Untargeted Diversity-Oriented Synthesis for the Discovery of New Antitumor Agents: An Integrated Approach of Inverse Virtual Screening, Bioinformatics, and Omics for Target Deconvolution

Abstract

Molecular diversity is one of the most pursued objectives in drug discovery, and diversity-oriented synthesis (DOS) perfectly responds to the achievement of this goal. In this paper, we describe a DOS approach applied to the antitumor field with the aim of identifying new anticancer structures and their associated targets. To accomplish this ambitious project, after an initial stage of phenotypic evaluation, we set up an integrated platform of inverse virtual screening (IVS), bioinformatics, and omics to predict the biological targets of the most promising compounds 31 and 63. Several proteins emerged from this study, and the most interesting ones were assessed by biophysical and in cellulo experiments, leading to the validation of six targets involved in calcium regulation, endoplasmic reticulum stress, and apoptosis. This work allowed us to identify two hit compounds with an interesting antitumor mechanism, but principally, to validate our platform as a fruitful tool for untargeted DOS campaigns.

1

Graphical representation of the present reagent-based UnDOS strategy.

1. Synthesis of Compounds 7, 8, 13, and 14

2. Synthesis of Compounds 22 and 23

3. Synthesis of Compounds 28–31, 33, 34, 37, and 38

4. Synthesis of Compound 41

5. Synthesis of Compounds 48 and 49

6. Synthesis of Compounds 52, 63, and 64

2

Funnel filtering approach used to narrow down the IVS results.

3

Predicted binding between 31 and 63 and CB2R (violet ribbon, PDB: 6PT0). (A,C) 3D view. Cyan dotted lines indicate π–π stacking and the interacting amino acids are labeled. (B,D) 2D view. Green lines represent π–π stacking, polar amino acids are in cyan, hydrophobic ones are in green, and positively charged residues are in blue.

4

Predicted binding between 31 and 63 and TRPM8 (lime ribbon, PDB ID: 8BDC). (A,C) 3D view. Cyan dotted lines indicate π–π stacking and yellow dotted lines represent hydrogen bonds. The interacting amino acids are labeled. (B,D) 2D view. Green lines represent π–π stacking, pink lines are hydrogen bonds, and π–cation interactions are reported with red lines. Polar amino acids are in cyan, hydrophobic ones are in green, and positively charged residues are in blue.

5

Predicted binding between 31 and 63 and BRD4 (blue ribbon, PDB ID: 6DJC and 7MRA, respectively). (A,C) 3D view. Cyan dotted lines indicate π–π stacking, and yellow dotted lines represent hydrogen bonds. The interacting amino acids are labeled. (B,D) 2D view. Green lines represent π–π stacking, and pink lines are hydrogen bonds. Polar amino acids are in cyan, hydrophobic ones are in green, and negatively charged residues are in red.

6

LFQ Proteomics results: Principal component analysis (A) showing clustering of 31 vs ctrl and (B) volcano plot of differentially expressed proteins, (C) STRING network and enrichment (D) of statistically significant proteins, and (E) gene ontology.

7

Untargeted lipidomics results: Principal component analysis (A,B) showing clustering of 31 and 63 vs CTRL and (C,D) volcano plot of differentially modulated lipids, and (E,F) subclass analysis with their average abundance between treatments and ctrl.

8

Untargeted metabolomics results: Principal component analysis (A,B) showing clustering of 31 and 63 vs ctrl and (C,D) volcano plot of differentially modulated metabolites, (E,F) SMPDB enrichments of 31 and 63.

9

The administration of 31 and 63 induces oxidative stress, protein misfolding and apoptosis. ROS assay revealed oxidative stress triggered by compounds 31 and 63. Quantitative analysis of ROS production was reported, after the exposure of compounds (A) 31 and (B) 63 (10, 5, 2.5 μM) for 2 to 8 h. (C,E) ThT test reported protein misfolding levels after exposure to compounds 31 (10, 5,2.5 μM) and 63 (20, 10, 5 μM) for 24 h. The fluorescence was observed by fluorescence microscope (ZOE Fluorescent Cell Imager microscope, Biorad. Magnification, 20×. Scale bar: 100 μm). (D,F) Quantitative analysis was reported. (G) Representative flow cytometry graphs with Annexin V-FITC staining for pro-apoptotic activity on A375 cells treated for 72 h with compounds 31 and 63 at 5, 2.5, and 1.25 μM. (H) Relevant quantitative analyses are shown. Results are expressed as the percentage of apoptotic cells. Data are expressed as mean ± SD of three different experiments performed in triplicate. *p < 0.05 vs Ctrl; **p < 0.01 vs Ctrl; ***p < 0.001 vs Ctrl.

10

Concentration–response curves of compound 63 in the cAMP Hunter assay enzyme fragment complementation chemiluminescent detection kit. The curves show the effect of increasing concentrations of the compound on NKH-477-induced cAMP levels in stable CHO cells expressing the human CB2R. The effect of JWH-133 is reported as a reference compound. Data are reported as means ± standard error of three independent experiments conducted in triplicate and normalized considering the NKH-477 stimulus alone as 100% of the response as indicated in the experimental session. EC₅₀ was determined by GraphPad Prism 9 as the concentration that provokes the response halfway between the baseline (bottom) and maximum response (top) of the fitted dose–response.