Thymidine phosphorylase and prostrate cancer cell proliferation inhibitory activities of synthetic 4-hydroxybenzohydrazides: In vitro, kinetic, and in silico studies

Over-expression of thymidine phosphorylase (TP) plays a key role in many pathological complications, including angiogenesis which leads to cancer cells proliferation. Thus in search of new anticancer agents, a series of 4-hydroxybenzohydrazides (1–29) was synthesized, and evaluated for in vitro thymidine phosphorylase inhibitory activity. Twenty compounds 1–3, 6–14, 16, 19, 22–24, and 27–29 showed potent to weak TP inhibitory activities with IC50 values in the range of 6.8 to 229.5 μM, in comparison to the standards i.e. tipiracil (IC50 = 0.014 ± 0.002 μM) and 7-deazaxanthine (IC50 = 41.0 ± 1.63 μM). Kinetic studies on selected inhibitors 3, 9, 14, 22, 27, and 29 revealed uncompetitive and non-competitive modes of inhibition. Molecular docking studies of these inhibitors indicated that they were able to interact with the amino acid residues present in allosteric site of TP, including Asp391, Arg388, and Leu389. Antiproliferative (cytotoxic) activities of active compounds were also evaluated against mouse fibroblast (3T3) and prostate cancer (PC3) cell lines. Compounds 1, 2, 19, and 22–24 exhibited anti-proliferative activities against PC3 cells with IC50 values between 6.5 to 10.5 μM, while they were largely non-cytotoxic to 3T3 (mouse fibroblast) cells proliferation. Present study thus identifies a new class of dual inhibitors of TP and cancer cell proliferation, which deserves to be further investigated for anti-cancer drug development.

Introduction Thymidine phosphorylase (TP) (EC 2.4.2.4) is an enzyme of pyrimidine salvage pathway, primarily responsible for maintaining nucleotide homeostasis. TP Over-expression has been reported in several pathological conditions, including rheumatoid arthritis, chronic inflammatory diseases, psoriasis, and tumor angiogenesis. Neoplastic tissues of bladder, gastric, cervical, lung, colon, prostate, esophageal, and breast cancers show hyperactivity of TP. Over-expression of TP in tumor cell lines is an indication of its role in angiogenesis [1]. Role of TP as angiogenic enzyme was proposed to be identical to human platelet derived endothelial cell growth factor (PD-ECGF), which promotes angiogenesis by facilitating endothelial cell proliferation, and migration [2]. TP Enzyme is present in many cells and tissues, such as platelets, stromal cells, macrophages, endothelial cells, reticulocytes, glial cells, and ovary [1].
TP Comprises of two identical subunits, with molecular weight between 90 to 110 kDa in Escherichia coli and mammals, respectively. Each subunit consists of a large α/β domain and a smaller α-helical domain, separated from each other by a cavity [3]. Active site of TP comprises of thymidine and phosphate binding sites. In addition to this, it also has a hydrophobic pocket near the substrate binding sites. In the presence of inorganic phosphate, it catalyzes the reversible cleavage of glycosidic bond of pyrimidine 2 0 -deoxynucleotides to 2 0 -deoxyribose-1-phosphate. Deoxyribose-1-phosphate then undergoes dephosphorylation to produce 2-Ddeoxyribose, which is used either as energy source for the cell or secreted out of the cell where it may act as an angiogenic growth factor [3,4].
Significant efforts have been focused on the development of TP inhibitors with possible therapeutic potential since 1960's. Some compounds were identified with excellent TP inhibitory activity, and tested pre-clinically and clinically. However, only one drug (Lonsurf 1 ) is in clinical use for the management of metastatic colorectal cancer. This drug was approved by US-FDA in 2015 and is combination of tipiracil (TP inhibitor) and trifluridine (a cytotoxin). This drugis associated with adverse side effects which demands the identification of new leads capable of inhibiting TP enzyme and thereby TP associated pathologies [5]. 7-Deazaxanthine (7-DX) and 6-amino-5-bromouracil (6A5BU) are widely used as reference compounds for in vitro studies against the TP enzyme [1,6].
Twenty-nine derivatives of 4-hydroxybenzohydarzide were subjected to an in vitro spectrophotometric TP inhibition assay. Some of the most active compounds were then subjected to kinetic and molecular docking studies in order to determine their mechanism of inhibition of TP enzyme. TP is particularly reported to be over-expressed in the prostate cancer, therefore, active compounds against TP were also evaluated for their effect on the proliferation of prostate cancer cells (PC3) using the (3-[4,5-dimethylthiazole-2-yl]-2,5-diphenyl-tetrazolium bromide) MTT colorimetric assay [2,30,31]. Interestingly, some of these compounds were also able to inhibit the PC3 cancer cells proliferation. Present study therefore identifies dual inhibitors of TP, and cancer cell proliferation.

Cell lines
Mouse fibroblast cell line (NIH/3T3 (ATCC 1 CRL-1658™)), and prostate cancer line (PC3 (ATCC1 CRL-1435™)) were purchased from American Type Culture Collection (ATCC), USA. As we have purchased these lines directly from ATCC, therefore we did not perform any quality control testing procedures.

Thymidine phosphorylase inhibition assay
Since human TP is not commercially available, we used commercially available recombinant TP (expressed in E. coli) enzyme. Primary sequence of TP is mostly conserved throughout the evolution and mammalian TP share 39% sequence similarity with TP of E. coli. The mammalian enzyme also shares 65-70% similarity with the active site residues of E. coli TP enzyme [2].
Assay for TP inhibition was performed spectrophotometrically, following the method of Bera et al. with some modifications [32]. Enzyme TP catalyzed the reversible phosphorolysis of thymidine to thymine and 2-deoxyribose-1-phosphate. Reaction mixture contained 150 μL of potassium phosphate buffer (pH 7.0, 50 mM), 20 μL of enzyme (0.058 U/well), and 10 μL of test compound (0.5 mM in dimethylsulfoxide). The reaction mixture was incubated for 10 min at 30˚C. 20 μL of substrate (1.5 mM) was then added and change in absorbance was recorded after every minute for a total time period of 10 minutes at 290 nm in ELISA plate reader (Spectra Max-340, Molecular Devices, CA, USA). Tipiracil and 7-deazaxanthine were used as standard inhibitors in this assay. We have used the 96-well UV-plates, purchased from Corning (Catalogue No. COSTAR 3635, Lot No. 00912050). In addition to this, we also carried out pre-read (blank reference) on a solution that does not contain the substrate, and this gave us zero reading. We then added the substrate (thymidine), and recorded the final readings. Therefore, the chances of getting erroneous results were minimized.
It is a colorimetric assay that measures the reduction of MTT by mitochondrial enzyme i.e. succinate dehydrogenase. The MTT enters into the mitochondria of cell, where it is reduced to an insoluble formazan salt. As reduction of MTT can only occur in metabolically active cells, the level of activity is actually a measure of the viability of the cells [34].
Mouse fibroblast normal cell line 3T3, and prostate cancer line (PC3) were cultured separately in DMEM, supplemented with 5% of FBS, 100 IU/mL of penicillin and 100 μg/mL of streptomycin, and kept at 37˚C in 5% CO 2 incubator. For the preparation of cell culture, 100 μL/well of cell solution (5 x 10 4 cells/mL) was added into 96-well plate. The plate was incubated for overnight, and fresh medium was added after the removal of old medium. The test compounds were also added in different concentrations into the plate, and again incubated for 48 h. After the completion of this incubation period, 200 μL MTT (0.5 mg/mL) was added, and plate was again incubated for 4 h, after this final incubation, 100 μL of DMSO was added to each well. The level of MTT reduction to formazan was evaluated by change in absorbance at 540 nm for 3T3, and 570 nm for PC3 using a micro-plate reader (Spectra Max plus, Molecular Devices, CA, USA). The anti-proliferative activity was recorded as concentration of the inhibitor causing 50% growth inhibition (IC 50 ) of cells.

Protocol for kinetic studies
Kinetic studies were carried out to identify the mechanism of inhibition by these compounds. Inhibitor could bind with the enzyme in multiple ways, such as in competitive, non-competitive, mixed or uncompetitive way. In kinetic assay, the enzyme (0.058 U/200 μL) was incubated with different concentrations of inhibitor for 10 min at 30˚C. The reaction was then initiated by adding different concentrations (0.1875-1.5 mM) of substrate (thymidine), and the resulting degradation of thymidine was measured continuously at 290 nm for 10 min on ELISA plate reader. Every experiment was run in triplicate.
Lineweaver-Burk plot was plotted to determine the type of inhibition. This was accomplished by plotting the reciprocal of the rate of reaction against the reciprocal of the substrate concentration. Ki values were determined by secondary re-plot of Lineweaver-Burk plot, and reconfirmed by Dixon plot.

Molecular docking studies
Molecular docking studies were carried out in order to understand the interaction of inhibitors (ligands) with TP (receptors). The structure of TP was taken from PDB (4LHM) which is the crystallographic structure of TP in E. coli. Protein was prepared using Protein Preparation Wizard and OPLS3 force field [35,36] in order to add missing hydrogens. The sulphate ions were replaced by phosphate so they did not affect the conformation of protein and occupy the same position as that of sulphate. Ligands were prepared to get the correct ionization and protonation states using Ligprep module [37] in Maestro Schrӧdinger2018-1.
Since all the inhibitors showed non-and uncompetitive mode of inhibition in kinetic studies, site map analysis [38,39] was performed to find out the best allosteric site available in TP. Five allosteric sites were observed and the one with highest score i.e. 1 was finally selected for the docking studies of these ligands. The dimensions of grid box were 15 Ǻ x 15 Ǻ x 15 Ǻ from the mass of centre of each docked ligand and extra precision mode of Glide [40][41][42][43] was used for the docking of ligands.

Statistical analysis
Results obtained for the enzyme inhibitory activity and MTT assay were analysed using Soft-Max Pro 4.8 software (Molecular Devices, CA, USA) and Microsoft Excel. Percent of enzyme inhibition, and growth inhibition of cells was calculated using the following formula: Results were presented as means ± standard error mean from triplicate (n = 3) observation. IC 50 values were determined by using EZ-FIT, Enzyme kinetics software by Perrella Scientific, Inc., USA. Grafit 7.0 version was used to determine the kinetics parameters. The software was purchased from the Erithacus Software Ltd. (Wilmington House, West Sussex RH19 3AU, UK).

General procedure for the synthesis of compounds 1-29
In a typical procedure, 4-hydroxylbenzohydrazones  were synthesized by mixing 4-hydroxylbenzohydrazide (1.5 mmol), substituted benzaldehydes (1.5 mmol) in ethanol (20 mL) with a catalytic amount of acetic acid (1 mL). The mixture was refluxed for 3 h, while progress of the reaction was monitored through thin layer chromatography. After completion of reaction, the reaction mixture was poured into China dish to let the solvent evaporate slowly at room temperature to afford crystals of the products. Structures of the compounds were deduced by using NMR and mass spectroscopic techniques.

Chemistry
For the study of thymidine phosphorylase inhibitory activity, twenty-nine derivatives of 4-hydroxybenzohydrazides were synthesized by reacting 4-hydroxybenzohydrazide with substituted benzaldehydes in ethanol, catalyzed by acetic acid (Scheme 1).

Scheme 1. Synthesis reaction: Synthesis of 4-hydroxybenzohydrazides 1-29
The structures of these derivatives were identified by EI-MS, and 1 H-NMR spectroscopy, and comparison with the data reported in literature. To confirm the stereochemical assignment of iminic bond, NOESY (nuclear overhauser enhancement spectroscopy) spectrum was recorded for a representative compound 29. Strong NOESY interactions of NH and 2 0 -OH with CH 2 of ethyl group revealed the geometry of the compound to be of (E)-configuration. Absence of NOESY interaction of NH with 2 0 -OH further confirmed the E stereochemistry (Fig 2).
Empirical structure-activity relationship (SAR) studies proposed that hydroxy and methoxy substitutions on phenyl ring play an important role in inducing TP inhibition. These groups may be involved in hydrogen bonding with the amino acid residues, present at the substratebinding site or hydrophobic pocket of TP enzyme.
Four halogen-substituted derivatives were evaluated (compounds 9, 11, 14, 15), and three were found to inhibit the TP activity with IC 50 values between 158.0-160.3 μM. Among these mono-halogenated derivatives, 9, 11, and 14 showed ability to inhibit the enzyme activity, while di-halogenated derivative 15 was found to be inactive with 20% inhibition. Based on the IC 50 values, halogens substitution were found to be more favorable in comparison to OH and OCH 3 substitutions. It was thus proposed that halogens might increase the ability of these compounds to interact via hydrogen bonding with amino acid residues present at substrate binding-site of TP.
Compound 16 with a methyl group at ortho position showed a moderate TP inhibition (IC 50 = 177.2 ± 0.5 μM), while compounds 17, and 18 with dimethylmine and methylsulfanyl groups, attached to para position, respectively showed less than 50% inhibition thus regarded inactive.
Three derivatives with hydroxyl-cum-methoxy substitutions 19, 23, 24 were evaluated, and all were found to be moderately active against TP enzyme (IC 50 values between 172.0-190.3 μM). Compound 19 with OH and OCH 3 groups at ortho (C-2 0 ) and meta (C-3 0 ) positions, respectively, showed a moderate TP inhibition (IC 50 = 181.5 ± 1.6 μM). Switching of hydroxy group to the other meta (i.e. C-5 0 ) position, as in compound 23, slightly increased the inhibition of TP (IC 50 = 172.0 ± 2.7 μM). Replacement of methoxy with an ethoxy group, as in compound 24, slightly decreased the TP inhibitory activity (IC 50 = 190.3 ± 0.4 μM). SAR proposed that when hydroxyl and methoxy groups are present adjacent to each other, they lower the ability of compounds to inhibit enzyme, as inferred from their IC 50 values.
Three derivatives with hydroxy-cum-halogen substitutions were also evaluated (compounds 20-22). Among them compound 22 with two bromo groups at meta (C-3 0 and C-5 0 ) positions, and an OH at ortho (C-2 0 ) showed a potent TP inhibitory activity (IC 50 = 6.8 ± 0.7 μM), in comparison to the standard used i.e., 7-deazaxanthine (IC 50 = 41.0 ± 1.63 μM). This is an unusual but reproducible behavior exhibited by this compound, whereas compound 20, which just lacks a bromo group at meta (C-3 0 ) position showed 38.6% inhibition. As TP has hydrophobic pocket near the substrate binding sites, it is possible that compound 22 with di-bromo  Thymidine phosphorylase inhibitors of synthetic origin substitution may be able to fit more appropriately at the hydrophobic pocket of the TP, which may not be possible for compound 20. Compound 20 might have a conformation which does not fit well in the hydrophobic pocket of enzyme.
Derivatizations were also made by replacing the benzylidene group with ethylidine and propylidine groups (Fig 1), in addition to OH substitutions on phenyl ring. Five derivatives 25-29 were evaluated, and among them compounds 27, 28, and 29 with two OH groups on phenyl ring, showed a moderate inhibition of TP (IC 50 = 176.9 ± 1.6, 183.4 ± 0.9, and 173.0 ± 1.4μM, respectively). This is consistent with the results we obtained for dihydroxylated derivatives with benzylidine group (Fig 4).

Kinetic studies
Kinetic study on compounds 3, 9, 14, 27, and 29 revealed that they inhibit the TP in an uncompetitive manner (Table 1), as deduced from the Lineweaver-Burk plot. Uncompetitive inhibitors interact with enzyme only when enzyme-substrate (ES) complex is formed. ES complex formation was proposed to induce conformational changes in the enzyme which facilitates the binding of the inhibitor. Uncompetitive inhibitors cause decrease in both K m and V max values of the enzyme (Fig 5). Compound 22 inhibited the enzyme in a non-competitive manner (Fig 6). This compound, therefore, interacted either with the amino acids of hydrophobic pocket of the  enzyme or at allosteric site of the enzyme. Noncompetitive inhibitors do not affect the K m value but changes the V max value. These compounds, therefore, do not competitively interact with the thymidine or phosphate-binding sites of TP when thymidine is used as the variable substrate. Values of dissociation constants (Ki) were determined by secondary re-plot of Lineweaver-Burk plot, and Dixon plot, these were in the range of 1.75-176 μM (Table 1).

Molecular docking studies
Molecular docking studies on selected inhibitors were performed in order to understand the ligand-receptor interactions at atomic level. Compounds 3, 9, 14, 22, 27, and 29 were found to be either uncompetitive or non-competitive inhibitors of TP. They showed binding to an allosteric site, located adjacent to the substrate binding site of thymidine phosphorylase. According to the crystal structure of TP (4LHM), it has a small α-helical domain containing residues   [44,45]. The molecular docking analysis of selected inhibitors showed that most of them were able to interact with the amino acid residues of large α/β domain (197 to 440) which play an important role in the catalytic activity of TP (Fig 7).
Compound 5 with dihydroxyl groups at meta and para positions were found to interact with Leu389 and Gln244 via H-bonds (Fig 8). Compounds 9, 14, and 22 showed slightly different docked poses, in comparison to compound 5. For instance, the OH group of phenyl ring in compound 9 was able to form H-bonds with Asp391, and Arg388 (Fig 9). Compound 14 was able to form H-bond interactions with Asp391 (Fig 10). Similarly, compound 22 was also able to interact with Asp391, Arg388, and Leu389 via H-bonds (Fig 11).
Compounds 27, and 29 possess CH 3 and C 2 H 5 , respectively, as R 2 substituents in addition to the hydroxyl group substituted phenyl ring. These alkyl groups further changed the docked poses of compounds 27, and 29. For instance, in compound 27 the carboxyl group of hydrazide was found to be involved in interacting with Arg271 and the OH group of phenyl ring interacted with water molecule via H-bond (Fig 12). The ortho substituted hydroxyl group in compound 29 interacted with carboxyl group of Leu389 via H-bond (Fig 13), while the meta substituted OH group interacted with the side chain of Arg271 via H-bond.
Compounds which showed anti-proliferation activity only towards 3T3 cell lines (such as 3, 7, 27-29) were cytotoxic to normal cells and therefore need to be investigated again for abovementioned activities after structural modification.

Conclusion
Role of TP in inducing angiogenesis and tumor growth makes it an important target for the discovery of anti-angiogenic (anti-cancer) agents. In this regards, a series of 4-hydroxybenzohydrazides  was evaluated for its in-vitro inhibitory activity against angiogenic enzyme TP. Twenty derivatives were found to inhibit the TP enzymatic activity. Among them, compound 22 showed a several fold more potent TP inhibitory activity (IC 50 = 6.8 ± 0.7 μM), when compared to the standard, 7-deazaxanthine (IC 50 = 41.0 ± 1.63 μM). In mechanistic studies, it showed a non-competitive mode of inhibition (Ki = 1.75 ± 0.009 μM). Compound 22 apparently interact with Arg271 and Pro270 of enzyme via H-bonds. Furthermore, it also showed a good anti-proliferative (cytotoxic) activity against prostate cancer (PC3) cell line. Present study thus identifies a new class of inhibitors against TP enzyme, and cancer cells proliferation. This class can be investigated further for anti-cancer studies at in-vivo level.