Research interests

Flavio Ballante, PhD

SUMMARY

The goal of my research is the discovery of ligands with improved pharmacology through studies based on computational techniques, organic chemistry, and biochemistry. My studies contributed to clarify the structural determinants of ligand binding for enhanced recognition at several targets involved in cancer, HIV, and parasitology, among others. Currently, I am contributing to multidisciplinary projects aimed at elucidating molecular recognition and identifying new GPCR ligands from unexplored areas of chemical space.

SCIENTIFIC ACHIEVEMENTS AND CONTRIBUTION TO SCIENCE

During my doctoral studies, one of the main problems in drug discovery was the limited amount of three-dimensional structural data of proteins. Structure-based approaches (e.g. molecular docking) couldn’t be applied for many targets, and when possible they were slow, biased to the few available complex structures, or simply inaccurate. Also, homology modeling wasn’t consistently reliable due to the low availability of proper template structures, and molecular dynamics simulation was computationally too expensive. Ligand-based approaches were commonly applied in drug discovery projects. Despite 3D quantitative structure-activity relationship (3D QSAR) was one of the most promising techniques, its development had stopped in the 1990s and the traditional 2D QSAR was still the method of choice. Since the modernization of 3D QSAR would have facilitated ligand discovery, I developed an improved 3D QSAR protocol.1 The method was implemented with a series of new features for guided model optimization, like the integration of different molecular interaction fields in one predictive model that proved its usefulness.2 Compared to the classic method, the new approach allowed generating more results in a fraction of the time, becoming an essential tool in my doctoral projects (most of these in collaboration with other research groups) which results were published in international peer-reviewed journals. Some examples are the discovery of a vascular endothelial growth factor receptor-2 (VEGFR-2) inhibitor as a new anti-angiogenic compound,3 and the discovery of new heat shock protein 90 (HSP90) inhibitors.4, 5 The new protocol has been used in several studies published by other authors.

Given the continuous increase of accessible experimental macromolecular 3D structures, the need for an enhanced structure-based 3D QSAR methodology, that could exploit structural data, arose. For this purpose, I successfully applied for a project grant from Sapienza University in Rome to support the development and application of such a tool in studies aiming at the discovery of HIV reverse transcriptase inhibitors,6, 7, 8 or at the design of isoform-selective histone deacetylase (HDAC) inhibitors.9

As a medicinal chemist, I aimed to strengthen my chemistry and biochemistry training to optimize the interplay between computation and experiments. I won National and EU mobility grants from different Institutions (LLP Erasmus, MIUR, and Sapienza University) to work at the bench. For six months I worked at the University of Lorraine in Metz (France) where I synthesized new VEGFR-2 inhibitors.10 Afterwards, I moved to the USA for three months and performed enzymatic measurements at Washington University School of Medicine (Saint Louis, MO).10 These experiences abroad helped to define my research vision and apply for additional funding. I was awarded a fellowship from Sapienza University for a project aiming at the implementation of computational methods for drug design, and a grant from the Pasteur Institute-Cenci Bolognetti Foundation for a multidisciplinary research project aimed at the discovery of selective HDAC inhibitors.

Despite the breakthrough in X-ray crystallography for relevant targets like HDACs or G protein-coupled receptors (GPCRs), many problems were still unsolved in drug discovery. For example, the determination of selective HDAC inhibitors was challenging, although many co-crystals were finally available. Specific types (called isoforms or isozymes) of HDACs have been found to play a primary role in different diseases, including cancer, neurodegenerative disorders, infections, and aging processes. No selective inhibitor for one of the 11 zinc-dependent HDAC isoforms had been discovered, only pan-HDAC inhibitors had been approved by the FDA and marketed for hematological diseases. As a postdoctoral researcher at Washington University in St. Louis, my primary objective was the discovery of new ligands that could inhibit specific HDAC isozymes and the elucidation of the molecular mechanisms underneath. The artillery was composed of computational chemistry, organic chemistry, and enzyme assays: I was in charge of designing and performing both in silico studies and biochemical assays. Regarding the computational part, I have worked first on the development of computational methods aimed at improving the docking performances,11 and on an optimized per-residue structure-based 3D QSAR method (WO patent).12 The new methodology allows computing the energy interactions between a ligand and each amino acid residue of the receptor in a complex, rationalizing isoform-specific key interaction and accounting for protein flexibility. Either structure-based or ligand-based strategies were applied to identify a hit by screening commercial chemical databases and in-house virtual libraries of diverse 3D shapes based on privileged semi-rigid scaffolds (including benzodiazepine and cyclic-tetrapeptides). I’ve designed and performed inhibition assays with a microfluidic high-throughput enzyme-assay device (more than 500 compounds and twelve positive controls have been profiled against different HDAC isozymes). Biochemical results were used to further refine the computational protocol and design optimized libraries of compounds for the next screenings. This interplay between modeling, chemistry, and biochemistry allowed the rationalization of HDAC specificity with ligands having macrocyclic headgroups,13 the discovery of novel benzodiazepine derivatives as specific human HDAC3 inhibitors,14 and highly potent and selective human HDAC6 inhibitors (low nanomolar range and more selective than the selective HDAC6 inhibitor Tubastatin A, unpublished); the discovery of a non-hydroxamic Schistosoma mansoni HDAC8 inhibitor;15 the development of a procedure to predict reliable binding poses of unresolved compounds,11 which proved to be a powerful tool in structure-based drug design, as demonstrated during the D3R Grand Challenge 216 competition. This approach has been also published as a step-by-step protocol for molecular docking validation in a book chapter.17 Furthermore, an international collaboration between our laboratory at Washington University in St. Louis and two research groups in Italy led to the discovery of a potent and selective HDAC6 inhibitor with a promising antiproliferative effect on uveal melanoma cell lines.18

The recent breakthroughs in structural biology, chemical databases, and computing technology have profoundly reshaped drug discovery, promoting structure-based approaches (e.g. molecular docking) as powerful tools for drug hunters. GPCRs drug discovery is one of the most striking examples: ~680 million compounds were virtually screened and >500 hits discovered in successful structure-based studies, which are discussed in a recent review I have authored19 during my ongoing research at Uppsala University. Despite the impact of such advances in GPCR biology, there are still many challenges to face, like the determination of target-selective ligands. One of my projects focused on determining if selective GPCR ligands could be identified by virtual screening of the “dark chemical matter” (drug-like compounds detected as non-active in ≥100 bioassays). The idea was that if ligands could be identified in this set of seemingly inert compounds, these would have excellent selectivity profiles. We demonstrated that virtual screening can serve as an efficient approach to explore the dark regions of chemical space: two high-affinity ligands of the A2A adenosine and D4 dopamine receptors were identified.20 These compounds represent excellent starting points for lead optimization because they are potent for the intended target and confirmed inactive at hundreds of other proteins, including several targets present in safety panels. Lead optimization initiated from dark compounds hence decreases the risk of false positives and could result in drug candidates with cleaner safety profiles.

The development of predictive agonist-bound GPCR structures using antagonist-bound templates is another difficult task. In a research project, our goal was to determine novel peptide agonists for the protease-activated receptor 2 protein (PAR2) which has been recently resolved with a small molecule inhibitor. By combining computational methods with mutagenesis experiments, we could build the first consistent agonist-bound state PAR2 model, allowing the identification of a novel orthosteric site, key binding site interactions, and the design of new potent PAR2 peptide agonists.21 This study (in collaboration with AstraZeneca) clarified the structural basis of PAR2 antagonism/agonism.

References (full list of publications is reported here)

1. Ballante, F.; Ragno, R. 3-D QSAutogrid/R: an alternative procedure to build 3-D QSAR models. Methodologies and applications. J Chem Inf Model. 2012, 52, 1674-1685, https://doi.org/10.1021/ci300123x

2. Friggeri, L.; Ballante, F.; Ragno, R.; Musmuca, I.; De Vita, D.; Manetti, F.; Biava, M.; Scipione, L.; Di Santo, R.; Costi, R.; Feroci, M.; Tortorella, S. Pharmacophore assessment through 3-D QSAR: evaluation of the predictive ability on new derivatives by the application on a series of antitubercular agents. J Chem Inf Model. 2013, 53, 1463-1474, https://doi.org/10.1021/ci400132q

3. Perspicace, E.; Jouan-Hureaux, V.; Ragno, R.; Ballante, F.; Sartini, S.; La Motta, C.; Da Settimo, F.; Chen, B.; Kirsch, G.; Schneider, S.; Faivre, B.; Hesse, S. Design, synthesis and biological evaluation of new classes of thieno[3,2-d]pyrimidinone and thieno[1,2,3]triazine as inhibitor of vascular endothelial growth factor receptor-2 (VEGFR-2). Eur J Med Chem. 2013, 63, 765-781, https://doi.org/10.1016/j.ejmech.2013.03.022

4. Ballante, F.; Caroli, A.; Wickersham, R. B., 3rd; Ragno, R. Hsp90 inhibitors, part 1: definition of 3-D QSAutogrid/R models as a tool for virtual screening. J Chem Inf Model. 2014, 54, 956-969, https://doi.org/10.1021/ci400759t

5. Caroli, A.; Ballante, F.; Wickersham, R. B., 3rd; Corelli, F.; Ragno, R. Hsp90 inhibitors, part 2: combining ligand-based and structure-based approaches for virtual screening application. J Chem Inf Model. 2014, 54, 970-977, https://doi.org/10.1021/ci400760a

6. Ballante, F.; Musmuca, I.; Marshall, G. R.; Ragno, R. Comprehensive model of wild-type and mutant HIV-1 reverse transciptases. J Comput Aided Mol Des. 2012, 26, 907-919, https://doi.org/10.1007/s10822-012-9586-6

7. Rotili, D.; Samuele, A.; Tarantino, D.; Ragno, R.; Musmuca, I.; Ballante, F.; Botta, G.; Morera, L.; Pierini, M.; Cirilli, R.; Nawrozkij, M. B.; Gonzalez, E.; Clotet, B.; Artico, M.; Este, J. A.; Maga, G.; Mai, A. 2-(Alkyl/aryl)amino-6-benzylpyrimidin-4(3H)-ones as inhibitors of wild-type and mutant HIV-1: enantioselectivity studies. J Med Chem. 2012, 55, 3558-3562, https://doi.org/10.1021/jm201308v

8. Rotili, D.; Tarantino, D.; Nawrozkij, M. B.; Babushkin, A. S.; Botta, G.; Marrocco, B.; Cirilli, R.; Menta, S.; Badia, R.; Crespan, E.; Ballante, F.; Ragno, R.; Este, J. A.; Maga, G.; Mai, A. Exploring the role of 2-chloro-6-fluoro substitution in 2-alkylthio-6-benzyl-5-alkylpyrimidin-4(3H)-ones: effects in HIV-1-infected cells and in HIV-1 reverse transcriptase enzymes. J Med Chem. 2014, 57, 5212-5225, https://doi.org/10.1021/jm500284x

9. Silvestri, L.; Ballante, F.; Mai, A.; Marshall, G. R.; Ragno, R. Histone deacetylase inhibitors: structure-based modeling and isoform-selectivity prediction. J Chem Inf Model. 2012, 52, 2215-2235, https://doi.org/10.1021/ci300160y

10. Ballante, F. Application of Medicinal Chemistry Methods on Different Classes of Drugs. PhD Thesis (http://hdl.handle.net/11573/918780 ), Sapienza University of Rome, Rome, 2014.

11. Ballante, F.; Marshall, G. R. An Automated Strategy for Binding-Pose Selection and Docking Assessment in Structure-Based Drug Design. J Chem Inf Model. 2016, 56, 54-72, https://doi.org/10.1021/acs.jcim.5b00603

12. Ragno, R.; Marshall, G.; Ballante, F. STRUCTURE-BASED MODELING AND TARGET-SELECTIVITY PREDICTION. WO/2015/002860, 2014.

13. Reddy, D. N.; Ballante, F.; Chuang, T.; Pirolli, A.; Marrocco, B.; Marshall, G. R. Design and Synthesis of Simplified Largazole Analogues as Isoform-Selective Human Lysine Deacetylase Inhibitors. J Med Chem. 2016, 59, 1613-1633, https://doi.org/10.1021/acs.jmedchem.5b01632

14. Reddy, D. R.; Ballante, F.; Zhou, N. J.; Marshall, G. R. Design and synthesis of benzodiazepine analogs as isoform-selective human lysine deacetylase inhibitors. Eur J Med Chem. 2017, 127, 531-553, https://doi.org/10.1016/j.ejmech.2016.12.032

15. Ballante, F.; Reddy, D. R.; Zhou, N. J.; Marshall, G. R. Structural insights of SmKDAC8 inhibitors: Targeting Schistosoma epigenetics through a combined structure-based 3D QSAR, in vitro and synthesis strategy. Bioorg Med Chem. 2017, 25, 2105-2132, https://doi.org/10.1016/j.bmc.2017.02.020

16. Gaieb, Z.; Liu, S.; Gathiaka, S.; Chiu, M.; Yang, H.; Shao, C.; Feher, V. A.; Walters, W. P.; Kuhn, B.; Rudolph, M. G.; Burley, S. K.; Gilson, M. K.; Amaro, R. E. D3R Grand Challenge 2: blind prediction of protein-ligand poses, affinity rankings, and relative binding free energies. J Comput Aided Mol Des. 2018, 32, 1-20, https://doi.org/10.1007/s10822-017-0088-4

17. Ballante, F. Protein-Ligand Docking in Drug Design: Performance Assessment and Binding-Pose Selection. Methods Mol Biol. 2018, 1824, 67-88, https://doi.org/10.1007/978-1-4939-8630-9_5

18. Nencetti, S.; Cuffaro, D.; Nuti, E.; Ciccone, L.; Rossello, A.; Fabbi, M.; Ballante, F.; Ortore, G.; Carbotti, G.; Campelli, F.; Banti, I.; Gangemi, R.; Marshall, G. R.; Orlandini, E. Identification of histone deacetylase inhibitors with (arylidene)aminoxy scaffold active in uveal melanoma cell lines. J Enzyme Inhib Med Chem. 2021, 36, 34-47, https://doi.org/10.1080/14756366.2020.1835883

19. Ballante, F.; Kooistra, J. A.; Kampen, S.; de Graaf, C.; Carlsson, J. Structure-based virtual screening for ligands of G protein-coupled receptors: What can molecular docking do for you? Pharmacol Rev. 2021,

20. Ballante, F.; Rudling, A.; Zeifman, A.; Luttens, A.; Vo, D. D.; Irwin, J. J.; Kihlberg, J.; Brea, J.; Loza, M. I.; Carlsson, J. Docking Finds GPCR Ligands in Dark Chemical Matter. J Med Chem. 2020, 63, 613-620, https://doi.org/10.1021/acs.jmedchem.9b01560

21. Kennedy, A. J.; Ballante, F.; Johansson, J. R.; Milligan, G.; Sundström, L.; Nordqvist, A.; Carlsson, J. Structural Characterization of Agonist Binding to Protease-Activated Receptor 2 through Mutagenesis and Computational Modeling. ACS Pharmacol Transl Sci. 2018, 1, 119-133, https://doi.org/10.1021/acsptsci.8b00019