Benzoxazole derivatives as new VEGFR-2 inhibitors and apoptosis inducers: design, synthesis, in silico studies, and antiproliferative evaluation

Abstract In this study, a set of novel benzoxazole derivatives were designed, synthesised, and biologically evaluated as potential VEGFR-2 inhibitors. Five compounds (12d, 12f, 12i, 12l, and 13a) displayed high growth inhibitory activities against HepG2 and MCF-7 cell lines and were further investigated for their VEGFR-2 inhibitory activities. The most potent anti-proliferative member 12 l (IC50 = 10.50 μM and 15.21 μM against HepG2 and MCF-7, respectively) had the most promising VEGFR-2 inhibitory activity (IC50 = 97.38 nM). A further biological evaluation revealed that compound 12l could arrest the HepG2 cell growth mainly at the Pre-G1 and G1 phases. Furthermore, compound 12l could induce apoptosis in HepG2 cells by 35.13%. likely, compound 12l exhibited a significant elevation in caspase-3 level (2.98-fold) and BAX (3.40-fold), and a significant reduction in Bcl-2 level (2.12-fold). Finally, docking studies indicated that 12l exhibited interactions with the key amino acids in a similar way to sorafenib.


Introduction
Cancer chemotherapy has been considered one of the most important medical advances in the past few decades 1 . However, the narrow therapeutic index besides the unpredictable effects were the major drawbacks of the primary introduced drugs 2 . In contrast, the recently developed targeted therapies gained the advantages of interfering with specific molecular targets almost located in the tumour cells with minimised effect on the normal cells 3 . Thus, these agents provide a high specific therapeutic window with limited non-specific toxicities.
Among the major vital cancer drug targets are tyrosine kinases (TKs) because of their potential role in the modulation of growth factor signalling 4,5 . Upon their activation, TKs increase both proliferation and growth of tumour cells with induction of apoptosis and reinforcement of angiogenesis and metastasis 6 . Thus, TKs inhibition by different inhibitors became a key approach in cancer management 7 . The evidenced drug ability as well as the safety profile of the FDA-approved TKs inhibitors emphasised the attractiveness of TKs as drug targets.
Owing to their significant participation in modulating angiogenesis, vascular endothelial growth factors (VEGFs) have been considered the key players over other TKs 8 . VEGFs action is performed after their binding to three different tyrosine kinase (TK) receptors, namely, VEGFR-1, VEGFR-2, and VEGFR-3 8 . VEGFR-2 receptor possesses the most crucial role among the rest subtypes as its activation leads to initiation of downstream signal transduction pathway via dimerisation followed by autophosphorylation of tyrosine receptor, a pathway resulting finally to angiogenesis 9 . Therefore, hindering VEGF/VEGFR-2 pathway or, even, weakening its response is of considered targets of the recent chemotherapeutic agents 10 . Despite a large number of small molecules with various chemical scaffolds being evidenced to tackle this pathway, resistance development in addition to different adverse effects still the main drawback of the current known VEGFR-2 inhibitors drugs 11 . Thus, the discovery of more effective and less dangerous VEGFR-2 inhibitors becomes an attractive therapeutic target for cancer drug discovery 12 . It has been discovered that VEGFR-2 inhibition in cancer cells causes and expedites apoptosis, which works in concert to enhance the antitumor effect. Hence, the most potent derivative has thoroughly discoursed in our work through the assessment of certain apoptotic markers such as caspase-3 (a crucial component in apoptosis that coordinates the destruction of cellular structures such as DNA and cytoskeletal proteins 13 , BAX and Bcl-2 (members of the Bcl-2 family and core regulators of the intrinsic pathway of apoptosis) 14 .
Over the last decade, we have built a project that is concerned with cancer management. Our high-throughput efforts gave us the opportunity to identify several small molecules that may serve as anti-angiogenic agents 9 . Most of these molecules exhibited VEGFR-2 inhibitory activity comparable to that of the FDA-approved inhibitor, sorafenib. These molecules were precisely designed to resemble the four main structural parts of sorafenib and other VEGFR-2 inhibitors [15][16][17] . Those parts were well-known to be a hydrophobic hinge binding head, a linker, a hydrogen-bonding moiety, and a hydrophobic tail ( Figure 1). These previously mentioned parts enabled the designed compounds to fit perfectly in the TK active pocket. Based on the promising biological results in our former published work in which we utilised benzoxazole moieties as a hinge-binding core 18 , we decided to continue our preliminary VEGFR-2 studies using the same three different scaffolds of benzoxazole but with two main considerable additional modifications; a) For the allosteric hydrophobic pocket, we used different terminal aliphatic hydrophobic moieties including cyclopentyl (compounds 12a-c) and ter-butyl moiety (compounds 12d-f). This allowed us to make a comparative study between aliphatic and aromatic derivatives of each scaffold and study the SAR of the obtained compounds as anticancer leads with significant VEGFR-2 inhibitory potentialities, as was planned in our design. b) The pharmacophore moiety was selected to be amide derivative (compounds 12a-l) or diamide derivatives (compounds 13a-c) to study which derivative is more preferred biologically.

Rationale and design
Forcing by the fact that molecular hybridisation is one of the most important drug discovery approaches, our team co-workers started the present work. Sunitinib, a multi-targeted receptor tyrosine kinase (RTK) inhibitor 19 , lucitanib, a dual VEGFRs and FGFRs inhibitor 20 , and compound A, a potent VEGFR-2 inhibitor were our guides for building a new anti-angiogenic hybrid 21 . Thus, the indolinylidene moiety of sunitinib was altered to be benzoxazole in the new hybrid to investigate its ability to modify the biological effects. In addition, we did another modification to the sunitinib structure via replacing the fluorine atom by either hydrogen, methyl, or chlorine atoms that allowed us to measure the biological effects of these atoms compared to the fluorine atom. In contrast, the carboxamide moiety of both sunitinib and lucitanib was kept or expanded to continue acting as a hydrogen bonding part. On the other side, the hydrophobic tail in the new hybrid was suggested to be either aliphatic (tert-butyl), alicyclic (cyclopentyl), or aromatic (methoxy or chloro phenyl) to get a diverse number of congeners with a higher chance to study the structureactivity relationship of the newly designed hybrid. However, an in silico study was also carried out through the docking tools to confirm the proposed design ( Figure 2).
The final target candidates 12a-l and 13a-c were furnished in dry DMF via heating the potassium salts 3a-c with the previously synthesised intermediates 7a-d and 11, respectively (Scheme 3). Infra-red (IR) spectra of compounds 12a-l indicated the presence of characteristic NH and C ¼ O groups stretching bands at a range of 3181-3412 and 1644-1688 cm À1 , respectively. Moreover, their 1 H NMR spectra showed the presence of the two NH amide group signals at a range of d 7.73-10.83 ppm. The formation of compounds 13a-c was confirmed by 1 H NMR spectra which showed the appearance of three singlet signals at a range of d 10.53-10.79 ppm corresponding to the NH protons.

Biological evaluation
2.2.1. In-vitro antiproliferative activities against MCF-7 and HepG2 cell lines The in vitro antiproliferative effects of the newly synthesised benzoxazole derivatives 12a-l and 13a-c were determined against hepatocellular cancer (HepG2) and breast cancer (MCF-7) cell lines employing the standard MTT assay protocol wherein sorafenib was applied as a reference. The cytotoxicity results were obtained as median growth inhibitory concentration (IC 50 ). As presented in Table 1, major members of the synthesised compounds displayed promising anticancer activity.
Observing the results of anti-proliferative activity, valuable data concerning the structure-activity relationships was determined.

Vegfr-2 inhibitory assay
VEGFR-2 inhibitory effect of the most cytotoxic candidates 12d, 12f, 12i, 12 l, and 13a was investigated and summarised in Table  2. Sorafenib was used as a reference.

Correlation study between cytotoxicity and VEGFR-2 inhibition
The VEGFR-2 inhibitory activities of the tested compounds were plotted against their corresponding cytotoxicity in a simple linear regression for the HepG2 cell line in order to confirm the relationship between VEGFR-2 inhibition and cytotoxicity. The calculated R 2 square value 0.6274) shows a significant correlation between the tested compounds' induction of cytotoxicity and inhibition of VEGFR-2. As a result, one possible mechanism of the established compounds' cytotoxicity in the established cell line is their inhibition of VEGFR-2 activity ( Figure 3).

Evaluation of in vitro cytotoxicity against normal cell line
The most potent members 12d, 12i, and 12l were assessed for their in vitro cytotoxicity against normal cell lines using WI-38 (a human lung cell line) and sorafenib as a reference The IC 50 values for compounds 12d, 12i, and 12l were 99.41, 76.78, and 37.97 M, respectively (Table 3). Such values were very high in comparison to the corresponding values on cancer cell lines, which reflect high safety profile of the tested candidates towards normal cell lines.

Cell cycle analysis
Compound 12I, achieved notable cytotoxic and VEGFR-2 inhibitory potencies was further studied mechanistically for cell cycle progression and induction of apoptosis in HepG2 cells. Cell cycle process was analysed after exposure of HepG2 cells to 12I with a concentration of 10.50 mM for 24 h. Flow cytometry data revealed that the percentage of cells arrested at Pre-G1 phase decreased from 0.93% (in control cells) to 0.79% (in 12I) treated cells. Additionally, a marked decrease in cell population was observed at the G1 phase (28.34%) comparing to control cells (51.07%). For the S phase compound 12l induced a significant increase in the cell population (38.68%) comparing to control cells (27.22%). Finally, compound 12I exhibited significant increase in the cell population (32.10%) at the G2/M phase, comparing to the control cells (20.78%). Such outputs verify that compound 12I arrested the HepG2 cancer cell's growth mainly at the Pre-G1 and G1 phases (Table 4 and Figure 4).

Apoptosis analysis
The most potent anticancer agent 12l was selected for the assessment of apoptosis in HepG2 cells using Annexin V/propidium iodide (PI) double staining assay method. In this method, HepG2 cells were incubated with compound 12l at the IC 50 concentration (10.50 mM) for 24 h. The results revealed that compounds 12l could induce apoptosis more than the untreated control cells by a ratio of 35.13%. In details, 32.45 and 2.86% for early and late apoptotic phases, respectively compared to control, (6.56%,5.34%,1.22%, respectively) ( Figure 5 and Table 5).

Evaluation of BAX and bcl-2 expressions
Compound 12lwas subjected to further cellular mechanistic study. The cellular levels of BAX and Bcl-2 were measured using the western blot technique after compound 12l was applied to HepG2 cells for 24 h. The results indicated that compound 12l increased the concentration of the pro-apoptotic factor BAX by 3.40-fold while decreasing the concentration of the anti-apoptotic protein Bcl-2 by 2.12-fold. Furthermore, a significant increase in the BAX/Bcl-2 ratio by 6.83-fold was observed. The obtained findings indicated that compound 12l was effective in the apoptosis cascade and may encourage the apoptotic pathway (Table 6 and Figure 6).

Caspase 3 assay
Caspase-3 has a key role in apoptosis initiation and execution 26,27 . The western blot technique was used to investigate the effect of compound 12 l, the most promising member, on the caspase-3 level. HepG2 cells were treated with 12l (10.50 mM) for 24 h. Comparing control HepG2 cells, compound 12l caused a significant increase in the cellular levels of caspase-3 (2.98-fold) as presented in Table 6 and Figure 6.

2.3.
In silico studies 2.31.1. Docking study To understand the pattern by which the synthesised compounds bound to the active site 28,29 , all compounds were subjected to a docking study into the VEGFR-2 ATP binding site (PDB: 4ASD, resolution: 2.03 Å). The native co-crystallized inhibitor, sorafenib, was adopted as a reference in the present work. Following the preparation of the downloaded protein, a validation step was carried out in which the native inhibitor, sorafenib, was re-docked against the catalytic VEGFR-2 site. Results of the previous step successfully reproduced an identical binding pattern to that of the co-crystallized ligand with an RMSD value of 0.71 Å Figure 7. Thus, the later findings supported the validity of the suggested docking protocol.
Observation of the kinds of interaction between sorafenib and the VEGFR-2 catalytic site revealed that it could form two interaction types (Figure 8). The 1 st type is an H-bonding interaction, as sorafenib formed two H-bonds with a critical amino acid (Cys919) in the hinge region in addition to three H-bonds with the DFG motif amino acids (Asp1046 and Glu885). The 2 nd interaction type included different p interactions between sorafenib and the hydrophobic amino acids among the active pocket.
Docking conformations of the synthesised derivatives revealed that they were stacked onto the VEGFR-2 catalytic site in a way similar to that of the original ligand. However, the predicted docking pose of compound 12l showed that its benzoxazole fragment was linked to the hinge region Cys919 amino acid via a strong Hbond. Additionally, compound 12l interacted by an H-bond with Glu885 and two H-bonds with Asp1046 in the DFG motif ( Figure  9). The later binding pattern gave a reasonable explanation for 12l of being the most active biologically among the tested  compounds. A superimposition poses of 12l and the native ligand, sorafenib, provided additional evidence to the obtained results. As presented in Figure 10, compound 12l and sorafenib generally overlapped well and had the same 3-D orientation. Niceties revealed that the pharmacophoric moieties of sorafenib represented by N-methylpicolinamide, phenoxy, urea, and 4-chloro-3-(trifluoromethyl)phenyl) moieties had the same orientation with the 5-methylbenzo[d]oxazol, N-phenylacetamide, amide, and 3chlorophenyl moieties, respectively of compound 12l.

Pharmacokinetic profiling study
In the current study, an in silico computational study of the tested candidates was conducted following the directions of Veber's and Lipinski's rule of five 30,31 .
The obtained findings presented in Table 7 showed that all tested compounds showed no contravention of Lipinski's and Veber's Rules and hence display a drug-like molecular nature. In detail, the LogP, molecular weight, number of H-bond donors, and number of H-bond acceptors of these fifteen compounds are within the accepted values of less than 5, 500, 5, and 10, respectively. Moreover, the number of rateable bonds and TPSA of such compounds are within the acceptable values of less than 10 and 140 Å 2 , respectively.

Swissadme study
To compute the physicochemical properties and the drug likeness properties of the most potent compounds 12d, 12i, and 12 l, SwissADME online web tool was applied. The obtained results  predicted that the physicochemical properties of the three candidates were in acceptable ranges, hence they may have good oral bioavailability. Also, they are expected to hve undesirable effects on CNS as they cannot pass BBB (Table 8). Furthermore, SwissADME revealed that compounds 12d, 12i, and 12l fulfilled Lipinsk ıs, Veber's, and Ghose's rules predicting that these compounds have promising drug-likeness profiles (Table 7). Moreover, the radar charts which involved the calculation of six parameters including lipophilicity, polarity, flexibility, size, saturation, and solubility showed that compounds 15 b and 17 b (represented by red lines and integrated into the pink area) are almost predicting acceptable oral bioavailability (Table 9).

Conclusion
In the present study, fifteen benzoxazole derivatives were designed, synthesised as potential anticancer and VEGFR-2 inhibitors.

4.2.1.
In vitro anti-proliferative activity MTT assay protocol 32 . This method was applied in accordance with the comprehensive description in Supplementary data.

In vitro VEGFR-2 kinase assay
The assay was applied by ELISA kits in accordance with the comprehensive description in 18,33 as described in Supplementary data.

Flow cytometry analysis for cell cycle
This assay was applied using propidium iodide (PI) staining in accordance with the comprehensive description in Supplementary data 34,35 4.2.4. Flow cytometry analysis for apoptosis Apoptotic effect was applied in accordance with the comprehensive description in Supplementary data 36,37

Western blot analysis
The western blot technique was applied in accordance with the comprehensive description in Supplementary data [38][39][40] 4.3. In silico studies 4.3.1. Docking studies Docking studies were applied using MOE 2014 [41][42][43] in accordance with the comprehensive description in were carried out against VEEGFR-2 (PDB ID: 4ASD, resolution: 2.03 Å) as described in Supplementary data.

Pharmacokinetic profiling study
This study was applied using Discover studio 4 in accordance with the comprehensive description in Supplementary data 44 . Table 9. Radar charts for prediction of oral bioavailability profile of compounds 12d, 12i, and 12 l Compounds 12d 12i 12l Radar images Table 10. Colours, yields, and meting points of the target compounds