Allosteric folding correction of F508del and rare CFTR mutants by elexacaftor-tezacaftor-ivacaftor (Trikafta) combination

Abstract

Based on its clinical benefits, Trikafta - the combination of folding correctors VX-661 (tezacaftor), VX-445 (elexacaftor), and the gating potentiator VX-770 (ivacaftor) - was FDA approved for treatment of patients with cystic fibrosis (CF) carrying deletion of phenylalanine at position 508 (F508del) of the CF transmembrane conductance regulator (CFTR) on at least 1 allele. Neither the mechanism of action of VX-445 nor the susceptibility of rare CF folding mutants to Trikafta are known. Here, we show that, in human bronchial epithelial cells, VX-445 synergistically restores F508del-CFTR processing in combination with type I or II correctors that target the nucleotide binding domain 1 (NBD1) membrane spanning domains (MSDs) interface and NBD2, respectively, consistent with a type III corrector mechanism. This inference was supported by the VX-445 binding to and unfolding suppression of the isolated F508del-NBD1 of CFTR. The VX-661 plus VX-445 treatment restored F508del-CFTR chloride channel function in the presence of VX-770 to approximately 62% of WT CFTR in homozygous nasal epithelia. Substantial rescue of rare misprocessing mutations (S13F, R31C, G85E, E92K, V520F, M1101K, and N1303K), confined to MSD1, MSD2, NBD1, and NBD2 of CFTR, was also observed in airway epithelia, suggesting an allosteric correction mechanism and the possible application of Trikafta for patients with rare misfolding mutants of CFTR.

Keywords: Cell Biology; Chloride channels; Drug therapy; Epithelial transport of ions and water; Pulmonology.

Figure 1. VX-445–mediated F508del correction is synergistic with type I and II correctors.

(A) Dose-response of VX-445 (24 hours, 37°C) in presence of 3 μM VX-661 for the correction of F508del-CFTR PM density in CFBE41o- cells expressed as percentage of the WT-CFTR (n = 3). EN1, enantiomer 1; EN2, enantiomer 2. (B) PM density of F508del-CFTR after type I corrector (VX-661, VX-809, ABBV-2222, or FDL169; 3 μM, 24 hours, 37°C), VX-445 (2 μM, 24 hours, 37˚C) or type I plus VX-445 corrector combination treatment expressed as percentage of WT-CFTR in CFBE41o- (n = 3). (C) Heatmap of the effect of corrector combinations on the PM density of F508del-CFTR expressed in CFBE41o- (n = 3). (D) Heatmap of the combinatorial profiling established by calculating the dual corrector effect in relation to the theoretical additivity of the compounds. Combinatorial profiles were subsequently used to cluster compounds by average linkage analysis, and the distance was determined by Spearman’s rank correlation. The underlying data are depicted as bar plots in Supplemental Figure 1B. Data in A and B are means ± SEM of 3 independent experiments. *P < 0.05, **P < 0.01, by 1-way ANOVA followed by Turkey’s post hoc test.

Figure 2. VX-445 binds to and changes the unfolding trajectory of CFTR-NBD1.

(A) Representative surface plasmon resonance (SPR) sensorgram for the binding of VX-445 (0–200 μM) to immobilized F508del–NBD1-1S. (B) Binding isotherms for VX-445 binding to immobilized F508del–NBD1-1S or WT–NBD1-1S as determined by SPR (n = 3). Curve fitting was performed as described in Methods. (C) Aggregation rates observed in NBD1 aggregation assays between 30 and 50 minutes for the different compounds were normalized by the rate observed for F508del–NBD1-1S in 1% DMSO (n = 4). Compound/chaperone concentrations were 50 or 100 μM VX-445, 10% glycerol, 100 μM VX-661, 100 μM 3151, 10 μM DnaK, or DnaK-DnaJ-GrpE at 10, 2, and 10 μM, respectively. (D) Protein secondary structure stability was studied by far-UV CD spectra of F508del–NBD1-1S. CD scans between 250 and 195 nm were taken every minute at 32°C in the presence of vehicle control (1% 1,4-Dioxane), 100 μM VX-661, or 100 μM VX-445. CD scans obtained at different time intervals of 1 representative experiment were overlaid. (E) Quantification of the ellipticity values (in mDeg) observed at 207 nm. Values were plotted as a function of time in the presence of vehicle control (1% 1,4-Dioxane); 50, 100, or 200 μM VX-445; or 100 μM VX-661 (n = 2–3). Continuous lines were derived by 4-point smooth iteration. Data in B, C, and E are means ± SEM of the indicated number of independent experiments. *P < 0.05, **P < 0.01 by 1-way ANOVA followed by Turkey’s post hoc test.

Figure 3. Trikafta mediated correction of CFTR^{F508del/F508del} in HNE.

(A) Effect of indicated single correctors or corrector combinations on the I_sc of human nasal epithelia with CFTR^{F508del/F508del} genotype (CF-HNE). CFTR-mediated currents were induced by sequential acute addition of forskolin (Fsk, 20 μM, arrow) and VX-770 (770, 10 μM, filled arrowhead), followed by CFTR inhibition with CFTR_inh-172 (Inh-172, 20 μM, open arrowhead) in intact monolayers with basolateral-to-apical chloride gradient. (B) Quantification of the CFTR_inh-172 inhibited current, after stimulation as in A, in CF-HNE isolated from 5 different homozygous F508del CF patients after single correctors (VX-809 and VX-661, 3 μM; VX-445, 2 μM; 24 hours, 37°C), corrector plus chronic potentiator (cVX-770, 1 μM, 24 hours, 37°C) or corrector combination treatment expressed as percentage of WT-CFTR currents in WT-HNE from 10 donors. **P < 0.01 by paired 2-tailed Student’s t test. (C) The fraction of potentiator-independent (forskolin-induced) current in HNE treated with single corrector or corrector combination. **P < 0.01 by paired 2-tailed Student’s t test followed by Bonferroni’s FDR correction.

Figure 4. Efficacy, combinatorial profiling, and clustering of mechanistic classes of correctors in rare CFTR folding mutants.

(A) PM density of the indicated CFTR mutants alone and after VX-661 (3 μM), VX-445 (2 μM), or combination treatment expressed as percentage of WT-CFTR in CFBE41o- (n = 3). Data are means ± SEM. *P < 0.05 and **P < 0.01 by 1-way ANOVA followed by Turkey’s post hoc test. (B) Heatmaps of the combinatorial profiling established by calculating the dual corrector effect in relation to the theoretical additivity of the compounds for S13F, S492F, V520F, L1077P, and M1101K-CFTR in CFBE41o-. Combinatorial profiles were subsequently used to cluster compounds by average linkage analysis, and the distance was determined by Spearman’s rank correlation. The underlying data are depicted as bar plots in Supplemental Figure 5, A–E.

Figure 5. Efficacy of Trikafta for the functional correction of rare CFTR folding mutants in HNE.

(A–E) Effect of indicated single correctors (VX-661, ABBV-2222, and FDL169, 3 μM; VX-445, 2 μM; 24 hours), corrector plus chronic potentiator (cVX-770, 1 μM, 24 hours), or corrector combinations on the I_sc of HNE with CFTR^G85E/G85E (A, n = 3), CFTR^{V520F/1717-1G->A} (B, n = 3), CFTR^Y569D/Y569D (C, n = 3), CFTR^{M1101K/M1101K} (D, n = 3), or CFTR^{N1303K/N1303K} (E, n = 3) genotype. Representative traces (top panels) and quantification of the CFTR_inh-172 inhibited current expressed as percentage of WT-CFTR currents in HNE from 10 donors (bottom panels). CFTR-mediated currents were induced by sequential acute addition of forskolin (Fsk, 20 μM) and VX-770 (10 μM) followed by CFTR inhibition with CFTR_inh-172 (Inh-172, 20 μM) in an intact monolayer with basolateral-to-apical chloride gradient. Data are means ± SEM of 3 measurements. *P < 0.05 and **P < 0.01 by 1-way ANOVA followed by Turkey’s post hoc test.