Therapeutic Targeting of Fumaryl Acetoacetate Hydrolase in Hereditary Tyrosinemia Type I

Abstract

Fumarylacetoacetate hydrolase (FAH) is the fifth enzyme in the tyrosine catabolism pathway. A deficiency in human FAH leads to hereditary tyrosinemia type I (HT1), an autosomal recessive disorder that results in the accumulation of toxic metabolites such as succinylacetone, maleylacetoacetate, and fumarylacetoacetate in the liver and kidney, among other tissues. The disease is severe and, when untreated, it can lead to death. A low tyrosine diet combined with the herbicidal nitisinone constitutes the only available therapy, but this treatment is not devoid of secondary effects and long-term complications. In this study, we targeted FAH for the first-time to discover new chemical modulators that act as pharmacological chaperones, directly associating with this enzyme. After screening several thousand compounds and subsequent chemical redesign, we found a set of reversible inhibitors that associate with FAH close to the active site and stabilize the (active) dimeric species, as demonstrated by NMR spectroscopy. Importantly, the inhibitors are also able to partially restore the normal phenotype in a newly developed cellular model of HT1.

Keywords: drug discovery; fumaryl acetoacetate hydrolase; metabolic rare disease; nuclear magnetic resonance; rare disease; tyrosine; tyrosinemia; tyrosinemia type I.

Figure 1

Compound induced stability experiments on fumarylacetoacetate hydrolase (FAH). (A) Typical profile for the stability screening, where the black circles correspond to the experimental data (converted to fraction folded) and the red line is the best fit to the dataset. (B) Reproducibility of the Tm values obtained by using the linear extrapolation model on the sigmoidal decays. The purple circles correspond to the experimental data, with an average value/uncertainty represented by the solid/dashed lines. Equivalent values for the reference protein (in the absence of compound) are shown in blue. (C) Thermal denaturation experiment for WT-FAH in the absence (black circles) and in the presence of increasing concentrations of the compound G1.30.B10 (light purple, 15 eq.; dark purple, 30 eq.; blue, 45 eq.). (D) Aggregation tendency of V166G-FAH over time (monitored by the ellipticity at 222 nm), in the absence (black circles) and in the presence (blue squares) of 15 eq. of G1.30.B10. The red lines correspond to the best exponential fittings to the experimental datasets.

Figure 2

Eukaryotic and in vitro assays for a selected set of compounds. (A–C) Representative fluorescence microscopy images for the eukaryotic cellular assay. M1 cells incubated with (A) the proteasomal inhibitor MG132, (B) a compound of the library showing no effect on FAH homeostasis and (C) a compound that increases protein accumulation in the cytosol. (D) Percentage of positive cells for a set of selected compounds and for the four different mutant constructs under consideration: V166G-FAH-GFP, purple bars; A35T-FAH-GFP, violet bars; W234G-FAH-GFP, brown bars and T294P-FAH-GFP, black bars. (E) Change in the thermal stability (T_m) for the selected set of compounds and percentage of inhibition or activation as compared to the reference reaction for the set of compounds in the enzyme activity assay (E.a.), in the presence of 15 and 30 equivalents of a given compound as indicated. Error bars are obtained from triplicate measurements. (F) Region of the HPLC chromatogram corresponding to the fumaryl acetoacetate (FAA) elution. Legend: control experiment without enzyme (blue line), enzymatic reaction (red line), enzymatic reaction in the presence of compound (green line).

Figure 3

Inhibitors and chaperones found in this study. (A) Molecular formulas for the compounds under consideration. (B) Representative examples of the binding affinities as determined by saturation transfer difference spectroscopy. Blue circles and black squares correspond to the experimental data for G1.30.B10 and G2.HT1.1, respectively. The solid lines reflect the best fitting of the experimental data to Equation (1). (C) Diffusion-ordered 2D NMR spectroscopy (DOSY) experiment performed on WT-FAH (blue), the unstable variant V166G-FAH (red), and the same protein in the presence of 10 eq. of G1.30.B10 (green), G2.HT1.1 (purple) or G2.HT1.3 (yellow). The diffusion coefficient is proportional to the molecular weight of the particle. (D) Structural model of the complex between FAH and a selected set of compounds including G1.30.B10 (blue), G2.HT1.2 (purple), and G2.HT1.3 (red). The longer compounds occupy both the active site and the approximation cavity, while G1.30.B10 independently binds to each of these sites. (E) Methionine methyl region of the ¹H,¹³C correlation spectrum and chemical shift perturbation in the presence of the compounds. Both insets correspond to M140, in the absence (light blue) and in the presence of compounds, using the same color code. Notice the signal splitting in the presence of G1.30.B10 (upper inset).

Figure 4

A cellular model of HT1. (A) Scheme for the generation of an HT1 cellular model, based on the genomic modification (G337S in the fah gene) using the CRISPR/Cas9 technology. (B) Regions of the ¹H-NMR spectrum of the extracellular media in contact with: WT-HEK cells (blue), G337S-FAH-HEK cells (red), or the latter but incubated with G1.30.B10 (green) or G2.HT1.1 (purple). The cellular model adequately reflects the phenotype of HT1 patients since it accumulates succinylacetone (SA) and hydroxyphenylpyruvate (HPA), among other metabolites. (C) Bar plot with the changes in SA and HPA, quantified from the 1H-NMR spectrum. P-values of < 0.01, 0.001 and 0.0001 are represented by **, *** and **** respectively.