Hit Recycling: Discovery of a Potent Carbonic Anhydrase Inhibitor by

Jun 29, 2015 - Sbarro Institute for Cancer Research and Molecular Medicine, Center for Biotechnology, College of Science and Technology, Temple ...
0 downloads 0 Views 2MB Size
Letters pubs.acs.org/acschemicalbiology

Hit Recycling: Discovery of a Potent Carbonic Anhydrase Inhibitor by in Silico Target Fishing Mattia Mori,†,‡ Ylenia Cau,† Giulia Vignaroli,† Ilaria Laurenzana,§ Antonella Caivano,§ Daniela Vullo,∥ Claudiu T. Supuran,∥,⊥ and Maurizio Botta*,†,# †

Dipartimento di Biotecnologie, Chimica e Farmacia, Università degli Studi di Siena, via Aldo Moro 2, I-53100 Siena, Italy Center for Life Nano Science@Sapienza, Istituto Italiano di Tecnologia, viale Regina Elena 291, I-00161 Roma, Italy § IRCCS-Centro di Riferimento Oncologico Basilicata (CROB), Laboratory of Preclinical and Translational Research, Via Padre Pio 1, Rionero in Vulture 85028 Potenza, Italy ∥ Dipartimento di Chimica, Laboratorio di Chimica Bioinorganica, Università degli Studi di Firenze, Polo Scientifico, Via della Lastruccia 3, 50019 Sesto Fiorentino (Firenze), Italy ⊥ Dipartimento NEUROFARBA, Sezione di Scienze Farmaceutiche, Università degli Studi di Firenze, Via Ugo Schiff 6, 50019 Sesto Fiorentino (Firenze), Italy # Sbarro Institute for Cancer Research and Molecular Medicine, Center for Biotechnology, College of Science and Technology, Temple University, BioLife Science Building, Suite 333, 1900 N 12th Street, Philadelphia, Pennsylvania 19122, United States ‡

S Supporting Information *

ABSTRACT: In silico target fishing is an emerging tool in drug discovery, which is mostly used for primary target or off-target prediction and drug repositioning. In this work, we developed an in silico target fishing protocol to identify the primary target of GV2− 20, a false-positive hit highlighted in a cell-based screen for 14−3−3 modulators. Although GV2−20 does not bind to 14−3−3 proteins, it showed remarkable antiproliferative effects in CML cells, thus raising interest toward the identification of its primary target. Six potential targets of GV2−20 were prioritized in silico and tested in vitro. Our results show that the molecule is a potent inhibitor of carbonic anhydrase 2 (CA2), thus confirming the predictive capability of our protocol. Most notably, GV2−20 experienced a remarkable selectivity for CA2, CA7, CA9, and CA12, and its scaffold was never explored before as a chemotype for CA inhibition, thus becoming an interesting lead candidate for further development.

O

rescue, (iii) drug repositioning, (vi) off-target prediction, and (v) drug polypharmacology detection.4−6 Moreover, advancement in computer performance and the large availability of biological data has significantly increased isTF predictive power and usage in drug discovery campaigns. A number of isTF protocols have been developed and used in retrospective examples of cross-target identification, off-target prediction, and polypharmacology detection,7−10 whereas only a few examples of prospective primary target identification studies have been reported so far.11 From a methodology standpoint, different isTF approaches have been developed, ranging from chemical similarity searching, data mining/machine learning, molecular docking, or bioactivity spectra comparison.12 Nevertheless, most of these methods exhibit several limitations.13 For example, the widely used 2D descriptor-based similarity searching suffers from poor

ne of the main challenges in drug discovery is the correct classification of true- or false-positive hit molecules from high-throughput screening (HTS). Indeed, a false-positive may provide unrealistic positive outcome by interacting with chemical probes or interfering with the assay conditions in biochemical evaluations, as the well-known PAINS compounds do.1 A false-positive may also provide positive outcomes in cellbased assays, without however modulating the expected target.2,3 In all these cases, false-positive are generally discarded even if representing a potential resource for drug discovery. Notwithstanding, the identification of their molecular target is a key prerequisite for further optimization and/or development. In this context, in silico target fishing (isTF) is an emerging tool that may significantly advance the field of drug discovery by predicting the putative target(s) of a bioactive compound. Given the relative speed of execution, the possibility to automate the protocol, and its high reliability, isTF has recently attracted much interest. Indeed, isTF proved to be a versatile and successful methodology in multiple and different approaches, including (i) primary target prediction, (ii) drug © XXXX American Chemical Society

Received: May 9, 2015 Accepted: June 29, 2015

A

DOI: 10.1021/acschembio.5b00337 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology

Figure 1. (a) Synthesis of GV2−20. Reagents and conditions: i = DIPEA, THF, r.t., 22 h. (b) Cell viability assay of K562, Jurl-MK1, and Jurl-MK2 after treatment with GV2−20 at 15 μM. K562 and Jurl-MK1 were treated for 96 h while Jurl-MK2 was for 120 h due to its slower doubling time. (c) Effect of GV2−20 on cell cycle distribution in K562, Jurl-MK1, Jurl-MK2, and CD34+ of CML patients compared with their control. Data are given as mean values ± SD of three independent experiments.

accuracy, mostly because the bioactive conformation of the query compounds is not accounted for, whereas molecular docking still suffers from well-known limitations associated with the scarce reliability of conventional scoring functions. In this respect, integrating information from different sources could be critical to increasing the likelihood of target identification and to refine the ranking order of target prediction.14 In a recent cell-based HTS aimed at identifying new 14−3−3 protein−protein interactions (PPIs) inhibitor against chronic myeloid leukemia (CML),15 the molecule GV2−20 (Figure 1a) was identified as active. Whether GV2−20 showed an antiproliferative effect in CML cells, it was unfortunately unable to bind the target of interest, namely 14−3−3σ (data not shown), thus becoming classified as a false-positive hit. However, given its remarkable antiproliferative activity, we decided to recycle GV2−20 and to investigate its mechanism of

action at the molecular level. To predict GV2−20 putative target(s), here we developed an isTF protocol by combing 3D molecular similarity search based on ligand bioactive conformation, literature analysis, and structure-based methods. A small number of putative target proteins were identified in silico, while further in vitro analysis has confirmed the reliability of the isTF protocol, showing that GV2−20 is a nanomolar affinity inhibitor of carbonic anhydrase 2 (CA2), notably endowed with a significant selectivity for some CA isoforms. On the basis of HTS results, GV2−20 was synthesized as summarized in Figure 1a for further investigations. 2Phenethylamine (2) was introduced by nucleophilic aromatic substitution on 4-chloro-3,5-dinitrobenzoic acid (1) using Hünig’s base and THF as a solvent.16 The total yield of the synthetic pathway was 80%. B

DOI: 10.1021/acschembio.5b00337 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology

molecules in complex with target proteins available in the Protein Data Bank (www.rcsb.org/pdb) was used. A total of 166 087 molecules were downloaded and subsequently filtered to remove metals, counterions, and all molecules having less than six heavy atoms (Supporting Information), thus providing a final library of 57 508 compounds. 2. Generation of ROCS (Rapid Overlay of Chemical Structures)18 queries. The ROCS program from OpenEye was used to generate a ROCS query on each single bioactive ligand included in the compound library delivered from step 1. GV2− 20 conformers were generated by the Omega2 program (OpenEye).19 3. ROCS screening. GV2−20 conformers were then screened against ROCS queries using the ROCS program, and the Tanimoto Combo score, corresponding to the algebraic sum of “ShapeTanimoto” and “ColorTanimoto” scores, was considered as the main scoring parameter. Since ROCS queries were built on the bioactive conformation of ligands in crystallographic complex with the respective macromolecular target, the higher the Tanimoto Combo score, the higher the probability of the protein to be a GV2−20’s target. For the purpose of our screening, only ROCS queries to which GV2− 20 aligned with a Tanimoto Combo score higher than 1 were kept, for a total of 203 ROCS queries. The huge amount of putative GV2−20 targets was efficiently reduced in a relatively short time based only on 3D similarity. It is worth mentioning that the Tanimoto Combo threshold value may be arbitrarily modified to survive a higher or lower number of ROCS queries. 4. Literature refinement. Since GV2−20 proved to impact on the proliferation of human CML cells, only ROCS queries corresponding to ligands cocrystallized with human proteins were retained. Accordingly, 70 ROCS queries survived this filter, which correspond to 31 different proteins (Supporting Information). Afterward, the SciFinder platform was interrogated for performing an additional refinement based on literature data, with the aim to discard all proteins not related to CML, and for which no records of expression in K562, JurlMK1, or Jurl-MK2 cells were available. After this step, 10 putative targets were identified (Supporting Information). 5. Molecular docking. The consistency of ROCS-based alignment was finally evaluated in a structure-based perspective, by comparing ROCS alignment with docking results. To this aim, the binding conformation of GV2−20 toward the 10 selected proteins was predicted by molecular docking. Multiple programs were used (Supporting Information), but only AutoDock20 was able to reproduce the crystallographic pose of bioactive ligands (self-docking) and was therefore used in this study. All proteins in which there was a very low correspondence between ROCS-based alignment and the docked pose of GV2−20 (RMSD > 3 Å) were discarded, thus providing six putative targets of GV2−20. The six final putative targets of GV2−20 identified by the isTF protocol were C-Jun N-terminal kinases 1 (JNK1, EC:2.7.11.24), cyclin-dependent kinase 2 (CDK2, EC:2.7.11.22), carbonic anhydrase 2 (CA2, EC:4.2.1.1), tyrosine-protein phosphatase 1B (PTP1B, EC:3.1.3.48), checkpoint kinase 1 (CHK1, EC:2.7.11.1), and rho-associated protein kinase 1 (ROCK1, EC:2.7.11.1). These proteins have the highest possibility to be inhibited by GV2−20, and all of them are enzymes, thus allowing a relatively easy evaluation of GV2−20 activity in vitro. It is worth mentioning that the sequence of steps performed in the isTF protocol herein developed may be modified, depending on user preference. On

GV2−20 antiproliferative activity was investigated in CML cell lines K562, Jurl-MK1, and Jurl-MK2.17 After treatment of K562 cells with 15 μM of GV2−20, preliminary MTS assay showed that the molecule decreased 50% of cells’ viability after 96 h. This result was confirmed in Jurl-MK1 and Jurl-MK2 cells where GV2−20 decreased cell viability of 45% after 96 h and of 40% after 120 h, respectively (Figure 1b). Longer treatment was necessary for the Jurl-MK2 cell line because of its slower doubling time. Unexpectedly, the GV2−20 antiproliferative effect was not driven by the induction of apoptosis (data not shown). Cell cycle analysis showed that GV2−20 induces cellcycle arrest in the G0/G1 phase in K562 cells, whereas it decreases the number of cells in the G2/M phase in Jurl-MK1 and Jurl-MK2 (Figure 1c). These results were further confirmed in mononuclear cells isolated from three CML patients. After incubation, cell cycle analysis on CD34+ cells showed a significant decrease of the G2/M phase and an increase in percentage of cells in the G0/G1 phase. Overall, cell cycle analysis clearly revealed that GV2−20 influences CML cell proliferation by interfering with normal cell cycle. To identify the putative target(s) of GV2−20, here we developed an isTF protocol (Figure 2) that relies on the

Figure 2. Workflow of the isTF protocol.

similarity between a small molecule (GV2−20 in our case) and active ligands in their bioactive conformation. Moreover, our isTF includes literature analysis and structure-based methods and is articulated into five main steps: 1. Preparation of the compounds library. The primary selection of putative target(s) was performed by comparing shape and chemical features of known bioactive molecules and GV2−20. To this aim, the bioactive conformation of small C

DOI: 10.1021/acschembio.5b00337 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology

GV2−20 toward CA9 and CA12 may be achievable for instance by developing GV2−20 analogs possessing membrane-impermeable chemical features.27 Being a carboxylic acid, it should be mentioned that at pH values >7.5, probably GV2−20 is in the anionic form, which may be not highly membrane permeable. Since preliminary crystallization trials were unsuccessful, the binding mode of GV2−20 toward CA isoforms was investigated by molecular modeling.20 Results of docking with AutoDock show that GV2−20 binds to CAs by coordinating monodentately the catalytic Zn(II) ion and establishing additional Hbonds and hydrophobic interactions with amino acids residues of the active site. In particular, in all CA isoforms studied herein, the benzoic acid moiety of GV2−20 was predicted to bind within the active site binding pocket and to coordinate the catalytic Zn(II) ion with a geometry resembling classical tetrahedral Zn coordination.28,29 The GV2−20 predicted binding mode within the CA7 active site, against which the molecule experienced the highest inhibitory activity, is reported in Figure 3. Besides Zn-coordination, a nitro group of GV2−20

the basis of the individual features of programs included in our in silico protocol, the authors are rather confident that the results of this study are independent from the order of execution of the five steps described above. This aspect, as well as the analysis of extended applicability of the isTF protocol and its automation, is the topic of currently ongoing studies, whose outcomes will be disclosed elsewhere. The predictive capability of the isTF protocol was proved in vitro. Preliminary enzymatic assays were performed to monitor GV2−20 efficacy against targets selected in silico. Results showed that GV2−20 at 15 μM strongly inhibits CA2 (more than 90% inhibition) and moderately inhibits JNK1 and other JNK isoforms (Supporting Information), thus confirming the predictive power of the isTF protocol and highlighting CAs as putative targets of GV2−20 activity in CML cells. Notably, JNK1 and CA2 do not share any common structural or functional features. CAs are a family of proteins that catalyze the reversible hydration of carbon dioxide to bicarbonate and protons. Despite the simplicity of this reaction, CAs are involved in many physiologic cellular processes but are also closely related to several human diseases, including cancer.21 To date, 16 CA isoforms have been characterized in mammals with different catalytic activities, subcellular distribution, and involvement in pathological processes.22 Accordingly, targeting selectively CA isoforms is one of the major challenges in CA inhibitor drug discovery. For these reasons, GV2−20 selectivity toward CA pharmacologically relevant isoforms was investigated. In contrast to what was observed for the pan-CAs inhibitor acetazolamide (AAZ), GV2−20 exhibited a noticeable selectivity for some CA isoforms such as CA2, CA7, CA9, and CA12 (Table 1). Table 1. Inhibitory Activity of GV2-20 and the Reference Inhibitor Acetazolamide (AAZ) against CA Isoforms isoform

GV2−20 Ki (nM)

AAZ Ki (nM)

CA1 CA2 CA3 CA4 CA5A CA6 CA7 CA9 CA12 CA13 CA14

352 67.3 >106 >105 895 576 8.7 42.3 9.6 764 570

250 12 2 × 105 74 63 11 2.5 25 5.7 17 41

Figure 3. GV2−20 docked pose on CA7. Protein is shown in green cartoon. Residues within 4 Å from GV2−20 are highlighted in stick, and GV2−20 is shown in cyan. Bonds between GV2−20 and CA7 are highlighted with black dashed lines.

established H-bond interactions with Gln91 and His63, this latter being conserved in other CAs and H-bonded by nitrocontaining sulfonamides,30 whereas the phenyl ring is projected toward the solvent and binds in a hydrophobic cleft surrounded by hydrophobic residues Trp4, Pro200, and Pro201 (Figure 3). One may note that recent examples of the alternative binding mode of carboxylate ions to CAs have been described.29,31 Even though the GV2−20 binding mode was herein predicted by molecular simulations, it is highly consistent with that of several CAs and Zn-protease inhibitors bearing a carboxylic ion.28,31−34 Moreover, GV2−20 theoretical free energy of binding to all CA isoforms finely correlates with experimental inhibitory activity data (R2 = 0.85; Supporting Information, Figure S1) thus reinforcing the consistency and reliability of predicted binding modes. Molecular docking may be a valuable tool in further optimization of potency, selectivity, and drug-likeness of GV2− 20. Finally, to the best of our knowledge, GV2−20’s scaffold was never explored before for CA’s inhibition, and GV2−20 is one of the few nanomolar inhibitors of CAs having a carboxylic acid,

It is worth mentioning that CA9 and CA12 are up-regulated in different tumor types and have been classified as cancerassociated isoenzymes.23,24 Several in vitro and in vivo studies supported the potential use of selective CA9 inhibitors in the treatment of hypoxic tumor cells that are generally resistant to traditional anticancer therapies.23 Moreover, the selective CA9 inhibitor, SLC-0111, has recently entered phase I clinical trials for the treatment of patients with advanced solid tumors, confirming the promising antitumor/antimetastatic application of CA inhibitors.25 Additionally, inhibitors of CA12 have been proposed in combination with CA9 inhibitors, affording a greater chance of treatment success.26 It is noteworthy, since CA9 and CA12 are membrane-associated CAs, while CA2 and CA7 are cytosolic isoforms, that the selectivity optimization of D

DOI: 10.1021/acschembio.5b00337 ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology



thus becoming a highly promising chemotype for further optimization. In this work, we developed an efficient isTF protocol that was used for identifying the target protein of GV2−20, a small molecule classified as false-positive in a cell-based HTS for inhibitors of 14−3−3 PPIs. The isTF protocol combines 3D chemical similarity searching, literature refinement, and molecular docking and could represent a key tool in primary target searches, drug repositioning, and off-target and toxicity predictions, which are emerging as lower-risk and cost-effective strategies in drug discovery.35,36 The key principle behind our isTF protocol is the similarity between GV2−20 (or a generic false-positive or test molecule) and the bioactive conformation of cocrystallized ligands. Therefore, starting from the structure of 166 087 available protein−ligand complexes, our isTF protocol has provided a shortlist of six putative targets of GV2−20 endowed with higher probability of success. Among them, CA2 and JNK1 proved to be inhibited by GV2−20, albeit with different potency. Indeed, GV2−20 is a weak inhibitor of JNK1 without isoform selectivity, whereas it is a potent inhibitor of CA2 also showing a certain degree of selectivity among CA isoforms. These results nicely confirm the predictive power of the isTF protocol herein developed and strongly support further tests on extended data sets. Enzymatic assays showed that GV2−20 inhibits with higher specificity CA2, CA7, CA9, and CA12. Interestingly, CA9 and CA12 are classified as tumor-associated CA isoforms and are promising targets for cancer detection, diagnostics, and therapy.37 Particularly, CA9 and CA12 are hypoxia-inducible isoforms that promote cancer cells’ survival in hypoxic and acidic microenvironments, and their combined gene silencing leads to a remarkable decrease of tumor growth,26,38 in agreement with the antiproliferative activity observed for GV2−20 in CML cells. Moreover, the GV2−20 scaffold was never explored before for CA inhibition, thus representing a promising starting point for the development of anticancer lead compounds. A medicinal chemistry project aimed at optimizing the GV2−20 scaffold is currently running; results will be provided soon.



REFERENCES

(1) Baell, J. B., and Holloway, G. A. (2010) New substructure filters for removal of pan assay interference compounds (PAINS) from screening libraries and for their exclusion in bioassays. J. Med. Chem. 53, 2719−2740. (2) Goudreau, N., Hucke, O., Faucher, A. M., Grand-Maitre, C., Lepage, O., Bonneau, P. R., Mason, S. W., and Titolo, S. (2013) Discovery and structural characterization of a new inhibitor series of HIV-1 nucleocapsid function: NMR solution structure determination of a ternary complex involving a 2:1 inhibitor/NC stoichiometry. J. Mol. Biol. 425, 1982−1998. (3) Bassilana, F., Carlson, A., DaSilva, J. A., Grosshans, B., Vidal, S., Beck, V., Wilmeringwetter, B., Llamas, L. A., Showalter, T. B., Rigollier, P., Bourret, A., Ramamurthy, A., Wu, X., Harbinski, F., Plonsky, S., Lee, L., Ruffner, H., Grandi, P., Schirle, M., Jenkins, J., Sailer, A. W., Bouwmeester, T., Porter, J. A., Myer, V., Finan, P. M., Tallarico, J. A., Kelleher, J. F., Seuwen, K., Jain, R. K., and Luchansky, S. J. (2014) Target identification for a Hedgehog pathway inhibitor reveals the receptor GPR39. Nat. Chem. Biol. 10, 343−U344. (4) Kharkar, P. S., Warrier, S., and Gaud, R. S. (2014) Reverse docking: a powerful tool for drug repositioning and drug rescue. Future Med. Chem. 6, 333−342. (5) Achenbach, J., Tiikkainen, P., Franke, L., and Proschak, E. (2011) Computational tools for polypharmacology and repurposing. Future Med. Chem. 3, 961−968. (6) Ashburn, T. T., and Thor, K. B. (2004) Drug repositioning: identifying and developing new uses for existing drugs. Nat. Rev. Drug Discovery 3, 673−683. (7) AbdulHameed, M. D., Chaudhury, S., Singh, N., Sun, H., Wallqvist, A., and Tawa, G. J. (2012) Exploring polypharmacology using a ROCS-based target fishing approach. J. Chem. Inf. Model. 52, 492−505. (8) Eric, S., Ke, S., Barata, T., Solmajer, T., Antic Stankovic, J., Juranic, Z., Savic, V., and Zloh, M. (2012) Target fishing and docking studies of the novel derivatives of aryl-aminopyridines with potential anticancer activity. Bioorg. Med. Chem. 20, 5220−5228. (9) Wale, N., and Karypis, G. (2009) Target fishing for chemical compounds using target-ligand activity data and ranking based methods. J. Chem. Inf. Model. 49, 2190−2201. (10) Schomburg, K. T., Bietz, S., Briem, H., Henzler, A. M., Urbaczek, S., and Rarey, M. (2014) Facing the challenges of structurebased target prediction by inverse virtual screening. J. Chem. Inf. Model. 54, 1676−1686. (11) Iorio, F., Bosotti, R., Scacheri, E., Belcastro, V., Mithbaokar, P., Ferriero, R., Murino, L., Tagliaferri, R., Brunetti-Pierri, N., Isacchi, A., and di Bernardo, D. (2010) Discovery of drug mode of action and drug repositioning from transcriptional responses. Proc. Natl. Acad. Sci. U. S. A. 107, 14621−14626. (12) Bender, A., Young, D. W., Jenkins, J. L., Serrano, M., Mikhailov, D., Clemons, P. A., and Davies, J. W. (2007) Chemogenomic data analysis: prediction of small-molecule targets and the advent of biological fingerprint. Comb. Chem. High Throughput Screening 10, 719−731. (13) Cereto-Massague, A., Ojeda, M. J., Valls, C., Mulero, M., Pujadas, G., and Garcia-Vallve, S. (2015) Tools for in silico target fishing. Methods 71, 98−103. (14) Wang, L., and Xie, X. Q. (2014) Computational target fishing: what should chemogenomics researchers expect for the future of in silico drug design and discovery? Future Med. Chem. 6, 247−249. (15) Mori, M., Vignaroli, G., Cau, Y., Dinic, J., Hill, R., Rossi, M., Colecchia, D., Pesic, M., Link, W., Chiariello, M., Ottmann, C., and Botta, M. (2014) Discovery of 14−3-3 protein-protein interaction inhibitors that sensitize multidrug-resistant cancer cells to doxorubicin and the Akt inhibitor GSK690693. ChemMedChem 9, 973−983. (16) Sykes, B. M., Atwell, G. J., Hogg, A., Wilson, W. R., O’Connor, C. J., and Denny, W. A. (1999) N-substituted 2-(2,6dinitrophenylamino)propanamides: Novel prodrugs that release a primary amine via nitroreduction and intramolecular cyclization. J. Med. Chem. 42, 346−355.

ASSOCIATED CONTENT

S Supporting Information *

Chemistry, antiproliferative assays and cell cycle analysis, details of the in silico target fishing protocol, and enzymatic assays. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acschembio.5b00337.



Letters

AUTHOR INFORMATION

Corresponding Author

*Tel.: +39 0577 234306. Fax: +39 0577 234333. E-mail: botta. [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors wish to thank the OpenEye Free Academic Licensing Program for providing a free academic license for molecular modeling and chemoinformatics software. I.L. and A.C. wish to thank F. D’Alessio for the kind gift of Jurl-MK1 and Jurl MK2 cells and F. La Rocca and L. Del Vecchio for the support with cellular studies. E

DOI: 10.1021/acschembio.5b00337 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology (17) Di Noto, R., Luciano, L., Lo Pardo, C., Ferrara, F., Frigeri, F., Mercuro, O., Lombardi, M. L., Pane, F., Vacca, C., Manzo, C., Salvatore, F., Rotoli, B., and Del Vecchio, L. (1997) JURL-MK1 (ckit(high)/CD30-/CD40-) and JURL-MK2 (c-kit(low)/CD30+/ CD40+) cell lines: ’two-sided’ model for investigating leukemic megakaryocytopoiesis. Leukemia 11, 1554−1564. (18) ROCS 3.2.0.4; OpenEye Scientific Software: Santa Fe, NM. http://www.eyesopen.com. (19) OMEGA 2.5.1.4; OpenEye Scientific Software: Santa Fe, NM. http://www.eyesopen.com. (20) Morris, G. M., Huey, R., Lindstrom, W., Sanner, M. F., Belew, R. K., Goodsell, D. S., and Olson, A. J. (2009) AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. J. Comput. Chem. 30, 2785−2791. (21) Supuran, C. T., and Scozzafava, A. (2007) Carbonic anhydrases as targets for medicinal chemistry. Bioorg. Med. Chem. 15, 4336−4350. (22) McKenna, R., and Supuran, C. T. (2014) Carbonic anhydrase inhibitors drug design. Sub-cellular biochemistry 75, 291−323. (23) Thiry, A., Dogne, J. M., Masereel, B., and Supuran, C. T. (2006) Targeting tumor-associated carbonic anhydrase IX in cancer therapy. Trends Pharmacol. Sci. 27, 566−573. (24) Hsieh, M. J., Chen, K. S., Chiou, H. L., and Hsieh, Y. S. (2010) Carbonic anhydrase XII promotes invasion and migration ability of MDA-MB-231 breast cancer cells through the p38 MAPK signaling pathway. Eur. J. Cell Biol. 89, 598−606. (25) See more at ClinicalTrials.gov: “Safety Study of SLC-0111 in Subjects With Advanced Solid Tumours.” ClinicalTrials.gov Identifier: NCT02215850. (26) Ulmasov, B., Waheed, A., Shah, G. N., Grubb, J. H., Sly, W. S., Tu, C., and Silverman, D. N. (2000) Purification and kinetic analysis of recombinant CA XII, a membrane carbonic anhydrase overexpressed in certain cancers. Proc. Natl. Acad. Sci. U. S. A. 97, 14212−14217. (27) Supuran, C. T., Scozzafava, A., Ilies, M. A., and Briganti, F. (2000) Carbonic anhydrase inhibitors: synthesis of sulfonamides incorporating 2,4,6-trisubstituted-pyridinium-ethylcarboxamido moieties possessing membrane-impermeability and in vivo selectivity for the membrane-bound (CA IV) versus the cytosolic (CA I and CA II) isozymes. J. Enzyme Inhib. Med. Chem. 15, 381−401. (28) Mori, M., Massaro, A., Calderone, V., Fragai, M., Luchinat, C., and Mordini, A. (2013) Discovery of a New Class of Potent MMP Inhibitors by Structure-Based Optimization of the Arylsulfonamide Scaffold. ACS Med. Chem. Lett. 4, 565−569. (29) D’Ambrosio, K., Carradori, S., Monti, S. M., Buonanno, M., Secci, D., Vullo, D., Supuran, C. T., and De Simone, G. (2015) Out of the active site binding pocket for carbonic anhydrase inhibitors. Chem. Commun. (Cambridge, U. K.) 51, 302−305. (30) D’Ambrosio, K., Vitale, R. M., Dogne, J. M., Masereel, B., Innocenti, A., Scozzafava, A., De Simone, G., and Supuran, C. T. (2008) Carbonic anhydrase inhibitors: bioreductive nitro-containing sulfonamides with selectivity for targeting the tumor associated isoforms IX and XII. J. Med. Chem. 51, 3230−3237. (31) De Simone, G., and Supuran, C. T. (2012) (In)organic anions as carbonic anhydrase inhibitors. J. Inorg. Biochem. 111, 117−129. (32) Di Fiore, A., Monti, S. M., Hilvo, M., Parkkila, S., Romano, V., Scaloni, A., Pedone, C., Scozzafava, A., Supuran, C. T., and De Simone, G. (2009) Crystal structure of human carbonic anhydrase XIII and its complex with the inhibitor acetazolamide. Proteins: Struct., Funct., Genet. 74, 164−175. (33) Boone, C. D., Tu, C., and McKenna, R. (2014) Structural elucidation of the hormonal inhibition mechanism of the bile acid cholate on human carbonic anhydrase II. Acta Crystallogr., Sect. D: Biol. Crystallogr. 70, 1758−1763. (34) Hakansson, K., Briand, C., Zaitsev, V., Xue, Y., and Liljas, A. (1994) Wild-type and E106Q mutant carbonic anhydrase complexed with acetate. Acta Crystallogr., Sect. D: Biol. Crystallogr. 50, 101−104. (35) Chong, C. R., and Sullivan, D. J. (2007) New uses for old drugs. Nature 448, 645−646. (36) Oprea, T. I., and Mestres, J. (2012) Drug Repurposing: Far Beyond New Targets for Old Drugs. AAPS J. 14, 759−763.

(37) Pastorekova, S., Zatovicova, M., and Pastorek, J. (2008) Cancerassociated carbonic anhydrases and their inhibition. Curr. Pharm. Des. 14, 685−698. (38) Chiche, J., Ilc, K., Laferriere, J., Trottier, E., Dayan, F., Mazure, N. M., Brahimi-Horn, M. C., and Pouyssegur, J. (2009) Hypoxiainducible carbonic anhydrase IX and XII promote tumor cell growth by counteracting acidosis through the regulation of the intracellular pH. Cancer Res. 69, 358−368.

F

DOI: 10.1021/acschembio.5b00337 ACS Chem. Biol. XXXX, XXX, XXX−XXX