HDAC inhibitors rescue multiple disease-causing CFTR variants

Abstract

Understanding the role of the epigenome in protein-misfolding diseases remains a challenge in light of genetic diversity found in the world-wide population revealed by human genome sequencing efforts and the highly variable response of the disease population to therapeutics. An ever-growing body of evidence has shown that histone deacetylase (HDAC) inhibitors (HDACi) can have significant benefit in correcting protein-misfolding diseases that occur in response to both familial and somatic mutation. Cystic fibrosis (CF) is a familial autosomal recessive disease, caused by genetic diversity in the CF transmembrane conductance regulator (CFTR) gene, a cyclic Adenosine MonoPhosphate (cAMP)-dependent chloride channel expressed at the apical plasma membrane of epithelial cells in multiple tissues. The potential utility of HDACi in correcting the phenylalanine 508 deletion (F508del) CFTR variant as well as the over 2000 CF-associated variants remains controversial. To address this concern, we examined the impact of US Food and Drug Administration-approved HDACi on the trafficking and function of a panel of CFTR variants. Our data reveal that panobinostat (LBH-589) and romidepsin (FK-228) provide functional correction of Class II and III CFTR variants, restoring cell surface chloride channel activity in primary human bronchial epithelial cells. We further demonstrate a synergistic effect of these HDACi with Vx809, which can significantly restore channel activity for multiple CFTR variants. These data suggest that HDACi can serve to level the cellular playing field for correcting CF-causing mutations, a leveling effect that might also extend to other protein-misfolding diseases.

Figure 1

PXD-101, LBH-589 & FK-228 rescue F508del CFTR trafficking in CFBE-F508del. Chemical structure of PXD-101 (A), LBH-589 (B) and FK-228 (C). Immunoblot analysis (upper) and quantifications (middle and lower) of CFTR expression following treatment of CFBE-F508del cells with different concentration of PXD-101 (D), LBH-589 (E) and FK-228 (F). Data are presented in middle panel as fold change relative to DMSO treatment (mean ± SEM, n ≥ 3). Lower panel represent a quantification of C/T ratio expressed as a % (mean ± SEM, n ≥ 3).

Figure 2

Establishment of F508del CFTR rescue following PXD-101, LBH-589 or FK-228 treatment. (A) Quantification of CFTR mRNA level following treatment of CFBE-F508del cells with PXD-101, LBH-589 or FK-228 after 2, 6, 12 and 24 h. mRNA was standardized by quantification of GUS mRNA, and all values were expressed relative to GUS (mean ± SEM, n ≥ 3). (B, C & D) Immunoblot analysis of CFTR expression following treatment of CFBE-F508del cells with 2 μm PXD-101, 10 nm LBH-589 and 5 nm FK-228, respectively.

Figure 3

PXD-101, LBH-589 & FK-228 rescue F508del CFTR function in F508del CFBE-F508del. Representative Fluorescent Imaging Plate Reader traces (upper) and quantification (lower) of YFP-quenching following treatment of CFBE-F508del-YFP cells with different concentrations of PXD-101 (A), LBH-589 (B) and FK-228 (C). Data are presented as % relative to baseline (mean ± SEM, n ≥ 3). In all panels, ^* indicates significant differences (P < 0.05) relative to DMSO treatment as determined by two-tailed t-test. (D) Representative FLIPR traces (left) and quantification (right) of YFP-quenching of CFBE-YFP, CFBE-F508del-YFP and CFBE-Wt-YFP cells. Data are presented as % relative to baseline (mean ± SEM, n ≥ 3).

Figure 4

Synergistic effect of PXD-101, LBH-589 or FK-228 in combination with Vx809. (A) Immunoblot analysis (upper) and quantification (lower) of CFTR expression following treatment of CFBE-F508del cells with 1 μm of PXD-101, 1.25 nm LBH-589 or 0.6 nm FK-228 with or without 3 μm of Vx809. Data are presented as fold change relative to DMSO treatment (lower left) or as a % of C/T ratio (mean ± SEM, n ≥ 3) (lower right). Representative FLIPR traces (upper) and quantification (lower) of YFP-quenching following treatment of CFBE-F508del-YFP cells with 1 μm of PXD-101 (B), 1.25 nm LBH-589 (C) or 0.6 nm FK-228 (D) with or without 3 μm of Vx809. Data are presented as % relative to DMSO (mean ± SEM, n ≥ 3). A calculated additive effect is shown next to the measured combinatorial effect. Abbreviations: 809 (Vx809), 101 (PXD-101), 589 (LBH-589) and 228 (FK-228). In all panels, * and # indicate significant differences (P < 0.05) relative to DMSO and Vx809 treatment, respectively, as determined by two-tailed t-test. + indicates significant difference between the same HDACi treatment with or without Vx809.

Figure 5

Impact of PXD-101, LBH-589 or FK-228 alone or in combination with Vx809 on F508del CFTR in hBE airway cells homozygous for F508del. (A) Immunoblot analysis (upper) and quantification (lower) of CFTR expression following treatment of F508del/F508del hBE primary cells with 2 μm of PXD-101, 5 nm LBH-589 or 5 nm FK-228 with or without 3 μm of Vx809. Data are presented as fold change relative to DMSO treatment (lower left) or as a % of C/T ratio (mean ± SEM, n ≥ 3) (lower right). Representative short-circuit current (Isc) traces for F508del/F508del hBE following treatment with 2 μm PXD-101 (B), 5 nm LBH-589 (C) and 5 nm FK-228 (D), with or without 3 μm of Vx809. Quantification of Isc for F508del/F508del hBE following treatment with 2 μm PXD-101, 5 nm LBH-589 and 5 nm FK-228 without (E) or with (F) 3 μm of Vx809, respectively (mean ± SEM, n ≥ 3). Abbreviations: 809 (Vx809), 101 (PXD-101), 589 (LBH-589) and 228 (FK-228). In all panels, ^* and # indicate significant differences (P < 0.05) relative to DMSO and Vx809 treatment, respectively, as determined by two-tailed t-test.

Figure 6

HDACi rescue CFTR trafficking and function through multiple pathways. Immunoblot analysis (A) and quantification of HDAC7, HDAC1 (B), HSF1, HSF1 phosphorylated (C), Eif3a, Eif2a and Eif2a phosphorylated (D) expression following treatment of F508del/F508del hBE primary cells with 2 μm of PXD-101, 5 nm LBH-589 or 5 nm FK-228. Data are presented as fold change relative to DMSO treatment (mean ± SEM, n ≥ 3). In all panels, ^* indicate significant differences (P < 0.05) relative to DMSO treatment, as determined by two-tailed t-test.

Figure 7

Characterization of CFTR2 variant trafficking. (A) Cryo-EM model of human CFTR (112) (blue: TMs of MSD1, yellow-green: NBD1, cyan: TMs of MSD2 and green: NBD2) showing the positions of the selected CFTR2 mutations in Magenta. (B) Quantification of C/T ratios following adenovirus transduction of CFTR2 variants into CFBE cells. C/T ratios are expressed as a % (mean ± SEM, n ≥ 3). (C) Quantification of YFP-quenching following adenovirus transduction of CFTR2 variants into CFBE-YFP cells. Data are presented as % normalized to WT (mean ± SEM, n ≥ 3).

Figure 8

Effect of LBH-589 or FK-228 alone or in combination with Vx809 on CFTR2 variant trafficking. (A to J) Immunoblot analysis of CFTR expression following treatment of CFBE transduced cells with 5 nm LBH-589 or 2.5 nm FK-228 with or without 3 μm Vx809 for P67L, G85E, E92K, S492F, F508del, G551D, R560T, L1077P, M1101K and N1303K variants, respectively. (K) Quantification of the C/T ratio for each of the 10 variants expressed as a % (mean ± SEM, n ≥ 3). Abbreviations: 589 (LBH-589) and 228 (FK-228).

Figure 9

Effect of LBH-589 or FK-228 alone or in combination with Vx809 on CFTR2 variant activity. Quantification of YFP-quenching following treatment with 5 nm LBH-589 or 2.5 nm FK-228 with or without 3 um Vx809 of CFBE-YFP transduced cells with P67L, G85E, E92K, S492F, F508del, G551D, R560T, L1077P, M1101K and N1303K variants. Data are presented as a % of YFP quenching normalized by CFBE-YFP identically treated cells and relative to CFBE-YFP cells transduced with WT CFTR and treated with DMSO.

Figure 10

Structural modeling of CFTR2 variants. (A to C) Cryo-EM model of CFTR2 variants using PyMOL software (Materials and Methods). Red shows impact of the mutant residue on other residues. Rotomers are predicted to minimize impact against other residues. (A) Cluster A: G85E, S492F, G551D, R560T, L1077P, M1101K and N1303K. (B) Cluster B: P67L and E92K variants. (C) Cluster C: G551D variant.