Discovery of new quinolines as potent colchicine binding site inhibitors: design, synthesis, docking studies, and anti-proliferative evaluation

Abstract Discovering of new anticancer agents with potential activity against tubulin polymerisation is still a promising approach. Colchicine binding site inhibitors are the most relevant anti-tubulin polymerisation agents. Thus, new quinoline derivatives have been designed and synthesised to possess the same essential pharmacophoric features of colchicine binding site inhibitors. The synthesised compounds were tested in vitro against a panel of three human cancer cell lines (HepG-2, HCT-116, and MCF-7) using colchicine as a positive control. Comparing to colchicine (IC50 = 7.40, 9.32, and 10.41 µM against HepG-2, HCT-116, and MCF-7, respectively), compounds 20, 21, 22, 23, 24, 25, 26, and 28 exhibited superior cytotoxic activities with IC50 values ranging from 1.78 to 9.19 µM. In order to sightsee the proposed mechanism of anti-proliferative activity, the most active members were further evaluated in vitro for their inhibitory activities against tubulin polymerisation. Compounds 21 and 32 exhibited the highest tubulin polymerisation inhibitory effect with IC50 values of 9.11 and 10.5 nM, respectively. Such members showed activities higher than that of colchicine (IC50 = 10.6 nM) and CA-4 (IC50 = 13.2 nM). The impact of the most promising compound 25 on cell cycle distribution was assessed. The results revealed that compound 25 can arrest the cell cycle at G2/M phase. Annexin V and PI double staining assay was carried out to explore the apoptotic effect of the synthesised compounds. Compound 25 induced apoptotic effect on HepG-2 thirteen times more than the control cells. To examine the binding pattern of the target compounds against the tubulin heterodimers active site, molecular docking studies were carried out.


Introduction
Cancer is the uncontrolled growth of abnormal cells 1 . There are about 100 types of cancer which need diagnosis and treatment 2 . In 2018, according to National Cancer Institute (NIH), approximately 1,735,350 new cases of cancer have been diagnosed in the United States and 609,640 people died from the disease. Cancer of breast, liver, colon, and rectum are common disease with high rate of incidence 3 . The current approaches of cancer treatment include surgery, radiation, chemotherapy; and hormonal treatment 4 .
Microtubules are important cytoskeletal structures that have a crucial role in cell division 5 , which make them attractive targets for design of new anticancer agents 6,7 . Microtubules are composed of two subunits of aand b-tubulin heterodimers. These subunits are arranged in slender-shaped filamentous tubes with many micrometres long 8 . Some natural products that target the tubulin and the microtubule system-also referred to as anti-mitotics-still important counterparts in combination chemotherapy for the treatment of several malignancies 9,10 .
The tubulin heterodimer comprises at least three binding sites: the paclitaxel, vinblastine, and colchicine binding sites ( Figure 1) 11 . There are many drugs used in clinical oncology acting on the paclitaxel and vinblastine binding sites 12,13 . These drugs are highly potent but there are some limitation in clinical use for many reasons: development of multi-drug resistance, high lipophilicity, low water solubility, intravenous administration due to poor water solubility 14 . Aforementioned drawbacks can largely overcome by use of tubulin inhibitors that bind to the colchicine binding site. Such inhibitors have therapeutic advantages: high water solubility so that it can be administered orally and there is no multi-drug resistance 15 .
Colchicine binding site inhibitors (CBSI) produce their biological activities by suppressing the vital process of tubulin assembly, and consequently suppressing microtubule formation 14 . Colchicine I, the most famous inhibitors in this category, binds to tubulin very tightly, but there is no any compound in this group has a significant use in treatment of cancer 16 . combretastatin A-1 (CA-1) II and combretastatin A-4 (CA-4) III are combretastatin analogs having microtubule inhibitory activity with limited value due to their low water solubility 17 . To improve their water solubility, such analogs were designed as prodrugs of monosodium phosphate salt. These prodrugs can be metabolised in vivo into CA-1 and CA-4 as active components 18,19 . CA-4P showed no bone marrow toxicity, stomatitis, and hair loss in phase II clinical trial 20 . One of CA-4 analogs is ombrabulin IV which has better water solubility, oral activity, enhanced anti-cancer activity and decreased adverse effects 14 .
Plinabulin V is an active molecule against non-small cell lung cancer 21 . It restricts tubulin polymerisation with immune-enhancing effects 22,23 . Indibulin VI has an effective antitumor activity with a lower side effects 24 . The antitumor activity of indibulin is believed to be related to its effects on microtubules 25 .
E7010 VII is an orally bioavailable tubulin-binding agent that was developed for cancer treatment. It is a sulphonamide derivatives having an antimitotic effect. It binds to the colchicine site on b-tubulin subunit, leading to cell cycle arrest at the G2/M phase, resulting in cellular apoptosis 26 . It exhibited a broad spectrum of antitumor activity in vitro and in vivo 26 . Its clinical trial indicated that it has dose-limiting toxicities included abdominal pain, constipation, and fatigue 14 .
Podophyllotoxin VIII, a naturally occurring anticancer agent, binds to the colchicine site of tubulin leading to inhibition of the tubulin assembly into microtubules 27 . It has severe toxicity which limits its clinical application as a cancer therapy. However, podophyllotoxin is still considered an attractive lead compound for the generation of new antitumor agents 28 .
Moreover, several derivatives (e.g. compounds IX 29 , X 30 , XI 31 , and XII 32 ) has been synthesised and evaluated as tubulin inhibitors targeting the colchicine binding site. These compounds were modified and tested to find highly potent agents for treatment of cancer.
In continuation to our previous efforts of design and synthesis of new anticancer agents [33][34][35][36][37][38][39][40][41][42] , a new series of quinoline derivatives were designed and synthesised. The synthesised derivatives have the same pharmacophoric features of CBSI to examine their effect as anticancer agents with potential tubulin inhibitory activities targeting the colchicine binding site.
Colchicine as a prototype of CBSIs is formed of three rings (A, B (linker), and C). some reports revealed that A-and C-ring encompasses the least pharmacophoric features required for binding to tubulin 45 . In addition, it was found that changes in the ring B (linker) could affect the antiproliferative activity the CBSIs 31 . Figure 4 shows the pharmacophoric points on colchicine and podophyllotoxin as representative examples of CBSIs.
The main target of this work was the synthesis of new quinoline derivatives having the same essential pharmacophoric features of the reported CBSIs ( Figure 5). The core of our molecular design rational comprised bioisosteric modification strategies of CBSIs at three different positions ( Figure 6).

Chemistry
The synthetic pathways employed to prepare the target compounds are outlined in Schemes 1-3. Firstly, p-aminoacetophenone 1 was refluxed in acetic anhydride/sodium acetate mixture to produce N-(4-methoxyphenyl) acetamide 2. Boiling compound 2 with POCl 3 in the presence of catalytic amount of DMF, produced aldehyde derivative 3. Chalcone derivatives 5, 6, and 7 were prepared through the reaction of compound 3 with the corresponding acetophenones 4 in the presence of alcoholic NaOH. The reaction occurred by ratio (1:1) to form chalcone which crystallised from ethanol, the structure of compounds 5, 6, and 7 were established on the basis of its elemental and spectral data. The IR spectra of compounds 5, 6, and 7 were characterised by strong absorption bands around 1655 cm À1 due to carbonyl ketone stretching, which appeared at low absorption value because of extended conjugation with the double bond. The absolute geometry of the a,b-unsaturated carbonyl linker was assigned to be in trans form based on the coupling constant alkene protons. In more details, the 1 H NMR spectrum of chalcone 5 showed two doublets, each equivalent to one proton, at d 8.1 ppm due to CH alkene (b-proton) and at d 7.9 ppm due to CH alkene (a-proton). Both protons have the same coupling constant value of 15.2 Hz, which confirms the E-configuration. Additionally, the 1 H NMR spectrum of compound 5 displayed two more doublets, each of two protons, at d 8.2 and 7.1 ppm, with coupling constant value of 8.4 Hz which is attributed to the p-disubstituted phenyl moiety (Scheme 1, Supplementary data).
The reactivity of chalcone can attributed to the fact that their molecules have two electron poor centres at C-1 and C-3 in addition to one electron rich centre at C-2, so the double bond in chalcone can be looked on as an electron rich bond that may enter in 1,3-dipolar cycloaddition reactions. Chalcone are heavily utilised synthons in 1,3-cycloadditon reactions to build a wide variety of heterocyclic systems such as pyrrolidines, oxazolines and pyrimidines. In this work, nine different 1,3-dipolar cyclo-addition reactions have been achieved. all reactions proceeded smoothly and final products were obtained in relatively good yields as detailed in the experimental part. The subsequent paragraphs point out those reactions in depth (Scheme 2).
First, to build a dihydroprazole ring system, a mixture of the chalcone (5, 6, and 7) with excess hydrazine hydrate in absolute ethanol was heated up at reflux temperature to finally give the desired corresponding compounds 8, 9, and 10. The structure of compound 8 was established on the basis of its elemental and spectral data. IR spectrum of compound 8 exhibited an important band at 3295 cm À1 due to NH stretching of the newly formed dihydropyrazole ring. The later NH was also appeared on the 1 H NMR spectrum as a broad deuterium-exchangeable singlet within the aromatic region at d 7.1 ppm. The hydrocarbon backbone of the newly built dihydropyrazole ring revealed three more signals; i.e. a triplet, of one proton at d 5.1 ppm due to pyrazole-H5 that  appeared downfield, as expected, because of the neighbouring nitrogen atom; on the other hand, there were two doublet of doublets, each of one proton at d 3.7 and 2.8 ppm with coupling constants of 5.6 and 16.8 Hz assignable for pyrazole-H4 axial proton, 9.6 and 16.4 Hz of pyrazol-H4 equatorial proton. In addition, mass spectrum of compound 8 showed a strong base peak with the proper chlorine-isotope distribution pattern (see experimental part). Formation of dihydropyrazole derivatives were postulated to pass through two steps. The first step involves Michael-type addition on carbonyl b-carbon, followed by protonation to afford b-hydrazinylpropanone intermediate. The terminal primary amine attacks the carbonyl group to finally give the dihydropyrazole final product after loss a molecule of water as detailed below (Supplementary data).
Next, to synthesise dihydroisoxazole derivatives 11, 12, and 13, a mixture of chalcone (5,6,7), hydroxylamine hydrochloride, and NaOH in ethanol was heated up till reflux to give the corresponding target compounds 11,12, and 13. The IR spectra of compound 11 characterised by a band at 1590 cm À1 due to C¼N, 1 H NMR spectra of compound 11 showed, a doublet of doublet of one proton at d 6 ppm due to isoxazole-H-5 which appeared downfield as expected because C-5 of isoxazole attached to oxygen atom, a two doublet of doublet, each equivalent for one proton at d 4.0 and 3.5 ppm which attributed to isoxazole-H4 axial proton and isoxazole-H4 equatorial proton with coupling constant (11.2, 16.8 Hz) of axial proton and (6.8, 17.2 Hz) of equatorial proton, respectively, the most characteristic feature of 1 H NMR spectrum of compound 11 is the disappearance of the olefinic protons (Supplementary data).
Next, to prepare pyrimidine-2(1H)-thione derivatives 26,27, and 28, a mixture of chalcone, thiourea, and NaOH in ethanol was heated to reflux to give the corresponding target derivatives. Taking compound 26 as an example, the IR spectrum was characterised by a band at 3308 cm À1 assignable to one NH stretching, the 1 H NMR spectrum shows, a singlet of one proton at 14.1 ppm due to NH which is D 2 O exchangeable, a singlet for one proton at d 8.6 ppm due to quinoline-H4, a doublet for one proton at d 8.3 ppm due to quinoline-H-8 with coupling constant 8. 8 Hz, a doublet of two protons at d 8.1 ppm due to phenyl-H2,H6 with coupling constant 8. 4 Hz, a doublet for one proton at 7.8 ppm due to quinoline-H7 with coupling constant 8. 8 Hz, a doublet of two protons at d 7.5 ppm due to phenyl-H3,H5 with coupling constant 8. 4 Hz, a singlet for one proton at 7.4 ppm due pyrimidine-H5, a singlet of one proton at d 7.3 ppm due to quinoline-H5, a singlet of three protons at d 3.8 ppm due to OCH 3 of quinoline (Scheme 3).
Similarly, pyrimidin-2-amine was synthesised. Briefly, a mixture of chalcone, guanidine hydrochloride, and NaOH in ethanol was heated to reflux to give the corresponding compounds 29, 30, and 31. The structure of the isolated product was confirmed by spectral data, taking compound 29 as a representative example, the IR spectrum characterised by a band at 3320 and 3284 cm À1 due to NH 2 stretching, the 1 H NMR spectrum revealed a singlet for one proton at d 8.4 ppm due to quinoline-H4, a doublet of one proton at d 8.3 ppm due to quinoline-H8 with coupling constant 8.8 Hz, a doublet for two protons at d 7.9 ppm due to phenyl-H2,H6 with coupling constant 8. 4 Hz, a singlet of one proton at d 7.5 ppm due to pyrimidine-H5, a singlet of one proton at 7.1 ppm due to quinoline-H5, a doublet of two protons at 7 ppm due to phenyl-H3,H5, a doublet for one proton at d 6.7 ppm due to quinoline-H7 with coupling constant 8.8 Hz, a singlet for two protons at 5.4 ppm which attributed to NH 2 , D 2 O exchangable, a singlet of three protons at 3.7 ppm due to OCH 3 of quinoline.
Unlike other reactions in this series, reaction of chalcone with urea was preceded first under basic conditions, but the yield was poor, therefore we shifted towards the acidic medium. In brief, compounds 32, 33, and 34 were prepared, in a good yield, by allowing the corresponding chalcone to react with urea in the presence of conc. hydrochloric acid at a reflux temperature. Under strong acidic media, an alternative product was also anticipated, where the chlorine atom was expected to be hydrolysed to afford the amide-containing structure. Later structure was excluded mainly base on the absence its molecular ion peak from the MS. Other spectral and elemental data confirm this assumption and confirm the structure of compounds 32, 33, and 34 (Supplementary data).

In vitro anti-proliferative activity
In vitro cytotoxic activities of the synthesised compounds were assessed using standard MTT method [46][47][48] . a panel of human cancer cell lines namely; hepatocellular carcinoma (HepG-2), colorectal carcinoma (HCT-116), and breast cancer (MCF-7) were used in this test utilising colchicine as a reference standard. The antiproliferative activity was expressed in IC 50 values as reported in Table 1.
From the cytotoxic screening, the tested compounds showed different degrees of activities against the tested cells. In general, some compounds were proven to be efficient anticancer candidates with promising activities against the tested cells. Comparing to colchicine (IC 50

Tubulin polymerisation assay
The most active antiproliferative members were further evaluated for their inhibitory effect against tubulin polymerisation. The inhibition assay on microtubule polymerisation was evaluated turbidimetrically using a fluorescent plate reader 49 . Colchicine and CA-4 were used as positive controls (Table 2).
Compounds 21 and 32 exhibited the highest tubulin polymerisation inhibitory effect with IC 50 values of 9.11 and 10.5 nM, respectively. Such members showed activities higher than that of colchicine (IC 50 ¼ 10.6 nM) and CA-4 (IC 50

Cell cycle analysis
The impact of the most promising compound 25 on cell cycle distribution was assessed. Such analysis gave a better insight into the effect of the synthesised compounds on cancer cell growth inhibition. HepG-2 cells were utilised in this test. The reported method described by Wand et al. 50 was followed. HepG-2 cells were treated with compound 25 at a concentration equals its IC 50 value (1.89 mM) for 24 h.
The results revelled that compound 25 induced an increase in the percentage of HepG-2 cells from 24.83% to 28.95% at S phase. It exhibited marked increase in the percentage of HepG-2 cells from 13.8% to 30.17% at G2/M phase. Also, it increased the percentage of HepG-2 cells from 2.57% to 21.46% at Pre-G1 phase. On the other hand, it decreased the percentage of HepG-2 cells from 61.37% to 40.88% at G0-G1phase. These findings indicated that compound 25 can arrest the cell cycle at G2/M phase, and can cause apoptosis at pre-G1 phase (Table 3 and Figure 8).

Annexin V-FITC apoptosis assay
Annexin V and PI double staining assay 51 was carried out to explore the proposed apoptotic effect of the synthesised compounds. Compound 25 as a representative example was tested against HepG-2 cells. In this test, HepG-2 cells were incubated with compound 25 at a concentration of 1.89 mM for 24 h.
As shown in Table 4 and Figure 9, compound 25 induced apoptotic effect equal 19.05% (7.66% and 11.39 at early and late apoptosis, respectively), which was thirteen times more than the control (1.39%).

Docking studies
The synthesised compounds were docked against tubulin heterodimers using MOE2014 software to determine the binding free energy and binding mode (Table 5). Molecular docking studies gives perception about the binding mode and the degree of affinity between the docked compounds and the prospective target. good biological effect is shown by lower binding free energy and similar binding mode to that of the reference co-crystallised ligand. 52 A docking study of the co-crystallised ligand, DAMA-colchicine provides a binding energy value of À12.88 Kcal/mol with four hydrogen bonds and six hydrophobic interactions. The trimethoxy phenyl (A-ring) ring occupied the first pocket of the colchicine  binding site forming one hydrogen bond with Cys241. Other interactions were observed between the ring-A and different essential amino acid residues as Cys241, Leu255, and Ala250. Additionally, the side chain of the B-ring (2-mercaptoacetamide moiety) occupied the second cavity of the colchicine binding site forming two hydrogen bonds with Ser178 and Leu248. Furthermore, the methoxytropone moiety (C-ring) occupied the third pocket of the receptor forming one hydrogen and one hydrophobic bonds with Lys352 ( Figure 10).
Compound 19 as a representative example showed a binding mode like that of DAMA-colchicine, with affinity value of À16.51 kcal/mol. The 3,4,5-trimethoxyphenyl moiety (A-ring) occupied the first cavity of the colchicine binding site forming three hydrophobic interactions with Ala250, Leu255, and Leu248. Additionally, the 2-acetyl-4,5-dihydro-1H-pyrazol moiety (B-ring) occupied the second cavity of the colchicine binding site forming one hydrogen bond with Asn258 and one hydrophobic interaction with Leu248. Furthermore, the 2-chloro-6-methoxyquinoline moiety (C-ring) occupied the third pocket of the colchicine binding site ( Figure 11).
The binding mode of compound 28 exhibited an affinity value of À12.74 kcal/mol. The 3,4,5-trimethoxyphenyl moiety (A-ring) occupied the first cavity of the colchicine binding site forming four hydrophobic interactions with Ala316, Ala250, Cys241, and Leu255. The pyrimidine-2(1H)-thione (B-ring) occupied the second cavity of the colchicine binding site forming one hydrogen one hydrophobic bonds with Lys352. Furthermore, the 2-chloro-6methoxyquinoline moiety (C-ring) occupied the third pocket of the colchicine binding site forming three hydrophobic interactions with Lys254 and Ala180. The methoxy group formed one hydrogen bond with Gln11 ( Figure 13).

Structure-activity relationship (SAR)
SAR as an aim of our work was based on the results of in vitro cytotoxic activities of the synthesised compounds. Initially, the effect of the A-ring on the activity was explored. Comparing the cytotoxic activity of compounds 22, 25, 28, 31, and 34 incorporating 4-methoxyphenyl as an A-ring with compounds 20, 23, 26, 29, and 32 incorporating 4-chlorophenyl as an A-ring, and compound 21, 24, 27, 30, and 32 incorporating 4-aminophenyl as an A-ring indicated that the cytotoxic activities decreased in the order of 4-methoxyphenyl > 4-chlorophenyl > 4-aminophenyl.

Conclusion
In summary, twenty-nine of quinoline derivatives were designed and synthesised. The synthesised compounds were evaluated for their anti-proliferative activities against a group of three human cancer cell lines including; colorectal carcinoma (HCT-116), hepatocellular carcinoma (HepG-2), and breast cancer (MCF-7). Compounds 20,21,22,23,24,25,26, and 28 exhibited superior cytotoxic activities with IC 50 values ranging from 1.78 to 9.19 mM. Additionally, the most promising members were tested for their tubulin polymerisation inhibitory effect. Compounds 21 and 32 exhibited the highest tubulin polymerisation inhibitory effect with IC 50 values of 9.11 and 10.5 nM, respectively. Structure-activity relationship of the synthesised compounds revealed that 4methoxyphenyl > 4-chlorophenyl > 4-aminophenyl as an A-ring. In the same time. The it was found that substituted five-membered ring > non-substituted six-membered ring > non-substituted five-membered ring as a B-ring. Moreover, compound 25 was approved to arrested the cell cycle at the G2/M phase and induced apoptosis in HepG-2 cells. Docking experiments assisted these findings by anticipating potential binding interactions between the target compounds and the active sites of tubulin heterodimers. The most effective candidates in the quest for strong and selective antineoplastic agents will serve as valuable lead compounds and merit further investigations.

Chemistry
1 H NMR spectra were run at 400 MHz and 13 C spectra were determined at 100 MHz in deuterated chloroform (CDCl 3 ), or dimethyl sulfoxide (DMSO-d 6 ) on a Varian Mercury VX-400 NMR spectrometer. Chemical shifts are given in parts per million (ppm) on the delta (d) scale. Chemical shifts were calibrated relative to those of the solvents. Flash chromatography was performed on 230-400 mesh silica. The progress of reactions was monitored with Merck silica gel IB2-F plates (0.25 mm thickness). The infra-red spectra were recorded in potassium bromide discs on pye Unicam SP 3300 and Shimadzu FT IR 8101 PC infra-red spectrophotometer. Mass spectra were recorded at 70 eV. High resolution mass spectra for all ionisation techniques were obtained from a Finnigan MAT XL95. Melting points were determined using capillary tubes with a Stuart SMP30 apparatus and are uncorrected. All yields reported refer to isolated yields.  To a stirred and ice-cooled aqueous solution of sodium hydroxide (10 mmole, 50% w/w) and absolute ethanol (12.5 mL), substituted acetophenone (10 mmole) namely, 4-chloroacetophenoe, 4-aminoacetophenone, and 3,4,5-trimethoxyacetophenone was added followed by 2-chloro-6-methoxyquinoline-3-carbadehyde 4 (10 mmole). The reaction mixture was vigorously stirred for 3-6 h while temperature was maintained below 25 C till the reaction mixture became thick. The reaction mixture was left in the refrigerator overnight. The formed precipitate was filtered off under vaccum and washed with copious amount of water until the filtrates became neutral to litmus paper, washed with ice-cold ethanol (20 mL), and then recrystallized from ethanol to give the titled compound 5, 6, and 7, respectively.

Cell cycle analysis
The effect of the most promising compound 25 on cell cycle was evaluated using flowcytometer 50,53,54 as illustrated in Supplementary data.

Annexin V-FITC apoptosis assay
The effect of compound 25 on apoptosis induction was analysed using Annexin V-FITC/PI apoptosis detection kit using flowcytometer 51,55,56 as illustrated in Supplementary data.

Docking studies
The Crystal structure of the target receptor (tubulin) [PDB ID: 1SA0, resolution 3.00 Å] was obtained from Protein Data Bank (http://www.pdb.org). The docking process was carried out using MOE2014 software. At first, the crystal structure of the target was prepared by removing water molecules and retaining the two essential chains and the co-crystallised ligand, N-deacetyl-N-(2mercaptoacetyl)-colchicine (DAMA-colchicine). Then, the protein structure was protonated, and the hydrogen atoms were hided. Next, the energy was minimised, and the binding pocket of the protein was defined.
The 2D structures of the synthesised compounds and reference ligand (DAMA-colchicine) were sketched using ChemBioDraw Ultra 14.0 and saved as MDL-SD format 57 . Then, the saved files were opened using MOE and 3D structures were protonated. Next, energy minimisation was applied. Before docking process, validation of the docking protocol was carried out by running the simulation only using the co-crystallised ligand (DAMA-colchicine) which showed low RMSD value. The molecular docking of the synthesised was performed using a default protocol against the target receptor. In each case, 30 docked structures were generated using genetic algorithm searches, London dG was used for scoring and forcefield (MMFF94) for refinement. The London dG scoring function estimates the free energy of binding of the ligand from a given pose. The output from of MOE was further analysed and visualised using Discovery Studio 4.0 software. [58][59][60][61][62][63]

Disclosure statement
No potential conflict of interest was reported by the author(s).