Research ArticleCystic Fibrosis

In utero and postnatal VX-770 administration rescues multiorgan disease in a ferret model of cystic fibrosis

See allHide authors and affiliations

Science Translational Medicine  27 Mar 2019:
Vol. 11, Issue 485, eaau7531
DOI: 10.1126/scitranslmed.aau7531

Tackling cystic fibrosis in the womb

Cystic fibrosis (CF) is a multiorgan disease caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. Abnormalities are observed early during development. Sun et al. reasoned that early intervention using the approved drug VX-770 (ivacaftor), a CFTR modulator, could prevent the development of such abnormalities. The authors tested the effect of in utero and early postnatal VX-770 administration in a ferret model of CF. Lung, gastrointestinal, pancreatic, and male reproductive pathologies were partially prevented in ferrets treated prior to and following birth. The results suggest that CFTR plays a critical role in early organ development and that prenatal and postnatal treatment might prevent developmental pathologies and onset of disease.


Cystic fibrosis (CF) is a multiorgan disease caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR). In patients with CF, abnormalities initiate in several organs before birth. However, the long-term impact of these in utero pathologies on disease pathophysiology is unclear. To address this issue, we generated ferrets harboring a VX-770 (ivacaftor)–responsive CFTRG551D mutation. In utero VX-770 administration provided partial protection from developmental pathologies in the pancreas, intestine, and male reproductive tract. Homozygous CFTRG551D/G551D animals showed the greatest VX-770–mediated protection from these pathologies. Sustained postnatal VX-770 administration led to improved pancreatic exocrine function, glucose tolerance, growth and survival, and to reduced mucus accumulation and bacterial infections in the lung. VX-770 withdrawal at any age reestablished disease, with the most rapid onset of morbidity occurring when withdrawal was initiated during the first 2 weeks after birth. The results suggest that CFTR is important for establishing organ function early in life. Moreover, this ferret model provides proof of concept for in utero pharmacologic correction of genetic disease and offers opportunities for understanding CF pathogenesis and improving treatment.


Cystic fibrosis (CF) is a recessive genetic disease caused by defects in the cystic fibrosis transmembrane conductance regulator (CFTR), an anion channel that moves chloride and bicarbonate across epithelia of many organs (1). Patients with CF have greatly benefited from the discovery of small molecules that increase the amount and/or function of many mutant CFTR forms (2). The two main classes of such drugs are potentiators and correctors, which rescue channel gating or processing defects in CFTR, respectively (2, 3). Alone and in combination, these drugs have shown clinical benefit in an expanding number of CFTR genotypes (2, 4). The best-characterized CFTR potentiator in people is VX-770 (ivacaftor), which restores channel gating of CFTR produced by CFTRG551D (5), a mutation found in 2 to 5% of patients with CF, as well as other mutant CFTR forms that produce mutant CFTR protein expression at the cell surface (6, 7). VX-770 has been evaluated in children with CF as young as 2 years of age (8) and is now approved for use in people with CF with specific CFTR mutations who are 1 year of age and older.

Studies in younger patients are important because early CF manifestations may affect disease progression. For example, the loss of CFTR function impairs proper development of the vas deferens and epididymis, such that males with CF are almost always infertile (9). In addition, gastrointestinal abnormalities and obstruction can occur before and after birth (10), and destruction of the exocrine pancreas occurs in most people with CF within the first 2 years of life (11). These and other early consequences of CF substantially affect the growth and health of infants with this disease (11). Despite intervention, growth of newborns with CF lags behind healthy controls within the first year of life (12). This can have long-term consequences because impaired growth during the early postnatal period correlates with earlier lung infections in infants with CF (13). Cumulatively, these and other findings suggest that CF disease starts in utero and affects postnatal health. Thus, it is very important to be able to study these early manifestations of CF to understand the pathogenesis of disease and the degree to which CFTR modulators may affect disease progression when administered early in life.

Animal models of CF offer a path to understand how CFTR function affects disease severity and progression in ways that would not be practical in humans. Although large animal models of CF such as the pig and the ferret mimic many aspects of early CF disease, mortality associated with gut and pancreatic abnormalities (1416) has hindered their utility to study CF pathogenesis. Like infants with CF, CF ferrets are prone to meconium ileus (MI, gut obstruction at birth) and are particularly fragile during the neonatal period (16). However, the incidence of MI in CF ferrets is ~4- to 5-fold greater than that of infants with CF (10, 16). Despite numerous interventions that prevent gut obstruction and promote growth (elemental diet, stool softeners, pancreatic enzymes, and proton pump inhibitors) and inhibit lung infection (multiple antibiotics), CF ferrets grow slower than non-CF controls and have a higher rate of neonatal mortality (1517).

To address this limitation, we reasoned that in utero pharmacological rescue with a CFTR modulator could improve survival and also provide the basis for studying disease-relevant processes afforded by precise control of CFTR function. This would facilitate evaluation of the importance of CFTR function during fetal development and the early postnatal period while also enabling studies of longer-term disease progression. The ferret model was chosen because it manifests many aspects of CF intestinal, pancreatic, and lung disease (1518) and is modestly sized compared with the pig.


Generating a knock-in ferret harboring the CFTRG551D allele

We generated a knock-in ferret harboring the CFTRG551D allele and studied the effects of in utero and postnatal administration of the potentiator VX-770 on survival and CF-associated pathologies. To establish the CFTRG551D model, we used recombinant adeno-associated virus (rAAV)–mediated homologous recombination in fetal ferret fibroblasts, followed by somatic cell nuclear transfer (fig. S1A). This generated CFTRG551D/WT founders for which a clean homologous recombination event was confirmed by Southern blotting (fig. S1B), polymerase chain reaction (PCR; fig. S1C), and sequencing (fig. S1D). A silent codon change at the target locus introduced a Bcl I restriction site for rapid genotyping (fig. S1E). These CFTRG551D/WT founders were bred to CFTRKO/WT and CFTRWT/WT ferrets for two generations to produce CFTRG551D/KO and CFTRG551D/G551D ferrets. Expression of mRNA produced from the G551D allele in the lungs and intestines of CFTRG551D/KO and CFTRG551D/G551D animals was only ~25 and ~50% that produced in CFTRWT/WT animals, respectively (fig. S1F). Neonate lung CFTR protein expression exhibited a similar gene dose-dependent relationship, where CFTRG551D/KO and CFTRG551D/G551D expression was ~23 and ~50% of wild type, respectively (fig. S1, G and H).

VX-770 activation of CFTRG551D in intestinal organoids, intestine, and pancreatic ductal epithelium

To assess the VX-770 responsiveness of the ferret CFTRG551D protein, we performed forskolin-induced swelling assays using intestinal organoids derived from newborn CFTRWT/WT CFTRKO/KO, CFTRG551D/KO, and CFTRG551D/G551D ferrets. Forskolin dose-response experiments demonstrated that VX-770 significantly (P < 0.001) increased CFTR-dependent swelling in both CFTRG551D/KO and CFTRG551D/G551D organoids, as compared to their vehicle [dimethyl sulfoxide (DMSO)] control treatment groups (Fig. 1A and fig. S2, A to D). CFTRWT/WT organoids gave large swelling responses that were maximal at 0.8 μM forskolin, whereas CFTRKO/KO organoids failed to swell at any forskolin concentration. In contrast to CFTRG551D organoids, there was no difference in swelling between vehicle- and VX-770–treated CFTRKO/KO or CFTRWT/WT organoids (Fig. 1A and fig. S2, A to D). Consistent with reduced expression from the CFTRG551D allele (fig. S1F), the VX-770–induced swelling response for CFTRG551D/KO and CFTRG551D/G551D organoids at 0.8 μM forskolin was ~30 and 53% that of CFTRWT/WT organoids, respectively (Fig. 1B).

Short circuit current (Isc) measurements on freshly excised intestine, from untreated CFTRWT/WT and CFTRG551D/G551D ferrets born to VX-770–treated jills, demonstrated that VX-770 treatment of CFTRG551D/G551D intestine produced a cyclic adenosine monophosphate (cAMP)–inducible chloride current ~28% that observed in CFTRWT/WT intestine, which was inhibited by the CFTR channel blocker GlyH101 (Fig. 1C and fig. S2E). Polarized pancreatic ductal epithelium (PDE) generated from CFTRWT/WT, CFTRKO/KO, and CFTRG551D/G551D pancreata demonstrated similar results to that of intestine. CFTRG551D/G551D PDE cultures differentiated in the presence of VX-770 gave cAMP-inducible chloride currents averaging 29% that of CFTRWT/WT PDE cultures (Fig. 1D and fig. S2F). By contrast, CFTRG551D/G551D PDE cultures differentiated in the presence of vehicle (DMSO) demonstrated no cAMP-inducible or GlyH101-inhibitable chloride currents, similar to CFTRKO/KO PDE (Fig. 1D). Thus, ferret CFTRG551D PDE retains little function to conduct chloride in the absence of VX-770.

Reduction of MI in CFTRG551D ferrets treated in utero with VX-770

We next determined whether VX-770 could rescue the severe MI phenotype in the CF ferret model. Because most CF ferrets die at birth due to MI (~80%) (16), we chose to administer VX-770 to the pregnant jills during the gestational period of 42 days. We chose embryonic day 28 (E28) as the gestational time to start VX-770 administration because CFTR expression was abundant in both the intestine and pancreas at this time (fig. S3, A to F). This timing would be roughly equivalent to the start of the third trimester in humans. Pharmacokinetic experiments determined the VX-770 dose required to obtain human therapeutic concentrations of VX-770 in fetal tissues of pregnant wild-type jills (see Supplementary Materials and Methods). Administration of VX-770 from E28 to E42 led to a significantly greater fraction of CFTRG551D/KO kits (63%) and CFTRG551D/G551D kits (96%) passing meconium following birth, as compared to CFTRKO/KO kits (22%) (P = 1.14 × 106) (Fig. 1, E and F). In the absence of in utero VX-770 administration (no treatment or vehicle-treated), the percentage of CFTRG551D/KO and CFTRG551D/G551D kits passing meconium was not significantly different compared to that observed for untreated and vehicle-treated CFTRKO/KO ferrets (Fig. 1, E and F) (no treatment, P = 0.3992; vehicle, P = 0.5398). Survival of CFTRG551D/KO and CFTRG551D/G551D ferrets in the first month was also significantly enhanced over that of CFTRKO/KO (Fig. 1G) (G551D/KO, P = 0.0062; G551D/G551D, P < 0.0001). Cumulatively, these results demonstrated that the ferret CFTRG551D protein is responsive to VX-770, and when administered in utero, this CFTR potentiator can protect CF ferrets from MI.

Fig. 1 Administration of VX-770 in utero enhances CFTRG551D-mediated intestinal fluid secretion and protects from development of MI.

(A) Swelling of intestinal organoids in response to various forskolin doses, for the indicated genotypes, and measured as area under the curve (AUC) in the presence of DMSO or VX-770 (3 μM). Each point represents the AUC for five to seven independent experiments, each from independent donor animals. Data represent the means ± SEM. Significant differences were determined by two-way analysis of variance (ANOVA) and Bonferroni post hoc tests, **P < 0.01 and ***P < 0.001 for comparison of VX-770 to DMSO within a genotype, and #P < 0.1, ##P < 0.01, and ###P < 0.001 for comparison of VX-770 treatment groups across genotypes. There was no significant difference between any time points when comparing VX-770 to DMSO treatment groups for WT/WT and KO/KO groups. ANOVA P values for the significance of the VX-770 effect were WT/WT (P = 0.6471), KO/KO (P = 0.9569), G551D/KO (P < 0.0001), and G551D/G551D (P < 0.0001). (B) Percentage of the WT AUC swelling response for the indicated G551D genotypes and forskolin concentration. The two-way ANOVA P values for the significance of the genotype effect are given below the legend. The average KO/KO AUC over all time points was subtracted from WT and G551D genotypes before making this calculation. (C and D) Isc measurements on (C) freshly excised intestine at the level of the jejunum and (D) polarized PDE grown at an air-liquid interface. Shown are the changes in IscIsc) in response to sequential addition of amiloride, 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS), forskolin (FSK) and 3-isobutyl-1-methylxanthine (IBMX), and/or GlyH101 (CFTR inhibitor), as indicated for the given genotypes. Data represent the means ± SEM [n in graphs indicates independent measurements from three to four animals per genotype in (C) and independent cultures derived from three to four donor animals per genotype in (D)]. The order of genotypes in (D) from left to right is WT/WT, KO/KO, G551D/G551D + DMSO, and G551D/G551D + VX-770. Significant differences were determined by two-way ANOVA and Bonferroni post hoc tests; P > 0.05 [ns (not significant)], *P < 0.05, **P < 0.01, and **P < 0.001 for comparison between genotypes in (C) and for comparison to KO/KO in (D). (E) Representative images of intestine from neonatal kits of the indicated genotypes and treatment conditions with VX-770 or vehicle administration to jills at E28. Asterisks mark intestinal dilatation and arrows mark the point of intestinal obstruction. (F) Percentage of newborn kits passing meconium at birth. Starting at E28, pregnant jills were treated with VX-770, vehicle only, or left untreated. Numbers in parentheses represent births for each genotype evaluated. Fisher’s exact test was used to assess differences between the following groups: (i) treatment groups across all genotypes: no treatment, P = 0.3992; vehicle, P = 0.5398; and VX-770, P = 1.14 × 10−6; (ii) vehicle versus VX-770: KO/KO, P = 0.356; G551D/KO, P = 0.003; and G551D/G551D, P = 6.16 × 10−6; and (iii) genotypes with VX-770 treatment: KO/KO versus G551D/KO, P = 9.70 × 10−4; KO/KO versus G551D/G551D, P = 5.10 ×10−7; and G551D/KO versus G551D/G551D, P = 5.84 × 10−4. (G) Survival rates of kits after 1 week of age for the various genotypes. G551D genotypes were born to VX-770–treated jills and maintained on VX-770 after birth, whereas KO/KO genotypes were not treated with VX-770. Numbers in parentheses represent the number of kits in each group starting at 1 week of age. Log-rank (Mantel-Cox) test was used to assess significance of differences in survival between genotypes: KO/KO versus G551D/KO (P = 0.0062), KO/KO versus G551D/G551D (P < 0.0001), and G551D/KO versus G551D/G551D (P = 0.1186).

In utero VX-770 protection of CFTRG551D/G551D male reproductive tract

We then evaluated whether other in utero developmental abnormalities could be rescued by VX-770. More than 99% of human adult males with CF have congenital bilateral absence of the vas deferens, and some also have abnormalities in the tail and body of the epididymis (9). Vas deferens abnormalities in newborn CFTRKO/KO ferret kits have been reported (16), but the phenotype of the epididymis remains unknown. Here, we show that newborn CFTRKO/KO kits lacked not only the vas deferens but also the epididymis (Fig. 2, A and B). These developmental abnormalities were rescued only in CFTRG551D/G551D kits born to VX-770–treated jills, whereas there was no effect in vehicle-treated jills for CFTRG551D/G551D kits or vehicle- and VX-770–treated jills for CFTRG551D/KO kits (Fig. 2, C to F). Histologic examination of the epididymis confirmed only remnant epithelial structures in all genotypes and treatment conditions except CFTRWT/WT and VX-770–treated CFTRG551D/G551D kits (Fig. 2, G to L). Thus, unlike the partial rescue of MI by VX-770 in CFTRG551D/KO kits, the development of these male reproductive structures requires greater CFTR expression and function than provided in CFTRG551D/KO kits.

Fig. 2 In utero administration of VX-770 rescues vas deferens and epididymis development only in ferrets homozygous for the G551D allele.

(A to F) Whole-mount photomicrographs depicting gross morphology of the testis, vas deferens (Vas), and corpus region of the epididymis in newborn kits of the indicated genotypes and in utero treatment conditions (vehicle or VX-770). The middle row shows enlargements of boxed regions in photomicrographs in the top row. Black arrows mark intact vas deferens and corpus of epididymis; red arrows mark regions where the vas deferens and corpus of the epididymis should reside, but are absent. (G to L) Representative hematoxylin and eosin (H&E)–stained sections of the epididymis (corpus) in newborn kits for the above-listed genotypes and treatment conditions. Black arrows mark the intact epididymis; red arrows mark degenerated epithelium or location where the epididymis should reside. The number of animals within the indicated phenotype evaluated for each genotype and condition was as follows: WT/WT, n = 6; KO/KO, n = 4; G551D/KO vehicle, n = 3; G551D/KO VX-770, n = 5; G551D/G551D vehicle, n = 3; and G551D/G551D VX-770, n = 3.

Restoration of growth and partial preservation of exocrine pancreas function in VX-770–treated CFTRG551D ferrets

To determine the postnatal therapeutic effects of VX-770 administration, we established a dosing regimen based on pharmacokinetic experiments on young kits (see Supplementary Materials and Methods). We then evaluated the impact of VX-770 administration on animal growth and fecal elastase (EL-1) concentration as a measure of pancreatic function during the first several months of life. For CFTRG551D/G551D and CFTRG551D/KO kits reared on VX-770, growth rates were similar to those of wild-type animals and significantly (P < 0.001) higher than those of CFTRKO/KO kits (Fig. 3A). Although the majority of VX-770–treated CFTRG551D/KO ferrets exhibited pancreatic insufficiency (PI) within the first month of life, they continued to grow at normal rates even without pancreatic enzyme supplementation while nursing (Fig. 3, A to C). Moreover, the majority of VX-770–treated CFTRG551D/G551D ferrets maintained pancreatic sufficiency (PS) with EL-1 concentrations in the normal range; only one of eight animals exhibited persistent PI, and this occurred during the third months of life (Fig. 3, B and C).

Fig. 3 Administration of VX-770 to CFTRG551D ferrets in utero and postnatally enhances growth and partially preserves pancreatic exocrine function.

In all experiments, treatment with VX-770 commenced at E28 and was sustained through the end of the experiment. (A) Body weight in animals of the indicated genotypes with and without VX-770 treatment. For each genotype, n = 6 to 11 animals. Significant differences in weight gain were observed between CFTRKO/KO versus CFTRG551D/KO and CFTRKO/KO versus CFTRG551D/G551D using two-way ANOVA and Bonferroni post hoc tests, P < 0.001. (B) EL-1 concentrations within the first 100 days of life in CFTRG551D/KO and CFTRG551D/G551D animals maintained on VX-770. The dotted lines correspond to 200 μg of EL-1 per gram of feces, below which is considered to indicate PI. Each unique symbol represents an independent animal (n = 9 CFTRG551D/KO; n = 8 CFTRG551D/G551D). (C) EL-1 concentrations for the indicated genotypes, treatment conditions, and age brackets. One animal in the CFTRG551D/G551D group exhibited PI after 2 months of age. For each genotype and age bracket, n = 6 to 20 animals. Significant differences in fecal EL-1 were observed between CFTRKO/KO versus CFTRG551D/G551D using two-way ANOVA and Bonferroni post hoc tests, **P < 0.01 and ***P < 0.001. In all graphs, data represent the means ± SEM. (D) Pancreas morphology in newborns of the indicated genotype and treatment, as assessed by H&E staining of sections. Arrows mark exocrine acini that are dilated and filled with eosinophilic secretory material. (E) Pancreas morphology in adult animals (>5 months) of the indicated genotype and treatment, as assessed by H&E staining. PI and PS, as judged by fecal EL-1 assay. The animal marked G551D/G551D + VX-770 (PI) is the single pancreatic insufficient CFTRG551D/G551D ferret shown in the bottom panel of (B). The animal marked G551D/G551D + VX (PS)➔−VX (PI) was an adult pancreatic sufficient CFTRG551D/G551D ferret removed from VX-770 for 67 days. Arrows mark islets. Scale bars, 100 μm (D) and 300 μm (E).

The pancreatic histopathology observed at birth in untreated CFTRG551D/G551D and CFTRG551D/KO kits was similar to that in newborn CFTRKO/KO kits and involved luminal dilatation of the exocrine acini (Fig. 3D). This pathology was largely prevented in CFTRG551D/G551D kits by in utero treatment with VX-770; the same was true, though to a lesser degree, for CFTRG551D/KO kits (Fig. 3D). Pancreatic histology of the pancreatic sufficient CFTRG551D/G551D ferrets was similar to that of CFTRWT/WT controls, both retaining the acinar cell compartment. By contrast, pancreatic insufficient CFTRG551D/KO and CFTRG551D/G551D ferrets were characterized by exocrine destruction and fibrosis resembling those observed in CFTRKO/KO ferrets (Fig. 3E). Collectively, these results revealed that VX-770 improved growth in both CFTRG551D/KO and CFTRG551D/G551D ferrets despite large differences in exocrine pancreatic disease.

VX-770–mediated preservation of glycemic stability only in pancreatic sufficient CFTRG551D/G551D ferrets

We next determined whether VX-770–mediated protection of the exocrine pancreas rescues greater glycemic control in CFTRG551D ferrets. Destruction of the exocrine pancreas in CFTRKO/KO ferrets leads to spontaneous hyperglycemia and impaired postprandial glucose tolerance following their normal liquid diet at ~1 to 2 months of age (Fig. 4, A and B) during a period in which pancreatic inflammation peaks and islet mass is reduced (19). As expected, CFTRG551D/KO ferrets with PI and maintained on VX-770 also underwent these glycemic disturbances, whereas CFTRG551D/G551D ferrets with PS did not (Fig. 4, A and B). One CFTRG551D/KO animal fluctuated between PS and PI over this period, and its postprandial glucose was the lowest for that genotype (Fig. 4B). By contrast, the one CFTRG551D/G551D ferret that later became pancreatic insufficient had the highest postprandial glucose of its group (Fig. 4B). We additionally performed formal mixed-meal tolerance tests (MMTTs) on 3- to 4-month-old animals with a body surface area–normalized caloric intake from solid and liquid food. This analysis confirmed that glucose tolerance was normal in CFTRG551D/G551D ferrets with PS and maintained on VX-770, whereas it was abnormal in CFTRG551D/KO and CFTRG551D/G551D with PI despite sustained VX-770 treatment (Fig. 4, C to F). Removal of CFTRG551D/KO ferrets from VX-770 for a 3-week period led to no change in glucose tolerance, as indexed by postprandial glucose excursions and AUC glucose (Fig. 4, C to D). These findings are consistent with previous work in the CFTR knockout ferret model implicating destruction of the exocrine pancreas and islet remodeling as the primary determinants of impaired glycemic status in CF ferrets (19).

Fig. 4 Administration of VX-770 to CFTRG551D/G551D ferrets in utero and postnatally improves postprandial glucose tolerance only in animals with PS.

(A) Blood glucose at baseline in nonfasted, actively nursing kits of the indicated genotypes at 1 to 2 months of age. All animals were treated with VX-770 in utero and maintained postnatally. (B) Blood glucose at 1 hour after gavage feeding an elemental diet for the animals shown in (A). Each data point in (A) and (B) represents the average of four to six measurements per kit on different days. (C) Blood glucose as assessed by MMTT in CFTRG551D/KO ferrets at 3 to 4 months of age. All animals were treated with VX-770 in utero; for the “On” group, treatment was sustained throughout the postnatal period and during testing, but for the “Off” group, treatment was terminated 3 weeks before testing. (D) Quantitation of blood glucose for animals in (C), by AUC analysis. (E) Blood glucose as assessed by MMTT in 3- to 4-month-old ferrets of the indicated genotypes. All animals were treated with VX-770 both in utero and throughout the postnatal period. For the CFTRG551D/G551D group, curves are provided for the entire cohort (n = 7), i.e., including both animals with PS (n = 6) and PI (n = 1), and the single CFTRG551D/G551D ferret that developed PI. (F) Quantitation of blood glucose excursions for animals in (E), by AUC analysis. Significant differences in (A), (B), and (F) were determined by one-way ANOVA and Bonferroni post hoc tests, *P < 0.05, ***P < 0.001, ns, for comparison to CFTRKO/KO, and for (E) two-way ANOVA and Bonferroni post hoc tests, **P < 0.01, ***P < 0.001 for comparison of the complete CFTRG551D/G551D and CFTRWT/WT cohorts to the CFTRKO/KO group. In all graphs, data represent the means ± SEM.

Impact of CFTR function during early postnatal development on growth and survival

After establishing that in utero and early postnatal administration of VX-770 could increase survival and protect exocrine and endocrine pancreatic function, we evaluated the impact of VX-770 treatment cessation postnatally on the health of CFTRG551D/KO kits by terminating treatment of CFTRG551D/KO kits born to VX-770–treated jills at birth or at 7, 14, or 25 days after birth. In the case of CFTRG551D/KO kits, when VX-770 treatment continued only to birth or 7 days of age, growth rates did not differ from those of CFTRKO/KO kits (Fig. 5, A and B). By contrast, when VX-770 treatment was continued to 14 or 25 days of age, their growth was substantially more rapid than that of CFTRKO/KO kits (Fig. 5, A and B). Rescue of growth rates to that of wild-type animals (Fig. 5A), however, was achieved only when VX-770 treatment was maintained to 25 days. In addition, in CFTRG551D/KO kits, increase in survival required VX-770 treatment beyond 7 days of life, and the greatest benefit with respect to these parameters was in the 25-day group (Fig. 5C). The primary reasons for early euthanasia of CFTRG551D/KO ferrets after termination of VX-770 administration were rectal prolapse, gut obstruction, and/or lung infection. Analysis of culturable bacteria in the lungs at the time of euthanasia identified bacterial taxa and titers similar to those found in CFTRKO/KO ferrets (Fig. 5D) (16, 17). Gross lung pathology showed evidence of atelectasis and air trapping in the lung and mucus accumulation in the trachea (Fig. 5, E to G). Histopathology demonstrated mucus accumulation in the large airways and submucosal glands (Fig. 5, H to Q, and fig. S4, A to H). Each of these pathologies has been observed in CFTRKO/KO ferrets (16, 17). Cumulatively, these findings demonstrate that VX-770 treatment in utero and during the early postnatal period had therapeutic effects in CFTRG551D/KO ferrets.

Fig. 5 Early postnatal administration of VX-770 to CFTRG551D/KO ferrets is required for marked improvements in growth, survival, and slowing of lung disease progression.

(A) Growth of CFTRWT/WT and CFTRKO/KO kits during the first 3 weeks of life. (B) Growth of CFTRG551D/KO kits born to VX-770–treated jills starting at E28, with VX-770 treatment terminated at the indicated age in days (d). (C) Survival curves for CFTRG551D/KO kits receiving VX-770 in utero and postnatally for various lengths of time. Log-rank (Mantel-Cox) test was used to assess significance of differences in survival between days 0 and 7 versus day 14 (P = 0.0009 and P = 0.0088, respectively) and between days 0 and 7 versus day 25 (P = 0.0004 and P = 0.0010, respectively). (D) Quantification of bacterial taxa (left y axis) and CFU (right y axis, red dots) in lung homogenates from different CFTRG551D/KO ferrets shown in (C) at the time of euthanasia. Bacterial taxons were identified using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) fingerprinting. Only the most abundant taxons are listed in the legend. (E to K) Representative lung images of a CFTRG551D/KO ferret removed from VX-770 at 14 days and euthanized at 174 days of age due to poor health, showing gross lung pathology (E to G) and H&E-stained sections (H to K). Boxed region in (E) is enlarged in (F) and shows air trapping (white arrow) and atelectasis (yellow arrow). (G) Thick mucus in the trachea. For histology of all five lobes from this CFTRG551D/KO animal, see fig. S4 (A to H). Boxed region in (J) is enlarged in (K) and shows mucus accumulation in submucosal glands (SMGs). (L) Representative SMG image from a wild-type ferret for comparison to (K). (M to Q) H&E-stained sections from another CFTRG551D/KO ferret removed from VX-770 at 25 days and euthanized at 177 days of age. Lung samples from this animal were not insufflated with fixative at the time of tissue collection. Boxed regions in (N) and (P) are enlarged in (O) and (Q), respectively. Histologic images (H to J, M, N, and P) are all different lobes from each animal with mucus accumulation in the airways marked by asterisks. Growth statistics (A and B): two-way ANOVA and Bonferroni post hoc tests were used to determine significant differences between groups when comparing the following dataset. For CFTRWT/WT (n = 11) versus (i) CFTRKO/KO (n = 11, days 11 to 21, P < 0.05), (ii) CFTRG551D/KO 0 days (n = 7, days 11 to 21, P < 0.05), (iii) CFTRG551D/KO 7 days (n = 5, days 15 to 21, P < 0.05), (iv) CFTRG551D/KO 14 days (n = 8, day 21, P < 0.01), and (v) CFTRG551D/KO 25 days (n = 6, all days were not significant). For CFTRKO/KO versus (i) CFTRG551D/KO 0 days (all days were not significant), (ii) CFTRG551D/KO 7 days (all days were not significant), (iii) CFTRG551D/KO 14 days (days 12 to 21, P < 0.05), and (iv) CFTRG551D/KO 25 days (days 10 to 21, P < 0.05).

Lung disease after cessation of VX-770 in adult CFTRG551D/KO ferrets

To evaluate the rate of lung colonization in healthy adult CFTRG551D/KO ferrets, we used the cohort that was reared on both VX-770 and multiple antibiotics to ~8 months of age, during which glucose tolerance testing was performed free of potentially confounding effects of lung infection. The triple-antibiotic regimen was previously shown to prevent bacterial colonization of the CFTRKO/KO lung but not mucoinflammatory disease resulting from airway dehydration (20). Lung infection in adult CFTRG551D/KO ferrets was monitored by serial bronchoscopy before and after termination of VX-770 and antibiotics (Fig. 6A). The time to detection of culturable bacteria in the lung ranged from 51 to 134 days after VX-770 removal, and colony-forming unit (CFU) titers increased progressively with time (Fig. 6B). CFTRG551D/KO ferrets removed from VX-770 eventually contracted terminal lung disease with mucus obstruction in the airways and submucosal glands (Fig. 6, C to J), as previously observed in adult CFTRKO/KO ferret (17).

Fig. 6 Withdrawal of VX-770 from adult CFTRG551D/KO ferrets leads to lung disease.

(A) Timeline for analysis of CFTRG551D/KO ferrets. Bronchoalveolar lavage fluid (BALF) was collected before and after withdrawal of VX-770 to follow bacterial colonization of the lung. (B) Quantitation of bacterial load (measured as total CFU and CFU/ml) in the BALF from CFTRG551D/KO ferrets, before and after VX-770 withdrawal at 267 days (d) of age. The number of total bronchoscopies is indicated in parentheses. Data represent the means ± SEM, using the average CFU counts for each animal at each time frame (n = number of animals). One-way ANOVA using Kruskal-Wallis and Dunn’s post hoc test was used to assess significant differences in CFU counts between baseline (on VX-770) and the two time points off VX-770, **P < 0.01. (C to I) Representative histopathology images of the trachea (C) and lung (D to I) from an adult CFTRG551D/KO ferret in which VX-770 treatment terminated and the animal was euthanized 160 days later because of poor health. Separate lung lobes are provided for each low-power image. (E, H, and I) Enlargement of boxed regions in (D), (G), and (H), respectively. (J) Representative lung histology from a wild-type ferret. Asterisks: basophilic fibrillar material (mucus) and cellular debris.

Fig. 7 Withdrawal of VX-770 from adult CFTRG551D/G551Dferrets leads to lung disease.

CFTRG551D/G551D ferrets were born to E28 VX-770–treated jills and maintained on VX-770 until adulthood, at which time pancreatic function and glucose tolerance testing was performed. VX-770 was then tempered off over a month. (A to F) Lung histopathology from a CFTRG551D/G551D ferret (animal 3) euthanized 187 days after VX-770 cessation. All low-power images represent different lobes. The pathology in this animal was characterized by distal airway disease (A, B, and D to F) and pneumonia in some but not all lobes (C). (B and F) Enlargement of boxed regions in (A) and (E), respectively. Insets in (B) and (F) are taken at the region of the black arrows and demonstrate neutrophil and accumulation in the distal airways. (G to P) Lung histopathology from an adult CFTRG551D/G551D ferret (animal 2) euthanized 67 days after VX-770 cessation. (G) Breath-hold computed tomography image on the day before sacrifice demonstrating an occluded lobe of the lung (asterisk). (H) Gross lung pathology showing the atelectatic lobe (asterisk) and abscesses in the distal lung [boxed region with enlargement shown in (I)]. (J) Bronchoscopy at 47 days after VX-770 cessation demonstrating thick viscous mucus with the indicated CFU titer of bacteria found in BALF. (K to P) Lung histopathology from animal 2 with (K) being the atelectatic lobe. (L and M) Enlargement of boxed regions in (K). All low-power images represent different lobes. Airway mucus. Yellow arrows mark bronchial-associated lymphoid tissue expansion.

Pancreatic and lung disease after cessation of VX-770 in adult CFTRG551D/G551D ferrets

The majority of CFTRG551D/G551D ferrets were pancreatic sufficient while maintained on VX-770 (Fig. 3B). To assess the effect of VX-770 removal from adult CFTRG551D/G551D ferrets, we reduced the VX-770 dose gradually by 2.5 mg/kg per week owing to concerns that severe acute onset of pancreatitis might be lethal. In addition to monitoring weight and fecal EL-1, we also assessed plasma pancreatic lipase (measure of pancreatitis). All the animals tested (n = 5) developed PI (EL-1 <200 μg/g feces) within 3 weeks of complete VX-770 withdrawal (fig. S5A). Although plasma concentrations of pancreatic lipase remained relatively constant within each animal (fig. S5B), weight declined 10 to 20% from maximum for four of the animals during the 3-week period after complete VX-770 withdrawal (fig. S5C). Histopathology of pancreatic sufficient CFTRG551D/G551D ferrets removed from VX-770 demonstrated a conversion from normal pancreatic exocrine architecture to a fibrotic pancreas with plugging of the ducts and depletion of acinar cell mass (Fig. 3E, lower right). MMTTs on these CFTRG551D/G551D before and after VX-770 withdrawal demonstrated a diabetic phenotype (blood glucose >200 mg/dl at 120 min) in only one of five ferrets (animal 4) at 1 month after cessation of VX-770 (fig. S5, D and E). By 3 to 5 months after cessation of VX-770, postprandial glucose in this animal (animal 4) normalized, but a third ferret (animal 3) progressed to a diabetic phenotype (fig. S5, D and E).

Progression of lung disease in these five CFTRG551D/G551D ferrets after VX-770 withdrawal was assessed by bronchoscopy and lung histopathology at the time of declining health. The average culturable bacteria (CFU) in the BALF before stopping VX-770 was 73 ± 63 CFU/ml (n = 4, ±SEM; range, 0 to 255 CFU/ml; one bronchoscopy could not be obtained). At 1 to 2 months after complete termination of VX-770, follow-up BALF bacterial titers in these animals averaged 1891 ± 821 CFU/ml (n = 5, ±SEM; range, 0 to 4366 CFU/ml). Only one animal did not culture bacteria in the BALF after termination of VX-770, and titers in the others exceeded 1000 CFU/ml (table S1). Two of these animals (animals 2 and 3) were euthanized for failure to thrive, and lung histopathology in animal 3 (Fig. 7, A to F) and animal 2 (Fig. 7, G to P) showed predominantly distal airway and alveolar disease dominated by neutrophil infiltration. However, the rapid decline of animal 2 after cessation of VX-770 was accompanied by the mucus occlusion of one lobe (Fig. 7, G, H, and K to M). Although the timing of terminal disease onset was variable across subjects, VX-770 offered protection from bacterial infections and lung pathologies in CFTRG551D/G551D ferrets.

Several CFTRG551D/KO and CFTRG551D/G551D ferrets were euthanized while maintained on VX-770 to evaluate whether treatment prevented lung disease. Histopathology of CFTRG551D/KO ferrets demonstrated single distal airways in most lobes with intermixed cellular debris and neutrophils (fig. S6, A to F). In contrast, lung histopathology of CFTRG551D/G551D ferrets was normal and submucosal glands lacked mucus accumulation (figs. S6, G to I, and S4, I to P). These findings suggest that CFTRG551D/G551D ferrets have sufficient CFTR expression to maintain normal lung clearance and innate immunity while on VX-770.

VX-770 cessation promotes CFTRG551D/KO lung inflammation

Mucus accumulation and inflammation are hallmarks of the CF lung (21, 22), and in CFTRKO/KO ferrets, mucoinflammatory lung disease can occur with and without bacterial infection depending on the antibiotic treatment regimen used (17, 20). We reasoned that termination of VX-770 treatment in CFTRG551D ferrets would enhance markers of CF lung disease, such as mucus accumulation and inflammation in the lung. To this end, we performed quantitative proteomics on BALF from three CFTRG551D/KO ferrets before and 1 month after cessation of VX-770. Results from this analysis identified 239 proteins, of which 62 were significantly changed (P < 0.05) after removal of VX-770 (Fig. 8A and sheet A in data file S1). Sixty-five percent of the differentially regulated proteins were increased in the BALF after VX-770 withdrawal. Proteins of lung epithelial origin that were shown to be significantly (P < 0.05) increased in the BALF after VX-770 withdrawal included mucin 5AC (MUC5AC) and kallikrein 1 (KLK1), a serine protease known to be associated with lung disease (23), whereas significantly (P < 0.05) decreased proteins included submucosal gland–secreted MUC5B and lactotransferrin (LTF) and surface airway club cell-secreted secretoglobin family 1A member 1 (SCGB1A1) (Fig. 8A). Pathway analysis using ReactomePA also demonstrates significant enrichment (adjusted P value, 3.2 × 10−35) in BALF proteins associated with neutrophil degranulation (R-HSA-6798695) (Fig. 8, A and B).

Fig. 8 Withdrawal of VX-770 from adult CFTRG551D/KO ferrets leads to enhanced lung inflammation.

BALF was collected from three CFTRG551D/KO ferrets before and 1 month after withdrawal of VX-770. The BALF proteome was then examined in a paired fashion using quantitative proteomics. (A) Volcano plot of 239 BALF proteins detected with the log2 (fold change) representing (OFF VX-770)/(ON VX-770) [black circles: not statistically significant or log2 (fold change) was <1.5 or >−1.5; red circles: statistically significant P < 0.05 with a log2 (fold change) ≥1.5 or ≤−1.5; n = 3 paired samples]. Proteins (gene names) found in neutrophil granules are marked in blue, whereas proteins secreted by airway epithelium and submucosal glands are marked in black. (B) Volcano plot of 63 BALF proteins enriched (P = 3.2 × 10−35) in the ReactomePA neutrophil degranulation pathway. (C) List of selected significant disease pathways discovered using MeSH analysis. (D and E) List of selected significant (D) canonical pathways and (E) diseases or functions pathways discovered using IPA. The number of proteins found in each pathway is given in parentheses, and pathways with near-significant (>1.9) or significant (>2) positive z scores are highlighted in red.

Pathway analysis of the proteomics data using Medical Subject Headings (MeSH) supported a rise in lung disease after VX-770 withdrawal, with enriched disease pathway including lung disease, CF, pneumonia, and infection, among others (Fig. 8C and sheet B in data file S1). Similarly, Ingenuity Pathway Analysis (IPA) confirmed enrichment in canonical pathways relating to acute phase response signaling, liver X receptor-retinoid X receptor (LXR/RXR), complement system, and phagosome maturation, among others (Fig. 8D and sheet C in data file S1). IPA diseases and functions pathways also demonstrated enrichment in processes involved in cellular infiltration of the lung, inflammation of the lung, and leukocyte migration and an enhancement of pathways (z score >2) involved in quantity of leukocytes, quantity of phagocytes, and quantity of antigen-presenting cells (Fig. 8E and sheet D in data file S1). Each of these analyses supports enhanced lung disease and inflammation after withdrawal of VX-770 from CFTRG551D/KO ferrets.


We have created a CFTRG551D ferret model that exhibits multiorgan disease and is responsive to the CFTR potentiator VX-770. We showed that VX-770 can be administered to pregnant ferrets to protect animals from developmental abnormalities caused by in utero CFTR defects, including the prevention neonatal mortality associated with MI. The G551D ferret response to VX-770 is remarkably similar to that of patients with CF (8), including partial recovery of pancreatic function, slowed lung disease, reduced infections, and improved weight gain. Furthermore, withdrawal of VX-770 from CFTRG551D ferrets leads to emergence of multiorgan disease similar to that observed in CFTRKO/KO ferrets (1518), and although there is little comparable information in CFTRG551D patients, there is one case study reporting the reemergence of lung disease upon VX-770 withdrawal (24).

Initial characterization of this model revealed similarity to the human condition, therefore providing a path to probe the relationship between CFTR function during development and early life and disease severity and progression later in life. For example, the in utero data suggest that there may be a window during development [E28 to P25 (postnatal day 25)] when CFTR function is critical for postnatal health and survival of CF ferrets. The similar postnatal growth rates observed for weaning pancreatic insufficient CFTRG551D/KO and pancreatic sufficient CFTRG551D/G551D ferrets maintained on VX-770 suggest that factors extrinsic to the pancreas may play a key role in CF health early in life and is consistent with data from patients with CF (8). In addition, we observed a clear association between abnormal glucose tolerance and exocrine pancreas insufficiency in both CFTRG551D/G551D and CFTRG551D/KO ferrets, consistent with exocrine pancreatic disease early in life being associated with a higher incidence of CF-related diabetes in humans (25). However, the variable and somewhat less severe postprandial glucose phenotypes in adult CFTRG551D/G551D removed from VX-770 were unexpected, with only one of the four surviving animals progressing to a CF-related diabetes phenotype at 3 to 5 months after cessation. This model is expected to prove useful for studying CFTR functions in the pancreas and the mechanisms by which age-dependent pancreatic remodeling events in CF lead to failure of islet function.

These data not only confirm the in vitro and/or in vivo efficacy of a CFTR modulator on airway and intestinal epithelia but also provide the first direct evidence that VX-770 increases CFTR function in pancreatic epithelial cells. The hypomorphic nature of the ferret CFTRG551D allele likely played a key role in the extent of VX-770–mediated protection of CF ferrets from lung disease, MI, and exocrine pancreatic dysfunction. The low expression of the knock-in allele is likely due to the presence of a neomycin expression cassette in the adjacent intron of the targeted locus. The LoxP sites flanking the neomycin have not been activated in the ferrets studied here. The current dosing regimen led to ferret VX-770 exposures similar to those associated with maximal efficacy in patients with the CFTRG551D mutation (26). Despite the ~50% reduction in mRNA and protein expression from the CFTRG551D knock-in allele, VX-770 exposures resulted in partial to complete protection from disease in the various organs studied in CFTRG551D ferrets.

Data from humans possessing hypomorphic CFTR intronic splice variants suggest that ~8 to 12% normal CFTR transcript is sufficient to prevent disease (27). On the basis of in vivo concentrations of VX-770 and the in vitro studies of epithelia from the CFTRG551D ferrets, the ferret model appears relatively consistent with human data. Future studies titrating VX-770 exposure and varying timing of administration will better elucidate the threshold amount of CFTR activity and treatment window necessary to prevent disease across organs in the ferret model.

This study has some limitations. Despite improved availability of viable CFTRG551D ferrets due to VX-770 treatment, the size of cohorts for several of the more labor-intensive experiments and end points were resource-limited. With this in mind, longitudinal studies that use less invasive translational end points, such as nasal potential difference to assess CFTR function and spirometry to quantitatively measure lung disease progression, could provide outcomes closer to those assessed in patients with CF. Regional fluctuations in CF lung colonization over time can also make the use of bronchoscopic BALF collection less reproducible for quantifying bacterial load longitudinally in animals, despite controlling for this by sampling the same lobe; there are limited data in patients with CF to compare our experience. In addition, we did not investigate the presence of nonbacterial organisms in the lung. Thus, we cannot comment on the contribution of either fungal or viral infections in CFTRG551D ferret lung disease progression. Last, the in vivo dose-response relationship between VX-770 and CFTRG551D function was not fully explored in the present study; thus, it remains unknown whether optimal efficacy was achieved in all CF animals.

Fundamental questions remain regarding CF pathogenesis that cannot be easily addressed in humans. The current VX-770–responsive CFTRG551D model affords a pharmacological switch for selective activation and inactivation of CFTR function and provides a path to address many of these questions. For example, what is the critical window to rescue organ function, and at what point does organ damage become irreversible? Another key question is how well rescue of CFTR function during development and early postnatal periods protects against initiation of lung disease later in life. Furthermore, widespread use of the CF ferret model has previously been limited by the high rate of neonatal mortality and the labor-intensive nature of rearing CF kits, two factors overcome by the G551D ferret model. The implications of this study for human therapy remain to be determined because examples of in utero treatment of a genetic disease are rare (28). However, the findings from this study emphasize the importance of early interventions in the treatment of CF and offer a model to understand the relation between CFTR function and disease pathophysiology and progression, which may inform the development and application of future CF therapies.


Study design

The objectives of this study were to create a CFTRG551D ferret model and to evaluate whether VX-770 could protect animals from disease progression. In vitro end points were used to assess the extent to which VX-770 could potentiate CFTRG551D and included the use of intestinal organoids, freshly excised intestine, and polarized PDE. In vivo end points for assessing VX-770 action included the following: (i) in utero VX-770 treatment and assessment of the extent of protection from developmental abnormalities of the gut, pancreas, and male reproductive tract; (ii) sustained postnatal VX-770 treatment and assessment of growth, survival, exocrine pancreatic function, glucose intolerance, and lung colonization; and (iii) withdrawal of VX-770 during the early postnatal period or in adulthood and analysis of disease progression in the pancreas and lung. The primary data are reported in table S1.

Generation of CFTRG551D ferrets using homologous recombination and somatic cell nuclear transfer

The strategy used to generate the CFTRG551D ferret was the same as previously described for the CFTRKO ferret (29). Targeting was initiated by infecting primary fetal ferret fibroblasts with rAAV harboring the G551D mutant exon 11, neomycin selection cassette, and flanking intronic homology (AAV2.G551D-TG; see fig. S1). On day 1 after infection, fibroblasts were serially diluted and cultured with G418 (300 μg/ml) for 15 days. G418-resistant clones were screened for homologous recombination PCR (see fig. S1); clones positive for both the left and right arms of the predicted homologous recombination event were expanded and animals were cloned by serial somatic cell nuclear transfer (29). The reconstructed embryos were then transferred into pseudopregnant recipient jills. The genotype of each G551D-targeted ferret founder was confirmed by Southern blotting.

Rearing of ferrets

All animal experimentation was approved by the Institutional Animal Care and Use Committee. VX-770 (20 mg/kg body weight, formulated in elemental liquid diet EleCare, Abbott) was administered orally once a day to pregnant jills, starting on day 28 of gestation (parturition typically occurs on day 42 of gestation). Newborn CF kits were gavaged daily with VX-770 (5 mg/kg) formulated in EleCare or EleCare plus the laxative NuLYTELY (Braintree Laboratories). The dose was increased to 10 mg/kg daily at 14 days of age and maintained at that dose thereafter. CF ferrets used for glucose tolerance testing were also maintained on three antibiotics from birth, as previously described (19), to ensure that the lungs would not be infected. These animals were later withdrawn from antibiotics and VX-770 to assess progression of lung disease.


At necropsy, the lung, trachea, pancreata, vas deferens, and epididymis were fixed in 10% neutral buffered formalin, routinely processed, embedded in paraffin, and sectioned at 4 μm. All sections were stained with H&E, and select sections were stained with Periodic Acid–Schiff for light microscopy. A board-certified veterinary pathologist performed the histopathological examination.

Statistical analysis

Statistical analysis was performed using GraphPad Prism software. For all experiments, data are expressed as means ± SEM. The statistical significance of the observed differences was calculated using one- or two-way ANOVA with Bonferroni post hoc tests. Results were considered significant when P < 0.05. Proteomics data were evaluated using Scaffold Q+S version 4.7 (Proteome Software) at 0% false discovery rate and with statistical testing accompanying the Scaffold algorithm. Experiments were not blinded or randomized.


Materials and Methods

Fig. S1. Generation of G551D-CFTR ferrets using AAV-directed mutagenesis and somatic cell nuclear transfer.

Fig. S2. Forskolin-induced swelling of intestinal organoids and Isc of intestinal epithelia and PDE.

Fig. S3. Expression of CFTR mRNA in the fetal intestine and pancreas.

Fig. S4. Lung histology of all lobes from a CFTRG551D/KO ferret removed from VX-770 and a CFTRG551D/G551D ferret maintained on VX-770.

Fig. S5. Withdrawal of VX-770 from adult CFTRG551D/G551D ferrets leads to PI and variable degrees of glucose intolerance.

Fig. S6. Lung histology of CFTRG551D/KO and CFTRG551D/G551D ferrets maintained on VX-770 for life.

Table S1. Primary data (provided as a separate Excel file).

Data file S1. Proteomics data (provided as a separate Excel file).

Data file S2. RNA tissue quantification and smFISH probe maps.

References (3035)


Acknowledgments: We acknowledge members of the Comparative Pathology Core at the University of Iowa who assisted in tissue processing and C. Blaumueller and J. Barr for editorial assistance. Funding: This work was supported by NIH grants R24 DK096518, R24 HL123482, P30 DK054759, R01 DK047967, R01 DK115791, and P01 HL051670 (to J.F.E.); the Cystic Fibrosis Foundation (CFF) ENGELH0XX0 (to J.F.E.); the Carver Chair in Molecular Medicine (to J.F.E.); a Vertex Pharmaceuticals Sponsored Research Agreement (to J.F.E.); and ZonMW 91214103 and NCFS HIT-CF2 (to J.M.B.). Funding contributions were as follows: (i) NIH and the CFF supported creation and generation of the G551D ferrets; (ii) NIH, CFF, and Vertex supported their characterization; and (iii) Vertex provided the VX-770 (ivacaftor). Author contributions: X.S., Y. Yi, Z.Y., B.H.R., M.C.W., D.F.F., P.N., F.V.G., and J.F.E. designed experiments and analyzed the data. X.S., Z.Y., B.L., Y. Yang, and S.Y.P. generated the ferret model and performed genetic characterization and genotyping. M.C.W., X.S., B.L., Y. Yang, and Y. Yi managed care of the animals. M.C.W., Y.Z., P.G.R., and X.S. performed the electrophysiologic analyses. B.H.R., M.C.W., Y.Z., J.S.G., Y. Yang, T.I.A.E., W.Z., and S.R.M. performed bronchoscopy sample collection and bacteriologic analyses. X.S., Y. Yi, Y. Yang, and B.L. performed MMTTs. X.S. and K.N.G.-C. performed the male reproductive organ characterization. M.C.W., X.S., and B.L. performed the MI and survival characterization. D.F.F., D.M.T., and L.J. performed pharmacokinetic analyses. D.F.F. and L.W. performed Western blotting analysis. Z.Y. and S.Y.P. performed mRNA analyses. T.I.A.E., A.M.V., and J.M.B. performed intestinal organoid assays. J.F.E. and D.F.F. performed statistical analyses. K.N.G.-C., X.S., Y. Yi, M.C.W., Y. Yang, Y.Z., B.L., and W.Z. performed all necropsies and tissue harvesting. K.N.G.-C. managed histologic evaluation of all tissue specimens and imaging and performed histopathologic assessments. X.S., Y. Yi, M.C.W., T.I.A.E., and J.F.E. wrote the original draft of the manuscript. X.S., Y. Yi, T.I.A.E., M.C.W., P.N., F.V.G., D.F.F., K.N.G.-C., and J.F.E. edited the manuscript. X.S., Y. Yi, T.I.A.E., M.C.W., D.F.F., and J.F.E. generated the figures. J.M.B., P.N., and J.F.E. provided funding. Competing interests: L.W., D.M.T., L.J., D.F.F., P.N., and F.V.G. are employees of Vertex Pharmaceuticals, a publicly traded company. J.F.E. has sponsored research with Vertex Pharmaceuticals but holds no equity in this company. All other authors declare that they have no competing interests. Data and materials availability: The ferret model described within this manuscript will be freely distributed under a material transfer agreement. All data are available in the main text or the Supplementary Materials.

Stay Connected to Science Translational Medicine

Navigate This Article