Targeting VEGFR-2 in breast cancer: synthesis and in silico and in vitro characterization of quinoxaline-based inhibitors

Abstract

A novel series of quinoxaline derivatives was designed and synthesized to target VEGFR-2, a receptor critical in cancer progression, with a focus on favorable pharmacophoric features. Among these derivatives, compound 11d emerged as a promising candidate, exhibiting potent cytotoxicity against MDA-MB-231 and MCF-7 cancer cell lines, with IC₅₀ values of 21.68 μM and 35.81 μM, respectively, while displaying significantly reduced toxicity in normal cell lines WI-38 and WISH (IC₅₀ values of 82.46 μM and 75.27 μM). Compared to standard treatments doxorubicin and sorafenib, compound 11d demonstrated a favorable therapeutic window. Inhibition assays showed that 11d inhibits VEGFR-2 with an IC₅₀ of 62.26 nM ± 2.77, comparable to sorafenib. Mechanistically, treatment with 11d upregulated pro-apoptotic markers BAX, caspase-8, and caspase-9, while downregulating the anti-apoptotic marker Bcl-2, resulting in a significant BAX/Bcl-2 ratio increase (16.11). A wound healing assay confirmed 11d's anti-migratory effects, limiting wound closure in MDA-MB-231 cells to 27.51% compared to untreated cells. Additionally, flow cytometry revealed that 11d induced both early (46.43%) and late apoptosis (31.49%) in MDA-MB-231 cells, alongside G1 phase cell cycle arrest, reducing S and G2/M phase progression. Molecular docking and dynamics simulations over 200 ns demonstrated stable binding of compound 11d to VEGFR-2, with docking scores superior and comparable to sorafenib. Density Functional Theory (DFT) calculations underscored 11d's stability and reactivity, while in silico ADMET analysis predicted a favorable safety profile over sorafenib, particularly with respect to carcinogenic and chronic toxicity risks. These findings indicate that quinoxaline derivative 11d holds potential as a selective and effective VEGFR-2 inhibitor with promising antitumor and anti-metastatic properties, warranting further investigation.

---

1. Introduction

Cancer remains one of the leading causes of mortality worldwide,¹ with breast cancer being one of the most prevalent and challenging forms to treat due to its potential for metastasis and resistance to standard therapies.² According to the latest global cancer statistics, breast cancer accounted for approximately 2.26 million new cases in 2020 and remains the leading cause of cancer-related deaths among women globally.³ Among the therapeutic targets in cancer treatment, vascular endothelial growth factor receptor 2 (VEGFR-2) has gained significant attention for its role in promoting angiogenesis, which supports tumor growth and metastatic spread.^4,5 Targeting VEGFR-2 with specific inhibitors can disrupt angiogenesis, thereby impeding tumor progression and offering a potential strategy for cancer management.^6,7

Quinoxaline derivatives have emerged as promising scaffolds in anti-cancer drug development,⁸ particularly in the context of breast cancer.⁹ Structurally, quinoxalines offer versatile chemical frameworks that can be modified to enhance selectivity and potency toward specific molecular targets, including VEGFR-2.¹⁰ Previous studies have shown that quinoxaline-based compounds can effectively inhibit cancer cell proliferation and metastasis, though further exploration is needed to optimize their efficacy and safety profiles.¹¹

Building on these findings and our ongoing efforts to develop novel anticancer agents,^12–17 particularly quinoxaline derivatives,^18–22 we designed, synthesized, and evaluated a new series of quinoxaline derivatives targeting VEGFR-2 as potential anti-breast cancer agents. These compounds were designed based on key pharmacophoric features known to interact with VEGFR-2, aiming to maximize binding affinity and biological activity. Among the synthesized derivatives, compound 11d showed particular promise, exhibiting selective cytotoxicity in cancer cells. In addition, 11d demonstrated significant anti-migratory activity, induced apoptosis, and caused cell cycle arrest, indicating its potential as a multi-faceted anti-cancer agent.

To provide a comprehensive evaluation, we complemented our experimental findings with in silico analyses, including molecular docking, molecular dynamics (MD) simulations, density functional theory (DFT) studies, and ADMET profiling. These computational approaches allowed us to investigate the stability, reactivity, pharmacokinetic, and toxicological properties of compound 11d, further supporting its potential as a targeted cancer therapeutic. The promising results from this study underscore the therapeutic potential of quinoxaline derivatives in the development of new VEGFR-2 inhibitors and highlight the need for further preclinical evaluation of compound 11d as an anti-cancer candidate.

1.1. Rationale

Several VEGFR-2 inhibitors have received FDA approval for the clinical treatment of cancer.^23–29 Fig. 1 shows some VEGFR-2 inhibitors as sorafenib I,^30,31 regorafenib II,³² and lenvatinib III,³³ which are urea-based derivatives.^34–38 These drugs are approved for various cancers, including prostate cancer, thyroid cancer, melanoma, renal cell carcinoma, and hepatocellular carcinoma.^39,40 However, their use is often associated with adverse effects such as cardiovascular complications, diarrhea, renal impairment, fatigue, hypertension, reduced appetite, nausea, thrombocytopenia, and proteinuria.^41–45

Fig. 1 Design rationale for the synthesized quinoxaline derivatives.

Our research group recently introduced compound IV, a bis([1,2,4]triazolo)[4,3-a:3′,4′-c]quinoxaline derivative, as a promising VEGFR-2 inhibitor with notable anti-proliferative activity against HepG-2 and MCF-7 cell lines. Compound IV effectively induced cell cycle arrest in HepG-2 cells at the G2/M phase and exhibited a strong pro-apoptotic effect. Additionally, it significantly increased the levels of caspase-3, caspase-9, and BAX while reducing the expression of Bcl-2 in treated cells.⁴⁶

The chemical structure of VEGFR-2 inhibitors must include four key pharmacophoric features for effective binding at the VEGFR-2 active site (Fig. 1). (i) The first feature is a heteroaromatic moiety, essential for forming hydrogen bonds with Cys917 in the hinge region of the ATP-binding site.⁴⁷ (ii) The second is a spacer group that occupies the space between the hinge region and the DFG domain (Asp–Phe–Gly motif located in the activation loop of protein kinases, which plays a critical role in regulating kinase activity and ATP binding).⁴⁸ (iii) The third feature is a pharmacophore comprising hydrogen bond donor and acceptor groups, enabling interactions with Glu883 and Asp1044 in the DFG domain.⁴⁹ (iv) The fourth is a terminal hydrophobic moiety, which engages in hydrophobic interactions with the allosteric hydrophobic pocket of the active site.^50–52

The rationale in this research focused on modifying compound IV to develop new anti-proliferative agents targeting VEGFR-2 with enhanced apoptotic potential. The newly designed compounds incorporate bis([1,2,4]triazolo)[4,3-a:3′,4′-c]quinoxaline and urea functional groups. The design strategy preserved the bis([1,2,4]triazolo)[4,3-a:3′,4′-c]quinoxaline moiety from compound IV as a heterocyclic core due to its established biological benefits. This moiety contains four nitrogen atoms that act as electron acceptors, facilitating hydrogen bonding in the hinge region. Additionally, its planar structure promotes hydrophobic interactions within the hinge region. Previous studies have highlighted its promising anti-proliferative activity.^53–55 The second modification line included a phenoxy group, as observed in compounds I, II, and III, which served as a linker. The third modification line incorporated the urea group from compounds I, II, and III, functioning as a pharmacophore. Lastly, various substituted aromatic groups were introduced as hydrophobic tails to explore the structure–activity relationship (Fig. 1).

2. Results and discussions

2.1. Chemistry

The chemical processes for furnishing the target molecules are shown in Scheme 1. Initially, o-phenylenediamine 1 reacted with oxalic acid 2 in the presence of 4 N HCl to get 2,3-(1H,4H)-quinoxalinedione 3.^56,57 Subsequent treatment of compound 3 with thionyl chloride yielded 2,3-dichloroquinoxaline 4. The reaction of compound 4 with hydrazine hydrate at ambient temperature yielded 2-chloro-3-hydrazinylquinoxaline 5.^56,57 Subsequently, compound 5 was subjected to heating with triethyl orthoformate, resulting in the formation of 4-chloro[1,2,4]triazolo[4,3-a]quinoxaline 7.^56,57 The subsequent step involved the reaction of isocyanate derivatives including phenyl isocyanate, 3-chlorophenyl isocyanate, 4-chlorophenyl isocyanate, and 3-methoxyphenyl isocyanate 9a–d, respectively with para-aminophenol 8 in acetonitrile, resulting in the formation of the key intermediates 10a–d.⁵⁸ Finally, a nucleophilic substitution reaction of the formed intermediates 10a–d with compound 7 in the presence of triethylamine (TEA) in THF resulted in the synthesis of target compounds 11a–d, respectively with commendably high yield.

Scheme 1 Synthesis of the final compounds 11a–d.

Spectral data were used to describe compounds 11a–d. These compounds' IR spectra revealed prominent NH bands at 3331–3103 cm⁻¹. Additionally, it had significant C=O absorption bands between 1654 and 1682 cm⁻¹. Moreover, the triazolo CH protons of the four derivatives were identified by singlet signals in ¹H NMR spectra with δ 10.13 and 10.17 ppm. These results were supported by ¹³C NMR spectra, which revealed distinctive peaks for the corresponding carbons.

2.2. Biological evaluation

2.2.1. Cytotoxicity and selectivity. The in vitro cytotoxicity (IC₅₀) values of the synthesized compounds 11a, 11b, 11c, and 11d against breast cancer cell lines MDA-MB-231 and MCF-7 are summarized in Table 1, with comparisons to sorafenib. Among the tested compounds, 11d showed the highest potency. Compound 11d showed notable cytotoxicity in cancer cell lines, with IC₅₀ values of 21.68 μM in MDA-MB-231 and 35.81 μM in MCF-7, though it is less potent than sorafenib (7.64 μM and 7.26 μM, respectively).

Table 1 In vitro cytotoxicity (IC₅₀) of 11a, 11b, 11c, 11d and sorafenib and against breast cancer and normal cell lines

Comp.	In vitro cytotoxicity IC50^a (μM)
	MDA-231	MCF-7	WI-38	WISH
a The values are expressed as the mean ± SEM from three independent experiments.
Sorafenib	7.64 ± 0.4	7.26 ± 0.3	10.65 ± 0.8	13.45 ± 1.1
11a	68.40 ± 3.7	77.98 ± 4.1	—	—
11c	76.23 ± 3.9	69.82 ± 3.6	—	—
11d	21.68 ± 1.5	35.81 ± 2.3	82.46 ± 4.2	75.27 ± 3.9
11b	33.56 ± 2.1	46.53 ± 2.7	—	—

---

In normal cell lines WI-38 and WISH, 11d has significantly higher IC₅₀ values (82.46 μM and 75.27 μM, respectively), indicating lower toxicity compared to sorafenib, which have IC₅₀ values of 10.65 μM and 13.45 μM, respectively. The selectivity index (SI), calculated as the ratio of IC₅₀ in normal cells (WI-38 or WISH) to that in cancer cells (MDA-231 or MCF-7), highlights the preferential cytotoxicity of compounds. For sorafenib, the SI ranges from 1.39 to 1.85, indicating limited selectivity between cancer and normal cells.

In contrast, compound 11d exhibited significantly higher SI values, ranging from 3.80 to 11.29, demonstrating greater selectivity for cancer cells over normal cells. This selective cytotoxicity of 11d suggests a promising therapeutic window, as it appears to be more effective against cancer cells than normal cells, making it a potentially safer candidate for further development.

Compound 11b also demonstrated moderate cytotoxicity, particularly in MDA-MB-231 cells (IC₅₀ = 33.56 μM), while 11a and 11c exhibited weaker activity, with IC₅₀ values above 68 μM in both cell lines. These findings suggested that 11d holds potential as a lead compound for further investigation, particularly due to its greater cytotoxic effect compared to the other compounds in the series.

2.2.2. VEGFR-2 inhibition. In order to examine the design, the ability of compound 11d to inhibited VEGFR-2 was evaluated, yielding an IC₅₀ of 62.26 nM ± 2.77 (Table 2). In comparison, sorafenib, a well-known VEGFR-2 inhibitor, showed a slightly lower IC₅₀ of 53.32 nM ± 2.52. While sorafenib exhibited marginally greater potency, the slight difference in IC₅₀ values suggests that 11d is a promising candidate for further development. The low SEM value indicates consistent and reliable results, underscoring the potential for optimizing 11d as a VEGFR-2 inhibitor.

Table 2 IC₅₀ values of 11d and sorafenib against VEGFR-2 kinase^a

Comp.	VEGFR-2, IC50 (nM) ± SEM
a Values are given as mean ± SEM of three independent experiments.
11d	62.26 ± 2.77
Sorafenib	53.32 ± 2.52

---

From cytotoxicity, selectivity, and VEGFR-2 assays, was can notice that compound 11d exhibited an IC₅₀ value of 62.26 nM against VEGFR-2, while sorafenib showed a slightly lower IC₅₀ of 53.32 nM. While this difference might seem small, its clinical relevance extends beyond mere numerical comparison.

A slight reduction in potency could necessitate a higher therapeutic dose of 11d compared to sorafenib to achieve the same level of VEGFR-2 inhibition. However, this must be weighed against the broader therapeutic window that 11d demonstrates, as indicated by its higher selectivity index (SI), meaning it spares normal cells more effectively than sorafenib. This could reduce the incidence of off-target toxicities, making dose escalation more feasible.

Sorafenib, while potent, is associated with several adverse effects, including hepatotoxicity, hypertension, and fatigue. The significantly lower toxicity of 11d in normal cells (WI-38 and WISH, IC₅₀ = 82.46 μM and 75.27 μM, respectively) compared to sorafenib (IC₅₀ = 10.65 μM and 13.45 μM) suggests that even if a slightly higher dose of 11d is required for equivalent VEGFR-2 inhibition, it may still offer a superior safety profile.

The balance between potency and selectivity is critical for drug development. Although 11d's IC₅₀ is slightly higher than sorafenib's, its higher selectivity and reduced toxicity could make it a preferable candidate for long-term therapy, especially in patients susceptible to sorafenib's side effects. Additionally, further formulation strategies (e.g., prodrugs, nano-formulations) could enhance 11d's bioavailability and therapeutic efficacy, mitigating the impact of its slightly higher IC₅₀.

2.2.3. Cancer cells migration and healing assay. The cancer cell healing assay was used to evaluate the migratory potential of MDA-231 cells, providing insights into the 11d's ability to inhibit cell migration, a key process in metastasis (Fig. 2). In this assay, a “wound” is created in a confluent monolayer of cells, and the closure of the wound over time is measured as an indicator of cell migration. The results of this assay for compound 11d demonstrate its anti-migratory potential in MDA-231 cells at its 0.5 IC₅₀ concentration. After 48 hours of treatment, as shown in Table 3, untreated cells showed a wound closure of 58.95%, reflecting typical cell migration activity. However, treatment with 11d reduced wound closure to 27.51%, indicating a significant inhibition of migration and suggesting that 11d interferes with the cells' ability to move across the wound area. These results highlight the compound's potential as an anti-metastatic agent.

Fig. 2 Microscopic images illustrating the inhibitory effect of compound 11d (11 μM) on the migration of the MDA-231 cell lines (B), compared to the control, untreated MDA-231 cell line, (A).

Table 3 Effect of 11d (0.5 IC₅₀) on migration and wound closure in MDA-231 cells after 48 hours

Sample	Quantitative closure (%) for scratched assay after 48 h
Control (MDA-231)	58.95
Compound 11d (11 μM)	27.51

---

2.2.4. Apoptosis assay. The data in Table 4 and Fig. 3 provide insights into the effects of compound 11d on cell viability and stages of cell death in MDA-MB-231 cells after 48 hours of treatment, compared with untreated cells. Treatment with compound 11d resulted in a significant reduction in viable MDA-MB-231 cells, with only 20.17% of cells remaining viable compared to 95.48% in the untreated control. Regarding apoptosis, compound 11d induced a notable increase in apoptotic cell populations. Early apoptosis was observed in 46.43% of cells treated with 11d. Late apoptosis was also induced at a high level (31.49%) by 11d. These findings indicate that 11d effectively promotes apoptotic cell death, with a substantial proportion of cells progressing to later stages of apoptosis. It's worthy to mention that minimal necrosis was observed in both treated groups, with 11d causing necrosis in only 1.91% of cells. The low levels of necrosis suggest that the primary mechanism of cell death induced by 11d is apoptotic rather than necrotic, which is often preferable in cancer treatment to avoid inflammatory responses associated with necrosis.

Table 4 Impact of 11d on different stages of the cell death process in MDA-MB-231 cells following 48 h of treatment

Sample	Viable^b (left bottom)	Apoptosis^b		Necrosis^b (left top)
		Early (right bottom)	Late (right top)
a Significant P value significant P value < 0.05 & by using one-way ANOVA followed by Tukey's post hoc multiple comparison tests.b Values represent stages of the cell death process in MDA-MB-231cells treated with or without 11d.
MDA-MB-231	95.48	3.84	0.45	0.32
11d/MDA-MB-231	20.17^a	46.43^a	31.49^a	1.91^a

---

Fig. 3 The effect of compound 11d on the cell viability of the MDA-231 cell lines.

2.2.5. MDA-MB-231 cell cycle analysis. MDA-MB-231 cell cycle analysis after 11d treatment (Table 5 and Fig. 4) indicated that compound 11d primarily inducing G1 phase arrest, with 64.06% of cells in the G1 phase after 48 hours, compared to 44.98% in untreated cells. This shift limited progression into the S and G2/M phases, as seen in the lower S phase (16.45%) and G2/M phase (3.74%) populations, which suggests compound 11d effectively halts cell cycle advancement. Furthermore, compound 11d showed a lower sub-G1 population (15.75%) compared to the untreated cells (19.48%). Overall, compound 11d's ability to arrest cells in G1 may offer an alternative pathway to reduce cell proliferation in cancer treatment, warranting further investigation.

Table 5 Impact of 11d on cell cycle progression in MDA-MB-231 cells after 48 h of treatment

Sample	Cell cycle distribution^b (%)
	% sub-G1	% G1	% S	% G2/M
a Significant P value significant P value < 0.05 & by using one-way ANOVA followed by Tukey's post hoc multiple comparison tests.b Cell cycle distribution (%) of MDA-MB-231 cells were treated with or without compound 11d.
MDA-MB-231	19.48	44.98	24.95	10.59
11d/MDA-MB-231	15.75	64.06^a	16.45^a	3.74^a

---

Fig. 4 Impact of compound 11d on cell cycle progression in MDA-MB-231 cells following 48 hour treatment.

2.2.6. Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) analysis. The mRNA expression levels of apoptosis-related genes, including BAX, Bcl-2, caspase-8, and caspase-9, were analyzed in control and compound 11d-treated MDA-231 cells (21.68 μM) using qRT-PCR. Treatment with 11d (Table 6) significantly increased the expression of pro-apoptotic genes BAX, caspase-8, and caspase-9, with fold changes of 5.48, 4.15, and 5.58, respectively, as well as a notable decrease in anti-apoptotic Bcl-2 expression (0.34). This led to a marked increase in the BAX/Bcl-2 ratio (16.11), a strong indicator of apoptotic potential. In comparison, sorafenib also upregulated BAX and caspases but to a lesser extent, with a BAX/Bcl-2 ratio of 6.10. These results suggest that 11d induced a more pronounced pro-apoptotic response, highlighting its potential as a promising therapeutic candidate for apoptosis induction in MDA-231 cells.

Table 6 Effect of compound 11d (21.68 μM) on levels of BAX, Bcl-2, caspase-8, and caspase-9 genes expression in MDA-231 cells

Sample	Gene expression (fold change)^b
	BAX	Bcl-2	BAX/Bcl-2 ratio	Caspases-8	Caspases-9
a Significant P value significant P value < 0.05 & by using one-way ANOVA followed by Tukey's post hoc multiple comparison tests.b Values are given as changes from the corresponding control (MDA-231 cells) group. Data from three independent experiments represent the mean ± SEM, as the fold changes, with control is set to ‘1’.
MDA-231 cells	1.00 ± 0.08	1.00 ± 0.07	1.00 ± 0.13	1.00 ± 0.09	1.00 ± 0.11
11d	5.48 ± 0.45^a	0.34 ± 0.01^a	16.11 ± 1.01	4.15 ± 0.48^a	5.58 ± 0.40^a

---

2.2.7. Statistical analysis section. GraphPad Prism 6 software (GraphPad, San Diego, CA, USA) was used to analyze all results. The data were represented as mean ± standard deviation (mean ± SD) from at least three independent experiments (n = 3). One-way analysis of variance (one-way ANOVA) followed by Tukey post hoc multiple comparison tests were elected to analyze the significant difference between all groups. P < 0.05 was considered statistically significant.

2.3. Computational evaluations

2.3.1. Molecular docking study. To further understand the biological activity of the newly synthesized derivatives (11a–d) and their binding interactions at the molecular level, docking simulations were conducted. The re-docking of the native ligand at the active site of VEGFR-2 has conclusively validated the docking procedure. The results indicated a significant overlap between the native and docked ligands, with an RMSD of 0.89 Å, as illustrated in Fig. 5A.

Fig. 5 (A) Overlay of the co-crystallized ligand (green) and the redocked ligand (turquoise) in the VEGFR-2 active site. 2D interaction of (B) sorafenib, (C) 11a, (D) 11b, (E) 11c, and (F) 11d, against VEGFR-2.

The docking simulations evaluated the compounds' ability to interact with the active site of VEGFR-2. The docking scores for derivatives 11a–d ranged from −22.71 to −23.65 kcal mol⁻¹, indicating strong binding affinities (Table 7). Notably, these scores were comparable and higher than those of sorafenib, a known VEGFR-2 inhibitor, suggesting that the new derivatives may exhibit enhanced binding efficiency. The binding score of sorafenib was determined to be −22.50 kcal mol⁻¹. An examination of sorafenib's binding pattern indicated alignment with the reported findings, where the hydrophobic moiety (3 chloro-4-trifluoromethylphenyl) was positioned within the allosteric site. Additionally, two significant hydrogen bonds were identified between the urea linker and Glu883 and Asp1044. Also, Cys917 forms two hydrogen bonds with the N-methylpicolinamide arm (see Fig. 5B).

Table 7 The S score values for the reference inhibitor (sorafenib) against the novel compounds 11a–d

Comp.	11a	11b	11c	11d	Sorafenib
Energy score (kcal mol^−1)	−22.98	−23.02	−22.71	−23.65	−22.50

---

The docking results showed that the four synthesized derivatives nearly matched the orientation and position of sorafenib in the active binding region of the VEGFR-2 enzyme (Fig. 5). The derivatives' amide groups maintained two necessary hydrogen bonds with Glu883 and Asp1044 in the DFG domain. Interestingly, the hydrophobic moieties of the synthesized candidates (phenyl, 3-chlorophenyl, 4-chlorophenyl, and 3-methoxyphenyl) behaved similarly to sorafenib and occupied the VEGFR-2 enzyme's allosteric site. Such hydrophobic moieties formed several hydrophobic bonds with Leu886, Leu887, Leu1017, and Ile890. Additionally, in the hinge area, the [1,2,4]triazolo[4,3-a]quinoxaline moiety of compounds 11a–d established a crucial hydrogen bond with Cys917 and shared a number of hydrophobic contacts with Ala864, Leu1033, Phe1045, and Phe916. Moreover, the central phenyl group of the synthesized derivatives formed several hydrophobic interactions with Val846, Val914, and Lys866 besides one electrostatic interaction with Cys1043 (Fig. 5C–F).

Molecular docking results indicate that 11d has a docking score of −23.65 kcal mol⁻¹, which is slightly better than sorafenib (−22.50 kcal mol⁻¹), suggesting a strong binding affinity for VEGFR-2. However, a deeper structural comparison with other VEGFR-2 inhibitors will enhance our understanding of its potential advantages.

2.3.1.1. Key binding interactions. Compound 11d interacts with Cys917 (hinge region), Glu883 and Asp1044 (DFG domain), and Leu886, Leu887, and Ile890 (hydrophobic pocket). These interactions mirror those of known VEGFR-2 inhibitors like sorafenib and lenvatinib, reinforcing the rationale for its design.

This good binding mode is attributed to the high number of hydrophobic interaction caused by the planar bis([1,2,4]triazolo)[4,3-a:3′,4′-c]quinoxaline moiety. This moiety contains four nitrogen atoms that act as electron acceptors, facilitating hydrogen bonding in the hinge region. Additionally, its planar structure promoted hydrophobic interactions within the hinge region.

2.3.2. MD simulations studies. MD simulations revealed distinct differences in the dynamics of VEGFR-2 upon binding 11d compared to sorafenib before 50 ns the VEGFR-2 backbone in the VEGFR-2_sorafenib complex (dark green line) (Fig. 6A) maintained stable RMSD values around 2 Å and VEGFR-2 backbone in the VEGFR-2_11d complex (red line) (Fig. 6A) maintained stable RMSD values around 4 Å. Throughout the simulation VEGFR-2 in the VEGFR-2_11d complex (red line) decreased to 3 Å and returned increase after 150 ns to 4 Å until the end of simulation. In contrast, VEGFR-2_sorafenib displayed a similar trend until initially but exhibited an RMSD increase after 80 ns, stabilizing around 3 Å for the remainder of the simulation. Ligand RMSD profiles (Fig. 6B) suggest stable binding in both systems. 11d (red) displayed a stable average 2.5 Å. Also, sorafenib showed a stable RMSD throughout the whole simulation with an average of 1.5 Å. Compound 11d's RMSD profile (Fig. 6B) displayed stable binding in both systems. Average RoG and SASA values (Fig. 6C and D) indicated no significant changes in the protein's overall structure. However, hydrogen bond analysis (Fig. 6E) revealed distinct interaction patterns. Sorafenib formed a relatively constant two hydrogen bond throughout the simulation, with a small number of frames showing three H-bonds. 11d, on the other hand, exhibited an increased number of frames with three H-bonds hydrogen bond analysis (Fig. 6E) revealed that the complex formed a relatively constant four hydrogen bonds throughout the simulation, with a few frames showing three H-bonds. RMSF profiles of C-alpha atoms (Fig. 6F) showed nearly identical fluctuations for both systems except for the LYS995 and LYS1060 displayed higher RMSF in the VEGFR-2_11d system compared to the reference system. LYS995 (up to 4.9 Å) and LYS1060 (maximum 7.68 Å), reflecting their increased mobility associated with LYS1060 is close to the next residue ASP110 at the end of the simulation. The center-of-mass distance analysis (Fig. 6G) indicated stable ligand binding in both systems. 11d maintained a slightly higher average distance (up to 9 Å) compared to sorafenib (up to 8 Å). Overall, these findings suggest that both systems retain overall structural stability. The ligands exhibit distinct binding dynamics, with 11d showing more flexibility compared to the tighter interactions of sorafenib.

Fig. 6 MD analysis parameters: (A) RMSD values from the trajectory for the VEGFR-2 protein in VEGFR-2_11d (red line) and VEGFR-2_sorafenib complex (dark green line), (B) shows the ligands RMSD values, (C) radius of gyration for the VEGFR-2 protein in VEGFR-2_11d (red line) and VEGFR-2_sorafenib complex (dark green line), (D) SASA for VEGFR-2 protein in VEGFR-2_11d (red line) and VEGFR-2_sorafenib complex (dark green line), (E) change in the number of hydrogen bonds between 11d (pink line) or sorafenib (brown line) and VEGFR-2, (F) RMSF for VEGFR-2 protein in VEGFR-2_11d (red line) and VEGFR-2_sorafenib complex (dark green line), (G) distance from the center of mass of sorafenib or 11d compound and VEGFR-2 protein.

2.3.3. MM-GBSA analysis. Fig. 7 unveils the intricate components of the predicted binding free energies calculated using the MM-GBSA method. Despite exhibiting a slightly lower overall binding affinity (−43.3 kcal mol⁻¹) compared to sorafenib (−51.3 kcal mol⁻¹), 11d demonstrates interesting similarities and dissimilarities in its interaction profile. Both ligands benefit from robust van der Waals interactions averaging around −53 kcal mol⁻¹. However, when considering electrostatic contributions, 11d shows a significantly more favorable electrostatic term (−21.16 kcal mol⁻¹) compared to sorafenib (−8.9 kcal mol⁻¹). To pinpoint the contribution of amino acid residues within 1 nm of each ligand, decomposition analysis was performed (Fig. 8). Interestingly, twelve residues exhibit enhanced contributions to 11d binding compared to sorafenib. These key amino acids are Leu838 (−1.5 kcal mol⁻¹ for 11d vs. −1.3 kcal mol⁻¹ for sorafenib), Val 846 (−1.07 kcal mol⁻¹ for 11d vs. −1.2 kcal mol⁻¹ for sorafenib), Ala864 (−1.09 kcal mol⁻¹), Leu887 (−1.5 kcal mol⁻¹ for 11d vs. −1.59 kcal mol⁻¹ for sorafenib), Val897 (−1.35 kcal mol⁻¹ for 11d vs. −1.06 kcal mol⁻¹ for sorafenib), Val914 (−1.5 kcal mol⁻¹ for 11d vs. −1.2 kcal mol⁻¹ for sorafenib), Phe916 (−2,5 kcal mol⁻¹ for 11d vs. −1 kcal mol⁻¹ for sorafenib), Cys917 (−2.2 kcal mol⁻¹ for 11d), Leu1033 (−1.6 kcal mol⁻¹ for 11d vs. −1.2 kcal mol⁻¹ for sorafenib), Cys1043 (−3 kcal mol⁻¹ for 11d vs. −2.3 kcal mol⁻¹ for sorafenib), Asp1044 (−1.6 kcal mol⁻¹ for 11d vs. −1.9 kcal mol⁻¹ for sorafenib), Phe1045 (−1.1 kcal mol⁻¹ for sorafenib).

Fig. 7 MM-GBSA analysis of the VEGFR-2-11d complex and VEGFR-2-sorafenib complex. Bars represent the standard deviations.

Fig. 8 VEGFR-2-11d complex and VEGFR-2-sorafenib complex binding free energy decomposition.

2.3.4. Protein–ligand interaction fingerprint (ProLIF) analysis. Protein–Ligand Interaction Fingerprints (ProLIF) analysis shed light on the specific amino acids involved in ligand interactions within the VEGFR-2 binding pocket. Fig. 9A–C and 10 represent amino acids that exhibited a high propensity for hydrophobic interactions (H.I.), van der Waals (VdW) contact, and hydrogen bond acceptor (HBA) with a ratio of >80% occurrence, respectively. For 11d (Fig. 9) these key amino acids are LEU (98%, H.I.), VAL846 (99.5%, H.I.), ALA864 (99.5%, H.I.), LYS866 (91%, H.I.), GLU883 (84%) H.I., ILE886 (90.3%) H.I., LEU887 (98.3% H.I.), VAL897 (89.2% H.I.), VAL914 (99.9% H.I.), GLU915(98.6%, VdW), and (97.4%, HBD), PHE916 (90.4%, H.I.), CYS917 (98.7%, VdW), and (92%, HBA), LEU1033 (99.8%, H.I.), CYS1043 (94.4%, H.I.), ASP1044 (93.8%, H.I.), (99.9%, VdW) and (99%, HBA). In contrast, For sorafenib (Fig. 10) 13 amino acids exhibited a high propensity for hydrophobic interactions (>80% occurrence). Key amino acids are Leu838 (91.2%), Val846 (98.6%), Lys866 (87.9%), Ile886 (92.9%), Leu887 (95.1%), Ile890 (95.2%), Val897 (88.9%), Val914 (99.7%), Leu1017 (95.4%), His1024 (86.6%), Leu1033 (91%), Cys1043 (99.8%), and Asp1044 (99.3%). Furthermore, Asp1044 contributed via hydrogen bonding, with a formation rate of 88.2%.

Fig. 9 ProLIF assay: the ProLIF Python library was used to analyze the amino acids involved, the types of interactions within compound 11d-VEGFR-2 complex, and their frequency throughout the entire simulation. Panel (A) represents hydrophobic interactions (H.I.), panel (B) depicts van der Waals (VdW) contacts, and panel (C) shows hydrogen bond acceptor (HBA) interactions.

Fig. 10 ProLIF assay: the ProLIF Python library was used to analyze the amino acids involved, the types of interactions within compound sorafenib-VEGFR-2 complex, and their frequency throughout the entire simulation.

2.3.5. Essential dynamics. Principal Component Analysis (PCA) was our preferred technique for determining the coordinates of the α-carbon atoms in the VEGFR-2 complex. The right dimensions of the reduced subspace must be carefully determined taking into account a number of the parameters mentioned in the ESI Methods section.† Among them was the cumulative variation that could be accounted for by a larger number of principal components (PCs), the scree plot, and the eigenvector distribution. The scree plot (Fig. 11A) showed that the slope began to significantly reduce at the third PC, indicating a possible point of inflection for the selection of dimensionality. As seen in Fig. 11A, the first eigenvector alone accounted for a significant 76.17% of the overall variance, while the first three PCs combined contributed roughly 86.24%. This suggested that these three PCs accounted for a sizable fraction of the total protein movements. The distributions of the first five PCs differed from the Gaussian distribution, which further supported this decision (Fig. 11B). A non-Gaussian distribution often indicates significant, non-random motions that these components can capture. Accordingly, we selected the top five PCs to illustrate the essential subspace using a combination of non-Gaussian eigenvector distributions, variance collected, and scree plot. We then examined the Root Mean Square Inner Product (RMSIP) and cosine content of the complex's first 10 principal components (PCs) to assess any potential redundancy and unpredictability within the indicated key subspaces. We subsequently examined the cosine content of the first ten principal components (PCs) for the VEGFR-2 system (Fig. 11C). Except for the second PCs (0.23 for VEGFR-2_sorafenib and 0.45 for VEGFR-2_11d) and the third PC of VEGFR-2_11d (0.10), all other cosine values for the first ten PCs remained below 0.2. This suggests that the identified essential motions captured by the PCs are not random. The RMSIP analysis indicated limited overlap between the two covariance matrices. Focusing on the initial three PCs, which capture the most significant variance, the RMSIP value was only 23.9%. This low value implies that the dominant collective motions represented by these initial PCs are largely non-overlapping between the two systems. Additionally, the overall RMSIP analysis revealed that the C matrices, encompassing the complete set of eigenvectors, were only 35% similar. This underscores the distinct sampling characteristics and dynamic landscapes experienced by VEGFR-2 in each complex.

Fig. 11 Essential dynamics analysis for the first ten eigenvectors of VEGFER-2-11d complex and VEGFER-2-sorafenib complex. (A) Eigenvalues changing, (B) the eigenvectors' distribution, and (C) cosine values.

2.3.6. Free energy landscape (FEL) analysis. The projected trajectories onto various two-dimensional planes defined by the chosen PCs are displayed in Fig. 12. Every graphic displays a different landscape with different basins that correspond to local minima on the FEL. These basins correspond to preferred conformations adopted by the protein–ligand complex during the simulation. For the projection on the first two PCs (Fig. 12A), the sampling starts from (white dots) basin and then moves to a transient basin (white and grey dots) before finally reaching the most stable one (black dots). There is a 2.090 kJ mol⁻¹ difference between the global minimum and the next minimum. Fig. 12B presents the projection on PC1–PC3, which shows a projection with three basins, the trajectories move from the transient basin in the middle (white dots) to the stable basin (white grey and dark dots). The difference in energy between the global and the next minima is calculated to be 0.27 kJ mol⁻¹. Similarly, Fig. 12C shows the projection on PC2 and PC3, revealing a wide basin and a narrow basin. The trajectory begins in the wide basin and then transitions to the narrow basins (the most stable basin) at the end of the simulation (black dots). The difference between the second and global minima is 0.0 kJ mol⁻¹.

Fig. 12 The 2D and 3D projections of the VEGFR-2_11d complex trajectories FEL on (A) the first two, (B) first and third and (C) second and third eigenvectors.

2.3.7. Density functional theory (DFT) calculations. DFT calculations for compound 11d were carried out using Gaussian 09 software with the 6-31+G(d, p) basis set to investigate its stability, reactivity, and electrical properties following its structural optimization. Under the same theoretical framework, the quantum investigations of global reactivity, energy gap (E_gap) and molecular electrostatic surface potential (MESP) were also carried out. In Fig. 13A, the optimized structure of 11d is shown. Following the optimization process, Table 8's computed dipole moment (D_m) value (7.301 debye) was comparatively high. A higher degree of dipole moment corresponds to a larger degree of charge separation in 11d, so the D_m value is used as an indicator of charge separation in 11d. Since dipole–dipole interactions are common among molecules with strong dipole moments, a higher DM value indicates a polar molecule and can provide information about its reactivity. These interactions can impact a molecule's reactivity in polar liquids.⁵⁹

Fig. 13 The optimized structure (A), the HOMO/LUMO distribution function and energy gap (B), molecular electrostatic potential map (C) and total density of state (D) at B3LYP/6-31+G(d,p) level for 11d.

Table 8 The global reactivity parameters for 11d calculated through the DFT

IP	EA	μ (eV)	χ (eV)	η (eV)	σ (eV)	ω (eV)	Dm (debye)	TE (eV)	ΔNmax	ΔE (eV)
5.904	2.335	−4.120	4.120	1.784	0.560	15.142	4.279	−39225.1	2.309	−15.142

---

As seen in Fig. 13B, the density of the LUMO function is mostly dispersed across 11d's right side, but the distribution of the HOMO function is virtually evenly distributed over the left side. This unique distribution raises the possibility that directionality in the electron transfer routes through 11d from HOMO to LUMO which is important in electron transfer and charge separation reactions which are relevant in drug interaction. Additionally, as the HOMO designates the area where an electron can be transferred, the target will be able to approach the right side of 11d electrophilically. It was discovered that the HOMO/LUMO energy gap (E_gap) was 3.569 eV. Table 8 displays the calculated global reactivity parameters. The electrophilicity index of 11d was determined to be 15.142 eV, indicating that it has a high electron-accepting capacity from biomolecules. Due to its high electrophilicity, the medication may interact effectively with protein nucleophilic sites. Also, 11d is soft (0.560 eV). The electrons' ability to escape is measured by chemical potential, which was found to be −4.12 eV. When a molecule loses electrons, its chemical potential is negative, indicating that it is relatively stable. A molecule is less able to donate electrons when its negative chemical potential is higher, which makes it more difficult to oxidize.⁶⁰

To identify the nucleophilic and electrophilic active sites, Fig. 13C presents the analysis and presentation of the molecular electrostatic potential, or MESP. The electron-rich site's negative potential, as indicated by the MESP map, can form hydrogen bonds with the target's electrophilic active sites by spreading over the C=O groups. The blue zones on hydrogens indicate sites that are electron-deficient have the ability to make hydrogen bonds with the target's nucleophilic sites. The target protein's hydrophobic regions can interact with the neutral green patches on the MESP map, Fig. 13C.

The overall density of states was examined using Multiwfn software, along with the number of states that might be created at each level. The HOMO energy line is shown as a dashed line in Fig. 13D. The electronic density of states is larger for orbitals higher than the LUMO, as can be seen from the total density chart. This suggests that 11d can readily receive more electrons from electron donors in biological settings. Because of its capacity to take electrons, 11d is more reactive in oxidative stress conditions, which are frequently present in cancer cells.

2.3.8. In silico ADMET analysis. To evaluate the pharmacokinetic profile of the synthesized compounds, Table 9 presents a summary of key properties for compounds 11a–d, including blood–brain barrier (BBB) permeability, solubility, absorption, CYP2D6 inhibition, and plasma protein binding (PPB), alongside sorafenib. Among the compounds, only 11a exhibited high BBB permeability, indicating potential central nervous system (CNS) activity, which could be advantageous or a limitation depending on the therapeutic application. The remaining compounds, including sorafenib, show very low BBB permeability, suggesting limited CNS exposure, which may reduce the risk of CNS-related side effects. Solubility remains a challenge, as compounds 11b, 11c, and sorafenib exhibited very low solubility, while 11a and 11d showed slightly improved but still low solubility. Despite these solubility limitations, all compounds demonstrate good absorption levels, potentially supporting sufficient bioavailability for systemic action. In terms of metabolic interaction, all compounds are predicted to be non-inhibitors of CYP2D6, reducing the risk of drug–drug interactions via this pathway. Furthermore, each compound has a PPB prediction greater than 90%, indicating high plasma protein binding, which could impact distribution and bioavailability. Overall, the good absorption and non-CYP2D6 inhibitory properties across these compounds are favorable for drug development. Compound 11a, with its unique high BBB permeability, may be especially suitable for CNS-targeted therapies, while the low CNS exposure of the other compounds may be beneficial for peripheral applications. Improving solubility could further enhance their therapeutic profiles.

Table 9 Computational pharmacokinetic profile of compounds 11a–d

Comp.	BBB level	Solubility level	Absorption level	CYP2D6 prediction	PPB prediction
11a	High	Low	Good	Non inhibitor	More than 90%
11b	Very low	Very low
11c
11d		Low
Sorafenib		Very low

---

2.3.9. In silico toxicity analysis. The results in Table 10 provide insight into the toxicological and safety profiles of compounds 11a, 11b, 11c, and 11d in comparison to sorafenib. Starting with carcinogenic potency and mutagenicity: the carcinogenic potency (TD₅₀) values indicate a general trend where compounds 11a to 11d exhibited variable carcinogenic potencies, with 11a having the highest TD₅₀ value of 38.76 mg kg⁻¹, suggesting a lower carcinogenic risk relative to sorafenib (TD₅₀ = 14.24 mg kg⁻¹). All tested compounds were predicted to be non-mutagenic by the Ames test, contrasting with sorafenib, which is a single carcinogen. This may imply a potentially safer profile in terms of mutagenic risk for these compounds. Regarding the tolerated and lethal doses (MTD and LD₅₀): in examining the maximum tolerated dose (MTD) and LD₅₀ values for these compounds in rats, the data shows that compounds 11a through 11d have MTD values ranging from 0.065 to 0.147 mg kg⁻¹, which are relatively comparable but slightly higher than sorafenib (MTD = 0.0885 mg kg⁻¹). The LD₅₀ values of 11-series compounds are lower than sorafenib (LD₅₀ = 0.822 mg kg⁻¹), which suggests that they have lower acute toxicity, although further studies are needed to confirm these findings.

Table 10 Computational toxicity profile of compounds 11a–d

Comp.	TD50 (rat)^a	Ames prediction	Mouse-female FDA	MTD	Rat oral LD50^b	Rat chronic LOAEL^b	SI	EI
a Unit: mg kg^−1 body weight per day.b Unit: g kg^−1 body weight.
11a	38.7633	Non-mutagen	Non-carcinogen	0.12016	0.203002	0.0213632	Non-irritant	Mild
11b	10.3763			0.147027	0.17387	0.017032
11c	10.3763			0.147027	0.275129	0.0158809
11d	3.12316			0.0655489	0.414131	0.0135571
Sorafenib	14.2442		Single-carcinogen	0.088543	0.822583	0.00482816

---

In the aspect of chronic exposure, the lowest observed adverse effect levels (LOAEL) indicate that compounds 11a to 11d present lower chronic toxicity levels compared to sorafenib. For instance, compound 11a, with a LOAEL of 0.021 mg kg⁻¹, is less toxic than sorafenib (LOAEL = 0.0048 mg kg⁻¹), suggesting a potentially improved safety profile under long-term exposure. Skin and (SI) Eye Irritancy (EI): none of the 11-series compounds were found to be skin irritants, and all were categorized as mild ocular irritants. This aligns with sorafenib's profile, indicating a low likelihood of dermal toxicity.

In conclusion, the toxicological profile of the 11-series compounds shows potential advantages over sorafenib, particularly in terms of carcinogenic risk, chronic toxicity, and irritancy. However, additional studies, particularly in vivo, would be necessary to confirm the safety and efficacy of these compounds as therapeutic candidates.

The computational predictions suggest that 11d exhibits notable pharmacokinetic advantages over sorafenib, particularly in terms of absorption, safety, and toxicity as follows. (i) Sorafenib is known for its poor aqueous solubility, which limits its bioavailability and necessitates high oral doses (400 mg twice daily) to achieve therapeutic plasma concentrations. The ADMET results indicate that compound 11d has slightly better solubility than sorafenib, though it remains in the low-to-moderate range. Improved solubility enhances intestinal absorption, which could allow for lower dosing requirements, reducing systemic exposure to toxic metabolites. (ii) The ADMET predictions suggest that 11d has good intestinal permeability, similar to sorafenib. (iii) Unlike compound 11a, which has high BBB permeability, 11d has low BBB penetration, similar to sorafenib. This is a desirable feature for anticancer drugs targeting VEGFR-2, as it minimizes potential neurological side effects, such as cognitive impairment, that are often associated with tyrosine kinase inhibitors (TKIs) with high CNS exposure. (iv) Both sorafenib and 11d exhibit strong plasma protein binding (>90%), meaning that most of the drug remains bound to albumin and other plasma proteins. While high PPB can limit free drug availability, it also prolongs the half-life and may contribute to sustained drug action. The comparable PPB suggests that 11d could have a half-life similar to, or potentially longer than, sorafenib, though this needs to be validated with in vivo pharmacokinetic studies. (v) Sorafenib is classified as a single carcinogen, meaning it has some potential for long-term oncogenic effects. In contrast, compound 11d is predicted to be non-mutagenic and non-carcinogenic, making it a potentially safer long-term therapy. (vi) The Maximum Tolerated Dose (MTD) of 11d (0.065 mg kg⁻¹) is slightly higher than sorafenib (0.0885 mg kg⁻¹), suggesting that 11d has a better safety margin. (vii) The Lowest Observed Adverse Effect Level (LOAEL) of 11d (0.0136 mg kg⁻¹) is higher than sorafenib (0.0048 mg kg⁻¹), reinforcing its lower chronic toxicity potential. (viii) Unlike some kinase inhibitors that cause severe dermatological toxicity, 11d is predicted to be non-irritant to the skin and a mild ocular irritant, similar to sorafenib. This suggests that 11d may be better tolerated in long-term use.

3. Conclusion

In this study, we synthesized and evaluated a series of quinoxaline derivatives as potential VEGFR-2 inhibitors, focusing on their anti-cancer properties against breast cancer cells. Compound 11d demonstrated notable selectivity and efficacy, exhibiting potent cytotoxic effects on MDA-MB-231 and MCF-7 cancer cell lines while showing reduced toxicity in normal cell lines. Mechanistic studies revealed that 11d effectively inhibited VEGFR-2, induced apoptosis by upregulating pro-apoptotic genes, and significantly impaired cancer cell migration and cell cycle progression. Furthermore, in silico analyses, including molecular docking, molecular dynamics simulations, ProLIF, PCA, FEL and DFT calculations, indicated strong and stable binding of 11d to VEGFR-2, along with a favorable ADMET profile, positioning it as a safer alternative to sorafenib. Collectively, our findings highlight 11d as a promising candidate for targeted anti-angiogenic therapy. Further preclinical studies are warranted to validate these effects in vivo and assess the therapeutic potential of 11d in combination with current cancer treatments. For future directions, we hope to carry out the following tasks. In vivo efficacy studies to evaluate the tumor suppression ability of 11d. Further studies should assess half-life, metabolism, and biodistribution in vivo. Further modifications could improve solubility and oral bioavailability. Evaluating 11d in combination with standard chemotherapies could determine synergistic effects.

4. Experimental

4.1. Chemistry

4.1.1. General procedure for synthesis of the key intermediates 10a–d. P-Aminophenol 8 (5 mmol) was added to a stirred solution of isocyanate derivatives 9a–d (13 mmol) in acetonitrile (100 mL). At room temperature, the reaction mixture was agitated, and TLC was used to track its development (DCM : methanol = 9 : 1). Compounds 10a–d were obtained by filtering, washing with acetonitrile, and drying the precipitate.⁶¹

4.1.2. General procedure for synthesis of the final target compounds 11a–d. 4-Chloro-[1,2,4]triazolo[4,3-a]quinoxaline 7 (2 mmol, 0.41 g) was added to a stirred solution of compounds 10a–d (2 mmol) and TEA (2 mmol) in tetrahydrofuran (20 mL) at reflux temperature. The reaction was then continued at room temperature and monitored using TLC. Following completion of the reaction, the precipitates were filtered and recrystallized from ethanol to provide the desired target compounds 11a–d.
4.1.2.1. 1-(4-([1,2,4]Triazolo[4,3-a]quinoxalin-4-yloxy)phenyl)-3-phenylurea 11a. White powder (yield, 75%); mp = 213–215 °C. FT-IR (ν_max, cm⁻¹): 3421, 3119 (NH), 3060 (C–H aromatic), 2925, 2830 (C–H aliphatic), 1682 (C=O); ¹H NMR (400 MHz, DMSO-d₆) δ 10.17 (s, 1H, CH triazolo), 8.88 (s, 1H, NH), 8.85 (s, 1H, NH), 8.37 (d, J = 8.1 Hz, 1H, Ar–H), 7.76–7.48 (m, 8H, Ar–H), 7.35 (dd, J = 8.6, 6.1 Hz, 4H, Ar–H); ¹³C NMR (101 MHz, DMSO-d₆) δ 153.03, 151.72, 146.74, 139.19, 138.77, 138.73, 137.75, 134.34, 129.11, 128.30, 128.21, 127.78, 125.85, 124.54, 122.70, 120.23, 120.04, 116.92; mass (m/z): 396 (M⁺, 57.84%), 381 (100%, base peak); anal. calcd for C₂₂H₁₆N₆O₂ (396.41): C, 66.66; H, 4.07; N, 21.20. Found: C, 66.92; H, 4.21; N, 21.43%.
4.1.2.2. 1-(4-([1,2,4]Triazolo[4,3-a]quinoxalin-4-yloxy)phenyl)-3-(3-chlorophenyl)urea 11b. White powder (yield, 77%); mp = 207–209 °C. FT-IR (ν_max, cm⁻¹): 3331, 3103 (NH), 3067, 3014 (C–H aromatic), 1682 (C=O); ¹H NMR (500 MHz, DMSO-d₆) δ 10.13 (s, 1H, CH triazolo), 8.91 (s, 1H, NH), 8.86 (s, 1H, NH), 8.33 (dd, J = 8.1, 1.4 Hz, 1H, Ar–H), 7.69 (q, J = 1.5 Hz, 1H, Ar–H), 7.64–7.58 (m, 2H, Ar–H), 7.56–7.50 (m, 3H, Ar–H), 7.32 (d, J = 2.2 Hz, 1H, Ar–H), 7.31–7.29 (m, 1H, Ar–H), 7.28–7.24 (m, 2H, Ar–H), 6.98 (dt, J = 6.7, 2.3 Hz, 1H, Ar–H); ¹³C NMR (126 MHz, DMSO-d₆) δ 153.06, 151.76, 147.00, 141.83, 138.81, 137.68, 134.44, 133.77, 130.91, 128.37, 128.31, 127.87, 124.64, 122.71, 122.04, 120.25, 118.23, 117.28, 116.99; mass (m/z): 430 (M⁺, 22.40%), 402 (100%, base peak); anal. calcd for C₂₂H₁₅ClN₆O₂ (430.85): C, 61.33; H, 3.51; N, 19.51. Found: C, 61.60; H, 3.79; N, 19.68%.
4.1.2.3. 1-(4-([1,2,4]Triazolo[4,3-a]quinoxalin-4-yloxy)phenyl)-3-(4-chlorophenyl)urea 11c. White powder (yield, 78%); mp = 200–202 °C. FT-IR (ν_max, cm⁻¹): 3320 (NH), 3056 (C–H aromatic), 2920 (C–H aliphatic), 1654 (C=O); ¹H NMR (500 MHz, DMSO-d₆) δ 10.13 (s, 1H, CH triazolo), 8.84 (s, 1H, NH), 8.81 (s, 1H, NH), 8.33 (dd, J = 8.2, 1.4 Hz, 1H, Ar–H), 7.65–7.58 (m, 2H, Ar–H), 7.56–7.50 (m, 3H, Ar–H), 7.49–7.45 (m, 2H, Ar–H), 7.32–7.28 (m, 4H, Ar–H); ¹³C NMR (126 MHz, DMSO-d₆) δ 153.14, 151.73, 147.05, 139.27, 138.82, 138.76, 137.79, 134.47, 129.12, 128.35, 128.32, 127.86, 126.08, 124.66, 122.62, 120.47, 120.20, 116.97; mass (m/z): 430 (M⁺, 34.27%), 48 (100%, base peak); anal. calcd for C₂₂H₁₅ClN₆O₂ (430.85): C, 61.33; H, 3.51; N, 19.51. Found: C, 61.58; H, 3.75; N, 19.59%.
4.1.2.4. 1-(4-([1,2,4]Triazolo[4,3-a]quinoxalin-4-yloxy)phenyl)-3-(3-methoxyphenyl)urea 11d. White powder (yield, 80%); mp = 208–210 °C. FT-IR (ν_max, cm⁻¹): 3327, 3103 (NH), 3067, 3014 (C–H aromatic), 2915 (C–H aliphatic), 1679 (C=O); ¹H NMR (400 MHz, DMSO-d₆) δ 10.16 (s, 1H, CH triazolo), 8.79 (s, 1H, NH), 8.74 (s, 1H, NH), 8.36 (d, J = 8.1 Hz, 1H, Ar–H), 7.68–7.54 (m, 6H, Ar–H), 7.34 (d, J = 8.5 Hz, 2H, Ar–H), 7.25–7.13 (m, 2H, Ar–H), 6.96 (d, J = 8.0 Hz, 1H, Ar–H), 6.61–6.52 (m, 1H, Ar–H), 3.74 (s, 3H, OCH₃); ¹³C NMR (101 MHz, DMSO-d₆) δ 160.17, 153.06, 151.72, 146.66, 141.36, 138.76, 138.73, 137.84, 134.33, 130.05, 128.33, 128.21, 127.80, 124.52, 122.68, 119.98, 116.91, 111.04, 107.75, 104.49, 55.41; mass (m/z): 426 (M⁺, 33.73%), 350 (100%, base peak); anal. calcd for C₂₃H₁₈N₆O₃ (426.44): C, 64.78; H, 4.25; N, 19.71. Found: C, 64.67; H, 4.48; N, 19.95%.

4.2. In vitro studies

4.2.1. Cytotoxicity. The cytotoxicity and selectivity of compounds 11a–d and the reference drug were evaluated against cell lines (MDA-231, MCF-7, WI-38, and WISH) using the MTT assay^62–64 (ESI†).

4.2.2. Enzyme inhibition. The inhibitory activities of compound 11d and reference drug against VEGFR-2 were tested using enzyme assay kits at concentrations ranging from 0.3 to 1000 nM^65,66 (ESI†).

4.2.3. RT-qPCR studies. The expression levels of apoptosis-related genes (BAX, Bcl-2, caspase-8, and caspase-9) in MDA-MB-231 cells treated with 11d (16.13 μM) were measured using qRT-PCR, following established protocols⁶⁷ (ESI†).

4.2.4. Flow cytometry. The effects of 11d on apoptosis and cell cycle progression in MDA-MB-231 cells were analyzed using flow cytometry^68,69 (ESI†).

4.3. Computational studies

4.3.1. Molecular docking. Molecular docking simulations for compounds 11a–d and reference drug against VEGFR-2 were performed as described in ESI.†

4.3.2. Molecular dynamics (MD) simulations. MD simulations were carried out for the VEGFER-2-11d complexe using the CHARMM-GUI web server and GROMACS 2021 software.^70,71 The binding free energies of 11d were calculated using the gmx_MMPBSA tool with the MM-GBSA method.^72,73 Principal component analysis (PCA) was used to analyze the dynamic motions of alpha carbons in specific amino acid sequences.⁷⁴ The free energy landscape (FEL) of the protein was studied in terms of two defined states, with their probabilities connected through an exponential relationship.⁷⁵ (ESI†).

4.3.3. ADMET and toxicity. The pharmacokinetic and toxicity profiles of compounds 11a–d were predicted using Discovery Studio 4.0 software⁷⁶ (ESI†).

4.3.4. Density functional theory (DFT). The structural optimization and reactivity analysis of 11d were performed using the Gaussian 09 (D.01) software with the DFT method, employing the B3LYP functional and the 6-31+G(d,p) basis set.⁷⁷ Detailed methods are available in the ESI.†

Data availability

Data are available in the manuscript and the ESI.†

Conflicts of interest

No conflict of interest to be declared.

Acknowledgements

This research was funded by Princess Nourah bint Abdulrahman University Researchers Supporting Project Number (PNURSP2025R116), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. The authors would like to thank Research Center at AlMaarefa University for funding this work.

References

1. S. M. Schwartz, Clin. Chem., 2024, 70, 140–149.
2. D. Trapani, O. Ginsburg, T. Fadelu, N. U. Lin, M. Hassett, A. M. Ilbawi, B. O. Anderson and G. Curigliano, Cancer Treat Rev., 2022, 104, 102339.
3. L. Wilkinson and T. Gathani, Br. J. Radiol., 2022, 95, 20211033.
4. A. A. Shah, M. A. Kamal and S. Akhtar, Curr. Drug Metab., 2021, 22, 50–59.
5. M. S. Taghour, H. Elkady, W. M. Eldehna, N. El-Deeb, A. M. Kenawy, A. E. Abd El-Wahab, E. B. Elkaeed, B. A. Alsfouk, A. M. Metwaly and I. H. Eissa, J. Biomol. Struct. Dyn., 2023, 41, 11535–11550.
6. X. J. Liu, H. C. Zhao, S. J. Hou, H. J. Zhang, L. Cheng, S. Yuan, L. R. Zhang, J. Song, S. Y. Zhang and S. W. Chen, Bioorg. Chem., 2023, 133, 106425.
7. E. B. Elkaeed, R. G. Yousef, H. Elkady, A. B. Mehany, B. A. Alsfouk, D. Z. Husein, I. M. Ibrahim, A. M. Metwaly and I. H. Eissa, J. Biomol. Struct. Dyn., 2023, 41, 7986–8001.
8. M. Montana, F. Mathias, T. Terme and P. Vanelle, Eur. J. Med. Chem., 2019, 163, 136–147.
9. V. Montero, M. Montana, M. Carré and P. Vanelle, Eur. J. Med. Chem., 2024, 271, 116360.
10. M. Zayed, Chemistry, 2023, 5, 2566–2587.
11. M. S. Nafie, S. H. Kahwash, M. M. Youssef and K. M. Dawood, Arch. Pharm., 2024, 357, e2400225.
12. M. A. Dahab, H. A. Mahdy, H. Elkady, M. S. Taghour, A. Elwan, M. A. Elkady, E. G. E. Elsakka, E. B. Elkaeed, A. A. Alsfouk, I. M. Ibrahim, A. M. Metwaly and I. H. Eissa, J. Biomol. Struct. Dynam., 2024, 42, 4214–4233.
13. A. M. Metwaly, H. Abd-El-Azim, M. Zewail, A. A. Alsfouk, E. B. Elkaeed and I. H. Eissa, J. Comput. Biophys. Chem., 2024, 1–17.
14. I. H. Eissa, R. G. Yousef, H. Elkady, A. A. Alsfouk, D. Z. Husein, I. M. Ibrahim, N. El-Deeb, A. M. Kenawy, W. M. Eldehna, E. B. Elkaeed and A. M. Metwaly, Mol. Divers., 2024, 28, 1153–1173.
15. I. H. Eissa, R. G. Yousef, E. B. Elkaeed, A. A. Alsfouk, D. Z. Husein, I. M. Ibrahim, A. Ismail, H. Elkady and A. M. Metwaly, ACS Omega, 2024, 9, 15861–15881.
16. E. A. Sobh, M. A. Dahab, E. B. Elkaeed, A. A. Alsfouk, I. M. Ibrahim, A. M. Metwaly and I. H. Eissa, J. Biomol. Struct. Dynam., 2024, 42, 2369–2391.
17. I. H. Eissa, R. G. Yousef, E. B. Elkaeed, A. A. Alsfouk, D. Z. Husein, I. M. Ibrahim, H. A. El-Mahdy, H. Elkady and A. M. Metwaly, Evol. Bioinf. Online, 2023, 19, 11769343231217916.
18. K. El-Adl, H. M. Sakr, R. G. Yousef, A. B. M. Mehany, A. M. Metwaly, M. A. Elhendawy, M. M. Radwan, M. A. ElSohly, H. S. Abulkhair and I. H. Eissa, Bioorg. Chem., 2021, 114, 105105.
19. R. G. Yousef, H. M. Sakr, I. H. Eissa, A. B. Mehany, A. M. Metwaly, M. A. Elhendawy, M. M. Radwan, M. A. ElSohly, H. S. Abulkhair and K. El-Adl, New J. Chem., 2021, 45, 16949–16964.
20. I. H. Eissa, A. M. Metwaly, A. Belal, A. B. M. Mehany, R. R. Ayyad, K. El-Adl, H. A. Mahdy, M. S. Taghour, K. M. A. El-Gamal, M. E. El-Sawah, S. A. Elmetwally, M. A. Elhendawy, M. M. Radwan and M. A. ElSohly, Arch. Pharm., 2019, 352, e1900123.
21. M. K. Ibrahim, M. S. Taghour, A. M. Metwaly, A. Belal, A. B. M. Mehany, M. A. Elhendawy, M. M. Radwan, A. M. Yassin, N. M. El-Deeb, E. E. Hafez, M. A. ElSohly and I. H. Eissa, Eur. J. Med. Chem., 2018, 155, 117–134.
22. M. K. Ibrahim, I. H. Eissa, A. E. Abdallah, A. M. Metwaly, M. M. Radwan and M. A. ElSohly, Bioorg. Med. Chem., 2017, 25, 1496–1513.
23. Y. Liu, Y. Li, Y. Wang, C. Lin, D. Zhang, J. Chen, L. Ouyang, F. Wu, J. Zhang and L. Chen, J. Hematol. Oncol., 2022, 15, 89.
24. S. Fogli, C. Porta, M. Del Re, S. Crucitta, G. Gianfilippo, R. Danesi, B. I. Rini and M. Schmidinger, Cancer Treat Rev., 2020, 84, 101966.
25. E. S. Duke, A. K. Barone, S. Chatterjee, P. S. Mishra-Kalyani, Y.-L. Shen, E. Isikwei, H. Zhao, Y. Bi, J. Liu and N. A. Rahman, Clin. Cancer Res., 2022, 28, 4173–4177.
26. N. Agarwal, A. Azad, J. Carles, S. Chowdhury, B. McGregor, A. S. Merseburger, S. Oudard, F. Saad, A. Soares and F. Benzaghou, Future Oncol., 2022, 18, 1185–1198.
27. S. M. Ferrari, M. Centanni, C. Virili, M. Miccoli, P. Ferrari, I. Ruffilli, F. Ragusa, A. Antonelli and P. Fallahi, Curr. Med. Chem., 2019, 26, 963–972.
28. B. Chaudhari, H. Patel, S. Thakar, I. Ahmad and D. Bansode, In Silico Pharmacol., 2022, 10, 10.
29. D. Pink, D. Andreou, S. Bauer, T. Brodowicz, B. Kasper, P. Reichardt, S. Richter, L. H. Lindner, J. Szkandera and V. Grünwald, Cancers, 2021, 13, 1223.
30. Y. Pang, A. Eresen, Z. Zhang, Q. Hou, Y. Wang, V. Yaghmai and Z. Zhang, Am. J. Cancer Res., 2022, 12, 2770.
31. A.-L. Cheng, S. Qin, M. Ikeda, P. R. Galle, M. Ducreux, T.-Y. Kim, H. Y. Lim, M. Kudo, V. Breder and P. Merle, J. Hepatol., 2022, 76, 862–873.
32. B. Liang, M. Tang, C. Huang, Y. Yang, Y. He, S. Liao and W. Shen, J. Gastrointest. Cancer, 2025, 56, 36.
33. J. Guo, J. Zhao, Q. Xu and D. Huang, Curr. Cancer Drug Targets, 2022, 22, 865–878.
34. L. J. Scott, Drugs, 2015, 75, 553–560.
35. P. Norman, Expert Opin. Orphan Drugs, 2015, 3, 445–455.
36. R. Elisei, M. J. Schlumberger, S. P. Müller, P. Schöffski, M. S. Brose, M. H. Shah, L. Licitra, B. Jarzab, V. Medvedev and M. C. Kreissl, J. Clin. Oncol., 2013, 31, 3639.
37. S. Wilhelm, C. Carter, M. Lynch, T. Lowinger, J. Dumas, R. A. Smith, B. Schwartz, R. Simantov and S. Kelley, Nat. Rev. Drug Discov., 2006, 5, 835–844.
38. S. M. Wilhelm, J. Dumas, L. Adnane, M. Lynch, C. A. Carter, G. Schütz, K. H. Thierauch and D. Zopf, Int. J. Cancer, 2011, 129, 245–255.
39. H. Y. Woo and J. Heo, Expet Opin. Pharmacother., 2012, 13, 1059–1067.
40. S. DiGiulio, Oncology Times, 2013.
41. Y. Li, Z. H. Gao and X. J. Qu, Basic Clin. Pharmacol. Toxicol., 2015, 116, 216–221.
42. S. K. Krishnamoorthy, V. Relias, S. Sebastian, V. Jayaraman and M. W. Saif, Ther. Adv. Gastroenterol., 2015, 8, 285–297.
43. C. Grüllich, in Small Molecules in Oncology, Springer, 2014, pp. 207–214.
44. C. Zhu, X. Ma, Y. Hu, L. Guo, B. Chen, K. Shen and Y. Xiao, Oncotarget, 2016, 7, 44545.
45. F.-W. Peng, D.-K. Liu, Q.-W. Zhang, Y.-G. Xu and L. Shi, Expert Opin. Ther. Pat., 2017, 27, 987–1004.
46. M. M. Alanazi, H. Elkady, N. A. Alsaif, A. J. Obaidullah, W. A. Alanazi, A. M. Al-Hossaini, M. A. Alharbi, I. H. Eissa and M. A. Dahab, J. Mol. Struct., 2021, 132220.
47. K. Lee, K.-W. Jeong, Y. Lee, J. Y. Song, M. S. Kim, G. S. Lee and Y. Kim, Eur. J. Med. Chem., 2010, 45, 5420–5427.
48. V. A. Machado, D. Peixoto, R. Costa, H. J. Froufe, R. C. Calhelha, R. M. Abreu, I. C. Ferreira, R. Soares and M.-J. R. Queiroz, Bioorg. Med. Chem., 2015, 23, 6497–6509.
49. Z. Wang, N. Wang, S. Han, D. Wang, S. Mo, L. Yu, H. Huang, K. Tsui, J. Shen and J. Chen, PLoS One, 2013, 8, e68566.
50. J. Dietrich, C. Hulme and L. H. Hurley, Bioorg. Med. Chem., 2010, 18, 5738–5748.
51. Q.-Q. Xie, H.-Z. Xie, J.-X. Ren, L.-L. Li and S.-Y. Yang, J. Mol. Graph. Model., 2009, 27, 751–758.
52. R. N. Eskander and K. S. Tewari, Gynecol. Oncol., 2014, 132, 496–505.
53. M. Ibrahim, M. Taghour, A. M. Metwaly, A. Belal, A. Mehany, M. Elhendawy, M. Radwan, A. Yassin, N. El-Deeb and E. Hafez, Eur. J. Med. Chem., 2018, 155, 117–134.
54. I. H. Eissa, A. M. Metwaly, A. Belal, A. B. Mehany, R. R. Ayyad, K. El-Adl, H. A. Mahdy, M. S. Taghour, K. M. El-Gamal and M. E. El-Sawah, Arch. Pharm., 2019, 352, 1900123.
55. E. M. Abbass, A. K. Khalil, M. M. Mohamed, I. H. Eissa and A. M. El-Naggar, Bioorg. Chem., 2020, 104, 104255.
56. N. A. Alsaif, M. A. Dahab, M. M. Alanazi, A. J. Obaidullah, A. A. Al-Mehizia, M. M. Alanazi, S. Aldawas, H. A. Mahdy and H. Elkady, Bioorg. Chem., 2021, 110, 104807.
57. N. A. Alsaif, M. S. Taghour, M. M. Alanazi, A. J. Obaidullah, A. A. Al-Mehizia, M. M. Alanazi, S. Aldawas, A. Elwan and H. Elkady, J. Enzym. Inhib. Med. Chem., 2021, 36, 1093–1114.
58. F. Azimian, M. Hamzeh-Mivehroud, J. S. Mojarrad, S. Hemmati and S. Dastmalchi, Eur. J. Med. Chem., 2020, 201, 112461.
59. D. Z. Husein, R. Hassanien and M. Khamis, RSC Adv., 2021, 11, 27027–27041.
60. R. G. Parr, L. v. Szentpály and S. Liu, J. Am. Chem. Soc., 1999, 121, 1922–1924.
61. R. G. Gieling, M. Babur, L. Mamnani, N. Burrows, B. A. Telfer, F. Carta, J.-Y. Winum, A. Scozzafava, C. T. Supuran and K. J. Williams, J. Med. Chem., 2012, 55, 5591–5600.
62. A. R. Kotb, D. A. Bakhotmah, A. E. Abdallah, H. Elkady, M. S. Taghour, I. H. Eissa and M. A. El-Zahabi, RSC Adv., 2022, 12, 33525–33539.
63. A. R. Kotb, A. E. Abdallah, H. Elkady, I. H. Eissa, M. S. Taghour, D. A. Bakhotmah, T. M. Abdelghany and M. A. El-Zahabi, RSC Adv., 2023, 13, 10488–10502.
64. A. Saha, S. Mohapatra, P. Kurkute, B. Jana, J. Sarkar, P. Mondal and S. Ghosh, RSC Adv., 2015, 5, 92596–92601.
65. H. A. Mahdy, H. Elkady, M. S. Taghour, A. Elwan, M. A. Dahab, M. A. Elkady, E. G. Elsakka, E. B. Elkaeed, B. A. Alsfouk and I. M. Ibrahim, Future Med. Chem., 2023, 15, 1233–1250.
66. A. E. Abdallah, R. R. Mabrouk, M. R. Elnagar, A. M. Farrag, M. H. Kalaba, M. H. Sharaf, E. M. El-Fakharany, D. A. Bakhotmah, E. B. Elkaeed and M. M. S. Al Ward, Drug Des., Dev. Ther., 2023, 587–606.
67. K. J. Livak and T. D. Schmittgen, Methods, 2001, 25, 402–408.
68. Z. Darzynkiewicz, E. Bedner and P. Smolewski, Flow cytometry in analysis of cell cycle and apoptosis, Seminars in hematology, Elsevier, 2001, pp. 179–193.
69. D. Bhunia, A. Saha, A. Adak, G. Das and S. Ghosh, RSC Adv., 2016, 6, 113487–113491.
70. B. R. Brooks, C. L. Brooks III, A. D. Mackerell Jr, L. Nilsson, R. J. Petrella, B. Roux, Y. Won, G. Archontis, C. Bartels and S. Boresch, J. Comput. Chem., 2009, 30, 1545–1614.
71. S. Jo, X. Cheng, S. M. Islam, L. Huang, H. Rui, A. Zhu, H. S. Lee, Y. Qi, W. Han and K. Vanommeslaeghe, Adv. Protein Chem. Struct. Biol., 2014, 96, 235–265.
72. T. Tuccinardi, Expet Opin. Drug Discov., 2021, 16, 1233–1237.
73. M. S. Valdés-Tresanco, M. E. Valdés-Tresanco, P. A. Valiente and E. Moreno, J. Chem. Theor. Comput., 2021, 17, 6281–6291.
74. A. Amadei, A. B. Linssen and H. J. Berendsen, Proteins: Struct., Funct., Bioinf., 1993, 17, 412–425.
75. E. Papaleo, P. Mereghetti, P. Fantucci, R. Grandori and L. De Gioia, J. Mol. Graph. Model., 2009, 27, 889–899.
76. A. M. Metwaly, E. B. Elkaeed, B. A. Alsfouk, A. M. Saleh, A. E. Mostafa and I. H. Eissa, Plants, 2022, 11, 1886.
77. D. S. Biovia, Discovery studio modeling environment, Release San Diego, CA, 2017.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ra00526d

---