Anti-proliferation evaluation of new derivatives of indole-6-carboxylate ester as receptor tyrosine kinase inhibitors

Abstract

Aim: The main goal was to create two new groups of indole derivatives, hydrazine-1-carbothioamide (4a and 4b) and oxadiazole (5, and 6a–e) that target EGFR (4a, 4b, 5) or VEGFR-2 (6a–e). Materials & methods: The new derivatives were characterized using various spectroscopic techniques. Docking studies were used to investigate the binding patterns to EGFR/VEGFR-2, and the anti-proliferative properties were tested in vitro. Results: Compounds 4a (targeting EGFR) and 6c (targeting VEGFR-2) were the most effective cytotoxic agents, arresting cancer cells in the G2/M phase and inducing the extrinsic apoptosis pathway. Conclusion: The results of this study show that compounds 4a and 6c are promising cytotoxic compounds that inhibit the tyrosine kinase activity of EGFR and VEGFR-2, respectively.

Plain language summary

Summary points.

• Two hydrazine-1-carbothioamide derivatives (4a and 4b), and oxadiazole-2-thione compound (5) were designed and synthesized to mimic the pharmacophoric properties of EGFR tyrosine kinase inhibitors.

• Several oxadiazole derivatives (6a–e) were designed and synthesized to mimic the pharmacophoric properties of VEGFR-2 tyrosine kinase inhibitors.

• Molecular docking revealed that compounds 4a and 6c were the best fits within the active sites of EGFR and VEGFR-2 tyrosine kinases, respectively.

• IC₅₀ values against HepG2, HCT-116 and A549 cancer cell lines were calculated.

• Compounds 4a and 6c were noted to be the most potent anti-proliferative compounds against cancer cells.

• Compound 4a with an unsubstituted phenyl moiety (R=H) exhibited the highest EGFR enzyme inhibitory activity. Compound 6c with a chloro group at the 4-position of the aromatic ring showed the highest VEGFR-2 enzyme inhibitory activity.

1. Background

Cancer is the second leading cause of death in developed countries after cardiovascular diseases. Cancer cells are defined by their ability to proliferate, invade and metastasize (spread to other bodily locations) [1]. Surgery, chemotherapy, radiation therapy, immunotherapy, monoclonal antibody treatment and other approaches can all be used to treat cancer. The location and grade of the tumor, stage of the disease and the patient's overall health all influence the decision of an appropriate treatment [2].

Since it is now recognized that neither radiation therapy, surgery or the combination of the two can effectively control metastatic cancer, chemotherapy has dominated cancer treatment efforts. Chemotherapeutics are most effective against rapidly dividing tumor cells, but they can also affect normal rapidly dividing cells, such as bone marrow cells and various types of epithelial cells [3,4].

The tyrosine kinase EGFR is a major mediator in cell signaling pathways promoting cell proliferation, apoptosis, angiogenesis and metastasis. Aberrant EGFR activation can cause uncontrolled cell division, ultimately developing a tumor [5]. On the other hand, the VEGF binds the tyrosine kinase vascular endothelial growth factor receptor-2 (VEGFR-2) to promote endothelial proliferation. VEGFR-2 activation promotes angiogenesis and raises vascular permeability [6,7]. In addition, VEGFR-2 is an essential regulator of tumor-induced angiogenesis [8].

Targeting oncogenic protein kinases, such as the dysregulated EGFR and VEGFR-2, has been shown to be an effective anticancer therapy. As of March 2019, the US FDA had approved 53 small-molecule kinase inhibitors to treat various types of cancer [9–11]. Most authorized kinase inhibitors target EGFR and VEGFR-2 [12–16]. EGFR inhibitors are frequently used to treat non-small-cell lung cancer (NSCLC) [17], while VEGFR-2 inhibitors are frequently used to treat thyroid and kidney malignancies [18]. However, existing EGFR and VEGFR-2 inhibitors cause serious side effects and toxicity in patients [19]. These undesirable side effects could be attributed to inhibition of either EGFR/VEGFR-2 and/or drugs off-targets [20]. Furthermore, continued use of these inhibitors may lead to the development of chemoresistance; for example, the EGFR T790M mutation renders EGFR inhibitors ineffective [21]. Thus, developing innovative treatments with fewer side effects and greater selectivity for cancer cells is a massive task.

The indole skeleton, present in a variety of active ingredients and natural products, is one of the most well-known structures with potent anticancer activity [22–24]. Several indole-containing compounds are already in clinical use as anticancer medications including vincristine, vinblastine, vindesine, mitraphylline, cediranib and panobinostat [25–27], with a sizable number of these indole-based compounds having tyrosine kinase (TK) inhibitory activity [28–30]. Derivatives of 2-oxoindoline-3-ylidene hydrazine-1-carbothioamide were found to be cytotoxic to different cancer cell lines [31]. Another study revealed that sulfonamide-hydrazine-1-carbothioamide derivatives were effective anticancer agents against MCF-7 breast cancer cells [32].

The frequently used pharmacophore 1,3,4-oxadiazole has been the focus of much research because of its metabolic profile and capacity to form hydrogen bonds with receptor sites. Because of the azole (-N=C–O-) group in the oxadiazole nucleus, which increases lipophilicity and affects the drug's ability to diffuse across membranes to reach its target, this pharmacophore is an excellent amide and ester bioisoster that can significantly contribute to pharmacokinetic properties. It has been reported that 1,3,4-oxadiazole derivatives target kinases, thereby increasing their anticancer activity [33].

Based on our previous work and others demonstrating that numerous indole derivatives target tyrosine kinases, and assisted by computer-aided drug design tools such as pharmacophore modeling and structure–activity relationship, the main goal of the current work was to create two new groups of indole derivatives, hydrazine-1-carbothioamide derivatives (4a and 4b) and oxadiazole derivatives (5 and 6a–e) that target the EGFR TK (4a–b and 5) or VEGFR-2 TK (6a–e). These compounds were characterized using a variety of chemical and spectroscopic techniques, including IR, ¹HNMR, ¹³CNMR and HRMS (electrospray ionization [ESI]). In vitro, the new derivatives were tested as anti-proliferative agents against A549, HCT-116 and HepG2 cancer cell lines.

2. Experimental

2.1. Materials & methods

All reagents were purchased from Sigma-Aldrich (Hamburg, Germany) and Hyper Chem (Hangzhou, China) and utilized without further purification. FT-IR spectra were obtained using a Shimadzu spectrophotometer (Kyoto, Japan) on a KBr disk with v = cm^-1. The Inova model Ultrashield (Bruker, MA, USA) at 400 MHz and 100 MHz was used to measure the ¹HNMR and ¹³CNMR spectra, respectively, with TMS as an internal standard. The chemical shift was expressed as (δ = ppm), and DMSO-d₆ was utilized as the solvent. The X500 QTOF Mass Spectrometer was used to determine mass spectra m/z. TLC solvent systems (chloroform: methanol of different ratios: chloroform 70%: methanol 30%; chloroform 80%: methanol 20%; and chloroform 60%: methanol 40% were used to monitor the reaction progression.

2.2. Chemistry

The intermediates and final compounds were synthesized according to Figure 1.

Figure 1.

Chemical synthesis of the new derivatives 4a–b, 5 and 6a–e.

2.2.1. Synthesis of methyl-1H-indole-6-carboxylate (1)

Compound 1 was synthesized and characterized according to reference [34]. More details can be found in Supplementary Data.

2.2.2. Synthesis of methyl-1-methyl-1H-indole-6-carboxylate (2)

A solution of compound 1 (0.19 g, 0.001 mol) and aq. KOH (0.168 g, 0.003 mol) in acetone (10 ml) were stirred at 20°C for 30°min. Methyl iodide (1.5 ml, 0.0011 mol) was added and stirred for 2h to yield compound 2. The product was purified using column chromatography (ethyl acetate/hexane 1:9) and recrystallized with 70% EtOH [35]. Golden yellow needles, yield 65%, m.p 155–157°C, R_f. = 0.62 (chloroform 70%: methanol 30%). IR (KBr) cm^-1: 3093, 3070 Ar(CH) str, 2953, 2943, 2839 (CH) str of aliph. (CH₃), 1693 of (C=O-ester) str, 1562, 1500, 1462 Ar (C=C) str, 1346 (C-N) str, 1265 (C-O) asym. str, 1188 (C-O) sym. str of ester, 771 (CH) out of plane bend of 6-substituted benzene ring. ¹HNMR (400 MHz, DMSO-d₆) δ: 8.10 (s, 1H, ArH), 7.65–7.56 (m, 3H, ArH), 6.53 (d, 1H, ArH), 3.87 (s, 6H, O-CH₃ and CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 167.69 (C=O), 136.06, 133.94, 132.18, 122.51, 120.62, 120.05, 112.15, 101.20 (Ar C), 52.25 (OCH₃), 33.11 (CH₃). (HRESI-MS) m/z: Calcd. for C₁₁H₁₁NO₂ [M+1]: 190.0790, found 190.0858.

2.2.3. Synthesis of 1-methyl-1H-indole-6-carbohydrazide (3)

Upon dissolving compound 2 (3.0 g, 0.0112 mol) in 25 ml of MeOH, excess hydrazine hydrate 80% (5.6 ml, 0.112 mol) was added. The mixture was refluxed for 9 h. A light yellowish precipitate appeared after the mixture was cooled to room temperature (RT), and it was refrigerated for an additional night. The precipitate was washed with distilled water (D.W.) and dried before being recrystallized with 70% EtOH [36]. Dark yellow needles, yield 78%, m.p 240–242°C, Rf = 0.23 (chloroform 70%: methanol 30%). IR (KBr) cm^-1: 3278 (NH) str of sec-NH. amide, 3163 (NH) str of prim. amine, 3059 (ArH) str, 2989, 2873 (CH str of CH₃), 1681 (C=O of amide) str (amide I band), 1624 (NH) bend, 1516, 1458, 1415 Ar (C=C) str, 1342 (C-N) str, 825, 671 (CH) bend of heterocyclic. ¹HNMR (400 MHz, DMSO-d₆) δ: 9.66 (s, 1H, NH), 8.00 (s, 1H, ArH), 7.58–7.52 (m, 2H, ArH), 7.48 (d, 1H, Ar-H), 4.47 (s, 2H, NH₂), 3.84 (s, 3H, CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 167.42(C=O), 136.21, 132.52, 130.52, 126.48, 120.25, 118.26, 109.59, 100.88, 33.01 (CH₃). (HRESI-MS) m/z: Calcd. for C₁₀H₁₁N₃O [M+1]: 190.0902, found 190.0962.

2.2.4. Synthesis of indole carbothioamide derivatives (4a, 4b)

In 15 ml of MeOH, compound 3 (0.3 g, 0.00118 mol) was dissolved, an equivalent amount (0.00118 mol) of each phenyl isothiocyanate derivatives was then added individually in a small volume of MeOH: phenyl isothiocyanate (0.16 g); 4-nitro phenyl isothiocyanate (0.21 g), to produce a clear mixture. The resulting mixture was subsequently stirred for 48 h at 55°C. Before filtration, the precipitate was left overnight at RT, rinsed with abs. MeOH, dried and recrystallized with 70% EtOH [37,38].

2.2.4.1. 2-(1-methyl-1H-indole-6-carbonyl)-N-phenylhydrazine-1-carbothioamide (4a)

White powder, yield 85%, m.p 238–240°C, R_f. = 0.6 (chloroform 70%: methanol 30%). IR (KBr) cm^-1: 3294 (NH) str of thioamide, 3159 indole (NH) str, 3051 (ArH) str, 1658 (C=O) str of amide (amide I band), 1546, 1496, 1469 Ar (C=C) str, 1608 (NH) bend of (amide II band), 1246(C-N) str, 1149(C=S) str, 725, 644 (CH) bend of heterocyclic. ¹HNMR (400 MHz, DMSO-d₆) δ: 10.41 (s, 1H, NH), 9.77 (s, 1H, NH), 9.66 (s, 1H, NH), 8.15 (s, 1H, ArH), 7.68–7.62 (m, 2H, ArH), 7.52 (m, 3H, ArH), 7.34 (t, 2H, Ar-H), 7.16 (t, 1H, Ar-H), 6.51 (s, 1H, Ar-H), 3.88 (s, 3H, CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 182.05 (C=S), 167.39 (C=O), 139.77, 139.62, 136.07, 133.09, 131.09, 128.62, 125.50, 125.36, 120.21, 119.13, 110.66, 101.05, 33.10 (CH₃). (HRESI-MS) m/z: Calcd. for C₁₇H₁₆N₄OS [M+1]: 325.1045, found 325.1115. 2-(1-methyl-1H-indole-6-carbonyl)-N-(4-nitrophenyl)hydrazine-1-carbothioamide (4b) Bright yellow powder, yield 84%, m.p 221–224°C, R_f. = 0.63 (chloroform 70%: methanol 30%). IR (KBr) cm^-1: 3282 (NH) str of sec, amide, 3101 (ArH) str, 2943, 2850 (CH str of CH₃), 1635 (C=O) str. of amide (amide I band), 1546 str, 1438 Ar (C=C) str, 1500, 1334 (asym/sym. str of NO₂ group, respectively), 1249 (C-N) str, 1215 (C=S) str, 844 (Ar- p-NO₂-substitution), 767, 663 (CH) bend of heterocyclic. ¹HNMR (400 MHz, DMSO-d₆) δ: 10.89 (s, 1H, NH), 10.56 (s, 1H, NH), 10.17 (s, 1H, NH), 8.26–8.15 (m, 3H, ArH), 7.94 (d, 2H, ArH), 7.90 (d, 1H, Ar-H), 7.69–7.63 (m, 2H, ArH), 7.55 (d, 1H, ArH), 6.51 (d, 1H, ArH), 3.88 (s, 3H, CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 181.31 (C=S), 167.41 (C=O), 156.44 (C-NO₂), 147.05, 140.92, 136.10, 133.22, 131.21, 126.02, 125.24, 124.08, 117.03, 110.71, 101.08, 33.11 (CH₃). (HRESI-MS) m/z: Calcd. for C₁₇H₁₅N₅O₃S [M+1]: 370.0896, found 370.0969.

2.2.5. Synthesis of 5-(1-methyl-1H-indol-6-yl)-1,3,4-oxadiazole-2-thione (5)

Following the addition of compound 3 (2.0 g, 0.0145 mol) to 20 ml of abs. EtOH, the mixture was cooled to 20°C in an ice bath. In 10 ml of abs EtOH, 0.818 g (0.0145 mol) of potassium hydroxide (KOH) was added after the mixture was stirred for 15°min. Then, 0.9 ml (0.0145 mol) of carbon disulfide (CS₂) was slowly added, a yellow precipitate was produced. After refluxing for 20 h, the mixture's yellow precipitate changed into a light-brown solution. After reduction of the solvent by a rotary evaporator, the solid product was dissolved in 25 ml of D.W., acidified with 10% HCl, and then recrystallized with 70% EtOH [39]. Beige powder, yield 55%, m.p 286–289°C, R_f. = 0.52 (chloroform 80%: methanol 20%). IR (KBr) cm^-1: 3116 (sec. thio-NH-amide) str, 3066 Ar(H) str, 2924, 2854 (CH) str of aliph. (CH₃), 1616 Ar (C=N) str, 1500, 1477 Ar (C=C) str, 1365 (C-N) str, 1234 (C-O-C) asym. oxadiazole str, 1176 (C-O-C) sym. oxadiazole str. ¹HNMR (400 MHz, DMSO-d₆) δ: 14.63 (s, 1H, NH-thioamide), 7.98 (s, 1H, ArH), 7.72 (d, 1H, Ar-H), 7.58–7.53 (m, 2H, ArH), 6.56 (d, 1H, ArH), 3.89 (s, 3H, CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 177.60 (C=S), 162.46 (C=N), 136.22 (Ar-C), 133.75 (Ar-C), 131.32, 121.78, 116.95, 115.17, 108.95, 101.59, 33.24 (CH₃). (HRESI-ESI) m/z: Calcd. for C₁₁H₉N₃OS [M+1]: 232.0466, found, 232.0529.

2.2.6. Synthesis of 1,3,4-oxadiazole-2-thione derivatives (6a–e)

10 ml of EtOH was used to dissolve compound 5 (0.25 g, 0.001395 mol) in a 150 ml (RB) flask. After that, triethylamine (0.2 ml, 0.00139 mol) was added, and the mixture was stirred for 15°minutes. Following that, a careful addition of each of the following: (A) 0.387 g, 0.001395 mol of 2,4′-dibromoacetophenone, (B) 0.34 g, 0.001395 mol of 2-bromo-4-nitroacetophenone, (C) 0.325 g, 0.001395 mol of 2-bromo-4′-chloroacetophenone, (D) 0.319 g, 0.001395 mol of 2-bromo-4′-methoxyacetophenone and (E) 0.412 g, 0.00412 mol of 2-bromoacetophenone to the mixture, and after being stirred at (RT) for approximately 2h, a suspended mixture was produced from the clear solution. The precipitate was filtered and recrystallized with 70% EtOH [40].

2.2.6.1. 1-(2,4-dibromophenyl)-2-((5-(1-methyl-1H-indol-6-yl)-1,3,4-oxadiazole-2-yl)thio)ethan-1-one (6a)

White powder, yield 58%, m.p 282–284°C, R_f. = 0.56 (chloroform 70%: methanol 30%). IR (KBr) cm^-1: 3059 (ArH) str, 2951, 2920 (CH) str of aliph. (CH₂), 1689 (C=O) str, 1504, 1477, 1446 Ar (C=C) str, 1392 (C-N) str, 1265 (C-O-C) asym. oxadiazole str, 1180 (C-O-C) sym. oxadiazole str, 636 (C-Br) str, 721, 675 (tri-substituted benzene ring). ¹HNMR (400 MHz, DMSO-d₆) δ: 8.04–8.01 (d, 3H, ArH), 7.82 (d, 2H, ArH), 7.71 (d, 1H, Ar-H), 7.59 (dd, 2H, Ar-H), 6.55 (s, 1H, ArH), 5.18 (s, 1H, CH₂), 3.87 (s, 3H, CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 192.14 (C=O), 166.53 (C=N), 162.17 (C=N), 135.83 (Ar-C), 134.09 (Ar-C), 133.07, 131.97, 130.65, 130.46, 128.15 (C-Br), 121.21(C-Br), 116.92, 115.29, 108.50 (Ar-C), 101.02, 40.31 (CH₂), 32.68 (CH₃). (HRESI-MS) m/z: Calcd. for C₁₉H₁₃Br₂N₃O₂S [M+1]: 505.9095, found 505.1320.

2.2.6.2. 1-(2-bromo-4-nitrophenyl)-2-((5-(1-methyl-1H-indol-6-yl)-1,3,4-oxadiazole-2-yl)thio)ethan-1-one (6b)

Yellow powder, yield 55%, m.p 276–279°C, R_f. = 0.52 (chloroform 60%: methanol 40%). IR (KBr) cm^-1: 3066 (ArH) str, 2951, 2870 (CH) str of aliph. (CH₂), 1681 (C=O) str, 1519, 1415 Ar (C=C) str, 1342 (C-N) str, 1238 (C-O-C) asym. oxadiazole str, 1184 (C-O-C) sym. oxadiazole str, 628 (C-Br) str, 740, 717 (tri-substituted benzene ring). ¹HNMR (400 MHz, DMSO-d₆) δ: 8.42 (d, 2H, ArH), 8.32 (d, 2H, ArH), 8.04 (s, 1H, Ar-H), 7.72 (d, 1H, ArH), 7.63–7.58 (m, 2H, Ar-H), 6.55 (s, 1H, ArH), 5.26 (s, 2H, CH₂), 3.87 (s, 3H, CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 192.24 (C=O), 166.62 (C=N), 162.02 (C=N), 150.28 (C-NO₂), 139.73 (Ar-C), 135.84 (Ar-C), 133.10, 130.68, 129.93, 123.95, 121.22, 116.95, 115.27 (C-Br), 108.56, 101.03, 40.65 (CH₂), 32.69 (CH₃). (HRESI-MS) m/z: Calcd. for C₁₉H₁₃BrN₄O₄S [M+1]: 472.9841, found, 472.9916.

2.2.6.3. 1-(2-bromo-4-chlorophenyl)-2-((5-(1-methyl-1H-indol-6-yl)-1,3,4-oxadiazole-2-yl)thio)ethan-1-one (6c)

White powder, yield 65%, m.p 266–269°C, R_f. = 0.52 (chloroform 70%: methanol 30%). IR (KBr) cm^-1: 3097 (Ar-H) str, 2951, 2916 (ArH) str of aliph. (CH₃ and CH₂), 1678 (C=O) str, 1504, 1469, 1419 Ar (C=C) str, 1396 (C-N) str, 1238 (C-O-C) asym. oxadiazole str, 1184 (C-O-C) sym. oxadiazole str, 663 (C-Br) str, 721, 690 (tri-substituted benzene ring). ¹HNMR (400 MHz, DMSO-d₆) δ: 8.11 (d, 2H, ArH), 8.02 (s, 1H, ArH), 7.73–7.56 (m, 5H, ArH), 6.55 (d, 1H, ArH), 5.18 (s, 2H, CH₂), 3.87 (s, 3H, CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 191.92 (C=O), 166.53 (C=N), 162.18 (C=N), 138.91 (C-Cl), 135.83 (Ar-C), 133.77, 133.05, 130.65, 130.41, 129.03, 121.21 (C-Br), 116.94, 115.29, 108.52, 101.02, 40.32 (CH₂), 32.68 (CH₃). (HRESI-MS) m/z: Calcd. for C₁₉H₁₃BrClN₃O₂S [M+1]: 461.9600, found 461.9749.

2.2.6.4. 1-(2-bromo-4-methoxyphenyl)-2-((5-(1-methyl-1H-indol-6-yl)-1,3,4-oxadiazole-2-yl)thio)ethan-1-one (6d)

Beige powder, yield 75%, m.p 272–274°C, R_f. = 0.62 (chloroform 70%: methanol 30%). IR (KBr) cm^-1: 3116, 3059 (ArH) str, 2912, 2843 (CH) str of aliph. (CH₂), 1662 (C=O) str, 1504, 1473, 1415 Ar (C=C) str, 1388 (C-N) str, 1261 (C-O-C) asym. oxadiazole, 1176 (C-O-C) sym. oxadiazole str, 636 (C-Br) str, 763, 717 (tri-substituted benzene ring). ¹HNMR (400 MHz, DMSO-d₆) δ: 8.09 (d, 2H, ArH), 8.02 (s, 1H, ArH), 7.71 (d, 1H, ArH), 7.62 (dd, 1H, ArH), 7.56 (d, 1H, ArH), 7.11 (d, 1H, ArH), 6.55 (d, 1H, ArH), 5.14 (s, 2H, CH₂), 3.87 (s, 6H, OCH₃+CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 191.53 (C=O), 166.95 (C=N), 164.24 (C=N), 162.91 (Ar-C), 136.33 (ArC), 133.53 (ArC), 131.43, 131.14, 128.39, 121.71 (C-Br), 117.44, 115.85, 114.61, 109.00, 101.52, 56.15 (OCH₃), 40.77 (CH₂), 33.19 (CH₃). (HRESI-MS) m/z: Calcd. for C₂₀H₁₆BrN₃O₃S [M+1]: 458.0096, found, 458.0171.

2.2.6.5. 1-(2-bromophenyl)-2-((5-(1-methyl-1H-indol-6-yl)-1,3,4-oxadiazole-2-yl)thio)ethan-1-one (6e)

Red powder, yield 55%, m.p 286–289°C, R_f. = 0.72 (chloroform 70%: methanol 30%). IR (KBr) cm^-1: 3097, 3059 Ar(H) str, 2951, 2924 (CH) str of aliph. (CH₂), 1678 Ar (C=O) str, 1508, 1469, 1419 Ar (C=C) str, 1388 (C-N) str, 1238 (C-O-C) asym. oxadiazole str, 1180 (C-O-C) sym. oxadiazole str, 686 (C-Br) str, 756, 717 (di-substituted benzene ring). ¹HNMR (400 MHz, DMSO-d₆) δ: 8.16 (d, 2H, ArH), 8.09 (s, 1H, ArH), 7.79–7.77 (m, 2H, ArH), 7.69–7.62 (m, 3H, ArH), 6.61 (d, 1H, ArH), 5.27 (s, 2H, CH₂), 3.93 (s, 3H, CH₃). ¹³CNMR (101 MHz, DMSO-d₆) δ: 193.26 (C=O), 166.99 (C=N), 162.81 (C=N), 136.33 (C-O-CH₂), 135.55 (C-CH₂-O), 134.51, 133.56, 131.14, 129.42, 128.99, 121.72 (C-Br), 117.44, 115.81, 109.03, 101.52, 40.98 (CH₂), 33.21 (CH₃). (HRESI-MS)) m/z: Calcd. for C₁₉H₁₄BrN₃O₂S [M+1]: 427.9990, found, 428.0065.

2.3. In silico studies

2.3.1. Generation of the 3D pharmacophore model

Structure-based pharmacophore design uses the ligand–protein complex's structural data to create a pharmacophore model, allowing for virtual screening and identification of compounds with desired pharmacophoric properties. The Discovery Studio program was used to create the 3D-pharmacophore model. The first phase entailed creating input files, which typically include the structure of the protein–ligand combination (erlotinib bound to EGFR and sorafenib bound to VEGFR-2) retrieved from x-ray crystallography data accessible in the Protein Data Bank (PDB). To produce the pharmacophore model, the software applies an algorithm that uses a ‘receptor–ligand pharmacophore auto’ technique that automatically recognizes key features and their spatial arrangement based on the protein–ligand interaction. These properties include hydrophobic, aromatic ring, hydrogen bond donor and hydrogen bond acceptor groups, which were chosen based on their relation to the observed binding interactions. The new compounds were evaluated by fitting them to the pharmacophoric queries, using the pharmacophore-fit score as a guideline. The co-crystallized ligand (a reference molecule such as erlotinib) has a 100% relative pharmacophore-fit score, and other compounds are compared with it based on how well they fit the pharmacophore model.

2.3.2. Molecular docking

The current study used computer-based methods to analyze the tested molecules' potential binding to EGFR and VEGFR-2 receptor tyrosine kinases (RTKs). EGFR and VEGFR-2 and the newly synthesized carbothioamide and oxadiazole derivatives were optimized using the MMFF94 force field, and the protein structures were obtained from the PDB (Protein ID: 4HJO) for EGFR and (Protein ID: 4ASD) for VEGFR-2. Molecular docking was then performed as previously described [41] generating 20 potential orientations with the best orientation selected based on Root Mean Square Deviation (RMSD) values and affinity scores.

2.3.3. Molecular similarity

Molecular similarity calculation is a computer technique that compares the structural and physicochemical characteristics of two ligands to determine how similar they are to one another. In this study, three ligands (4a, 4b and 5) were analyzed for molecular similarity with erlotinib, a standard EGFR inhibitor and five ligands (6a–e) were analyzed for molecular similarity with sorafenib, a standard VEGFR inhibitor, using the Discovery Studio software. Rotatable bonds, cyclic rings, aromatic rings, partition coefficient (logp), molecular weight (MW), hydrogen bond donors (HBD), hydrogen bond acceptors (HBA) and molecular fractional polar surface area (MFPSA) were among the molecular attributes examined [42].

2.3.4. Density functional theory

The dataset includes several key parameters such as total energy, binding energy, highest occupied molecular orbital (HOMO) energy, lowest unoccupied molecular orbital (LUMO) energy, dipole moment magnitude and band gap energy. These parameters provide critical insights into the electronic structure, energetics and reactivity of compounds.

2.3.5. Adsorption, distribution, metabolism, elimination & toxicity studies

Erlotinib and sorafenib were employed as reference drugs in adsorption, distribution, metabolism, elimination and toxicity (ADMET) assessments using the Discovery Studio 2019 Software for the synthesized compounds. The blood–brain barrier (BBB) grade was identified as either very low (4), low (3), medium (2), high (1) or very high (0). The extent of solubility was classified as optimal (4), good (3), low (2) or very low (1). The rate of absorption was classified as very poor (3), poor (2), moderate (1) or good (0). True or False was assigned to hepatotoxicity. There were two classifications for CYP2D6: inhibitor (True) and non-inhibitor (False).

2.3.6. Toxicology

Toxicology assessments for the newly synthesized compounds were generated using the Discovery Studio 2019 Software.

2.3.7. Molecular dynamics simulation

Molecular dynamics (MD) simulation of protein–ligand complexes was carried out with GROMACS 2021.1 and the Linux 5.4 package. The GROMOS96 54a7 forcefield was used for the protein, and ligand topologies were constructed using the PRODRG server [43]. In a rectangular box, all compounds were solvated using simple point charge (SPC) water molecules. To achieve electrical neutrality in the simulated system, Na⁺ and Cl^- ions were supplied, and salt concentrations of 0.15 mol/l were set for all systems. Using the steepest descent method, all solvated systems were subjected to energy minimization over 5000 steps. The MD simulation was then used to perform the NVT (constant number of particles, volume and temperature) and NPT (constant number of particles, pressure and temperature) series, as well as the production run. The NVT and NPT series were tested at 300 K temperature and 1 atm pressure for 300 ps. The simulation included the V-rescale thermostat and Parrinello-Rahman barostat. Finally, the production run was carried out at 300 K for 100 ns. To assess the stability of the molecules, a comparative study was performed using RMSD, Root Mean Square Fluctuation (RMSF), radius of gyration (Rg), Solvent Accessible Surface Area (SASA) and hydrogen bonds. The Xmgrace application was utilized to represent the analyses in the format of plots.

2.4. Biological activity

2.4.1. Cell culture

Three cancer cell lines were selected for the current work because they overexpress both EGFR and VEGFR-2 [44–48]. A549 human lung cancer cell line was grown in Ham's F-12 medium (Capricorn scientific, MA, USA) completed with 10% fetal bovine serum (FBS), HCT-116 colorectal cancer cell line and HepG2 hepatocyte carcinoma cell line were grown in high glucose Dulbecco's Modified Eagle Medium (DMEM; Capricorn scientific) supplemented with 10% FBS. These cancer cell lines were chosen as model cell lines for EGFR/VEGFR-2 overexpression [38–40]. MCF-10A normal breast epithelial cell line was grown in DMEM/F12 media (Invitrogen, MA, USA) supplemented with horse serum (Invitrogen), penicillin-streptomycin, EGF (Sigma, Kawasaki, Japan), hydrocortisone (Sigma), cholera toxin (Sigma) and insulin (Sigma). For cell culture maintenance, trypsin-EDTA solution (0.05%) (Capricorn scientific) was used to passage cells. Cells were kept in a CO₂ incubator at 37°C with 5% CO₂ and 95% air.

2.4.2. In vitro viability & proliferation detection (MTT assay)

Cells were seeded in 96-well plates at a density of 10,000 cells/well in triplicates. The MTT reagent (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) prepared as a 5 mg/ml stock solution in Ham's F12 or high glucose-DMEM medium, was added to control and 72-h treated cells in eight different concentrations (0.03, 0.1, 0.3, 1, 3, 10, 30, 100 μM) of the experimental compounds, and incubated for 3 h [49]. Erlotinib and sorafenib (positive controls) [50] were included in each experiment depending on the target. Proliferation was assessed by measuring the absorbance at 570 nm with a 96-well plate reader (Biotek Synergy HT Multi-Mode Microplate Reader, VT, USA).

2.4.3. Cell cycle analysis

HCT-116 and A549 cells were seeded in 60 mm dishes for 24 h. Cells were subsequently treated with the experimental compounds 4a and 6c, respectively, at their IC₅₀ concentrations. After 72 h, cells were harvested, washed with 1X PBS, and fixed for 30 min using 70% cold EtOH at 4°C. Following washing and centrifugation, the cell pellet was treated with 50 μl of a 100 μg/ml stock of RNase (Zymo Research, CA, USA) and 200 μl propidium iodide (from 50 μg/ml stock solution) (R&D Systems, MN, USA). Fluorescence emitted from the PI–DNA complex was detected by MACS Miltenyi Biotec, and MACSQuantify software was employed for the analysis of cell distribution across stages of the cell cycle. The experiment was performed twice in duplicates.

2.4.4. Detection of apoptosis/necrosis

Using TACS annexin V-FITC apoptosis detection kit (R&D Systems), compound 4a-treated HCT-116 and compound 6c-treated A549 cells were assessed for apoptosis and necrosis induction by MACS Miltenyi Biotec. Data were analyzed using MACSQuantify software, and statistical analysis was conducted to assess differences between control and treated cell populations. The experiment was performed twice in duplicates.

2.4.5. Quantitative reverse transcription-polymerase chain reaction

Total RNA was extracted from compound 4a-treated HCT-116 cells, compound 6c-treated A549 cells (IC₅₀, 72 h), and control cultures using the Direct-zol RNA Microprep Kit (Zymo Research). The Power cDNA synthesis kit (Intron Biotechnology, Seoul, South Korea) was used to reverse transcribe RNA samples into cDNA. Later, cDNA samples underwent amplification of apoptotic markers (caspase 3, caspase 4, caspase 8 and caspase 9) using KAPA SYBR FAST qPCR Master Mix (KAPA BIOSYSTEMS, MA, USA) on Line Gene 9680 BioER instrument. β-actin was used as internal control to normalize the relative gene expression. Supplementary Table S1 shows the primers sequences (forward and reverse) that were used in the current study.

2.4.6. Western blot analysis

Control, compound 4a-treated HCT-116 culture, and compound 6c-treated A549 cells were lysed in cold SDS-lysis buffer. Cell extracts were boiled in a water bath for 5 min and sonicated before protein concentration determination via Bradford assay [51,52]. The protein extracts were separated by 10% SDS-PAGE, transferred into nitrocellulose membrane, and then probed with specific antibodies against caspase 3 (1:500; Invitrogen), caspase 9 (1:500; Invitrogen), GAPDH (1:2000; Invitrogen) and tubulin (1:1000; Sigma). After incubation with horseradish peroxidase-conjugated secondary antibodies, the protein bands were detected using FluorChem R system (Oxford, UK).

2.4.7. EGFR & VEGFR-2 inhibitory assay

The kinase activity of EGFR was assessed using the EGFR tyrosine kinase assay (Promega, WI, USA), and the inhibitory activity of the experimental compounds toward the kinase domain of VEGFR-2 was assessed using the KDR kinase enzyme system kit (Promega). Compounds tested for EGFR binding and inhibition were compared with erlotinib, while the other experimental compounds that were tested for VEGFR-2 inhibitory activity were compared with sorafenib.

2.4.8. Statistical analysis

Statistical analysis was performed using GraphPad 9 software using the Student's t-test, and p < 0.05 was considered significant. The number of experiments for each assay is indicated in each figure/table caption.

3. Results & discussion

3.1. Chemistry

Indole methyl ester (compound 1) was synthesized using MeOH and conc. H₂SO₄ (as a catalyst), as described previously [37]. The structure was validated by a peak at ύ = 1689 cm^-1 in the FT-IR spectrum (Supplementary Figure S1), which was attributed to ester (C=O) stretching. Compound 1 was reacted with methyl iodide to produce compound 2 that showed the disappearance of indole NH stretching peak (3275 cm^-1). Compound 2 was refluxed with hydrazine hydrate to generate methyl indole hydrazide derivative (compound 3). The IR spectrum (Supplementary Figure S1) is distinguished by bands of hydrazide secondary amide (NH) at ύ = 3278 cm^-1, 3163 (NH₂) str of primary amine, as well as a distinct peak for (C=O) stretching of hydrazide at ύ = 1681 cm^-1. The methyl indole carbothioamide derivatives (4a–b) were produced by treating compound 3 with various aryl isothiocyanate derivatives. The FT-IR spectra (Supplementary Figure S1) showed distinct chemical moieties of the final compounds, which are characterized by the absence of (NH₂) primary amine bands. The oxadiazole-2-thione derivative (compound 5) was synthesized by treating compound 3 with carbon disulfide (CS₂) and (KOH). Compound 5 was stirred with different phenacyl bromides to produce compounds 6a–e. The FT-IR spectra (Supplementary Figure S1) illustrated the following functional groups for compounds 6a–e distinguished by the presence of new bands at ύ = 1689, 1681, 1678, 1662 and 1678 cm^-1, due to C=O stretching, respectively.

¹HNMR spectroscopy (Supplementary Figure S2) revealed distinct signals for the produced compounds. All compounds have a signal at δ = (3.84–3.94) ppm, as a singlet integrated for three protons related to CH₃-N-indole. The aromatic protons of the compounds 4a, 4b, 5 and 6a–e were displayed in the predicted aromatic area, as shown in the experimental section. Compounds 4a and 4b have a signal at δ = 10.41, 9.77, 10.89 and 10.56 ppm, as a singlet integrated for two protons related to NH-NH of hydrazide. Compound 5 has a signal at δ = (14.63) ppm, as a singlet attributed to one proton related to NH-thioamide. Compounds 6a–e have a signal at δ = 5.18, 5.26, 5.18, 5.14 and 5.27 ppm, as a singlet owing to two protons of aliphatic S-CH₂.

¹³CNMR spectra (Supplementary Figure S3) revealed that all compounds have a signal at δ = 32.68–33.21 ppm, due to CH₃-N-indole. Compounds 4a and 4b¹³CNMR spectra exhibited typical signals of C=S carbon atoms at δ = 181.31 and 182.05 ppm, respectively. Compounds 5 and 6a–e¹³CNMR spectra displayed typical signals of oxadiazole (C=N) at δ = 162.46, 166.53, 166.62, 166.53, 166.95 and 166.99 ppm, respectively. Compounds 6a–e¹³CNMR spectra showed typical signals of aliphatic (S-CH₂) at 40.31, 40.65, 40.32, 40.77 and 40.98 ppm, respectively. Simultaneously, the remaining carbon atoms were displayed at their precise location.

The HRESI-MS spectra (Supplementary Figure S4) confirmed that the newly synthesized compounds were successfully prepared.

3.2. In silico studies

3.2.1. Generation of 3D-pharmacophore model

The basic goal of the pharmacophore-based drug design is to create a 3D pharmacophore model using the receptor–ligand pharmacophore generating technique. In this technique, the ligand's binding to the target protein (erlotinib to EGFR and sorafenib to VEGFR-2) is studied to show a number of properties required for the activity. As a result, any new molecule exhibiting these properties is likely to be active.

The current investigation demonstrated that the 3D-pharmacophore model for the erlotinib-EGFR complex had three characteristics: two terminal aromatic rings, a central hydrogen bond acceptor and a central hydrogen bond donor (Supplementary Figure S5). The created model was then utilized to investigate the tested compounds as potential EGFR TK inhibitors (Supplementary Figure S5). The 3D-pharmacophore model for the sorafenib-VEGFR-2 complex, on the other hand, revealed these characteristics: a terminal hydrogen bond donor, two internal aromatic rings, an internal hydrogen bond acceptor and a terminal aromatic ring (Supplementary Figure S5). The developed model was then utilized to investigate the tested ligands as potential VEGFR-2 TK inhibitors (Supplementary Figure S5).

The ligand pharmacophore-fit scores provide insights into the potential pharmacophoric similarities and differences between the reference compounds (erlotinib or sorafenib) and their analogs. The results suggest that erlotinib and sorafenib and their analogs (compounds 4a, 4b, 5 and 6a–e, respectively) exhibited varying degrees of conformational compatibility with the pharmacophore model (Supplementary Figure S5 & Supplementary Table S2). Compounds 4a and 4b demonstrated high pharmacophore-fit scores of 3.10 and 2.99, respectively. The results obtained imply that such compounds along with erlotinib (pharmacophore-fit score of 3.01), have pharmacophoric properties closely aligned with the pharmacophore model. Therefore, compounds 4a and 4b are potential candidates for further investigation. On the other hand, compound 5 showed a significantly lower pharmacophore-fit score of 0.75. This significant divergence from the pharmacophore model implies a possible lack of critical pharmacophoric properties, raising concerns about its efficiency in interacting with the target protein relevant to the desired therapeutic function (Supplementary Figure S5 & Supplementary Table S2). Compounds 6b and 6d exhibited high pharmacophore-fit scores of 3.28 and 3.45, respectively. This suggests that these compounds possess pharmacophoric properties that are similar to sorafenib (pharmacophore-fit scores of 3.56). Compounds 6a, 6c and 6e displayed slightly lower pharmacophore-fit scores of 3.12, 2.99 and 3.05, respectively. These lower scores may indicate few differences in their pharmacophoric properties compared with sorafenib (Supplementary Figure S5 & Supplementary Table S2).

3.2.2. Docking study

The binding energy values of erlotinib and the tested compounds 4a–b, and 5 with EGFR TK, and sorafenib and compounds 6a–e with VEGFR-2 TK are presented in Supplementary Table S3.

3.2.2.1. EGFR TK binding

The binding energy of the crystal ligand erlotinib (PDB codes: 4HJO) against EGFR TK was -7.16 kcal/mol. Erlotinib generated seven hydrophobic binding sites with Leu694, Val702, Ala719, Lys721 and Leu820, it also formed H-bond with Met769 (1.98 Å) (Figure 2A). Compound 4a‘s binding energy against EGFR TK was comparable to that of erlotinib (-7.66 kcal/mol). Compound 4a generated eight hydrophobic contacts with Leu694, Val702, Ala719, Met742, Leu753, Leu764 and Leu820, and three H-bonds with Lys721, Thr766 and Asp831 (1.96, 2.67 and 3.01 Å), respectively (Figure 2B). Compound 4b's binding energy against EGFR TK was -7.53 kcal/mol, which is close to that of erlotinib. It produced four hydrophobic contacts with Leu694, Val702 and Lys721, and two H-bonds with Lys704 and Met769 (Supplementary Table S6A). Compound 5 showed a binding energy of -5.72 kcal/mol against EGFR TK, which is unfavorable and significantly different from that of erlotinib. It established five hydrophobic contacts with Leu694, Val702, Ala719, Leu768 and Leu820, and one H-bond with Met769 (2.12 Å) (Supplementary Table S6B). The three EGFR-targeting derivatives (4a, 4b and 5) established hydrophobic interactions and hydrogens bonds with the same amino acids in EGFR active site as erlotinib.

Figure 2.

Mapping surface and 3D orientation. (A) Erlotinib. (B) Compound 4a docked in EGFR tyrosine kinase (TK) ATP-binding site. (C) Sorafenib. (D) Compound 6c docked in VEGFR-2 TK ATP-binding site.

3.2.2.2. VEGFR-2 TK binding

The binding energy of the crystal ligand sorafenib (PDB codes: 4ASD) against VEGFR-2 TK was -8.40 kcal/mol. It formed 19 hydrophobic interactions with Leu840, Val848, Ala866, Lys868, Ile888, Leu889, Ile892, Val899, Val916, Phe918, Leu1019, Leu1025 and Cys1045, additionally it bound Glu885, Val899, Cys919 and Asp1046, by four H-bonds (Figure 2C). The binding affinity of compound 6a displayed -8.21 kcal/mol binding energy with VEGFR-2 TK. 16 hydrophobic contacts were formed by compound 6a with Leu840, Ala866, Val899, Val916, Phe918, Cys919, Leu1035, Cys1045 and Phe1047, additionally, it formed one H-bond with Cys919 (2.23 Å) (Supplementary Figure S6C). The binding affinity of compound 6b showed -8.38 kcal/mol binding energy against VEGFR-2 TK. Thirteen hydrophobic contacts were established by compound 6b with Val848, Ala866, Lys868, Leu889, Val899, Val916, Leu1019, Cys1024, His1026 and Cys1045, and it formed one H-bond with Lys868 (2.26 Å) (Supplementary Figure S6D). The binding affinity of compound 6c displayed a binding energy of -8.58 kcal/mol against VEGFR-2 TK. 13 hydrophobic contacts were formed by compound 6c with Val848, Ala866, Lys868, Ile888, Leu889, Val899, Val916, Cys1024, Arg1027 and Cys1045, additionally, it created one hydrogen bond with Asp1046 (2.17 Å) (Figure 2D). The binding affinity of compound 6d displayed -8.35 kcal/mol binding energy against VEGFR-2 TK. 13 hydrophobic contacts were established by compound 6d with Leu840, Val848, Ala866, Lys868, Leu889, Ile892, Val916, Cys919 and Leu1035 (Supplementary Figure S6E). Finally, the binding affinity of compound 6e displayed -8.27 kcal/mol binding energy against VEGFR-2 TK. 13 hydrophobic contacts were generated by compound 6e with Leu840, Val848, Ala866, Lys868, Leu889, Val899, Val914, Val916, Phe918, Cys919 and Leu1035, additionally, it displayed one H-bond with Cys919 (2.63 Å) (Supplementary Figure S6F). The binding energies (docking scores or G) of the VEGFR-2-targeting compounds (6a–e) were remarkably close to sorafenib's, and they formed hydrophobic interactions and hydrogens bonds with the same amino acids in VEGFR-2 active site as sorafenib.

3.2.3. Molecular similarity

According to the results displayed in Supplementary Figure S7 & Supplementary Table S4, compounds 4a and 4b showed similar molecular properties to erlotinib, and compounds 6a, 6c, 6d and 6e showed similar molecular properties to sorafenib. In contrast, compounds 5 and 6b are categorized as dissimilar to the standard molecules (erlotinib and sorafenib, respectively) due to differences in certain physicochemical properties, such as molecular weight, hydrogen bond acceptors and donors.

3.2.4. Pharmacokinetic study (adsorption, distribution, metabolism, elimination & toxicity studies)

The in silico pharmacokinetic features of the newly synthesized compounds and the reference compounds erlotinib and sorafenib were determined based on their ADMET results. These features include the degree of penetration of the BBB, solubility and the ability to bind to plasma proteins (PPB). The new compounds 4a, 4b and 5 and erlotinib showed similar ADMET properties (Supplementary Figure S8 & Supplementary Table S5) such as low solubility, good absorption and hepatotoxicity, however, compounds 4a, 4b and 5 showed lower ability to bind to plasma proteins in comparison with erlotinib. Furthermore, compounds 4a, 4b and 5 have very low to medium BBB levels, indicating that they are less likely to cross the BBB and so enter the central nervous system, perhaps leading to fewer CNS adverse effects [53].

The new compounds 6a, 6c and 6e, and sorafenib showed similar ADMET properties (Supplementary Figure S8 & Supplementary Table S5) such as very low solubility, good absorption, hepatotoxicity and the ability to bind to plasma proteins. Compounds 6a, 6c, 6d and 6e have high BBB levels, indicating that they are likely to pass the BBB. Compound 6b was a notable exception with BBB level of 4, signifying very low chances to enter the CNS. Compound 6c was the only compound that was predicted to inhibit cytochrome P2D6 (CYP2D6). Furthermore, compound 6b had exceptionally moderate absorption, which may have an impact on its bioavailability and potential efficacy.

3.2.5. Toxicity studies

The BIOVIA Discovery Studio software, erlotinib and sorafenib (as reference compounds) were used to conduct virtual toxicity studies against various toxicity models. Supplementary Table S6 displays an extensive examination of the in silico toxicity features of the newly synthesized compounds. These toxicity studies play a role in the initial stages of drug development, assisting in recognizing compounds that have beneficial safety profiles, and those that might necessitate more investigation or modification. Compounds 4a, 4b, 5 and erlotinib are categorized as non-carcinogenic and non-toxic. In comparison to sorafenib, compounds 6a, 6b and 6c exhibited relatively low TD₅₀ values, indicating higher carcinogenic potency in terms of their capacity to induce cancer in rats. These compounds may require careful consideration for safety, particularly in long-term exposure scenarios. The Rat Maximum Tolerated Dose values suggest that compounds 6a, 6b and 6c are associated with non-toxic maximum tolerated doses when administered through feed. All compounds, including sorafenib and erlotinib, exhibited a ‘Non-Toxic’ developmental toxicity potential. The Rat Oral LD₅₀ values for 6a, 6b and 6c confirmed their non-toxic nature when administered orally. On the other hand, sorafenib is classified as toxic based on its low LD₅₀ value, suggesting that it poses a higher acute toxicity risk upon oral exposure. The Rat Chronic LOAEL values indicated mild or non-irritant chronic toxicity for compounds 6a, 6b, 6c and sorafenib. This suggests that they have low chronic toxicity levels. All compounds, including sorafenib and erlotinib, are rated as ‘Mild’ in terms of ocular irritancy and ‘Non-Irritant’ for skin irritancy in rat models. This indicates their low potential to cause irritation when in contact with the eyes or skin (Supplementary Table S6) [53].

3.2.6. Molecular orbital analysis

The electronic profiles of the synthesized compounds (4a–b, 5 and 6a–e) were determined using density functional theory (DFT) calculations (Supplementary Figure S9 & Supplementary Table S7). The factors investigated included total energy, binding energy, HOMO energy, LUMO energy, dipole magnitude and band gap energy. The following are the main findings and their implications [54].

Total energy values represent the overall energy of each compound's electronic system. Compounds 4a and 4b had similar total energy to erlotinib, while compound 5 had a lower total energy, indicating that it may be less stable than erlotinib and the other synthetic compounds. Compounds 6a–e had very high total energy values when compared with sorafenib, indicating remarkable stability.

The binding energy values represents the energy released when the molecules are bound. The binding energy is computed as the difference between the total energy of the bound state and the total energy of the individual, isolated components. Compounds 4a and 4b showed favorable binding energy, although lower than that of erlotinib. Whereas compound's 5 binding energy was unfavorable indicating weak binding interactions with EGFR. Compounds 6a–e exhibited favorable and very close binding energy values to that of sorafenib, suggesting strong binding interactions with their target (VEGFR-2).

Variations in HOMO and LUMO energies reveal differences in electron density distribution and electronic structure between substances. Higher HOMO energy indicates better electron-donating capacity, and lower LUMO energy indicates better electron-accepting ability. Compounds 4a and 5 have equivalent HOMO and LUMO energy levels to erlotinib, indicating that they have similar electronic structures. Compounds 6a–e, on the other hand, showed higher LUMO energy levels than sorafenib, indicating that they can accept electrons more efficiently.

The dipole moment magnitude describes the distribution of electric charge within a molecule. A higher dipole moment indicates greater polarity. Compounds 4b and 6b had the highest dipole magnitude, indicating a high level of polarity, which contributes to their molecular strength and ability to form ionic and hydrogen bonds. The band gap energy is the difference in energy between the HOMO and LUMO levels. It measures a compound's electrical conductivity and optical characteristics. Compound 4b had a lower band gap energy than erlotinib, whereas all of the VEGFR-2-targeting compounds (6a–e) had a lower band gap energy than sorafenib, indicating that they required more energy to make electronic transitions and thus had a better fit with the target, EGFR and VEGFR-2, than the reference molecules.

To summarize, DFT analysis sheds light on the electronic structure, reactivity and energetics of drug candidates. Variations in total energy, binding energy, HOMO and LUMO energies, dipole moment magnitude and band gap energy highlight the compounds' different electronic characteristics and reactivity potentials. Understanding these features is critical for predicting their behavior in chemical reactions, optical qualities and potential for a variety of applications, including drug discovery.

3.2.7. Molecular dynamics simulation studies

MD simulations were run for 100 ns to assess the stability of the compounds at the active sites of EGFR TK and VEGFR-2 TK, where they had the highest docking scores. Two compounds, 4a complexed with EGFR TK and 6c complexed with VEGFR-2 TK, were chosen for MD simulations. The acquired RMSD for the complexes and ligands in relation to their initial locations inside the active site were documented and examined. The Cα atoms of the proteins were used to track the conformational stability of the proteins with regard to their starting positions. Supplementary Figure S10A shows that EGFR TK complexed with compound 4a fluctuated a little within the first 20 ns, but thereafter stabilized. On the other hand, Supplementary Figure S10B demonstrates the remarkable stability of the VEGFR-2 TK-compound 6c complex, with an RMSD value that is within an acceptable range. To acquire a better understanding of the protein areas that were affected throughout the simulation, the flexibility of each residue was analyzed in terms of RMSF (Supplementary Figure S10C & D). The 4a–EGFR TK complex fluctuated at the 150–180 amino acid area and then showed greater stability during the rest of the simulation period (Supplementary Figure S10C), whereas the 6c–VEGFR-2 TK complex demonstrated long-term stability at the active site, with small fluctuations in the 140–150 amino acid area (Supplementary Figure S10D) [55].

The radius of gyration (Rg) serves as a representation of the complex's compactness; a system is more compact when its degree of volatility is lower during the course of the simulation. The 4a–EGFR TK complex's Rg was found to be marginally greater than the beginning period (Supplementary Figure S11A), however, the 6c–VEGFR-2 TK complex's Rg was shown to be lower (Supplementary Figure S11B). SASA was utilized to determine how solvents and protein–ligand complexes interacted over the course of the simulation. The complexes' SASA were determined in order to assess the degree of conformational changes that occurred throughout the interaction. The surface area of the 6c–VEGFR-2 TK complex was reduced (Supplementary Figure S11B) resulting in a lower SASA value than the initial period compared with the 4a–EGFR TK complex (Supplementary Figure S11A).

3.2.8. Heat map analysis

Compound 4a created hydrophobic interactions with Val702 (∼10%), Lys721 (∼15%), Leu753 (∼10%), Leu764 (∼15%), Cys773 (∼20%) and Leu820 (∼30%). Additionally, compound 4a created H-bond interactions with the following residues: Thr766 (23%), Arg817 (10%) and Asp1831 (170%), and it had the ability to create several water-bridged H-bond interactions with Lys721 and Asp831 (Supplementary Figure S12A). On the other hand, compound 6c formed H-bond interactions with the following residues: His1026 (∼5%), Cys1045 (∼5%) and Asp1046 (∼30%), as presented in Supplementary Figure S12B. Compound 6c established water-bridged H-bonds with Glu885 (∼10%), Val899 (∼20%), His1026 (∼10%) and Ile1044 (∼35%), and was able to form hydrophobic interactions with residues Val848 (∼30%), Ala866 (∼10%), Leu888 (∼20%), Leu889 (∼15%), Ile892 (∼15%), Val899 (∼20%), Val916 (∼90%), Phe918 (∼20%) and Phe1047 (∼10%).

Another method used to monitor these interactions involves plotting the number of interactions with respect to time, or a heat map (Supplementary Figure S12C & D), which shows the number of interactions at each frame. The dark color implies more interactions in the 4a–EGFR TK complex (Supplementary Figure S12C) and the 6c–VEGFR-2 TK complex (Supplementary Figure S12D). From the heat maps figures, it was observed that the highest number of conformations of the protein formed up to three hydrogen bonds for the 4a–EGFR TK complex, and five hydrogen bonds for the 6c–VEGFR-2 TK complex.

3.3. Biological study

3.3.1. The new hydrazine-1-carbothioamide & oxadiazole derivatives exhibit anti-proliferation activities against cancer cells

The new hydrazine-1-carbothioamide and oxadiazole derivatives were evaluated for their cytotoxicity against A549 (lung), HCT-116 (colorectal) and HepG2 (hepatic) cancer cell lines. The cytotoxicity of the compounds was expressed as IC₅₀ (μM) or half maximal inhibitory concentration and compared with erlotinib and sorafenib as reference drugs. As shown in Table 1, most tested compounds showed good IC₅₀. Compound 4a was the most potent in the EGFR-targeting group, while compound 6d, followed by compound 6c were the most potent in the VEGFR-2-targeting group.

Table 1.

MTT assay results (IC₅₀, μM) of the new compounds after 72 h.

Compound	HepG2	HCT-116	A549	MCF10A
4a	45.2 ± 6.6	6.3 ± 1.1	37.4 ± 9.8	118.1 ± 13.3
4b	304.9 ± 34.9	22.1 ± 5.1	571.4 ± 89.1
5	2.1 ± 0.4	116.4 ± 19.7	181.4 ± 23.6
Erlotinib	13.8 ± 2.8	8.9 ± 2.0	5.9 ± 1.1	38.4 ± 6.6
6a	747.2 ± 65.0	130.6 ± 27.1	13.1 ± 2.8
6b	13.9 ± 2.8	66.8 ± 17.3	156.2 ± 28.1
6c	36.9 ± 7.1	275.2 ± 37.7	7.2 ± 1.0	123.8 ± 15.8
6d	2.2 ± 0.3	6.9 ± 1.0	391.3 ± 34.8
6e	75.3 ± 23.7	21.3 ± 5.5	89.8 ± 13.7
Sorafenib	8.3 ± 2.0	33.2 ± 7.1	8.6 ± 1.1	27.8 ± 5.9

Results are presented as a mean ± SEM of three experiments performed in triplicates.

SEM: Standard error of mean.

Furthermore, the cancer selectivity of compounds 4a and 6c was assessed and compared with that of standard tyrosine kinase inhibitors (TKIs; erlotinib and sorafenib). The cancer-selectivity index (SI) was computed using the following formula: SI = IC₅₀ against normal cells/IC₅₀ against cancer cells (Table 1). According to this equation, the SI for compound 4a is 18.7 compared with 4.3 for erlotinib, and 17.2 for compound 6c compared with 3.2 for sorafenib, indicating that the new derivatives are more selective for cancer cells than the two standard TKIs.

3.3.2. Compound 4a targets EGFR while compound 6c targets VEGFR-2 to inhibit the proliferation of cancer cells

To validate the molecular docking results, the newly synthesized experimental compounds were tested for their in vitro EGFR and VEGFR-2 kinase inhibitory activities (Supplementary Figure S13), and the EC₅₀ values were obtained and compared with the reference drugs erlotinib and sorafenib (Table 2). The results showed that compound 4a targets EGFR and has the lowest EC₅₀ among that particular group, whereas compound 6c targets VEGFR-2 with the lowest EC₅₀. The EC₅₀ values of both compounds were comparable to that of the reference drug. Based on the results of the MTT assay as well as the EGFR/VEGFR-2 enzyme inhibition assay, compound 4a with an IC₅₀ value of 6.3 μM against HCT-116 and compound 6c with an IC₅₀ value of 7.2 μM against A549 were chosen for further biological analysis.

Table 2.

EGFR and VEGFR-2 kinase enzyme inhibition activities of the different compounds presented as EC₅₀ values.

Compound	Target	EC50 (nM)
4a	EGFR	8.2 ± 1.4
4b	EGFR	122.7 ± 17.5
5	EGFR	10.6 ± 2.0
Erlotinib	EGFR	15.9 ± 3.8
6a	VEGFR	10.3 ± 2.0
6b	VEGFR	98.9 ± 18.0
6c	VEGFR	3.4 ± 0.7
6d	VEGFR	51.7 ± 8.1
6e	VEGFR	35.6 ± 7.5
Sorafenib	VEGFR	2.2 ± 0.4

Results are presented as mean ± SEM from two independent experiments performed in triplicates.

SEM: Standard error of mean.

3.3.3. The effect of compounds 4a & 6c on cell cycle progression of HCT-116 & A549 cancer cells

To investigate the mechanism of action of compounds 4a and 6c, the ability of these compounds to affect cell cycle progression was investigated using flow cytometry (Supplementary Figure S14A). This assay employs a fluorescent DNA-binding dye (propidium iodide; PI) allowing the determination of the relative amount of DNA during the different phases of the cell cycle phases (G₀/G1, S, G2/M). In both control samples of HCT-116 and A549 cancer cells, the highest percentage of cells existed in the G₀/G1 phase, while for both treated samples (4a-treated HCT-116 and 6c-treated A549 cells) the highest percentage of cells was in the G2/M phase (Table 3). This indicates that both experimental drugs (4a and 6c) caused cell cycle arrest at the G2/M phase of the cell cycle (Table 3).

Table 3.

Distribution of cell populations in the different cell cycle phases using propidium iodide staining and flow cytometry.

	G0/G1 (%)	S (%)	G2/M (%)
Control HCT-116	48.1 ± 7.7	23.7 ± 2.5	28.2 ± 5.8
4a-treated HCT-116	14.7 ± 2.5‡	9.0 ± 1.9†	76.3 ± 12.8‡

	G0/G1 (%)	S (%)	G2/M (%)
Control A549	48.5 ± 10.3	21.4 ± 2.0	30.1 ± 5.5
6c-treated A549	22.0 ± 1.1‡	18.6 ± 2.2	59.4 ± 10.3†

HCT-116 colorectal and A549 lung cancer cells were treated with compound 4a and compound 6c, respectively, at the IC₅₀ concentration for 72 h. Each value represents the mean ± SEM determined from a minimum of two independent experiments.

^†p < 0.05.

^‡p < 0.01 compared with control.

3.3.4. Compounds 4a & 6c induce apoptosis & necrosis in HCT-116 & A549 cancer cells

To precisely determine the mode of cell killing (apoptosis versus necrosis) of compounds 4a and 6c, Annexin V-FITC apoptosis detection kit was used. Compound 4a-treated HCT-116 cells showed a more prevalent necrotic cell death that is visible in the dot plot (Supplementary Figure S14B) and high percentage of necrotic cells (Table 4) in comparison to control HCT-116 cells, while compound 6c-treated A549 cell culture showed almost equal apoptotic and necrotic cell death (Supplementary Figure S14B & Table 4).

Table 4.

Percentage of cells (control and treated) undergoing apoptosis or necrosis (%) in the different groups.

	Viable	Necrosis	Apoptosis
Control HCT-116	92.5 ± 14.3	1.2 ± 0.3	6.3 ± 1.1
4a-treated HCT-116	8.5 ± 1.2§	60.5 ± 11.8§	31.0 ± 4.7†
Control A549	93.6 ± 18.7	0.9 ± 0.1	5.5 ± 2.3
6c-treated A549	14.2 ± 2.3‡	40.1 ± 9.4‡	45.7 ± 8.8‡

Each value represents the mean ± SEM determined from a minimum of two independent experiments.

^†p < 0.05.

^‡p < 0.01.

^§p < 0.001 compared with control cells in each treated group.

3.3.5. Caspase 3 is activated downstream of compounds 4a & 6c

To determine the apoptotic pathway induced by compounds 4a and 6c that led to inhibition of cancer cell proliferation, qRT-PCR was used to measure gene expression of caspase 3, caspase 4, caspase 8, caspase 9 and β-actin (as an internal control). The results of compound 4a-treated HCT-116 showed a significant increase in caspase 3 and caspase 9 gene expression in comparison to untreated HCT-116 cells (Figure 3A). On the other hand, the results of compound 6c-treated A549 cells showed an increase in caspase 3 gene expression in comparison to untreated A549 cells (Figure 3B).

Figure 3.

Analysis of apoptosis markers by quantitative reverse transcription-polymerase chain reaction (qRT-PCR) and immunoblotting. (A & B) Fold change in mRNA expression evaluated by qRT-PCR in HCT-116 and A549 cells treated with (A) Compound 4a; and (B) Compound 6c, respectively, after 72 h treatment with the IC₅₀ concentration of each compound. *p < 0.05; ***p < 0.001 vs. control. Results are presented as mean ± SEM from two independent experiments. (C & D) Representative immunoblots of pro-and cleaved (active) caspases in (C) control and compound 4a-treated HCT-116 cells, and (D) control and compound 6c-treated A549 cells. The experiment was repeated twice.

SEM: Standard error of mean.

In addition, the activation of caspase 9 and 3 in compound 4a-treated HCT-116 cells (Figure 3C), and caspase 3 in compound 6c-treated A549 cells (Figure 3D) was analyzed by immunoblotting. Caspase 3 (an executioner caspase) was activated downstream of caspase 9 (an initiator caspase) in compound 4a-treated HCT-116 cells. Similarly, caspase 3 was activated in 6c-treated A549 cells, however, based on the qRT-PCR results, none of the initiator caspases studied in the current study (caspase 4, 8 and 9) were probed for by immunoblotting. This suggests that the execution of apoptosis in compound 4a-treated cells took place through the activation of the intrinsic pathway [56], however, in compound 6c-treated cells caspase 3 was activated through an initiator caspase that was not studied in the current study such as caspase 2 or caspase 10 [56].

3.4. Structure–activity relationships

Useful information regarding the structure–activity relationships (SARs) of the newly synthesized hydrazine-1-carbothioamide derivatives (4a and 4b) and the oxadiazole derivative (5) as potential EGFR TK inhibitors, and the oxadiazole derivatives (6a–e) as potential VEGFR-2 TK inhibitors, can be highlighted by comparing their structures to erlotinib and sorafenib, respectively, specifically those moieties on the reference compounds that make them successful anticancer medications (Figure 4). For erlotinib, these moieties are: a hydrophobic aromatic center, a linker, a terminal hydrophobic moiety and a terminal substituent, while for sorafenib, the moieties are: a hydrophobic aromatic center, a linker, a central HBA/HBD, a terminal hydrophobic moiety and a terminal substituent. Additionally, the SAR was created based on the outcomes of the EGFR TK and VEGFR-2 TK enzyme assay (Table 2).

Figure 4.

Structure–activity relationships. SAR of erlotinib and sorafenib, as inhibitors of EGFR and VEGFR-2, respectively. SAR of the new derivatives based on EGFR enzyme inhibitory activity (EC₅₀) order of compounds 4a, 4b and 5, and VEGFR-2 enzyme inhibitory activity order of compounds 6a–e.

SAR: Structure–activity relationship.

Comparing EGFR TK inhibitory activity of compounds 4a, 4b and 5 indicated that compound 4a containing unsubstituted phenyl moiety (R=H) is more potent than compound (5) containing oxadiazole moiety, while compound 4b had lower inhibitory enzyme activity than both compounds (Figure 4). Regarding derivatives 6a–e containing oxadiazole linkers targeting VEGFR-2 TK, the analysis of the enzyme inhibitory activity revealed that the compounds 6c (2-bromo-4-chloro phenyl moiety) and 6a (2,4-dibromophenyl moiety), containing terminal hydrophobic tail substituted with halogens were more active than the other derivatives (6e, 6d and 6b). The introduction of a chloro group at the 4-position in compound 6c led to a higher inhibitory enzyme activity when compared with compound 6a containing a bromo group at the 4-position. The introduction of an electron donating (ED) group (OCH₃) for compound 6d at the 4-position showed moderate inhibitory activity, while the introduction of an electron withdrawing (EWD) group (NO₂) showed the lowest VEGFR-2 enzyme inhibitory assay. Generally, all compounds with electron withdrawing groups (6c, 6a and 6e) showed more VEGFR-2 TK enzyme inhibitory activity than compound 6d containing an electron donating group (Figure 4).

In summary, according to the current findings, an ideal EGFR TKI contains a hydrophobic aromatic center, a linker, a terminal hydrophobic moiety and a terminal unsubstituted phenyl moiety (R=H), whereas an ideal VEGFR-2 TKI contains a hydrophobic aromatic center, a linker, a central HBA/HBD, a terminal hydrophobic moiety and a terminal aromatic ring substituted with a chloro group at the 4-position.

4. Conclusion

According to computer-aided drug design approaches, compounds 4a, 4b and 5 indole hydrazine-1-carbothioamide derivatives have high affinity for EGFR tyrosine kinase ATP-binding site, while compounds 6a–e indole oxadiazole derivatives have high affinity for VEGFR-2 tyrosine kinase ATP-binding site. Compounds 4a, 4b and 5 are not carcinogenic or toxic, whereas compounds 6a–e are carcinogenic but nontoxic, with the exception of compound 6d. There was an agreement between the in silico and in vitro results, which showed that compounds 4a and 6c are the best candidates to serve as EGFR and VEGFR-2 tyrosine kinase inhibitors, respectively. Compounds 4a and 6c caused cytotoxicity in cancer cells by inducing both apoptosis and necrosis. The findings of this study lay the groundwork for future research into the differences between the newly synthesized compounds and the reference compounds erlotinib and sorafenib. However, more testing will be required to examine all of these variations and identify the most promising molecules for future research.

Funding Statement

This work was supported partially by a grant (Award no. 286/2022) from the Deanship of Scientific Research, The Hashemite University, Jordan.

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/17568919.2024.2347084

Financial disclosure

This work was supported partially by a grant (Award no. 286/2022) from the Deanship of Scientific Research, The Hashemite University, Jordan. The authors have no financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending or royalties.

Competing interests disclosure

The authors have no competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.