Gabriele Macari

Carfilzomib is a last generation proteasome inhibitor (PI) with proven clinical efficacy in the treatment of relapsed/refractory multiple myeloma. This drug is considered to be extremely specific in inhibiting the chymotrypsin-like activity of the 20S proteasome, encoded by the β5 subunit, overcoming some bortezomib limitations, the first PI approved for multiple myeloma therapy which is however burdened by a significant toxicity profile, due also to its off-target effects. Here, molecular approaches coupled with molecular docking studies have been used to unveil that the Insulin-Degrading Enzyme, a ubiquitous and highly conserved Zn²⁺ peptidase, often found to associate with proteasome in cell-based models, is targeted by carfilzomib in vitro. The drug behaves as a modulator of IDE activity, displaying an inhibitory effect over 10-fold lower than for the 20S. Notably, the interaction of IDE with the 20S enhances in vitro the inhibitory power of carfilzomib on proteasome, so that the IDE-20S complex is an even better target of carfilzomib than the 20S alone. Furthermore, IDE gene silencing after delivery of antisense oligonucleotides (siRNA) significantly reduced carfilzomib cytotoxicity in rMC1 cells, a validated model of Muller glia, suggesting that, in cells, the inhibitory activity of this drug on cell proliferation is somewhat linked to IDE and, possibly, also to its interaction with proteasome.

Motivation: Bringing a new drug to the market is expensive and time-consuming. To cut the costs and time, computer-aided drug design (CADD) approaches have been increasingly included in the drug discovery pipeline. However, despite traditional docking tools show a good conformational space sampling ability, they are still unable to produce accurate binding affinity predictions. This work presents a novel scoring function for molecular docking seamlessly integrated into DockingApp, a user-friendly graphical interface for AutoDock Vina. The proposed function is based on a random forest model and a selection of specific features to overcome the existing limits of Vina’s original scoring mechanism. A novel version of DockingApp, named DockingApp RF, has been developed to host the proposed scoring function and to automatize the rescoring procedure of the output of AutoDock Vina, even to nonexpert users. Results: By coupling intermolecular interaction, solvent accessible surface area features and Vina’s energy terms, DockingApp RF’s new scoring function is able to improve the binding affinity prediction of AutoDock Vina. Furthermore, comparison tests carried out on the CASF-2013 and CASF-2016 datasets demonstrate that DockingApp RF’s performance is comparable to other state-of-the-art machine-learning- and deep-learning-based scoring functions. The new scoring function thus represents a significant advancement in terms of the reliability and effectiveness of docking compared to AutoDock Vina’s scoring function. At the same time, the characteristics that made DockingApp appealing to a wide range of users are retained in this new version and have been complemented with additional features.

The anticoagulant therapy is widely used to prevent and treat thromboembolic events. Until the last decade, vitamin K antagonists were the only available oral anticoagulants; recently, direct oral anticoagulants (DOACs) have been developed. Since 55% to 95% of DOACs are bound to plasma proteins, the in silico docking and ligand‐binding properties of drugs apixaban, betrixaban, dabigatran, edoxaban, and rivaroxaban and of the prodrug dabigatran etexilate to human serum albumin (HSA), the most abundant plasma protein, have been investigated. DOACs bind to the fatty acid (FA) site 1 (FA1) of ligand‐free HSA, whereas they bind to the FA8 and FA9 sites of heme‐Fe(III)‐ and myristic acid‐bound HSA. DOACs binding to the FA1 site of ligand‐free HSA has been validated by competitive inhibition of heme‐Fe(III) recognition. Values of the dissociation equilibrium constant for DOACs binding to the FA1 site (ie, ^calcK_DOAC) derived from in silico docking simulations (ranging between 1.2 × 10⁻⁸ M and 1.4 × 10⁻⁶ M) agree with those determined experimentally from competitive inhibition of heme‐Fe(III) binding (ie, ^expK_DOAC; ranging between 2.5 × 10⁻⁷ M and 2.2 × 10⁻⁶ M). In addition, this study highlights the inequivalence of rivaroxaban binding to mammalian serum albumin. Given the HSA concentration in vivo (~7.5 × 10⁻⁴ M), values of K_DOAC here determined indicate that the formation of the HSA:DOACs complexes in the absence and presence of FAs and heme‐Fe(III) may occur in vivo. Therefore, HSA appears to be an important determinant for DOACs transport.

Butylated hydroxytoluene (BHT) is one of the most commonly used synthetic antioxidants in food, cosmetic, pharmaceutical and petrochemical products. BHT is considered safe for human health; however, its widespread use together with the potential toxicological effects have increased consumers concern about the use of this synthetic food additive. In addition, the estimated daily intake of BHT has been demonstrated to exceed the recommended acceptable threshold. In the present work, using BHT as a case study, the usefulness of computational techniques, such as reverse screening and molecular docking, in identifying protein–ligand interactions of food additives at the bases of their toxicological effects has been probed. The computational methods here employed have been useful for the identification of several potential unknown targets of BHT, suggesting a possible explanation for its toxic effects. In silico analyses can be employed to identify new macromolecular targets of synthetic food additives and to explore their functional mechanisms or side effects. Noteworthy, this could be important for the cases in which there is an evident lack of experimental studies, as is the case for BHT.

In this work, the information contained in the contacts between fragments of small-molecule ligands and protein residues has been collected and its exploitability has been verified by using the scoring of docking simulations as a test case for bringing about a proof of concept. Contact statistics between small-molecule fragments and binding site residues were collected and analyzed using a dataset composed of 200,000+ binding sites and associated ligands, derived from the database of the LIBRA ligand binding site recognition software, as a starting point. The fragments were generated by applying the decomposition algorithm implemented in BRICS. A simple “potential” based on the contact frequencies was tested against the CASF-2013 benchmark; its performance was then evaluated through the rescoring of docking poses generated for the DUD-E dataset. The results obtained indicate that this approach, its simplicity notwithstanding, yields promising results that are comparable, and in some cases, superior, to those obtained with other, more complex scoring functions.