Structure-based discovery of CFTR potentiators and inhibitors

Abstract

The cystic fibrosis transmembrane conductance regulator (CFTR) is a crucial ion channel whose loss of function leads to cystic fibrosis, whereas its hyperactivation leads to secretory diarrhea. Small molecules that improve CFTR folding (correctors) or function (potentiators) are clinically available. However, the only potentiator, ivacaftor, has suboptimal pharmacokinetics and inhibitors have yet to be clinically developed. Here, we combine molecular docking, electrophysiology, cryo-EM, and medicinal chemistry to identify CFTR modulators. We docked ∼155 million molecules into the potentiator site on CFTR, synthesized 53 test ligands, and used structure-based optimization to identify candidate modulators. This approach uncovered mid-nanomolar potentiators, as well as inhibitors, that bind to the same allosteric site. These molecules represent potential leads for the development of more effective drugs for cystic fibrosis and secretory diarrhea, demonstrating the feasibility of large-scale docking for ion channel drug discovery.

Keywords: ABC transporter; anion channel; inhibitors; large-scale docking; ligand discovery; potentiators.

Figure 1:. Ultra-large docking screen identifies CFTR potentiators.

(A) The workflow of this study. (B) Compound Z2075279358 (‘358) potentiates ΔF508 CFTR. CFBE41o⁻ cells homozygous for ΔF508 CFTR were cultured with 1 μM lumacaftor to facilitate surface-expression and pre-treated with 20 μM forskolin to activate ΔF508 CFTR by PKA-phosphorylation. The relative potentiation was calculated as the ratio of flux rates with and without 1 μM potentiator. Data points represent the means and standard errors (SEs) of 6 to 8 measurements (each shown as a dot). (C) Potentiation activity of 10 μM GLPG1837 or 5 μM compound against WT CFTR fused to a carboxy-terminal GFP tag. Inside-out membrane patches containing WT CFTR were excised from CHO cells and then fully phosphorylated by protein kinase A (PKA) in the presence of 3 mM ATP. The fold stimulation is defined as the ratio of the current in the presence and absence of added compound. Data represent means and SEs of 3–33 patches with individual measurements shown as dots. Statistical significance relative to absence of compound was tested by two-tailed Student’s t-test with Benjamini-Hochberg correction (ns: not significant; *P < 0.05; **P < 0.01; ****P = 4.2×10⁻¹⁴). (D) The 2D structures of the potentiators GLPG1837, ivacaftor, and the positive hits from the initial screen. (E) Representative macroscopic current trace and dose-response curve of WT CFTR in response to perfusion with ‘358. CFTR-containing membrane patches were fully phosphorylated by PKA. The current in the presence of 3 mM ATP before titration was used to normalize the current potentiated by different concentrations of ‘358. The EC₅₀ is estimated to be 2.2 ± 0.6 μM by fitting the dose-responses with the Hill equation. Data represent means and SEs from 3 patches.

Figure 2:. New modulators identified through an analog screen.

(A) Potentiation activity of ‘358 analogs. All compounds were tested at a single concentration of 10 μM in inside-out membrane patches containing fully phosphorylated WT CFTR. The current stimulation levels of GLPG1837 and ‘358 are indicated as dashed lines. (B) Reported structures versus the NMR-determined structures. (C) Docked poses of the reported structure of ‘853 (left) versus the NMR-determined structure of ‘853 (right). (D) (S)-’853 potentiates WT CFTR currents, while (R)-’853 mildly inhibits them in inside-out patches. Both enantiomers were perfused at 100 μM concentration. Data represent means and SEs of 20 ((S)-’853) or 18 ((R)-’853) patches. Statistical significance relative to no effect was tested by two-tailed Student’s t-test (****P = 2.3×10⁻⁹ for (S)-’853 and P = 1.6×10⁻⁷ for (R)-’853). (E) Competition assay showing that the presence of (R)-’853 (100 μM) diminishes the potentiating effect of (S)-’853 (100 μM). (F) Dose-response curve of (S)-’853 versus the racemic mixture (±)-’853, as described for Figure 1E. The EC₅₀ for (S)-’853 was estimated to be 2.1 ± 0.9 μM. Data represent means and SEs of 2–20 patches. 3 mM ATP was used in all panels.

Figure 3:. Z1834339853 binds to the same site as ivacaftor and GLPG1837.

(A) Cryo-EM structure of phosphorylated and ATP-bound CFTR (E1371Q) in complex with ‘853. (B) Zoomed-in views of the density of ‘853 (top) and a comparison between the docked pose (salmon) and the cryo-EM pose (magenta). Compared to the experimentally determined pose, the docked ‘853 shifted towards R933. This difference is likely due to the presence of an unknown density between R933 and the potentiator (Liu et al., 2019), which was not modeled for docking. (C) Representative macroscopic current traces and dose-response curves of fully phosphorylated WT, S308A, and Y304A CFTR in response to perfusion of (S)-’853 onto inside-out excised membrane patches. 3 mM ATP was used. Each data point represents the mean and SEs determined from 3 to 12 patches.

Figure 4:. Medicinal chemistry leads to novel CFTR potentiators and inhibitors.

(A) General formula of newly synthesized ‘853 analogs. (B) Effects of ‘853 analogs on currents measured in inside-out excised membrane patches containing fully phosphorylated WT CFTR. Measurements were made with 3 mM ATP. Data represent means and SEs of 2 to 46 patches. (C) The chemical structures (top), dose-response curves (middle), and docked poses (bottom) of three of the most efficacious potentiators. Data represent means and SEs of 5–11 patches. (D) The chemical structures (top), dose-response curves (middle), and docked poses (bottom) of the two most efficacious inhibitors. The dose-responses of the Y304A and S308A variants in response to I1412 perfusion were also shown (left). Data represent means and SEs of 2–8 patches. (E) Pharmacological potentiation and inhibition of iodide flux rates in CFBE41o⁻ cells homozygous for ΔF508 CFTR. 1 μM lumacaftor was used in the culture to facilitate surface-expression. A pre-treatment with 20 μM forskolin was used to activate ΔF508 CFTR. The relative potentiation was calculated as the ratio of flux rates with and without modulator at the indicated concentrations. Data points represent the means and SEs of 2 (CFTR_inh-172) or 3 (all other conditions) experiments. Statistical significance relative to a DMSO only treatment was tested by one-way analysis of variance (ANOVA) (*P = 0.049 for I1412, P = 0.041 for I1422, and P = 0.015 for CFTR_inh-172; ****P = 9.0×10⁻⁷ for GLPG-1837, P = 7.8×10⁻⁶ for Ivacaftor, P = 3.0×10⁻⁶ for I1421, P = 1.0×10⁻⁵ for I1408, P = 6.1×10⁻⁷ for (S)-SX263, and P = 3.1 ×10⁻⁶ for BBG3). (F) Representative macroscopic current traces of fully phosphorylated WT CFTR in response to perfusion of GLPG-1837, I1421, and Elexacaftor onto inside-out excised membrane patches. 3 mM ATP, 10 μM GLPG-1837, and 1 μM Elexacaftor were used. (G) Relative potentiation of WT CFTR currents by Elexacaftor without or with GLPG1837 or I1421. Data represent means and SEs of 6 to 46 patches. Some data points are replotted from panel B. Statistical significance was tested by one-way ANOVA (*P = 0.026; ****P = 6.3×10⁻⁶ for Elexacaftor + GLPG-1837 versus Elexacaftor only, P = 2.6×10⁻¹² for Elexacaftor + GLPG-1837 versus GLPG-1837 only, and P = 9.3×10⁻⁵ for Elexacaftor + I1421 versus I1421 only).

Figure 5:. The activity of I1421 against 10 CF-causing mutations.

(A) The positions of the mutations mapped onto dephosphorylated and ATP-free CFTR (PDB 5UAK). (B) Representative macroscopic current traces in response to I1421 (10 μM) perfusion onto inside-out membrane patches excised from CHO cells. 3 mM ATP was used. (C) Potentiation activity of I1421 versus GLPG1837. The mean and SE values were determined from 2 to 7 patches. (D) Pharmacokinetic analysis of compound I1421. Plasma concentration-time profiles in male C57BL/6N mice following a single subcutaneous (SC), intraperitoneal (IP), per-oral (PO) (dose 10 mg/kg) or intravenous (IV) (3 mg/kg) administration. Data represent means and SDs. (E) Selected pharmacokinetic parameters of I1421. ${\text{C}}_{\max }$: peak plasma concentration; ${\text{T}}_{\max }$: the time when the peak plasma concentration was observed; ${\text{AUC}}_{\text{last}}$: the areas under the concentration time curve; ${\text{T}}_{1/2}$: terminal half-life; CL: clearance, ${\text{V}}_{\mathrm{ss}}$: steady-state volume of distribution; %F: %bioavailability.

Scheme 1.

Synthesis of the enantiopure derivatives of ‘853, modified at the phenoxy site, reagents and conditions: (a) 1. NaH, DMF, 0 °C, 30 min, 2. (R)-3-bromo-2-methylpropan-1-ol or (S)-3-bromo-2-methylpropan-1-ol, DMF, rt, 18–31 h, (50–72% crude); (b) 1. CrO₃, H₂SO₄, H₂O, acetone, 0 °C, 3–5 h, 2. iPrOH, 0 °C -> rt (48–64%, crude); (c) 1. EDC × HCl, HOAt, DMF, rt, 2. 6-fluoro-1H-indazole-3-amine (13), rt, 1–3 h (15–53%).

Scheme 2.

Synthesis of the ‘853 derivative renouncing the phenoxy site, reagents and conditions: (a) 1. isobutyryl chloride, HOAt, DIPEA, DMF, 0 °C, 1 h, 2. 6-fluoro-1H-indazole-3-amine (13), microwave irradiation (90%).

Scheme S3.

Synthesis of the ‘853 derivatives with modifications in positions 3 or 4, reagents, and conditions: (a) EDC × HCl, HOAt, DMF, rt (3a: 39%, 3b: 57%, 3c: 29%, 4: 31%); (b) AcCl, pyridine, DMAP, 0 °C to rt, 2 h (72%), (c) NaH, MeI, DMF, rt, 1 h (77%); (d) 1.25 M HCl in MeOH, 115 °C, 2 h (100% crude); (e) N₂H₄, EtOH, 70 °C, 17 h (15%).

Scheme 4.

Synthesis of the ‘853 derivatives with alkoxy substituents in positions 6, reagents and conditions: (a) PyBOP, HOAt, DMF, microwave irradiation (5a: 56%, 5b: 54%); (b) EDC × HCl, HOAt, DMF, rt (5c: 23%, 5d: 54%, 5e: 44%).

Scheme 5.

Synthesis of novel indazole scaffolds and the corresponding acyl derivatives, reagents and conditions: (a) NH₄Cl, PyBOP, DIPEA, DMF, rt, 12 h (83%); (b) EDC × HCl, HOAt, DMF, rt, 3 h (5f: 64%, 5g: 53%, F-5g: 45%); (c) N₂H₄, BuOH, 120 °C, 4 h (69%).

Scheme 6.

Synthesis of the ‘853 derivatives with modifications in the positions 3 and 6, reagents and conditions: (a) EDC × HCl, HOAt, DMF, rt (7a: 64%, 7b: 14%).

Scheme 7.

Synthesis of the ‘853 derivatives further modified with a formyl group, reagents and conditions: (a) 1. HCO₂H, Ac₂O, THF, 60 °, 2 h, 2. 5e, 0 °C, 30 min (42%); (b) 1. HCO₂H, Ac₂O, THF, 60 °C, 2 h, 2. 5e or 7b, 0 °C to rt, for 6b: 60 min, for 8: 90 min (6b: 37%, 8: 28%).

Scheme 8.

Synthesis of the fused acylindazole derivative (a) (acetonitrile)[2-biphenyl)di-tert-butylphosphine]gold(I) hexafluoroantimonate, toluene, 0 °C (crude); (b) H₂/Pd(OH)₂/C, MeOH, rt, 2 h (crude); (c) N-bromosuccinimide, CH₃CN, 0 °C (23% over three steps); (d) 1. BuLi, THF, −79 °C. 2. dimethylmalononitrile, THF, −79 °C (54%); (e) N₂H₄, BuOH, 120 °C, 22 h (76%); (f) EDC × HCl, HOAt, DMF, microwave irradiation (61%).

Scheme 9.

Synthesis of the ‘853 derivatives with a modified spacer, reagents and conditions: (a) 1. EDC × HCl, HOAt, DMF, rt, 10 min, 2. 6-fluoro-1H-indazole-3-amine (13), rt, 10a and 10b: 3 h, 10c: 2 h, 10d: 1 h, 10e: 1.5 h (10a: 27%, 10b: 45%, 10c: 43%, 10d: 30%, 10e: 42%).