A Selective Competitive Inhibitor of Aldehyde Dehydrogenase 1A3 Hinders Cancer Cell Growth, Invasiveness and Stemness In Vitro

Simple Summary The aldehyde dehydrogenases enzymes (ALDHs) are promising drug targets in cancer therapy. ALDHs are members of an enzymatic superfamily composed by 19 isoforms involved in the oxidation of aldehydes, with a scavenger role. Among them, the isoform ALDH1A3 is a cancer biomarker since it is highly expressed in cancer stem cells characterized by a marked drug resistance and the capacity to promote self-renewal, clonogenic growth and tumour-initiating capacity. In this paper, we present the first highly potent and selective ALDH1A3 inhibitor able to induce cytotoxic effects and to reduce cell migration and stemness of ALDH1A3-positive cancer cells. We propose the targeting of the ALDH1A3 enzyme as a promising approach for improving the treatments outcomes of patients affected by ALDH1A3-positive cancers. Abstract Aldehyde dehydrogenase 1A3 (ALDH1A3) belongs to an enzymatic superfamily composed by 19 different isoforms, with a scavenger role, involved in the oxidation of a plethora of aldehydes to the respective carboxylic acids, through a NAD+-dependent reaction. Previous clinical studies highlighted the high expression of ALDH1A3 in cancer stem cells (CSCs) correlated to a higher risk of cancer relapses, chemoresistance and a poor clinical outcome. We report on the structural, biochemical, and cellular characterization of NR6, a new selective ALDH1A3 inhibitor derived from an already published ALDH non-selective inhibitor with cytotoxic activity on glioblastoma and colorectal cancer cells. Crystal structure, through X-Ray analysis, showed that NR6 binds a non-conserved tyrosine residue of ALDH1A3 which drives the selectivity towards this isoform, as supported by computational binding simulations. Moreover, NR6 shows anti-metastatic activity in wound healing and invasion assays and induces the downregulation of cancer stem cell markers. Overall, our work confirms the role of ALDH1A3 as an important target in glioblastoma and colorectal cells and propose NR6 as a promising molecule for future preclinical studies.


Expression and Purification of Human ALDHs
A common protocol to express and purify all three members of the ALDH1A subfamily has been optimized [2,3]. Briefly, induced cells, collected by centrifugation, were re-suspended in an appropriate volume of lysis buffer (50 mM Na2HPO4, 300 mM NaCl, 1 mM β-mercaptoethanol, 20 mM imidazole, pH 7.5) added of benzonase nuclease (250 U/L) and Protease inhibitor cocktail from SIGMA. After ultracentrifugation, the clarified sample was purified by a His-tag affinity chromatography followed by sizeexclusion chromatography, using an AKTA FPLC system, at 4°C. To evaluate the purity and homogeneity of the protein, after each purification step, eluted fractions were analyzed by SDS-PAGE and the protein quantification was always determined by Bradford protein assay. This procedure allowed us to obtain about 20 mg of pure and active human ALDH1A3, ALDH1A2 and ALDH1A1 used for all crystallization trials and kinetic analyses.

High-throughput screening to determine Inhibitory Efficacy of Imidazo[1,2-a]pyridine Derivatives
All the three imidazo[1,2-a]pyridine derivatives were screened as inhibitors of ALDH1A3, ALDH1A1, and ALDH1A2 isoenzymes, using the previously reported continuous spectrometric assay for ALDH1A3 optimized to suit Greiner Bio-One 96-UV-Transparent Microplate [4]. Tests were carried out by using 100 μL of reaction mixture, containing 20 mM Tris HCl pH 8.0, 1 mM β-mercaptoethanol, 150 mM KCl, 15% DMSO 500 μM NAD + , and 2.6 μM pure recombinant isoenzymes. Reactions were started by the addition of 20 mM acetaldehyde. Changes in absorbance at 340 nm (εNADH = 6220 M-1 cm-1) were monitored for 30 min in Tecan Sunrise spectrophotometer, at 25 °C. The activity of the three enzymes was tested in the presence of each synthesized compound, at 25 μM concentration in triplicates. To avoid any possible error in the data analysis, in every plate was tested the natural NAD + reduction, using the same mix without the addiction of acetaldehyde. The intrinsic signal was subtracted from every data analysis.

Enzyme Kinetic Analysis to Calculate IC50 and Ki Values of the compounds
Selected compounds showing the best inhibitory activity at 25 M were further investigated for their inhibitory efficacy, to calculate their IC50 values. The enzymatic inhibition assays were performed in a total volume of 200 μL of 20 mM Tris HCl pH 8.0, 1 mM β-mercaptoethanol, 150 mM KCl, 15% DMSO, 500 μM NAD + , 2.6 μM of pure recombinant isoenzymes in the presence of different inhibitors concentrations from 100 μM to 0.78125 μM and preincubated for 5 min. To calculate the Ki value of NR6 in complex with all the three ALDH1A isoenzymes, the activity was tested in presence of different concentrations of the substrate and inhibitor, to understand the mechanism of inhibition and calculate the Ki value. The enzymatic inhibition assays were performed in a total volume of 100 μL of 20 mM Tris HCl pH 8.0, 1 mM β-mercaptoethanol, 150 mM KCl, 15% DMSO, 500 μM NAD + , 2.6 μM of pure recombinant isoenzymes in the presence of different inhibitors concentrations from 200 μM to 1 μM and different concentrations of acetaldehyde from 20 mM to 1 mM for ALDH1A2 and ALDH1A3 and from 1 mM to 0.2 mM for ALDH1A1. The kinetic parameters were determined by fitting the measured data to a Michaelis-Menten curve [5] by using SigmaPlot (SigmaPlot Extract Graphs and Data Analysis (StatSys v12.3)).

Western Blotting Analysis
Cells were lysed in lysis buffer composed of 20 mM HEPES, 100 mM NaCl, 5 mM EDTA, 1% Nonidet P-40+ Protease & Phosphatase Inhibitor Cocktail (Sigma). Proteins quantification was performed with Bradford Reagent (Sigma-Aldrich), and proteins were resolved on SDS-PAGE and transferred with TurboBlot system (BioRad). Densitometry analysis was performed with ImageLab program (Bio-Rad).

Gene Expression Analysis
Cells were lysated with Trizol reagent (Life-Technologies) and extracted with chloroform. 1μg mRNA was reverse transcribed with SENSIFAST kit as manufacturer's protocol (Aurogene) and cDNA expression determined with qPCR using SYBR-green (Bio-Rad) and detected by the CFX96 Real-Time System (Biorad). Expression data were normalized to actin expression.

Wound Healing Assay
When confluent monolayers of U87, HCT116 and 4T1 cells were established, we performed a cross-shaped scratch with tip, as reported in [6]. Then, the cells were washed twice with PBS to remove residual cell debris. Cells were then incubated with NR6 (10 nM; concentrations that occurs less mortality in 24h) for 24 h and pictures of a defined wound spot were made at different time points. The area of the wound in the microscopic pictures was measured and analysed using Image J software (National Institutes of Health, MD, USA).    [7] and ALDH1A2 (6B5G) [8]. All the three domains are highlighted (catalytic, NAD + and oligomerization).

Figure S11. (A)
Sequence alignment between all the nineteen isoforms. The tyrosine is not conserved in any other ALDH isoenzymes. None of the residues close to the same region present a hydrophobic side chain that could supply the absence of the tyrosine to coordinate the heterocycle with the same kind of interaction (edge-to-face π-π stacking).  Table:   Table S1. Data collection and refinement statistics of the three-dimensional structure of human ALDH1A3 in complex with NR6.