snRNP (U1) binding to SD sites to suppress premature 3′-end cleavage and polyadenylation^18,24,25 coupled with the reported weakness of CFTR intron 22 SD site relative to other SDs sites.²⁶ This may also explain our observation of low-level expression of e22 trunc mRNA (∼5%–10% WT levels) in a variety of primary airway, nasal, intestinal, and immortalized cells.

From these observations, we hypothesized that inhibition of the intron 22 SD site, and more broadly intron 22/23 splicing, may result in further increased usage of intron 22 ApA sites and upregulation of e22 trunc mRNA.

Previous studies have shown that a truncated CFTR protein (CFTR_trunc-1281), missing most of NBD2, retains partial function that can be augmented by CFTR potentiator treatment.^5,6 Both heterologous expression of the e22 trunc cDNA in FRT cells and genetic deletion of exons 23-3′UTR in 16HBE14o- cells demonstrated that e22 trunc protein behaved similarly to the CFTR_trunc-1281 protein. These unique features of e22 trunc mRNA and protein led us to hypothesize that modulation of e22 trunc mRNA/protein may have therapeutic potential for the treatment of CF caused by 3′ terminal PTCs in CFTR.

ASOs have recently garnered interest as a therapeutic modality for treatment of certain intractable human diseases. Notably, nusinersen, an ASO designed to promote inclusion of exon 7 in SMN2 mRNA was approved for use in patients with spinal muscular atrophy.²⁷ In vitro, proof of concept has shown that ASOs could provide therapeutic benefit to pwCF. These include ASOs targeting a cryptic splice site in intron 22 of the 3,849 + 10 kb C>T CFTR variant,²⁸ an ASO designed to induce exon 23 skipping for CFTR W1282X,²⁹ and ASOs designed to prevent binding of exon junction complexes downstream of exon 23 to attenuate NMD in a gene-specific manner for CFTR W1282X.³⁰

Our data show that steric-blocking ASOs targeting the intron 22 splice donor (SD) and splice acceptor (SA) sites promote upregulation of the e22 trunc mRNA, leading to expression of truncated CFTR protein and restoration of regulated CFTR channel function. This effect was observed in W1282X immortalized airway epithelial cells and W1282X^+/+ primary human bronchial epithelial (hBE) ALI cultures treated with IVA/ELX/TEZ, achieving approximately 20% and 25% of wild-type (WT) function, respectively. Notably, these levels of functional rescue surpass the widely accepted threshold of 10% improvement in an established in vitro assay predictive of clinical benefit across a cohort of people with CF (pwCF).³¹

This approach would apply to pwCF that have variants that reside in exons 23–27 such as (1) PTCs, (2) indels that result in in-frame PTCs, or (3) loss-of-function variants that are non-responsive to available highly effective modulator therapy (HEMT). However, we would not expect this approach to be appropriate for pwCF that have 1 or more HEMT-responsive alleles. Additionally, while the ASO targeting the 3,849 + 10-kb C>T variant restores FL-WT mRNA²⁸ and normal functional CFTR protein, our data suggest that upregulation of e22 trunc protein will require co-administration of CFTR modulators, at minimum a potentiator, to provide therapeutic benefit.

This study provides proof-of-concept that ApA can be modulated for therapeutic benefit in 3′ terminal PTC variants of CFTR. Importantly, this strategy may apply to other genes harboring 3′ terminal loss-of-function variants implicated in human disease. One such example is the dystrophin gene, a large gene comprising 79 exons and producing a 13.8-kb transcript. Dystrophin variants can cause Duchenne muscular dystrophy (DMD). Notably, 3′ terminal PTCs in dystrophin (or DMD gene) are associated with Becker muscular dystrophy (BMD), a clinically milder phenotype.³²

This observation has informed the investigations of C-terminal truncated microdystrophin (ΔC-microdystrophin, ∼3.8 kb), for the treatment of DMD. In the utrophin/dystrophin double knockout (u-dko) mouse model of DMD, adeno-associated virus (AAV)-mediated delivery of ΔC-microdystrophin in neonatal mice led to significant therapeutic outcomes, including reduced interstitial fibrosis and macrophage infiltration, decreased proportion of centrally nucleated myofibers, and normalization of muscle weight at 2 months post-treatment.³³

Building on the strategy developed for CFTR, ASO-mediated inhibition of SD and SA sites may enable the production of C-terminal truncated dystrophin isoforms. A naturally occurring 3′ truncated dystrophin mRNA isoform has been previously reported, implicating the use of ApA sites within intron 63 Ensembl database: ENST00000681870.1. Notably, targeted induction of ApA within intron 63 would generate a 3′ terminally truncated dystrophin transcript that retains the neuronal nitric oxide synthase (nNOS)-binding domain encoded by exons 42–46. Preservation of this domain is hypothesized to enhance muscle perfusion through nitric oxide production, offering potential therapeutic benefit.¹⁵ Additionally, intron 71 of DMD contains four consensus ApA sites whose usage is expected to result in a truncation of exons 71–79 which code for the C-terminal domain that is excluded from some mini-dystrophin designs used in gene therapy.³³ Ultimately, the ability to generate these dystrophin isoforms by ASO treatment, and their potential therapeutic benefit, would need to be experimentally tested, but, in principle, the strategy applied to CFTR could be applied to DMD.

Given the limited success in the development of therapies for pwCF harboring PTC variants, novel and/or combination approaches may be required for their treatment. Here, we show that inhibition of splicing via gene editing or ASO blockage increases e22 trunc mRNA, protein, and function to the range that is expected to provide therapeutic benefit. We speculate that further improvement of functional response may be achieved by improving e22 trunc mRNA half-life by promoting usage of downstream ApA site and increasing the 3′ alternative UTR length. Additionally, next-generation potentiators could increase the magnitude of functional restoration achievable with the e22 trunc protein. This work provides a foundation for a promising approach for the treatment of pwCF harboring 3′ terminal PTCs that might warrant further development.

Materials and methods

Cell culture

16HBE14o- parental cells, and CFF-16HBEge CFTR W1282X (referred to as 16HBEge W1282X) and CFF-16HBEge CFTR delta exon 23-3′UTR clonal lines (16HBEge-ΔE23-3′UTR) were grown as described previously.²⁹ FRT cells³⁰ and primary hBE cells³¹ were cultured as previously described. IOs were cultured as previously described IntestiCult Organoid growth medium: https://cdn.stemcell.com/media/files/pis/10000003510-PIS_08.pdf.

Generation of gene-edited cell lines

16HBE14o- parental cells were edited with CRISPR-cas9 and the resulting 16HBEge W1282X and 16HBEge-ΔE23-3′ UTR cells were cloned and cultured as previously described.⁵ The following targeting component of the crRNA (Crisper RNA) guide RNAs were used to generate the exon 23 to 3′ UTR deletion clones: 5′-TGCTCAGTTATAGTATATAA-3′ and 5′-TTAGTTATCTGTTTAAACTA-3’.

Free uptake ASO treatments of 16HBEge W1282X and 16HBE14o- cells

Steric blocking ASOs were designed and purchased from IDT (Coralville, Iowa) with phosphorothioate (PS) backbone and 2-O-methoxyethyl (2′-MOE) modifications at each nucleotide (Table S4). Lyophilized steric blocking ASOs were resuspended in PBS +Mg²⁺+Ca²⁺ to 1 mM stock concentration and stored in frozen aliquots. 16HBEge W1282X and 16HBE14o- cells were seeded at 150,000 cells/well in 24-well plates and grown as previously described. After 24 h, the media was refreshed with ASO-supplemented media. RNA was harvested after 48 h of treatment. (−) ASO_Scrambled and (−) ASO_CEP290, which target a non-CF-related gene, were used as negative ASO controls.

Free uptake ASO treatments in primary hBE cells

2′-MOE ASOs were resuspended as above and diluted into the cell culture media. Undifferentiated basal cells were treated with ASO-supplemented media for 48 h and resupplied with every media change throughout differentiation for 21 days. (−) ASO_TNMD targeting the TNMD (tenomodulin) gene was used as a negative control. TNMD is a cartilage-specific glycoprotein chosen because it is not expressed/detectable in 16HBE or hBE cultures.

ddPCR isoform quantification

RNA was isolated with the Aurum Total RNA Mini Kit (Bio-Rad). cDNA synthesis was performed with the iScript Advanced cDNA Synthesis Kit for RT-qPCR (Bio-Rad). Droplets were generated and quantitated on the QX200 Droplet Digital PCR System (Bio-Rad). Data were analyzed with Quantasoft software v.1.7.4. ddPCR probes designed to the junction of exon 25/26 were used as a proxy of full-length WT CFTR (FL-WT) or FL-W1282X mRNA and probes designed to exon 22/intron 22 junction including the first 140 bp of intron 22 were used as a proxy of exon 22 truncated CFTR mRNA (Table S5). All ddPCR measurements were run in triplicate.

Long read 3′ RACE analysis

Reverse transcription was completed using Oligo d(T)23 VN_T3 = 5′-GCAATTAACCCTCACTAAAGGTTTTTTTTTTTTTTTTTTTTTTTVN-3′ and ProtoScript II First Strand cDNA Synthesis Kit (New England Biolabs) according to manufacturer’s recommendations. PCR amplification used Q5 High-Fidelity DNA Polymerase (New England Biolabs) with CFTR exon 8 forward primer (5′-TTTCTGTTGGTGCTGATATTGCCTCAGGGTTCTTTGTGGTGT-3′) and T3 reverse primer (5′-ACTTGCCTGTCGCTCTATCTTCGCAATTAACCCTCACTAAAGG-3′) according to manufacturer’s recommendations. Amplicons were barcoded using SQK-LSK109 (Nanopore), purified with AMPure (Beckman Coulter), and sequenced on Minion (Nanopore).

mRNA half-life (t_1/2) analysis

mRNA stability of FL-WT, FL-W1282X, and e22 trunc was determined with actinomycin D time courses. 16HBE14o- and 16HBEge W1282X cultures were grown in MEM supplemented with 1% penicillin/streptomycin at 37°C and 5% CO₂ and treated at t = 0 with actinomycin D (5 μg/mL) to block transcription, and FL-WT, FL-W1282X, and e22 trunc mRNA were measured using ddPCR with specific probes targeting each mRNA isoform. Fraction of remaining mRNA R(t) was calculated relative to t = 0 at 2,4,6, and 8. Non-linear least-squares fit with exponential decay function was performed using R statistical software³⁴ as R(t) = exp(-t/τ)+ϵ for each mRNA species. The half-lives t_1/2and their confidence intervals were then calculated from the model coefficients τ and their errors reported in the fitted models.

Western blot analyses

Western blot analyses were performed as described previously,⁸ except the loading control, ACTB was detected with beta actin antibody C4 (sc-47778) from Santa Cruz Biotechnology. Blots were imaged using the ChemiDoc MP Imaging System (Bio-Rad) or ChemiDoc XRS+ Gel Documentation System (Bio-Rad) and analyzed with the accompanying image analysis software.

Electrophysiology

Conductance (FRT cells) and equivalent current (I_eq) studies (CF hBE cells at ALI and 16HBEge W1282X cells) with the TECC-24 system were performed as previously described.^7,8,9

Statistics

Statistical analysis of ddPCR data was performed using Microsoft Excel and R statistical software. The data from the instrument (averages of n = 3 technical repeat wells) were converted to log-FCs of counts under treatment vs. control count for each mRNA isoform studied; technical errors of log-FCs, ${\sigma }_{\text{tech}}^2$, were estimated by propagating ddPCR measurement errors as reported by the instrument software based on Poisson count statistics and technical repeats; then for each treatment and mRNA species, the average log-FC LFC_avg was obtained by calculating weighted average across independently run experiments (biological replicates, n = 2) with weights $\sim 1/{\sigma }_{\text{tech}}^2$, and variance component analysis (VCA) was performed to adjust the sample variance, ${\hat{\sigma }}_{\text{vca}}^2$, for each treatment/species combination. Two-tailed unpaired t test for significant differences of log-FCs from 0 was applied by calculating t-statistic as LFC_avg/SE with standard error estimated as ${SE}^2=\left({\hat{\sigma }}_{\text{vca}}^2/n\right)$. The JT test was applied to the sample of log-FC values across replicates, treated with vehicle or SD-18 only, and grouped into categories by SD-18 concentration (0, 10, 100 μM). Weighted linear model used to fit log-FCs was logFC ∼I_control+logSD18+I_SA08+I_SA08·logSD18+ϵ, where I_control is an indicator equal to 1 for control measurements (ASO dose = 0), SD18 is the ASO dosage (when >0, set to 0 otherwise), and I_SA08 is the categorical variable (indicator) for SA08 dosage (0 or 10 μM).

ANOVA with Tukey post-hoc test was performed on TECC-24 l_eq and conductance data. p values were calculated using Origin v.10.0.0.154 (OriginLab Corporation).

Data availability

Data will be made available for no less than 3 years after publication upon reasonable request.

Acknowledgments

The authors thank Michelle Hasting for discussions about delivery options for ASOs and the role of telescripting in repression of ApA, Garry Cutting and Karen Raraigh for assessment of fraction of pwCF who harbor variants that may be amenable to e22 trunc therapy, and Ali Sue Patterson and Genevieve Maul for the graphical abstract. The work was supported by the Cystic Fibrosis Foundation, international application no. PCT/US2023/075747; priority: US 63/412771 filed October 3, 2022.

Author contributions

N.E.A. and J.S.Y. conceived of the original concept of modulating ApA usage for therapeutic benefit of 3′ PTCs. H.J.B. conceived of gene-edited model to force intron 22 ApA usage. H.C.V. designed guide RNAs for gene-edited model. N.E.A., M.S.A., and J.S.Y. designed, analyzed, and interpreted experiments. M.S.A., J.S.Y., N.E.A., M.C.W., P.B., J.M.H., Y.C., and C.M.M. carried out experiments. J.E.M. suggested increased ASO dosing. A.S. designed and completed statistical analysis. N.E.A. wrote the manuscript. M.M., J.P.C., and C.U.C. provided critical feedback and edited the manuscript.

Declaration of interests

The authors declare no competing interests.