Photoaffinity Ligand of Cystic Fibrosis Corrector VX-445 Identifies SCCPDH as an Off-Target

Abstract

Cystic fibrosis (CF) pharmacological correctors improve the cystic fibrosis transmembrane conductance regulator (CFTR) protein trafficking and function. Several side effects of these correctors and adverse drug interactions have been reported, emphasizing the need to understand off-targets. We synthesized VU439, a functionalized photoaffinity ligand (PAL) of VX-445. Chemoproteomics analysis by mass spectrometry (MS) was used to identify cross-linked proteins in CF bronchial epithelial cells expressing F508del CFTR. We identified saccharopine dehydrogenase-like oxidoreductase (SCCPDH), an uncharacterized putative oxidoreductase, as a VX-445-specific off-target. We also characterized changes in the metabolomic profiles of cells overexpressing SCCPDH to determine the consequence of binding of VX-445 to SCCPDH. These data show dysregulation of amino acid metabolism and a potential inhibitory activity of VX-445 on SCCPDH. The identified off-target may explain the exacerbation of psychological symptoms observed in the clinic, thus emphasizing the need for further optimization of correctors.

1

Photoaffinity ligand retains rescue of F508del CFTR trafficking and function. (A) Molecular structures of VX-445 and photoaffinity ligand VU439. VU439 contains an alkyl diazirine moiety for UV-induced photo-cross-linking and a terminal alkyne for click-modifications. (B) Representative immunoblot image showing F508del CFTR correction by VX-445 and VU439 in HEK293T cells transiently transfected with CFTR plasmids. Cells were treated with respective compound concentrations or vehicle (dimethyl sulfoxide, DMSO) for 24 h before collection. Quantification of CFTR C bands is shown as mean ± standard deviation (SD) (n = 5). VU439 shows retained correction of fully glycosylated F508del CFTR. Quantified band C values in all conditions were normalized to F508del treated with VX-445 at 3 μM. Tubulin is shown as a loading control. Statistical differences were computed via one-way analysis of variance (ANOVA) with Geisser–Greenhouse correction and post hoc Dunnett’s multiple comparisons testing against F508del treated with vehicle. p-values as shown. (C) Representative electrophysiology traces measured by whole cell patch clamp on HEK293 cells transiently expressing wild-type (WT) or F508del CFTR. Data show whole cell current (pA) during stimulation with forskolin (20 μM) and inhibition by CFTR inh-172 (20 μM). VX-445 (5 μM) restored F508del CFTR ion current when stimulated with forskolin. VU439 (5 μM) restored ion current, showing retained functional correction compared to DMSO. (D) Representative I–V curve from (C).

2

SCCPDH was identified as an off-target of VX-445 via affinity purification mass spectrometry. (A) Off-target identification workflow. Cells are treated with VU439, UV cross-linked, and click-modified with TAMRA desthiobiotin azide (TMR). Streptavidin-enriched proteins are digested and analyzed by LC-MS/MS. Competition with the parent compound results in loss of specific targets. (B) Representative SDS-PAGE gel image showing streptavidin-enriched proteins in doxycycline-inducible CFBE F508del cells treated with vehicle (DMSO) or PAL probe (VU439 at 1 μM) or competition (VU439 at 1 μM and VX-445 at 10 μM). Addition of probe allows enrichment of cross-linked proteins modified by TMR. (C) Workflow A was followed and analyzed by data-independent acquisition (DIA) mass spectrometry (n = 8). Volcano plot shows log₂ fold change of enriched protein abundance of probe condition compared to that of vehicle. Enriched proteins with a standard deviation of at least one from the normal distribution of all identified proteins were selected (red dots) for comparison against the competition condition. (D) log₂ fold changes of filtered proteins were compared to competition conditions to identify probe-specific targets (red). SCCPDH showed clear enrichment effectively competed by the parent compound as a specific off-target of VX-445 (blue).

3

Validation of SCCPDH as an off-target of VX-445. (A) Representative immunoblot showing enrichment of PAL-modified SCCPDH by desthiobiotin–streptavidin pulldown. HEK293T cells transiently expressing SCCPDH-Myc construct were treated with probe, UV cross-linked, and clicked with TMR. SCCPDH-Myc was enriched dose-dependently. Competition with VX-445 at 5- or 10-fold excess led to dose-dependent loss of enrichment. (B) Quantification of labeled SCCPDH normalized to 3 μM VU439 treatment reiterates the specificity of VU439 binding (n = 3). Statistical differences were computed via one-sample t test comparing against 3 μM VU439 treatment. Statistical differences were computed via one-way ANOVA with Geisser–Greenhouse correction and post hoc Dunnett’s multiple comparisons testing against SCCPDH-Myc expressing cells treated with 3 μM VU439. p-values as shown. VX-445 outcompeted VU439 in binding to SCCPDH.

4

Global untargeted metabolomics reveal a role of SCCPDH in amino acid metabolism. (A) Schematic of global untargeted metabolomics study. HEK293T cells transfected with GFP or SCCPDH were treated with vehicle (DMSO) or VX-445 (3 μM) (n = 5). Metabolites were extracted from cell pellets for global untargeted metabolomics analysis. Ultrahigh-performance liquid chromatography-MS/MS (UHPLC-MS/MS) data was processed and analyzed to identify key metabolites related to SCCPDH. (B) Global principal components analysis (PCA) plot showing distribution of sample groups as labeled. Samples overexpressing SCCPDH form a cluster separate from GFP ± VX-445. (C) Globally, 1985 metabolites were detected and have a coefficient of variation (CV) < 25%. Venn diagrams show moderate overlap of metabolites identified with p value <0.05 cutoff between all comparisons. (D) Volcano plots showing comparison. Metabolites were further filtered by false discovery rate (FDR) (q value) < 5% and metabolites with high confidence of ID (L1, L2) are labeled and colored red or blue representing increased or decreased abundance, respectively. See also Table S3.

5

Pathways change with SCCPDH overexpression and VX-445 drives these changes in the opposite direction. (A) Heatmap shows metabolites significantly altered across all conditions (FDR < 5%, L1 and L2 annotation only). Metabolites were hierarchically clustered using Pearson distance and average clustering. Among the metabolites that drastically change with SCCPDH overexpression, VX-445 treatment further enhanced the observed changes. See also Figure S3. (B) Pathway analysis performed with the 19 metabolites from (A). Cysteine and methionine metabolism (1), and arginine biosynthesis (2) were the most significant at p < 0.01. Alanine, aspartate, and glutamate metabolism (3), glutathione metabolism (4), glycine, serine, and threonine metabolism (5), and arginine and proline metabolism (6) were observed at p < 0.05. Other pathways at p > 0.05 are listed in Table S4 in order of decreasing significance. Node color and size represent the p value and pathway impact score, respectively. (C) Cysteine and methionine metabolic pathway (1) from (B). Overexpression of SCCPDH resulted in serine accumulation and a decrease of downstream metabolite 2-aminoacrylic acid. Addition of VX-445 in GFP control further highlighted this finding. Metabolites detected in our data set were filled into the pathway as annotated in the key. See also Figure S4. (D) Arginine biosynthesis pathway (2) from (B). Overexpression of SCCPDH resulted in the loss of glutamic acid and glutamine with a small increase in ornithine and arginine. Addition of VX-445 in GFP control further highlighted this finding. Metabolites detected in our data set were filled into the pathway as annotated in the key. See also Figure S5.

6

Inhibition of SCCPDH by VX-445 may result in amino acid metabolism dysregulation resulting in side effects. Amino acids such as glutamate and glutamine were decreased upon overexpression of SCCPDH and VX-445 mitigated this decrease, which may be the underlying cause of psychological symptoms experienced by patients on VX-445. Additionally, metabolites reported to be dysregulated in CF were observed to change in identical directions upon SCCPDH overexpression. VX-445 may inhibit SCCPDH to counteract these dysregulations resulting in unforeseen outcomes such as side effects.