Design and synthesis of 4-piperazinyl quinoline derived urea/thioureas for anti-breast cancer activity by a hybrid pharmacophore approach

Abstract In an attempt to improve anti-breast cancer activity, a new series of 4-piperazinylquinoline derivatives based on the urea/thiourea scaffold were designed and synthesised by a pharmacophore hybrid approach. We then examined for their antiproliferative effects on three human breast tumor cell lines, MDA-MB231, MDA-MB468 and MCF7, and two non-cancer breast epithelial cell lines, 184B5 and MCF10A. Among those 26 novel compounds examined, 5, 9, 17, 18, 21, 23 and 29 showed significantly improved antiproliferative activity on breast cancer cells. Compound 23 (4-(7-chloro-quinolin-4-yl)-piperazine-1-carbothioic acid (2-morpholin-4-yl-ethyl)-amide) (RL-15) is especially desirable, since its antigrowth/cell-killing activity is 7-11 fold higher on cancer than non-cancer cells. Data from cell biological studies demonstrated that cancer cells compromised plasma membrane integrity in the presence of compound 23. The cancer cell-specific property of compound 23 shown in cell culture stands in vivo test, this compound can be an excellent lead for effective and safe anticancer drug.


Introduction
Chloroquine (CQ), which contains a 4-aminoquinoline scaffold (Figure 1), is a well-known antimalarial drug. Based on repurposing concept, we previously demonstrated that the combination of CQ with radiation or Akt inhibitors not only significantly increases antigrowth/cell-killing effects but also enhances the selectivity towards cancer over non-cancer cells [1][2][3] . CQ is well known for their lysosomotropic property and accrued in the lysosomes and elevates intra-lysosomal pH; and inhibits with autophagosome degradation in the lysosomes. This unique characteristic of CQ and its analogs may be imperative for the enhancement of cellkilling by cancer therapeutic agents in different tumor models 4,5 . Based on this interesting note, we synthesised several 4-aminoquinoline analogs (Figure 1, I) and examined their cytotoxic effects on breast cancer cell lines. We found that some of these compounds are very effective and show selective cytotoxic effects on cancer cells 6,7 . In continuation of our efforts to develop more effective CQ analogs (Figure 1, II and III) by merging 4-piperazinylquinoline ring structure with an isatin ring by a hybrid approach, and found that 4-piperazinylquinoline exhibited promising antibreast cancer activity [8][9][10] .
Small heterocyclic molecules have enrich potential for the discovery of drug candidates, among which urea and thiourea groups are privileged pharmacophores found in many medicinally active compounds. They have been shown to have promise against cancer cells, HIV-1 protease, hypercholesteromia and atherosclerosis 11,12 . The current literatures evident that molecules containing urea and thiourea pharmacophores (Figure 2(a), IV-VII) are potent inhibitors of human DNA-topoisomerase II and active against various cancer cells 11,[13][14][15][16][17][18] . Based on these prior annotations, we surmised that appropriate hybridization on these pharmacophores (i.e. 4-piperazinylquinoline and urea/thiourea) could possess an effective anticancer activity. Similar approaches of combining two molecules have previously been exploited with very good results [19][20][21][22][23][24] . Encouraged by these results, we designed and synthesised hybrid compounds by linking the main structural unit of the 4-piperazinylquinoline ring system with the urea/thiourea functionality, and examined their cytotoxic effects on three human breast tumor and two matching non-cancer cell lines ( Figure 2(b), Scheme 1). For the most desirable compound, we carried out cell-based experiments to gain the mechanism of function.

Chemistry
Melting points (mp) were taken in open capillaries on the Complab melting point apparatus. Elemental analysis was performed on a Perkin-Elmer 2400 C, H, N analyzer and values were within the acceptable limits of the calculated values. The 1 H NMR spectra were recorded on a DPX-500 MHz Bruker FT-NMR spectrometer using deuterated chloroform (CDCl 3 ) and dimethyl sulfoxide (DMSO)-d 6 as solvent. The chemical shifts were reported as parts per million (d ppm) tetramethylsilane (TMS) as an internal standard. Mass spectra were obtained on a JEOL-SX-102 instrument using fast atom bombardment (FAB positive). The progress of the reaction was monitored on readymade silica-gel plates (Merck) using chloroform-methanol (9:1) as solvent. Iodine was used as a developing agent or by spraying with the Dragendorff's reagent. Chromatographic purification was performed over a silica

Cell lines
The human breast cancer cell lines MDA-MB468, MDA-MB231 and MCF-7 were purchased from ATCC and maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum (Hyclone, Logan UT) and 2 mM L-glutamine. MCF10A and 184B5 immortalised breast cells were purchased from ATCC and maintained in mammary epithelial basal medium supplemented with an MEGM mammary epithelial singlequot kit (Cambrex). Cells were grown at 37 C with 5% CO 2 , 95% air under the humidified conditions. Cell line authentication was carried out by Genetica DNA Laboratories (Burlington, NC) using a short tandem repeat profiling method (March 2015; July 2015; September 2016).

SRB assay
Cell cytotoxic effects were determined by a SRB-based protocol 2,25 .
Cell staining procedure Lysosomal and AO staining were carried out as described previously 26 . Mitochondrial staining with MitoTracker was carried out as described previously 26 . Briefly, cells were plated on a poly-Dlysine pre-coated 1.8-cm sterile glass coverslip that had been placed in a well of a 6-well clustered dish. Cells were then incubated with 50 nM MitoTracker Red (ThermoFisher) for 30 min in a cell culture incubator. Plasma membrane staining was carried out with lipid-specific CellMask (Molecular probes, Life Technologies), according to the manufacturer's instruction. For actin staining, cells fixed with 4% formaldehyde were incubated with phalloidin conjugated with Fitc (green) for 30 min. For nuclear staining, cells were incubated with 5 mM of DRAQ5 (blue) for 5 min.

Results and discussion
Chemistry Twenty-six novel hybrid compounds (5-30) were synthesised in the present study as outlined in Scheme 1. The intermediate compounds of 7-substituted-4-piperazin-1-yl-quinoline (3 and 4) were prepared by aromatic nucleophilic substitution on 7-substituted-4chloro-quinoline (1 and 2) with excess of piperazine in the presence of triethylamine, and the products were obtained with a simple standard workup procedure in excellent yields. Substituted-4piperazinylquinoline derived urea (5-16) or thiourea (17-30) analogs were obtained from a "one pot-two component protocol" in which an appropriate secondary amine, substituted isocynate or thiocynate in the presence of triethylamine in anhydrous dimethyl formamide (DMF) at room temperature. The electrophilic character of the carbonylic carbon (5-16) is stronger than that of the thiocarbonylic carbon (17-30), thus facilitating the nucleophilic attack of the secondary amines. The products were purified by recrystallization with ethanol, and yields were in the range of 60-98% for 4-piperazinylquinoline derived urea analogs (5-16) and 55-98% for the 4-piperazinylquinoline derived thiourea analogs (17-30). The higher yield was observed for 4-(7-chloro-quinolin-4-yl)piperazine-1-carboxylic acid phenylamide (5) due to the higher reactivity of phenylisocyanate. The infrared (IR) spectra for the 4-piperazinylquinoline derived urea and thiourea analogs showed the absorption stretching band values in the range of 1645-1620 cm À1 and 1367-1348 cm À1 for C ¼ O and C ¼ S groups, respectively. The values obtained in the 1 H NMR chemical shifts permitted the characterization of the hydrogens showing similar values for both urea and thiourea analogs with the same substituents. However, deshield effects on H-1 and H-a NMR chemical shift values of thiourea analogs were observed (c.a. d 0.2-0.7) due to the heavy atom effect of sulfur. The signals of aromatic hydrogens occurred as multiplets in the characteristic region. The 1 H NMR chemical shift of the N-H in the urea and thiourea analogs were presented a strong deshielding when compared with the N-H values of the amines, for example, d 2.31 (3) and d 6.63 (5), where X ¼ Cl. 13 C NMR spectral analysis of the urea and thiourea analogs showed the typical absorptions for aliphatic carbons in the expected region, such as the signals for the carbonylic and thiocarbonylic carbons, in the range of d 154.39-156.34 and d 182.46-185.78, respectively. The compounds reported in this study have also been thoroughly characterised by elemental analysis and mass spectral data.

Antiproliferative effects of the compounds on cancer and noncancer cells
The newly synthesised 4-piperazinylquinoline compounds  were evaluated for their antiproliferative effects using three breast cancer cell lines, for which the compounds were diluted to achieve seven different concentrations ranging from 100 to 1.625 mM. Followed by 48 h incubation with the compound, cells were treated with sulforhodamine B (SRB) to measure their growth/viability (% of the untreated control) by spectrophotometer 25 . The total cellular macromolecules levels known to accurately reflect by UV reading of SRB staining. The 50% growth inhibition (GI 50 ) concentration for each derivative was calculated with reference to a standard curve (control cells), which represents the concentration that results in a 50% decrease in cell growth after 48 h of incubation. The resultant data are shown in Table 1. The reference compounds CQ and cisplatin were included for the comparison.
Among the 26 novel compounds synthesised and examined, sixteen compounds showed GI 50 in the range of 3.0-19.8 mM, 11 compounds at 20.6-50.6 mM against MDA-MB231 triple-negative breast cancer cells. In the case of MDA-MB468 cells, 18 compounds showed GI 50 range between 4.6 and 19.8 mM, eight compounds between 21.7 and 57.6 mM. As for MCF-7 cells, 21 compounds showed GI 50 in the range of 4.0-18.6 mM, five compounds at 20.4-30.2 mM. The differences in the GI 50 values may be attributable to a variety of factors such as nature of the seventh substitution at the 4-piperazinylquinoline ring system, the nature of urea/thiourea substitution and the genetic and biochemical background of the cell lines.
Compounds derived from the 7-chloro-4-piperazinylquinoline ring system hybridised with urea pharmacophore having a phenyl (5) or naphthyl (9) substitution showed increased antiproliferative effects on MDA-MB231, MDA-MB468 and MCF7 cells in comparison with those derived from the 7-trifluoro substituted 4-piperazinylquinoline ring system linked with urea pharmacophore having a phenyl (11) or naphthyl (15) substitution. Compounds derived from the 7-trifluoro-4-piperazinylquinoline ring system hybridised with a 2,5-dialkyl phenyl (12) or a 4-trifluoromethyl phenyl (13) urea analog showed increases of antiproliferative activity in all of the three breast cancer cell lines examined, compared to those derived from the 7-chloro substituted on the 4-piperazinylquinoline ring system with a 2,5-dialkyl phenyl (6) or a 4-trifluoromethyl phenyl (7) urea analog. Similarly, compounds derived from the 7trifluoro-4-piperazinylquinoline ring system hybridised with a 2,4,6-trichloro phenyl (14) or a cyclohexyl (16) substituted urea analog showed increased antiproliferative activities on all three breast cancer cell lines, compared to those derived from the 7chloro substituted 4-piperazinylquinoline ring system with a 2,4,6trichloro phenyl (8) or a cyclohexyl (10) urea substitution. However, all of these compounds showed lower antiproliferative activity than compound 5. Structure-activity relationship studies indicates that the introduction of small to bulky groups of liphohphilic substitution such as dialkyl (6), halogenated (7 and 8) poly aromatic hydrocarbon (naphthyl 9) or a cyclic alkyl (cycloalkyl 10) moiety leads to an increase in lipophilicity and decrease in the antiproliferative activity, compared to the unsubstituted phenyl compound (6). This further indicates that the 7-chloro-4-piperazinylquinoline ring system is favorable for antiproliferative activity. It is also clear that the replacement of the 7-chloro group with their bioisoteric functional group of the 7-trifluoromethyl substitution leads to the decrease in the antiproliferative activity on breast cancer cells 7 . The bioisoteric replacement of the chloro atom with stronger electron-withdrawing group of trifluoromethyl resulted in less potent analogs for the anticancer activity 7,27-29 .
Among the series, thiourea compounds 19, 21, 24, 28 and 29 showed stronger activities against all three different breast cancer cell lines. In this case, the more lipophilic property of thiourea may play an important role. For example, the Log p values of the thiourea compounds 28, which has a higher Log p values (6.7), is more effective than the corresponding urea compound 10, which has lower Log p values (4.2). In fact, all of the thiourea containing compounds with higher Log p values are more active, compared to the corresponding urea analogs. It may be due, at least in part, to the differences in electronegativitiy of sulfur and oxygen atoms on thiourea and urea analogs.
Compounds 5, 9, 23 and 29 show 4-8-fold higher activity than the reference compounds cisplatin and CQ. Among them, compound 23 is the most effective as its GI 50 values are 3.0, 4.6 and 4.5 mM against MDA-MB231, MDA-MB468 and MCF7 cells, respectively.
We further assessed their antiproliferative effects on two noncancer immortalised cell lines (184B5 and MCF10A). As shown in Table 1, the antigrowth effects of compounds 8, 10,11,14,15,16,21,22,25,26 and 27 are comparable against cancer and non-cancer cells, although some of them are more active against certain cell lines (e.g. compounds 8 and 10). Remarkably, the following compounds show higher activity against non-cancer cells than cancer cells: 6, 7, 12, 13, 19, 20 and 28. In contrast, compounds 5,9,17,18,23,24,29 and 30 showed much stronger activities against cancer than non-cancer cells. Among these compounds, compound 23 (RL-15) is the most desirable since its GI 50 values against non-cancer cell lines are 34.4 lM (184B5) and 32.3 mM (MCF10A) (Table 1; Figure 3). The antiproliferative effects of compound 23 on cancer cells are thus 7-fold (MDA-MB468 vs MCF10A) to 11-fold (MDA-MB231 vs 184B5) higher than matching non-cancer breast cells, which may indicate that compound 23 could potentially be safer than other compounds. Therefore, we examined it further to gain insights into its molecular mechanism of action.

Cancer cells lose plasma membrane integrity in the presence of compound 23 (RL-15)
In contrast to the Sham control, MCF7 breast cancer cells treated with compound 23 readily uptake ethidium bromide (EtBr), suggesting that the plasma membrane may have been compromised (Figure 4(a), left panels). This conclusion is further strengthened by data obtained from acridine orange (AO) staining, which also showed abnormal (swollen) cell morphology (Figure 4(a), right panels). However, data from phalloidin stained MCF7 cells indicated that the structure of actin filaments at the plasma membrane may largely be intact, even though some cells are apparently enlarged (Figure 4(b)). Interestingly, the entire cytoplasm was stained by the lipid-specific CellMask in the cells treated with compound 23 (Figure 4(c)). Thus, these data are consistent with the notion that the lipid structure on the plasma membrane has been compromised in response to compound 23 (Figure 4(c)). However, it is currently unclear how lipid molecules are found throughout the entire cytoplasm. We found that 42.4 ± 4.8% of 500 cells examined showed compromised plasma membrane phenotypes by 24-h post treatment. In addition, cells treated with compound 23 are often enlarged and flattened, and contained multi-nuclei. In addition, many small vacuoles are found in the cytoplasm of these cells (Figure 4(a-c)). All of these phenomena are likely leading to eventual cancer cell demise in the presence of compound 23.

The mitochondria and lysosome are increased in volumes and disorganised
The MCF7 cells treated with compound 23 showed substantial increases in mitochondrial and lysosomal volumes, often in a disorganised form (Figures 5 and 6). It should be noted that these abnormalities could occur if the membranes of these organelles are compromised. Thus, the molecular target of compound 23 is likely a common element found in the membranes of cytoplasm, mitochondria and lysosomes, but not in the nuclear membrane as it is not compromised in the presence of compound 23 ( Figures  4-6). Finally, our data also show that nucleus is often pushed to one side ( Figure 5(b)), probably due to the increase in volumes and disorganization of mitochondria and lysosomes.

Conclusion
We here describe the hybrid pharmacophore design and synthesis of a new series of 4-piperazinylquinoline derivatives. Data from an in vitro study showed that most of the compounds derived from 4-piperazinylquinoline exhibited promising anticancer activity against human breast cancer cells. Among the 26 novel compounds, compounds 5, 9, 17, 18, 23 and 29 showed significantly improved antigrowth/antiproliferative activity against MDA-MB231, MDA-MB468 and MCF7 breast cancer cells. Comparing those hybrid compounds containing thiourea or urea, the former is generally more active than the latter as compounds 19, 21, 24, 28 and 29 showed stronger activities than their corresponding urea analog compounds 7, 8, 11, 14 and 15 against all three different breast cancer cell lines. Since all of these compounds required higher drug concentration to achieve the same GI 50 value against the 184B5 and MCF10A non-cancer breast cell lines, they could be safer than other compounds. Among them, compound 23 (4-(7chloro-quinolin-4-yl)-piperazine-1-carbothioic acid (2-morpholin-4yl-ethyl)-amide) is considered to be the most desirable, since its antiproliferative activity is 7-11 fold higher on cancer than noncancer cells. Data from cell morphology study demonstrated that the membrane integrity of the cytoplasm, mitochondria and lysosome is compromised in the cells treated with compound 23, perhaps through the disruption of the lipid structure in these membranes. However, the nuclear membrane is not compromised, suggesting that the target molecule of compound 23 may be a common element found on the membranes of the cytoplasm, mitochondria and lysosome, but not in the nuclear membrane. Overall, our data suggest that hybrid compounds containing the 4-piperazinylquinoline pharmacophore and thiourea can be promising leads, and this new hybrid approach can be an excellent way of developing effective anti-breast cancer agents.