2-Arylquinolines as novel anticancer agents with dual EGFR/FAK kinase inhibitory activity: synthesis, biological evaluation, and molecular modelling insights

Abstract In this study, different assortments of 2-arylquinolines and 2,6-diarylquinolines have been developed. Recently, we have developed a new series of 6,7-dimethoxy-4-alkoxy-2-arylquinolines as Topoisomerase I (TOP1) inhibitors with potent anticancer activity. Utilising the SAR outputs from this study, we tried to enhance anticancer and TOP1 inhibitory activities. Though target quinolines demonstrated potent antiproliferative effect, specifically against colorectal cancer DLD-1 and HCT-116, they showed weak TOP1 inhibition which may be attributable to their non-coplanarity. Thereafter, screening against kinase panel revealed their dual inhibitory activity against EGFR and FAK. Quinolines 6f, 6h, 6i, and 20f were the most potent EGFR inhibitors (IC50s = 25.39, 20.15, 22.36, and 24.81 nM, respectively). Meanwhile, quinolines 6f, 6h, 6i, 16d, and 20f exerted the best FAK inhibition (IC50s = 22.68, 14.25, 18.36, 17.36, and 15.36 nM, respectively). Finally, molecular modelling was employed to justify the promising EGFR/FAK inhibition. The study outcomes afforded the first reported quinolines with potent EGFR/FAK dual inhibition.


Introduction
Cancer is a major health obstacle in the world threating the life of millions of people annually 1,2 . The universal burden of cancer has been expected to rise to 21.6 million in 2023 compared to 14.1 million in 2012 and was predicted to increase to 28.4 million in 2040 with 47% increment relative to 2020 [3][4][5] . In 2020, 19.3 million new cancer cases have been diagnosed and 10 million cancer patients passed away. As established by WHO in 2019, cancer was estimated to be the first or second dominant cause of death for the ages <70 years in 112 countries, while it was projected to be the third or fourth death cause in 23 countries. In general, the incidence and mortality of cancer are growing rapidly worldwide 4,6 . Subsequently, enormous attempts have been implemented to develop potent anticancer drugs through investigations of diverse scaffolds against numerous potential chemotherapeutic targets [7][8][9] .
Epidermal growth factor receptor (EGFR) is a member of tyrosine kinase family in which the endogenous ligand binds to the extracellular domain leading to conformational changes and dimerisation of EGFR resulting in its activation which subsequently stimulates its intrinsic intracellular protein-tyrosine kinase activity 10 . EGFR is over-expressed in many solid tumours and is related to cancer cell proliferation, angiogenesis and metastasis, so it has a critical role in cancer growth. Therefore, EGFR has been validated as an efficient target for anticancer drug discovery. In the last two decades, different 4-anilinoquinazoline-based EFGR inhibitors, such as Gefitinib, Erlotinib, Afatinib, and Dacomitinib ( Figure 1), have been FDA-approved for clinical use in treatment of non-small cell lung cancer 11,12 .
Focal adhesion kinase (FAK) is a cytoplasmic non-receptor tyrosine kinase involved in signal transductions from cell adhesions to regulate different biological cell functions including survival and cell migration 13,14 . Also, it is activated and overexpressed in diverse cancer types controlling cancer proliferation, survival and metastasis. Thus, FAK has been identified as a promising druggable target for targeted cancer therapy. Currently, several FAK inhibitors, such as 2,4-diaminopyridine derivative GSK2256098 and 2,4-diaminopyrimidine derivative Defactinib (Figure 1), are currently being evaluated in clinical trials for cancer treatment, in addition to the 2,4-diaminopyrimidine derivative TAE-226 ( Figure  1) which displayed potent antitumor impact in different cancer types in vivo and in vitro and usually used as a reference drug 7,15,16 . Noteworthy, it was established that the most affected colorectal cancer expressed high levels of EGFR and FAK that particularly correlated with tumour angiogenesis, cancer aggressiveness and poor prognosis 17,18 .
Thus, dual EGFR/FAK inhibition mechanism is an efficient strategy to fight cancer that could be attributed to a non-overlapping downstream signalling/inhibition 19,20 . For example, the kinase inhibitor APG-2449 ( Figure 1) was reported to improve the antitumor effect of Ibrutinib via EGFR/FAK inhibition mechanism in oesophageal squamous cell carcinoma 19 . Also, combined EGFR/ FAK inhibition caused higher radiosensitization than either approach alone 21 . Interestingly, few studies have succeeded to develop dual EGFR/FAK small molecule inhibitors. In 2020, Ai et al. has exploited a fragment-based drug design approach to identify novel series of 2,4-diaminopyrimidines as potent dual EGFR/FAK inhibitors with good in vitro and in vivo antitumor effects 20 .
Quinoline is an outstanding planar heterocyclic motif playing a distinctive role in anticancer drug discovery. So far, assortments of quinoline-based small molecules have been developed and investigated against numerous biological targets for cancer treatment displaying exquisite outcomes [22][23][24][25] . It is worth stressing that plenty of quinoline derivatives provoked their anticancer impact through different mechanisms of action, such as inhibition of DNA repair, tubulin polymerisation, and inhibition of various enzymes implicated in critical cancer cell proliferation prominently kinases enzymes (EGFR, VEGFR, pim-1 kinase, c-Met factor, and PI3K) which stood out as one of the most significant targets implemented in cancer therapy due to their functions in cellular signal transduction [26][27][28][29][30] . Of special interest, Pelitinib (EKB-569, Figure 2) is a 4-anilinoquinoline derivative which is a potent irreversible inhibitor of EGFR in the clinical trials as an anticancer candidate 24 . In addition, several 4-aminoquinoline derivatives, such as compounds I-III (Figure 2), were reported as promising EGFR inhibitors endowed with effective anticancer activities. Accordingly, quinoline stands out as a significant privileged scaffold in anticancer drug discovery to develop many efficient kinases inhibitors 24,[31][32][33][34][35] .
Recently, we have developed a new series of 6,7-dimethoxy-4alkoxy-2-arylquinolines as potential Topoisomerase I (TOP1) inhibitors 36 . The TOP1-mediated DNA cleavage assay was utilised to assess the ability of the reported compounds to stabilise TOP1-DNA cleavage complexes (TOP1ccs). The assay outcomes revealed a moderate TOP1 inhibitory activity of compounds IVa,b ( Figure  2). Interestingly, the developed quinolines showed outstanding anti-proliferative profile upon evaluation at the Developmental Therapeutics Program (DTP) of the NCI-USA. Noteworthy, the weightiness of incorporation of p-substituted phenyl at C-2, as well as propyl linker at C-4 of the quinoline scaffold, was highlighted by the SAR study.
As a continuation for our previous study 36 novel five sets of 4propoxy-2-arylquinolines (6a-o, 8a,b, 10a,b, 12a-d, 16a-d), and 4-propoxy-2,6-diarylquinolines (20a-f) are herein designed and synthesised, exploiting the deduced SARs from the previous study, with the aim to afford more potent anticancer TOP1 inhibitors ( Figure 2). Different structural modification strategies were adopted seeking to enhance both anticancer and TOP1 inhibitory activities of the lead compounds IVa,b ( Figure 3).
First, diverse secondary alicyclic and aromatic amines, in addition to different primary amines were appended to the propyl linker to illuminate the influence of these moieties on the activity. Also, we introduced the electron donating methyl group as well as electron withdrawing Cl and CF 3 for more elucidation of the electronic impact of p-substituent of the 2-phenyl. Likewise, the heterocycles: 2-furyl and 2-thienyl were appended to para position of the C2-phenyl moiety to explore their electronic and size impact on the desired activity. Afterward, the 6,7-dimethoxy groups were removed in some synthesised analogs to confirm their significance. Moreover, the ring system was extended via fusion of the quinoline motif with 1,3-dioxolane in attempt to enhance the planar structure requisite for DNA intercalation which may potentiate both anticancer and TOP1 poisoning effects. Finally, a structural extension approach was utilised via grafting the HBA-bearing 2-furyl and 4-methoxyphenyl moieties at C6 of quinoline scaffold, hoping to enhance the hydrophobic interactions ( Figure 3). The designed 4-propoxy-2-arylquinolines were prepared employing different synthetic procedures, and then investigated for their anticancer and TOP1 inhibitory activities.
Although the target quinolines demonstrated potent antiproliferative effect against different cancer cell lines, they showed no or weak TOP1 poisoning influence. Accordingly, the promising anticancer activity prompted us to search for the plausible molecular mechanism for herein reported quinolines.

Chemistry
The synthetic routes adopted for the synthesis of the target 4-propoxy-2-arylquinolines (6a-o, 8a,b, 10a,b, 12a-d, 16a-d) and 4propoxy-2,6-diarylquinolines 20a-f are illustrated in Schemes 1-3. Regarding Scheme 1, the benzamides 3a-c were prepared by reacting 4,5-dimethoxy-2 0 -aminoacetophenone 1 with p-substituted benzoyl derivatives 2a-c in dry THF and Et 3 N. The latter were cyclized in refluxing dioxane and NaOH to afford the corresponding quinolones 4a-c in excellent yields which then were subjected to O-alkylation with 1-bromo-3-chloropropane using our previously confirmed procedure to yield the respective 4-propoxy key intermediates 5a-c 36 . Finally, these key intermediates 5a-c were converted to the target 4-prpoxy-2-arylquinolines 6a-o in good to excellent yields (70-80%) through nucleophilic substitution with the appropriate amine in dry DMF and potassium carbonate anhydrous using catalytic amount of potassium iodide at 90 C.
In Scheme 2(A), the bromo analog of 4-propoxy-N-methylpiperazine-2-arylquinoline 7 was obtained based on the same pathway for 6a-o using p-bromobenzoyl chloride. Then, this bromo analog 7 has been transformed into the target heterocyclic derivatives 8a,b in good yields (70-72%) under Suzuki cross coupling condition through its reaction with the appropriate boronic acid derivative in dioxane using tetrakis catalyst and 2 M sodium carbonate under N 2 at 90 C. In Scheme 2(B,C), syntheses of the target demethoxylated analogs 10a,b and 1,3-dioxolo analogs 12a-d have been accomplished in good to excellent yields (72-87%) adopting the same synthetic routes described for 6a-o using the respective starting compounds, intermediates, and amines.
Concerning Scheme 3, initiated from 5-bromo-2 0 -aminoacetophenone 13, the target 6-bromo analogs 16a-d have been afforded in good to excellent yields (78-87%) applying the synthetic routes utilised for 6a-o. Moreover, 5-bromo-2 0 -aminoacetophenone 13 was transferred to the 6-aryl derivatives 17a,b via Suzuki coupling reaction by its reaction with the appropriate boronic acid derivative in dioxane/water in the presence of tetrakis catalyst and potassium carbonate under N 2 at 100 C. Subsequently, these 6-aryl derivatives 17a,b have been similarly converted to the target 4-propoxy-2,6-diarylquinolnes 20a-f in good to excellent yields (74-89%) exploiting the same experimental pathways adopted for 6a-o.
Regarding breast cancer cell lines, GI% ranged from 75.34 to 84.76% and 71.97 to 87.36% for MDMBA-231 and MCF-7 cell lines, respectively. While five compounds showed good inhibitory activity towards MDMBA-231, ten compounds possessed good inhibition to the growth of MCF-7 cancer cell line. For HeLa cell line, seven compounds had good GI% from 65.73 to 90.78%. Based on the preceding screening results, it was revealed that colorectal cancer cell lines (DLD-1 and HCT-116) were the most sensitive to the tested compounds, therefore they were selected for further antiproliferative assay at six doses levels ( Table 1).
Concerning HCT-116 cell line, it was found that compounds 6f, 6h, 16d, and 20f established twice inhibitory activity compared to Gefitinib (IC 50  It is noteworthy that appending p-CF 3 to the 2-phenyl enhanced the antiproliferative activity compared to p-Cl and p-CH 3 , except for the cyclohexylamine analogs 6b, 6g, and 6l in which the p-CH 3 substituent is preferred for antiproliferative activity. Remarkably, the incorporation of p-(2-furyl) to the 2-phenyl abolished the growth inhibitory activity, while the introduction of p-(2-thienyl) increased the activity compared to p-Cl and p-CH 3 analogs. In the context of impact of amine substituents on the antiproliferative activity, it was proved that 4-amino-N-methylpiperidine is preferred for anticancer activity, then piperazine, tetrahydrofurfurylamine and cyclohexylamine. Notably, grafting imidazole along with 6,7-dimethoxy substituents abolished the antiproliferative activity for all derivatives.
Besides, the removal of 6,7-dimethoxy groups from the quinoline scaffold decreased or abolished activity. While the fusion of 1,3-dioxolo to the quinoline scaffold along with p-CF 3 substitution on phenyl dramatically potentiated the anticancer activity of the imidazole derivatives, the replacement of imidazole with morpholine decreased the activity of p-CF 3 analogs and markedly increased the activity of p-Cl counterparts. On the other hand, appending of electron withdrawing (Br) to position 6 of quinoline parallel with p-Cl substitution on 2-phenyl greatly enhanced the activity of the imidazole derivatives while, p-F substitution on the 2-phenyl extremely decreased the anticancer activity of imidazole derivatives. Also, the replacement of imidazole with morpholine almost had no impact on the p-Cl analogs, but tremendously elevated the anticancer activity of p-F derivatives.
Furthermore, the incorporation of 4-methoxyphenyl or 2-furyl to position 6 of quinoline along with p-Cl or F on the 2-phenyl enhanced the activity of the imidazole derivatives compared to the dimethoxy analogs, but the 2-furyl derivatives exhibited better activity. Moreover, the replacement of imidazole with morpholine elevated the antiproliferative activity of p-Cl analogs while, diminished the activity of p-F derivatives.
Finally, the deduced structure activity relationships indicated that the substitution pattern on positions 6 and 7 of quinoline and position 4 of the 2-phenyl, in addition to the amine substitution on the 4-propoxy linker are crucial elements for the anticancer activity. In general, incorporation of dimethoxy groups at positions 6 and 7 of quinoline along with p-CF 3 at 2-phenyl and 4-amino-N-methylpiperidine, piperazine or tetrahydrofurfurylamine on the 4-propoxy linker, in addition to grafting of electron withdrawing (Br) or 2-furyl at position 6 of quinoline along with p-F or Cl on the 2-phenyl and morpholine on the propoxy linker resulted in the most potent antiproliferative agents in this study.

Annexin V-FITC/propidium iodide apoptosis assay (AV/PI)
The apoptotic impact of the most potent antiproliferative agents 6f, 6h, 6i, 16d, and 20f on DLD-1 colorectal cancer cell line was investigated exploiting AV/PI dual staining assay. The assay outcomes proved that the tested compounds elicited apoptosis of such cell line as indicated by significant rise in the total percentage of AV positive apoptotic DLD-1 cells compared to the control ( Figure 4). Compound 6h increased the total percentage of apoptotic cells from 7% for the control to 90.33%. Also, compounds 6f, 6i, 16d, and 20f exerted potential apoptotic effect elevating the total percentage of apoptotic cells to 84.33, 71.66, 51.66, and 80.66%, respectively.

Topoisomerase I-mediated DNA cleavage assay
TOP1 poisoning activity of all target compounds has been estimated utilising TOP1-mediated DNA cleavage assay that determines the TOP1 poisoning activity relative to 1 mM Camptothecin (CPT) 37 . The tested compounds were incubated at 0.1, 1, 10, and 100 mM with recombinant human TOP1 enzyme and a 3 0 -[ 32 P]labeled 117-bp DNA oligonucleotide 38 . TOP1 poisoning agents specifically trap TOP1ccs leading to their stabilisation and DNA cleavage. The drug-induced stabilised TOP1ccs are visualised by gel electrophoresis demonstrating specific DNA cleavage patterns. Then a semiquantitative scoring system by visual comparison between lanes induced by the target compounds and 1 mM CPT was used to score the compounds (Table 3, see Supplementary Data for the results of gel electrophoresis) [39][40][41] .
Compounds 6a-e, 6h, 6l, 6m, 12a, and 12b exhibited weak TOP1 poisoning effect (-/þ) displaying DNA cleavage activity equal to 0-25% of the activity of 1 mM CPT, while compound 16c demonstrating activity (þ) equals to 25-50% of the activity of 1 mM CPT. The rest of compounds possessed no cleavage activity. Despite the target compounds showing promising antiproliferative activity against cancer cell lines, their TOP1 poisoning activities were not encouraging for further development as potent TOP1 inhibitors. Accordingly, the synthesised compounds have been evaluated for the plausible mechanism by which they provoked the antiproliferative activity.

Kinase profiling
The non-significant TOP1 poisoning activities of target compounds obtained from Topoisomerase I-mediated DNA cleavage assay motivated us to search for the plausible molecular mechanism for herein reported 4-propoxy-2-arylquinolines.
The potential inhibitory activity of the target 4-propoxy-2-aryl- diarylquinolines (20a-f) was explored against a panel of nine kinases representing different signalling pathways; EGFR, FAK, FRK, IGF-1R, BTK, c-Src, VEGFR-1, VEGFR-2 and HER-2 (see Supplementary Data, Table S1). The half maximal inhibitory concentration (IC 50 ) values were calculated for each kinase and presented in Table S1 and Table 4. Strikingly, the screening outcomes revealed that the investigated quinolines exhibited promising dual inhibitory effect towards EGFR and FAK kinases.
It is worth stressing that 4-propoxy-2-arylquinolines 6f, 6h, 6i, and 20f emerged not only as the most potent dual EGFR/FAK inhibitors in this study, but also as the most efficient anti-proliferative agents towards the examined colorectal cancer (DLD-1 and HCT-116) cell lines.

2.5.
In silico molecular docking 2.5.1. Docking into EGFR binding site The molecular docking approach was utilised to investigate the potential binding of herein reported 4-propoxy-2-arylquinolines to EGFR binding site (PDB: 1M17). The docking procedure was validated through the redocking of the co-crystalised ligand. The correct pose was predicted accurately with RMSD of 1.498 between the docked and co-crystalised ligand using DockRMSD server ( Figure 5(a)) 42 . In addition, docking was able to maintain hydrogen bonding seen in the co-crystalised ligand with NH of M769 (2.7 Å) and with the NH of G772 (3.2 Å). Furthermore, hydrophobic interactions with residues in the active site were also maintained. These include interactions between K721 and ethyne benzene moiety, and L694, L768, and L820 with hydrophobic part of the quinazoline ring ( Figure 5(b)).
Docking scores of the tested compounds with EGFR are shown in Table 5. The docking score of the co-crystalised ligand was À7.2 kcal/mol, whereas all the tested quinolines have shown better docking scores than that of the co-crystalised ligand (À7.9 to À9.7 kcal/mol). Best docking scores were seen with quinolines form the series 20. Compound 20c showed the best binding energy to EGFR with docking score of À9.7 kcal/mol and its binding pose is shown in Figure 5(c). The compound has formed 2 hydrogen bonds between imidazole ring and both K721 (3.5 Å) and T766 (3.1 Å). Also, several hydrophobic interactions have been seen, which included interactions between quinoline ring and side chains of V702 and T830. The phenyl ring at position 2 formed hydrophobic interactions with L694 and L820 which are also common with the co-crystalised ligand. In addition, p-p stacking between the ring at quinoline position 6 and F699.
Another compound from this series is compound 20e which was selected as a representative example because it has shown good biological results with both EGFR and FAK. The docking pose of this compound is shown in Figure 5(d) showing a similar docking pose to 20c. The same hydrogen bonds with K721 (3.5 Å) and T766 (3.1 Å) were maintained as well as hydrophobic interactions with V702 and T830 as well as with L694 and L820. The p-p stacking was also seen between F699 and the furan ring of 20e. This compound, in complex with EGFR, was subjected to further investigation using molecular dynamics to study the stability of its complex with EGFR as will be discussed later.

Docking into FAK binding site
Potential binding of target quinolines to FAK was also investigated using docking studies (PDB: 2JKM). Initially, the docking procedure was validated through the redocking of the co-crystalised ligand (AZW592). The docking searching algorithm was able to correctly predict the binding pose with acceptable accuracy with RMSD of 1.318 between the docked and co-crystalised ligands as predicted by DockRMSD server (Figure 6(a)) 42 . The docked structure was able to maintain same hydrogen bonds that are in the crystal structure including those between the sulphamoyl moiety oxygen and the terminal amino group of K454 (Å) and the hydrogen bond with the a-carbonyl group of C502. In addition, several hydrophobic interactions have been also seen with residues in the active site including I428, V436, V484, L501, and L553 ( Figure 6(b)).
Next, target synthesised quinolines were docked in the active site of the FAK after their preparation. The docking scores of tested compounds are shown in Table 5. The docking score of co-crystalised ligand (AZW592) was found to be À8.5 kcal/mol. Some of tested compounds have shown scores that are comparable to the co-crystalised ligand. Best results were seen with 12 and 20 and some of the 16 series. Docking poses of these compounds were similar with most of the docked compounds as can be seen in Figure 6(c) which shows docking pose of some compounds from these series.
Docking pose of compound 20e which was chosen as representative example is shown in Figure 6(d). The compound was able to form 3 hydrogen bonds with N551 (3.1 Å), D564 (3.1 Å), and E506 (3.3 Å). In addition, several hydrophobic interactions were also seen, such as the hydrophobic interaction between quinoline ring and L553 and between phenyl ring at position 2 and I428 which are common with the co-crystallized ligand. This pose was selected for further investigation of the complex stability using molecular dynamics study.

Molecular dynamics (MD) simulation
The stability of compound 20e complexes with both EGFR and FAK was investigated using 100 ns molecular dynamics studies. With each complex, the results were compared with the 0 a Scoring: 0: no activity; À/þ: 0-25% 1 mM CPT; þ: 25-50% 1 mM CPT. co-crystalised ligand complex as a control and with the apoprotein (the protein alone with no ligands). The missing loops in both targets were built using Swiss-Model server 43 before starting the dynamics to ensure correct results. All complexes were equilibrated under NVT then NPT conditions for 1 ns each and the analysis was done on the production run.
Analysis of the production runs trajectories for 20e in the active site of EGFR demonstrated stability comparable to the cocrystalised ligand. Radius of gyration (R g ) is a measure of the compactness of the complexes. Stable R g suggested the stability of the protein or complex under investigation. Figure 7(a) shows a plot of R g of 20e, co-crystalised ligand and apoprotein. The average R g was found to be 2.01 ± 0.02, 2.03 ± 0.02, and 2.02 ± 0.01 nm for apoprotein, control, and 20e, respectively. In addition, Root mean square fluctuation (RMSF) of protein residue (Figure 7(b)) for all the three complexes showing similar patterns. The average RMSF for co-crystalised ligand and 20e was found to be 0.19 and 0.18 nm, respectively which is slightly higher than that of the apoprotein (0.15 nm). Although root mean square deviation (RMSD) of ligand heavy atoms for 20e is slightly higher than that of the cocrystalised ligand (Figure 7(c)), the value is <1 nm for most of the trajectory. This value cannot be calculated for the apoprotein as it has no ligand in the system. Finally, the number of hydrogen bonds between ligands and protein are shown in Figure 7(d), which showed that 20e formed extra hydrogen bonds during at least 50% of the production run time. These results suggested that 20e complex with EGFR is at least of comparable stability when compared to the complex with the co-crystalised ligand; Erlotinib.
Complexes of FAK with 20e, its co-crystalised ligand and the apoprotein showed similar pattern to that of the EGFR ( Figure  8). This includes the radius of gyration (R g ), which showed an average of 2.00 ± 0.01, 1.99 ± 0.01, and 1.96 ± 0.01 nm for apoprotein, co-crystalised ligand, and 20e, respectively ( Figure  8(a)). Also, RMSF of protein residues was found to follow similar patterns for all the three studied systems (Figure 8(b)). The average RMSF for the three systems was found to be 0.12 ± 0.07, 0.13 ± 0.09, and 0.12 ± 0.07 nm for apoprotein, cocrystalised ligand, and 20e, respectively. In addition, plotting of RMSD of ligand heavy atoms (Figure 8(c)) showed minimal fluctuation for both with and average RMSD of 0.19 ± 0.05 and  0.57 ± 0.16 nm for co-crystalised ligand and 20e, respectively. Although, the value for 20e is higher but it is within acceptable range (<1 nm). Finally, plotting of hydrogen bonds between ligands and target protein (Figure 8(d)) showed that the number of hydrogen bonds is higher in case of the co-crystalised ligand compared to 20e. Being said, 20e was still able to maintain an average of 1.93 ± 0.45 hydrogen bonds during the 100 ns production run.  These results collectively suggested the stability of compound 20e complexes with both EGFR and FAK compared to the corresponding co-crystalised ligands in each target. This in general supported the dual mechanism similar to the enzymatic inhibition data.

General procedure for the synthesis of benzamides (3a-c)
In dry THF (8 ml) and Et 3 N (2 ml), 1-(2-amino-4,5-dimethoxyphenyl)ethan-1-one 1 (0.976 g, 5 mmol) was dissolved and cooled in ice bath. Then, a solution of the respective p-substituted benzoyl chloride 2a-c (5.1 mmol) in dry THF (2 ml) was added dropwise while cooling in ice bath. The reaction mixture was stirred in ice bath for 30 min and then overnight at room temperature. After that, reaction mixture was poured into ice/water and the resulting solid was filtered off and washed excessively with water and methanol to afford the corresponding 3a-c.
The spectral characterisation of the benzamides 3a,b were reported in our previous study 36 .

General procedures for synthesis of the quinolones (4a-c)
Under N 2 atmosphere, the benzamides 3a-c and three equivalents NaOH were refluxed in dry dioxane at 110 C for 4 h and then cooled to room temperature. Small amount of water and excess amount of hexane were added to the reaction mixture. The resulting mixture was subjected to sonication for 2 min and then neutralised using 1 M HCl. The separated solid was filtered off and washed excessively with water to give the corresponding quinolones 4a-c.
The spectral data of the quinolones 4a,b have been reported in our previous study 36  4.1.5. General procedure for synthesis of the key intermediates (5a-c) The quinolones 4a-c (3 mmol), KI (0.498 g, 3 mmol) and KOH (1.009 g, 18 mmol) were stirred for 2 h in dry DMF (30 ml) and then 1-bromo-3-chloropropane (2.361 g, 15 mmol) was added to the mixture and stirred at room temperature for 24 h. Thereafter, the mixture was poured into ice/water and the separated solid was filtered off and washed with water then hexane to furnish the key intermediates 5a-c which were used without further purification.
The spectral data of 5a,b have been reported in our previous study 36

In vitro antiproliferative assay
The antiproliferative assay against five cancer cell lines representing three different tumour subpanels, including colorectal (DLD-1, HCT-116), breast (MDMBA-231, MCF-7), and cervical (HeLa) cell lines was conducted using MTT assay. Cells were seeded at 1 Â 10 4 cells/well and cultured overnight in a 96-well plate. The cells were treated with either 10 m of tested quinolines 6a-o, 8a,b, 10a,b, 12a-d, 16a-d, and 20a-f, or DMSO as a negative control. After 24 h, the cells washed with phosphate buffered saline (PBS, Invitrogen Gibco) and incubated with 20 ml of MTT solution (2 mg/ml) for 4 h at 37 C. Then, 150 ml DMSO was used to solubilise MTT formazan crystals. Finally, the plates were shaken, and the optical density was determined at 570 nm using ELISA plate reader. At least, three independent experiments were performed. Percentage of growth inhibition was determined as (1-[OD of treated cells/OD of control cells]). On the other hand, using the MTT assay, we tested the effect of different concentrations (0.5, 1, 10, 30, 50, and 100 mM)) of the synthesised compounds on colorectal cancer cell lines (DLD-1 and HCT-116), using DMSO as a negative control, whereas Gefitinib and TAE226 were used as positive controls. The IC 50 values were calculated using Prism v0.8 software (GraphPad Software Inc., La Jolla, CA).

Apoptosis assay
The apoptotic effect of the most potent antiproliferative agents 6f, 6h, 6i, 16d, and 20f on DLD-1 colorectal cancer cell line was investigated using the annexin V/propidium iodide (AV/PI) staining kit (BioLegend, San Diego, CA, USA) according to the manufacturer's instructions. DLD-1 cells were treated with 3 mM of the most potent antiproliferative compounds 6f, 6h, 6i, 16d, and 20f or DMSO as a negative control then incubated for 24 h. Flow cytometric analysis was conducted using FACS Calibur flow cytometer (BD Biosciences, San Jose, CA, USA).

Topoisomerase I-mediated DNA cleavage assay
A 3 0 -[ 32 P]-labeled 117-bp DNA substrate oligonucleotide was prepared as described previously 39 . Radiolabeled DNA was incubated with recombinant human TOP1 in 20 mL reaction buffer (10 mmol/ L Tris-HCl, pH 7.5, 50 mmol/L KCl, 5 mmol/L MgCl2, 0.1 mmol/L EDTA, and 15 mg/mL BSA) at 30 C for 20 min in the presence of the indicated drug concentrations. Reactions were terminated by adding SDS (0.5% final concentration) followed by the addition of two volumes of loading dye (80% formamide, 10 mmol/L sodium hydroxide, 1 mmol/L sodium EDTA, 0.1% xylene cyanol, and 0.1% bromophenol blue). Aliquots of reaction mixtures were subjected to 20% denaturing PAGE. Gels were dried and visualised by using PhosphorImager and Image Quant software (Molecular Dynamics).

In silico molecular docking
Ligands were converted to 3D structures and minimised using Avogadro 48 . Ligands were prepared and converted to pdbqt files using PyRx 49 . Protein targets were downloaded from the protein data bank under the codes 1M17 for EGFR and 2JKM for FAK 50 . Co-crystalised ligands were extracted form pdb files and prepared similar to the tested ligands. Docking was done using Autodock Vina 51 in a grid box of 25 3 Å 3 centred on the co-crystalised ligand with exhaustiveness of 16. Visualisation and 3D images were prepared using PyMol 52 .

Molecular dynamics (MD) simulation
Missing loops in the 3D structures of the protein were constructed using Swiss-Model 43 before starting molecular dynamics steps. All atom molecular dynamics simulations were performed using GROMACS 2020.3 53 for the selected protein-ligand complexes as reported earlier 54 . In brief, SwissParam server 55 was used for ligands parameterisation while Charmm36 all-atom force field 56 was used to generate topology files for the protein. Ligand coordinates obtained from docking studies were used to build complexes which were boxed in a dodecahedron box and then solvated with TIP3P 57 explicit water. Systems were neutralised by the addition of required number of Na þ or Cl À ions. Systems energy was minimised with a maximum force of 1000 kJ mol À1 nm À1 using steepest descent algorithm. Equilibration using NVT and NPT ensembles for 1 ns each was done afterward then production run was done for 50 ns. Temperature was kept at 300 K using the V-rescale algorithm 58 while pressure was controlled using the Parrinello-Rahman barostat 59 as required. The LINear Constraint Solver (LINCS) algorithm 60 and Particle mesh Ewald (PME) method 61 were used for bond's length constraints and long-range electrostatics calculations, respectively. Two femtosecond timestep was used for all simulations. Van der Waals distance cut-off (rvdw) was set to 1.2 nm. Trajectories from the production run were used for analysis using trajconv after correction of periodic boundary condition (PBC).