Chlorothiophene‐Substituted N‐Phenylpyrazolines: Synthesis, Anticancer Evaluation, and Computational Insight into COX‐2 Inhibition

ABSTRACT

Cancer remains a leading cause of mortality worldwide, highlighting the need for novel anticancer agents. Pyrazoline derivatives have attracted considerable interest due to their structural flexibility, enabling modifications, including bioisosteric replacement, to enhance anticancer potential. In this study, eight N‐phenylpyrazoline analogs bearing chlorothiophene moieties were synthesized via cyclocondensation of chalcone intermediates with phenylhydrazine, yielding 59%–94%. Structural characterization was confirmed using ATR–IR, GC–MS, ¹H‐, and ¹³C–NMR analyses. The synthesized derivatives (2a‐2h) were assessed for cytotoxicity against the T47D, MCF7, and WiDr cancer cell lines, along with normal Vero cells, using the MTT assay. Selectivity was determined by comparing cytotoxic effects on cancer versus normal cells. Among the compounds, 5‐(benzo[d][1,3]dioxol‐5‐yl)‐3‐(5‐chlorothiophen‐2‐yl)‐1‐phenyl‐4,5‐dihydro‐1H‐pyrazole (2g) exhibited the most potent cytotoxicity against WiDr cells (IC₅₀ = 19.63 µg/mL) and the highest selectivity index (SI = 665.88). Molecular docking revealed strong binding affinities of compound 2g to COX‐2 (−10.5 kcal/mol), supported by favorable pharmacokinetic properties.

Eight novel N‐phenylpyrazoline derivatives bearing chlorothiophene moieties were synthesized and evaluated for anticancer activity against T47D, MCF7, and WiDr cell lines. Compound 2g showed the strongest cytotoxicity against WiDr cells (IC₅₀ = 19.63 µg/mL) with the highest selectivity index (665.88), highlighting its potential as a selective anticancer candidate.

1. Introduction

Cancer remains one of the leading causes of mortality worldwide, with approximately 19.9 million new cases reported in 2022 and projections reaching 28.4 million cases by 2040 [1]. In Indonesia, an estimated 408,661 new cancer cases and 242000 cancer‐related deaths were recorded in 2022, with breast cancer as the most frequently diagnosed malignancy [2]. Despite significant advancements, nowadays, chemotherapy remains a viable option in cancer treatment due to its ability to inhibit the proliferation of rapidly dividing cancer cells [3, 4]. However, therapeutic resistance in cancer continues to increase. This phenomenon is recognized as a consequence of complex signaling networks rather than isolated molecular events. One prominent mechanism underlying resistance is pathway transactivation, whereby inhibition of a single oncogenic pathway is caused through compensatory activation of interconnected signaling cascades [5, 6]. Epidermal growth factor receptor (EGFR) signaling is central to tumor proliferation, survival, and metastasis. However, its inhibition is often limited by resistance arising from cross‐talk with inflammatory and survival pathways such as cyclooxygenase‐2/prostaglandin E2 (COX‐2/PGE2), phosphoinositide 3‐kinase/protein kinase B (PI3K/Akt), and nuclear factor κB (NF‐κB) [7, 8]. This pathway transactivation underscores the need for alternative targets or multitarget strategies to overcome therapeutic resistance effectively.

Among these interconnected pathways, the COX‐2/PGE2 inflammatory axis has emerged as a key upstream regulator of EGFR‐driven oncogenic signaling [9]. COX‐2 is frequently overexpressed across a wide range of solid tumors and is associated with enhanced tumor growth, angiogenesis, immune evasion, and metastasis [10]. Elevated COX‐2 expression increases PGE2 production, which in turn promotes EGFR transactivation through both ligand‐dependent mechanisms and Src‐mediated ligand‐independent pathways [7, 9]. This COX‐2–driven network promotes cancer cell survival and contributes to resistance against conventional therapies [8]. Together, these observations identify COX‐2 as a key upstream regulator of oncogenic signaling networks and a promising complementary target for anticancer intervention.

Several selective COX‐2 inhibitors, including celecoxib, rofecoxib, parecoxib, and valdecoxib, share a characteristic butterfly‐shaped triaryl scaffold containing electron‐rich systems, such as nitrogen–nitrogen (N─N) or nitrogen–oxygen (N─O) bonds (Figure 1a). However, safety concerns have limited their clinical use. Rofecoxib and valdecoxib, in particular, were withdrawn from the market due to increased cardiovascular risk [11, 12], while celecoxib continues to carry a risk of serious cardiovascular adverse events [13]. These limitations have restricted the long‐term use of COX‐2 inhibitors in cancer therapy. Consequently, there is an urgent need for novel COX‐2–targeting agents with improved safety profiles and retained anticancer efficacy. In our ongoing efforts to address antiproliferative resistance associated with EGFR signaling, we have explored N‐phenylpyrazoline‐based hybrid molecules designed with promising anticancer activities [14, 15, 16]. Guided by the butterfly‐like architecture of selective COX‐2 inhibitors [17], we employed a bioisosteric replacement strategy to mimic key structural features while maintaining a favorable safety profile [18] (Figure 1b). The pyrazoline core was selected due to its structural versatility and electron‐rich nitrogen atoms, which are associated with diverse biological activities, including antimalarial [19], antioxidant [20], antibacterial [21], and anticancer effects [22, 23]. Supporting this concept, our in silico studies on similarly designed analogues demonstrated that N‐phenylpyrazoline derivatives exhibit moderate binding affinity toward the COX‐2 protein [24]. Furthermore, introducing an N‐phenyl substituent has been reported to enhance anticancer activity by improving hydrophobic interactions and steric complementarity within the enzyme binding site [25].

FIGURE 1.

(a) known selective COX‐2 inhibitor, (b) our continuous strategy in anticancer development [14, 15, 16], (c) the structural design of this study.

In this study, we employed a bioisosteric modification strategy by replacing phenyl rings with chlorinated thiophene moieties (Figure 1c). Thiophene substitution can modulate electronic properties and improve target binding [26]. However, thiophene metabolism is known to generate reactive intermediates that may contribute to toxicity [27]. To mitigate this risk, chlorine substitution at the 2‐position of the thiophene ring was introduced to favor safer metabolic pathways through glutathione conjugation [28]. Additionally, chlorinated pyrazoline derivatives have been reported to exhibit enhanced cytotoxic potency, reflected by lower IC₅₀ values in cancer cells [14]. Prabudheva et al. previously synthesized a series of pyrazoline derivatives bearing a chlorothiophene moiety, incorporating a chloro substituent and a 3,4‐dimethoxy‐substituted benzaldehyde variation. However, the biological evaluation in that study was limited to antimicrobial, anti‐inflammatory, and antioxidant activities, with no further exploration of anticancer potential [28, 29, 30, 31]. In contrast, the present study employed a more varied range of benzaldehyde derivatives, varying the number and position of methoxy substituents, along with benzo[1,3]dioxole and dimethylaminobenzaldehyde variations as electron‐donating group (EDG) substituents. The resulting compounds were evaluated specifically for anticancer activity.

As illustrated in Figure 1b, our group has progressively explored a bioisosteric replacement strategy across several scaffolds, beginning with chloroacetophenone [14], followed by furan [16], thiophene [15], and now chlorothiophene. Because all previous studies used the same panel of cancer cell lines, we similarly evaluated the present study against MCF‐7, T47D, and WiDr cancer cell lines to continue this ongoing investigation. This approach allows for direct comparative analysis of anticancer activity across this bioisosteric series. Herein, we report the rational design and synthesis of chlorothiophene‐substituted N‐phenylpyrazoline derivatives and evaluate them as potential anticancer agents. We performed molecular docking and dynamics simulations to elucidate COX‐2 binding interactions and inhibitory mechanisms and to develop safer multitarget agents that target COX‐2 and overcome therapeutic resistance.

2. Results and Discussion

2.1. Binding Pocket of COX‐2 and Structural Design of Target Compound

Cyclooxygenase (COX), a central enzyme in prostaglandin biosynthesis, exists as three isoforms (COX‐1, COX‐2, and COX‐3). COX‐1 is constitutively expressed in normal tissues to maintain physiological function, whereas COX‐2 is inducible and associated with inflammation and cancer development [32]. Elevated COX‐2 expression has been reported in multiple malignancies, including colorectal (60%), breast (40%), pancreatic, esophageal, and lung cancers, as well as melanoma [33, 34]. Both COX‐1 and COX‐2 function as membrane‐associated homodimers, with each monomer consisting of (i) an N‐terminal epidermal growth factor (EGF)‐like domain, (ii) a membrane‐binding domain, and (iii) a large C‐terminal catalytic domain. Although the two isoforms share ∼60%–65% amino acid sequence identity and highly similar three‐dimensional structures, clear differences in their catalytic sites enable selective targeting [35, 36]. The COX‐2 binding pocket is approximately 20% larger than that of COX‐1 due to substitutions (Ile523→Val523 and Ile434→Val434), which generate an additional secondary pocket adjacent to the main hydrophobic channel. The His513→Arg513 substitution further alters the electrostatic environment, while residues such as His90, Gln192, and Tyr355 regulate ligand access [37, 38]. These structural distinctions form the molecular basis for designing selective COX‐2 inhibitors.

The selective COX‐2 inhibitors (coxibs) were rationally designed to exploit the enlarged secondary pocket of COX‐2 through a rigid diaryl heterocyclic scaffold with a characteristic butterfly‐like geometry. In this arrangement, two aryl “wings” extend from a central heterocyclic core, typically bearing sulfonyl or sulfonamide functionalities, allowing deep penetration into the V‐shaped catalytic channel and selective engagement of the side pocket [39, 40]. Celecoxib interacts within the COX‐2 binding pocket via strong hydrogen bonds with several amino acid residues, including Leu388, Phe504, Arg499, Ser399, and Arg106 [41]. Although this structural framework provides COX‐2 selectivity, it may also pose safety risks. Rofecoxib and valdecoxib were withdrawn due to increased cardiovascular risk linked to off‐target human Ether‐à‐go‐go‐Related Gene (hERG) channel interactions [42]. In particular, rofecoxib undergoes slow oxidative conversion to a reactive maleic anhydride intermediate, which has been implicated in hERG affinity and pro‐arrhythmic effects [43].

Triaryl pyrazoline systems have been explored as alternative COX‐2 selective scaffolds. Qiu et al. [44] synthesized a series of novel pyrazoline derivatives that showed low cytotoxicity toward normal 293T cells (CC₅₀ > 100 µM), and their antiproliferative effect correlated with cellular COX‐2 expression. Docking analysis revealed five hydrogen bonds between the sulfonamide group and key COX‐2 residues (Gln178, Leu338, Ser339, Arg499, and Phe504). Similarly, Yan et al. [45] developed diaryl and heteroaryl dihydropyrazole derivatives bearing sulfonamide and benzodioxole moieties that also demonstrated superior anticancer activity compared to celecoxib against SW‐620 cells (IC₅₀ = 0.86 ± 0.02 vs. 1.29 ± 0.04 µM) and exhibited broad potency across A549, HeLa, MCF‐7, and HepG2 cell lines. The 1,3‐benzodioxole scaffold contributed to enhanced activity, and docking studies supported strong COX‐2 binding (−45.96 kcal/mol) through hydrogen bonding with His75, Phe504, and Val102, along with stabilizing non‐covalent interactions.

However, both studies retained the sulfonamide moiety, which may still present potential cardiotoxicity concerns. Therefore, our design strategy avoided highly polar sulfonamide groups and instead focused on maintaining key binding interactions through alternative hydrophobic and heteroaryl substituents. In our previous in silico investigations of related designs, halogen introduction, particularly chlorine, was employed to enhance hydrophobic interactions within the active site [24, 46]. Supporting this rationale, halogenated systems such as the bromo‐substituted thiophene in Dup‐697 highlight the favorable role of electronegative substituents in COX‐2 inhibition [47]. Furthermore, several studies have shown that highly electronegative substituents at the para position of the aryl ring significantly improve COX‐2 inhibitory activity [38, 48]. Based on this rationale, we introduced methoxy and chloro substituents at different positions to evaluate how their nature and placement influence protein–ligand interactions and overall inhibitory potency.

2.2. Chemistry

The synthesis of the targeted chlorothiophene‐substituted N‐phenylpyrazoline compounds was accomplished in two steps, as illustrated in Figure 6. First, chalcone intermediates (1a–1h) were prepared via Claisen–Schmidt condensation using 2‐acetyl‐5‐chlorothiophene as the aromatic ketone and various substituted benzaldehydes as the aromatic aldehydes. The reactions were carried out in the presence of a strong base catalyst (NaOH or KOH) for 24 h, affording the corresponding chalcone derivatives as solids. Most compounds were known [49, 50, 51], and their characterization was confirmed by GC–MS and IR spectroscopy, as shown in Figures S2–S25. The GC–MS analysis showed the expected molecular ion (m/z) peaks, while IR spectra displayed characteristic absorption bands at 1627–1651 cm⁻¹ (C═O stretching) and 972–1026 cm⁻¹ (out‐of‐plane bending of trans‐alkenes), collectively confirming the chalcone framework.

FIGURE 6.

Synthesis scheme of 2a–2h.

Subsequently, the chalcone intermediates underwent cyclocondensation with equimolar phenylhydrazine under basic conditions, using NaOH as the catalyst. The reaction mixture was refluxed for 1 h, yielding the corresponding N‐phenylpyrazoline derivatives (2a–2h), as colored solids in yields ranging from 59% to 94% (Table 1). The GC–MS, IR, ¹H‐NMR, and ¹³C‐NMR spectroscopy were employed to elucidate the structure of the synthesized compounds. GC analysis revealed that the synthesized pyrazoline derivatives were highly pure, with values ranging from 95.58% to 99.31%, and the mass spectra showed molecular ion peaks (m/z) consistent with the expected molecular structures. IR spectra further verified the formation of the pyrazoline ring, showing a characteristic absorption band at 1589–1597 cm⁻¹ corresponding to C═N stretching, along with an additional band at 1365–1388 cm⁻¹ attributed to C─N stretching vibration. The disappearance of the C═O stretching band, previously observed in the chalcone intermediates, provided additional evidence for successful cyclization [52]. To further support these findings, ¹H‐NMR spectra of all synthesized derivatives exhibited three distinct proton signals within the pyrazoline moiety, forming the characteristic AMX spin system [53]. In the ¹H NMR spectra, three characteristic doublets of doublets were observed at δ 5.11–5.57, 3.70–3.79, and 2.91–3.07 ppm, with geminal coupling constants of 16.7–16.9 Hz. In the ¹³C NMR spectra, the C–N carbon resonated at 58.60–65.53 ppm, while the C═N carbon appeared at 144.39–144.93 ppm, further confirming the formation of the pyrazoline scaffold. Complete characterization spectra are available in Figures S26–S65, and the proposed mechanism of pyrazoline formation is shown in Figure S1.

TABLE 1.

Structure, yield, and melting point of 2a–2h.

Compound	Structure	Yield (%)	Melting point (°C)
2a		73	140–141
2b		59	103–104
2c		94	119–120
2d		62	112–113
2e		78	118–119
2f		93	117–118
2g		88	139–140
2h		88	130–132

2.3. Anticancer Activity

The cytotoxic activity of the synthesized compounds was evaluated using the MTT assay against several human cancer cell lines, including MCF‐7 and T47D (breast cancer) and WiDr (colorectal cancer), to determine their IC₅₀ values, assess the effect of functional group variations on anticancer activity, and explore the potential of chlorothiophene‐containing derivatives as anticancer candidates. Cell viability was evaluated using the MTT assay and expressed as a percentage relative to untreated control cells. IC₅₀ values, defined as the concentration of compound required to reduce cell viability by 50%, were determined from nonlinear regression analysis of dose–response curves. Table 2 summarizes the IC₅₀ values for each compound against the four cancer cell lines. Doxorubicin and 5‐fluorouracil (5‐FU) were utilized as reference drugs.

TABLE 2.

Anticancer activity of 2a–2h.

Compounds	IC50 (µg/mL)				Selectivity index WiDr to Vero
	MCF‐7	T47D	WiDr	Vero (Normal cell)
2a	>1000	188.02	135.39	n.db	—
2b	154.21	43.40	>1000	2282.36	1.92
2c	>1000	374.47	572.39	n.db	—
2d	179.38	51.16	132 92	1168.38	8.79
2e	997.27	>1000	130 09	4968.59	38.19
2f	n.aa	>1000	720.17	n.db	—
2g	238.15	82.49	19.63	13 069.27	665.88
2h	56.25	74.74	62.60	2045.48	32.67
Doxorubicin	14.63	50.84	—	—	—
5‐Fluorouracil	—	—	14.44	—	—

^an.a = not active (>10 000).

^bn.d = not detected.

Among the synthesized series, compounds 2g and 2h exhibited the lowest IC₅₀ values across the three evaluated cancer cell lines, with compound 2g showing the most potent activity against WiDr colorectal cancer cells (IC₅₀ = 19.63 µg/mL). At this point, 2g, incorporating a benzo[1,3]dioxole moiety, demonstrated strong cytotoxicity (IC₅₀ < 20 µg/mL) and exhibited the highest selectivity index (SI = 665.88). The enhanced activity of 2g may be attributed to the electron‐donating character of the benzo[1,3]dioxole group, which can modulate the electronic distribution of the pyrazoline scaffold and potentially strengthen interactions within the target binding pocket. This observation is consistent with previous reports indicating that electron‐rich substituents improve cytotoxic potency in pyrazoline derivatives [42, 54]. Notably, comparison with a previous study [15], which evaluated the anticancer activity of nonchlorinated thiophene‐based N‐phenylpyrazoline derivatives bearing a 3,4‐dimethoxybenzaldehyde substituent against WiDr colorectal cancer cells, revealed that the IC₅₀ value of that analogue was not markedly different from that of compound 2g in the present study. The key structural distinction lies in the incorporation of a chlorine substituent on the thiophene ring and the cyclic dimethoxy configuration (benzo[1,3]dioxole) in place of the open‐chain 3,4‐dimethoxy substituent. While the IC₅₀ value remained within an active range, the selectivity index markedly improved, from a non‐selective value of 2.74 in the previous analog to a substantially higher value of 665.88 for compound 2g. This finding suggests that chlorine substitution on the thiophene ring, combined with the cyclic benzo[1,3]dioxole configuration, plays a critical role in enhancing selectivity toward cancer cells over normal cells, while preserving potent cytotoxic activity. Beyond compound 2g, further evaluation of the substituent effect across the series showed that compound 2h, bearing a strongly electron‐donating dimethylamino group (–N(CH₃)₂), also showed superior anticancer activity compared to methoxy‐substituted derivatives in this series, likely due to its stronger electron‐donating effect, enhancing target interactions. The overall structure–activity relationship (SAR) of the synthesized N‐phenylpyrazoline bearing chlorothiophene series is summarized in Figure 2.

FIGURE 2.

Structure–activity relationship (SAR) of the synthesized N‐phenylpyrazoline derivatives bearing a chlorothiophene moiety.

2.4. Molecular Docking

To investigate the potential COX‐2 inhibitory mechanism, we evaluated the most active compound (2g), which showed the strongest cytotoxicity against WiDr cells, for its COX‐2 binding affinity. Before docking simulations, the protocol was validated by redocking the native ligand, celecoxib, into the COX‐2 active site. Docking accuracy was assessed by calculating the root‐mean‐square deviation (RMSD) between the crystallographic and predicted poses. The obtained RMSD value of 0.732 Å (well below the 2 Å acceptance threshold) [55] confirmed that the protocol reliably reproduces the native binding mode and is suitable for subsequent docking studies. A comparative docking analysis of compound 2g and celecoxib within the COX‐2 active site showed that 2g exhibited a slightly higher binding affinity (−10.5 kcal/mol) than celecoxib (−12.7 kcal/mol), indicating that celecoxib binds more strongly to the active site. However, beyond binding energy values, the nature of protein–ligand interactions is also crucial for inhibitory activity [56]. Despite lacking strong polar functionalities such as the sulfonamide group present in celecoxib, compound 2g displayed a similar interaction pattern within the catalytic pocket (Table 3). Although 2g forms fewer hydrogen bonds, the key amino acid residues that hydrogen‐bond with celecoxib also interact extensively with 2g via hydrophobic and π‐based interactions, including Leu338 and Phe504. Such π‐alkyl and π‐hydrogen bond interactions have been reported to facilitate ligand binding within the active site by providing aromatic stacking stabilization, while the combination of polar and nonpolar contacts is likely to promote more stable and selective binding within the target pocket [57].

TABLE 3.

Binding affinity value and considerable interaction profile of 2g and celecoxib with COX‐2.

	2g	Celecoxib
Binding affinity (kcal/mol)	−10.5	−12.7
Interactions
Hydrogen bond	Arg106	Arg106, Gln178, Leu338, Ser339, Arg499, Phe504
Hydrophobic e.g: Pi‐Sigma, Pi‐Pi T shaped; Amide‐Pi Stacked, Alkyl, Pi‐Alkyl	Val102, Val335, Leu338, Leu345, Trp371, Trp373, Ala502, Phe504, Val509, Leu517, Ala513	Val335, Tyr341, Leu345, Leu370, Trp371, Trp373, Val509, Gly512, Ala513

Figure 3a,b depict the specific atomic interactions between the ligands and surrounding amino acid residues, showing the 2D interaction profiles of celecoxib and 2g, respectively. Although 2g lacks the sulfonamide moiety responsible for multiple hydrogen bonds in celecoxib, it preserves the key hydrophobic and aromatic interactions within the COX‐2 active site. As illustrated in Figure 3c, the superimposition of 2g (blue) and celecoxib (green) reveals a high degree of spatial overlap, indicating that 2g adopts a comparable binding orientation within the catalytic pocket. Importantly, the benzo[1,3]dioxole moiety significantly enhances hydrophobic contacts in the side pocket region of COX‐2, reinforcing ligand accommodation within the predominantly hydrophobic channel. A weak polar interaction, identified as a halogen‐associated hydrogen bond, was observed between the chloro substituent of the thiophene ring and Arg106. In ligand–protein complexes, halogen atoms can participate in non‐covalent interactions by acting as weak hydrogen‐bond acceptors toward N–H or O–H donors through their electronegative lone pairs. Among halogens, fluorine and chlorine are most frequently involved in such interactions due to their relatively high electronegativity and partial negative charge in organic frameworks [58]. Although weaker than classical hydrogen bonds, these interactions can contribute to additional stabilization within the binding pocket. The interaction of the chloro atom with Arg106 is particularly interesting, as it occurs at a position that spatially overlaps with the CF₃ group of celecoxib, which engages the same residue (Figure 3c). Arg106 has been reported as a critical amino acid involved in mediating arachidonic acid binding within the COX‐2 active site. Consequently, ligand occupancy at this critical region may impede the accommodation of arachidonic acid and thereby disrupt the catalytic process [29, 59]. Despite exhibiting a slightly lower binding energy than celecoxib, the preservation of essential non‐covalent interactions, particularly π–π stacking and hydrophobic contacts with critical residues, suggests that 2g effectively mimics the binding mode of the reference ligand. These findings support its potential as a promising COX‐2 inhibitor candidate.

FIGURE 3.

2D interaction visualization of (a) celecoxib, (b) 2g, and (c) their 3D superimposed poses within the COX‐2 binding pocket.

2.5. Molecular Dynamics Simulation

Following docking, we performed 120 ns molecular dynamics simulations of both the celecoxib–COX‐2 and 2g–COX‐2 complexes in an explicit aqueous environment to assess structural stability and dynamic behavior. Backbone RMSD and root mean square fluctuations (RMSF) analyses monitored global stability and residue‐level fluctuations throughout the trajectory. As shown in Figure 4a, RMSD values for both complexes remained below 3 Å, consistent with stable protein–ligand systems. Although minor fluctuations were observed around 25 ns for celecoxib and 95–105 ns for compound 2g, both systems quickly re‐equilibrated. Extended simulations would further clarify complex stability. The RMSF stability profile (Figure 4b) indicates that key active‐site residues exhibit comparable structural rigidity in both systems, suggesting that ligand binding does not markedly perturb local conformational features. This observation is consistent with previous reports that lower RMSF values indicate more stable and constrained protein regions, particularly those involved in crucial inhibitor interactions.

FIGURE 4.

Superimposition of (a) root mean square deviation (RMSD), (b) solvent accessible surface area (SASA), (c) radius of gyration (RoG), and (d) root mean square fluctuations (RMSF) of complex native ligand–protein (black) and complex 2g–COX‐2 (red) backbones during the MD simulations.

We also analyzed solvent‐accessible surface area (SASA) and radius of gyration (RoG) to evaluate global structural changes. We used SASA to assess differences in protein–solvent exposure, while RoG assessed whether the protein became more compact or relaxed during the simulation. The SASA (Figure 4c) and RoG (Figure 4d) profiles showed comparable trends for both complexes throughout the simulation. SASA values remained within a similar range (∼23 000–24 500 Å²), with only minor fluctuations reflecting normal equilibration. The close overlap of the trajectories indicates that 2g does not significantly alter protein compactness or solvent exposure. Overall, the MD results demonstrate that 2g preserves the structural integrity and dynamic stability of COX‐2 in a manner comparable to celecoxib, supporting its potential as a stable COX‐2‐binding ligand. Consistent with the RMSF trajectory, the key interacting residues (Val335, Leu338, Ser339, Phe504, Val509, and Ala513), as observed from the per‐residue protein secondary structure analysis of the protein–ligand 2g complex (Figure 5), retained similar secondary structural conformations across both systems. These findings collectively indicate that the binding of compound 2g does not induce significant local structural perturbations within the active site.

FIGURE 5.

Per‐residue secondary structure evolution of the protein–ligand 2g complex throughout the MD simulation.

2.6. Pharmacokinetic Prediction (ADMET)

The ADMET prediction (Table 4) indicates that compound 2g possesses a competitive pharmacokinetic profile relative to celecoxib and rofecoxib. For absorption, 2g showed excellent human intestinal absorption (HIA, 94.55%) and favorable Caco‐2 permeability (1.063), comparable to celecoxib and slightly below rofecoxib. Overall, 2g exhibits a competitive oral absorption profile. The predicted total clearance (0.149) was substantially lower than that of celecoxib (0.435) and rofecoxib (0.888), suggesting a potentially longer half‐life that may allow less frequent dosing. Regarding safety, 2g was predicted to be non‐inhibitory toward hERG I, indicating a low risk of severe cardiotoxicity. Its predicted acute oral toxicity (LD₅₀ = 2.905) was higher than both comparator drugs, suggesting lower acute toxicity, while the chronic LOAEL value was comparable to celecoxib and more favorable than rofecoxib.

TABLE 4.

ADMET of compound 2g compared to commercial drug.

	2g	Celecoxib	Rofecoxib
Absorption
HIA (%)	94.552	92.995	98.413
Caco‐2 cell permeability (log Papp in 10‐6 cm/s)	1.063	0.839	1.289
Water solubility (log mol/L)	−6.22	−4.45	−4.217
Distribution
VDss (log L/Kg)	0.382	−0.273	−0.619
BBB (log BB)	0.278	−0.931	0.013
CNS (log PS)	−1.301	−2.052	−2.297
Metabolism
CYP inhibitor
1A2	Yes	Yes	Yes
2C19	Yes	Yes	Yes
2C9	Yes	Yes	Yes
2D6	No	No	No
3A4	Yes	Yes	No
Excretion
Total clearance (Log/mL/min/kg)	0.149	0.435	0.888
Renal OCT2 substrate	No	No	No
Toxicity
herG I inhibitor	No	No	No
herG II inhibitor	Yes	No	No
Oral Rat Acute Toxicity (LD50) (mol/kg)	2.905	1.975	2.387
Oral Rat Chronic Toxicity (LOAEL) (log mg/kg_bw/day)	1.072	1.126	1.65

However, several limitations were identified. Compound 2g exhibited markedly poor aqueous solubility (−6.22), lower than celecoxib (−4.45) and rofecoxib (−4.217), which may compromise formulation feasibility and bioavailability. From a distribution perspective, 2g showed a higher predicted volume of distribution than both reference drugs. Additionally, 2g showed relatively high predicted BBB permeability (0.278), exceeding that of celecoxib and rofecoxib, which may increase the risk of unintended central nervous system exposure. The compound was also predicted to inhibit hERG II, a potential cardiac safety concern, although it was not predicted to inhibit hERG I. Furthermore, 2g was predicted to inhibit multiple CYP isoenzymes (1A2, 2C19, 2C9, and 3A4), a pattern that is partially similar to that of celecoxib. Overall, compound 2g demonstrates a broadly comparable ADMET profile relative to celecoxib and rofecoxib, particularly in absorption, distribution, and predicted acute toxicity. Nevertheless, 2g exhibits a promising and largely comparable ADMET profile, with advantages in distribution and predicted safety, but requires structural optimization to address solubility, potential CNS exposure, and cardiotoxicity.

3. Conclusions

In this study, eight N‐phenylpyrazoline derivatives incorporating chlorothiophene moieties were successfully synthesized and evaluated for their anticancer potential. Among them, compound 2g, bearing a benzo[1,3]dioxole group, showed the most potent and selective cytotoxic activity, particularly against WiDr colorectal cancer cells, supported by its high selectivity index. Molecular docking and MD simulation analyses further confirmed strong, stable interactions of 2g with COX‐2, indicating its potential as a molecular target. Additionally, ADME predictions revealed favorable pharmacokinetic properties, strengthening its suitability for further drug development. Nevertheless, these computational predictions are based on simplified, isolated systems and may not fully reflect the complexity of a living biological environment. Cytotoxicity was evaluated using only the MTT assay, which reflects overall cell viability without elucidating the specific mechanism of growth inhibition; further experimental validation is still needed. Overall, compound 2g is a promising lead candidate and warrants further mechanistic studies (e.g., immunocytochemistry to assess COX‐2 expression) and in vivo evaluation to substantiate its therapeutic potential.

4. Experimental Section

4.1. The General Procedure of Chlorothiophene‐Based N‐phenylpyrazoline

The target N‐phenylpyrazolines were synthesized via chalcone intermediates (1a–1h), as shown in Figure 6. After all the chalcone intermediates were established, in a three‐necked round‐bottom flask, chalcone (1a–1h), 1 mmol, was dissolved in 10 mL of ethanol/methanol. Sodium hydroxide pellets (400 mg) were added to the reaction mixture, then stirred and heated until the mixture became completely homogeneous. Subsequently, phenylhydrazine (0.1 mL, 1 mmol) was added dropwise to the mixture under continuous stirring. The resulting mixture was refluxed, and reaction progress was monitored by thin‐layer chromatography (TLC) using n‐hexane/ethyl acetate (7:3) as the eluent. Upon completion of the reaction, the mixture was poured into an Erlenmeyer flask placed in an ice bath, with stirring continued. The solid product was filtered and dried in a desiccator. All chemicals and reagents were obtained from Sigma–Aldrich Co., LLC and Shanghai Macklin Biochemical Co., Ltd. and used without further purification. The dried final products were then thoroughly characterized using GC–MS (Shimadzu QP‐2010S), IR (Attenuated Total Reflected Infrared, Shimadzu IRTracer‐100), ¹H‐(500 MHz) and ¹³C‐(125 MHz) (JEOL JNMECA 500). The melting point was analyzed using the Electrothermal IA9100.

4.1.1. 3‐(5‐Chlorothiophene‐2‐yl)‐5‐(2‐Methoxyphenyl)‐1‐Phenyl‐4,5‐Dihydro‐1H‐Pyrazole (2a)

Compound 2a (1 mmol) was prepared by reacting chalcone 1a (1 mmol) in ethanol according to the general procedure. The resulting product yielded 73%, with a purity of 99.02%, a yellow solid, and m.p. 140°C–141°C. ¹H NMR (500 MHz, CDCl₃) δ 7.23 (m, 1H), 7.17 (m, 2H), 7.10 (dd, 7.6; 1.7 Hz, 1H), 6.97 (m, 2H), 6.93 (dd, 8.3; 1.1 Hz, 1H), 6.83 (t, 1H), 6.79 (d, 3.9 Hz, 1H), 6.76 (m, 1H), 6.73 (d, 3.9 Hz, 1H), 5.57 (dd, 6.7; 12.3 Hz, 1H), 3.92 (s, 3H), 3.79 (dd, 12.3; 16.9 Hz, 1H), 2.93 (dd, 6.7; 16.9, 1H). ¹³C NMR (125 MHz, CDCl₃) δ 156.05, 144.43, 142.96, 135.83, 130.89, 128.70, 129.41, 129.01, 126.66, 126.46, 124.79, 121.10, 119.17, 113.28, 110.62, 58.89, 55.56, 42.24. IR: νmax 1589, 1496, 1450, 1388, 1234, 1134, 748. MS (70 eV) m/z (%): 368 (M⁺, 70), 261 (30), 91 (100), 77 (30), 64 (20), 51 (10).

4.1.2. 3‐(5‐Chlorothiophene‐2‐yl)‐5‐(3‐Methoxyphenyl)‐1‐Phenyl‐4,5‐Dihydro‐1H‐Pyrazole (2b)

Compound 2b (1 mmol) was prepared by reacting chalcone 1b (1 mmol) according to the general procedure, using ethanol. The resulting product yielded 59%, with a purity of 98.32%, a yellow solid, and m.p. 103°C–104°C. ¹H NMR (500 MHz, CDCl₃) δ 7.24 (m, 1H), 7.15 (m, 2H), 7.00 (m, 2H), 6.89 (dd, 7.6; 1.3 Hz, 1H), 6.83 (m, 1H), 6.78 (m, 3H), 6.73 (d, 3.9 Hz, 1H), 5.20 (dd, 7.5; 12.4 Hz, 1H), 3.76 (dd, 12.4; 16.9 Hz, 1H), 3.75 (s, 3H), 3.06 (dd, 7.5; 16.9 Hz, 1H). ¹³C NMR (125 MHz, CDCl₃) δ 160.36, 144.57, 143.90, 142.27, 135.47, 131.18, 130.43, 129.01, 126.51, 124.96, 119.58, 118.15, 113.57, 113.10, 111.46, 64.84, 55.35, 43.68. IR: νmax 1589, 1496, 1458, 1381, 1249, 1126, 740. MS (70 eV) m/z (%): 368 (M⁺, 70), 261 (50), 91 (100), 77 (40), 64 (20), 51 (10).

4.1.3. 3‐(5‐chlorothiophene‐2‐yl)‐5‐(2,4‐dimethoxyphenyl)‐1‐phenyl‐4,5‐dihydro‐1H‐pyrazole (2c)

Compound 2c (1 mmol) was prepared by reacting chalcone 1c (1 mmol) according to the general procedure, using ethanol. The resulting product yielded in 94%, with a purity of 99.00%, a yellow solid, and m.p. 119°C–120°C. ¹H NMR (500 MHz, CDCl₃) δ 7.15 (m, 2H), 6.98 (m, 3H), 6.76 (m, 2H), 6.73 (d, 3.9 Hz, 1H), 6.50 (d, 2.4 Hz, 1H), 6.34 (dd, 8.5; 2.4 Hz, 1H), 5.50 (dd, 6.6; 12.2 Hz, 1H), 3.89 (s, 3H), 3.74 (m), 3.75 (s, 3H), 2.91 (dd, 6.6; 16.9 Hz, 1H). ¹³C NMR (125 MHz, CDCl₃) δ 160.35, 157.06, 144.42, 142.96, 135.92, 130.80, 128.99, 127.31, 126.45, 124.73, 121.84, 119.11, 113.28, 104.34, 98.91, 58.60, 55.59, 55.46, 42.37. IR: νmax 1589, 1496, 1450, 1388, 1257, 1111, 740. MS (70 eV) m/z (%): 398 (M⁺, 80), 261 (20), 91 (100), 77 (30), 64 (20), 51 (10).

4.1.4. 3‐(5‐Chlorothiophene‐2‐yl)‐5‐(2,5‐Dimethoxyphenyl)‐1‐Phenyl‐4,5‐Dihydro‐1H‐Pyrazole (2d)

Compound 2d (0.5 mmol) was prepared based on the general procedure of the reaction with chalcone 1d, using methanol. The resulting product yielded 62%, with a purity of 95.58%, a pale yellow solid, and m.p. 112°C–113°C. ¹H NMR (500 MHz, CDCl₃) δ 7.17 (m, 2H), 6.98 (m, 2H), 6.85 (d, 8.8 Hz, 1H), 6.77 (m, 3H), 6.73 (d, 3.7 Hz, 1H), 6.70 (d, 3.1 Hz, 1H), 5.53 (dd, 7.0; 12.3 Hz, 1H), 3.88 (s, 3H), 3.78 (m, 1H), 3.61 (s, 3H), 2.93 (dd, 7.0; 16.9 Hz, 1H). ¹³C NMR (125 MHz, CDCl₃) δ 154.04, 150.13, 144.56, 142.96, 135.75, 130.95, 130.77, 129.00, 126.46, 124.86, 119.31, 113.38, 112.96, 112.69, 111.66, 55.75, 56.04, 59.04, 42.30. IR: νmax 1597, 1496, 1450, 1373, 1211, 1041, 740. MS (70 eV) m/z (%): 398 (M⁺, 70), 275 (50), 261 (30), 91 (100), 77 (40), 64 (20), 51 (10).

4.1.5. 3‐(5‐Chlorothiophene‐2‐yl)‐5‐(2,3,4‐Trimethoxyphenyl)‐1‐Phenyl‐4,5‐Dihydro‐1H‐Pyrazole (2e)

Compound 2e (0.5 mmol) was prepared based on the general procedure of the reaction with chalcone 1e, using methanol. The resulting product yielded 78%, with a purity of 98.34%, a yellow solid, and m.p. 118°C–119°C. ¹H NMR (500 MHz, CDCl₃) δ 7.17 (m, 2H), 7.00 (m, 2H), 6.79 (m, 3H), 6.74 (d, 3.9 Hz, 1H), 6.54 (d, 8.7 Hz, 1H), 5.45 (dd, 6.9; 12.5, 1H), 3.76 (dd, 12.5; 16.8 Hz, 1H), 3.99 (s, 3H), 3.89 (s, 3H), 3.80 (s, 3H), 2.99 (dd, 6.9; 16.8 Hz, 1H). ¹³C NMR (125 MHz, CDCl₃) δ 153.38, 150.60, 144.45, 142.74, 142.35, 135.75, 130.92, 129.02, 127.29, 126.49, 124.83, 121.21, 119.26, 113.35, 107.52, 61.05, 60.94, 59.41, 56.0, 42.78. IR: νmax 1597, 1489, 1450, 1365, 1288, 1087, 748. MS (70 eV) m/z (%): 428 (M⁺, 70), 261 (40), 91 (100), 77 (30), 64 (20), 51 (10).

4.1.6. 3‐(5‐Chlorothiophene‐2‐yl)‐5‐(3,4,5‐Trimethoxyphenyl)‐1‐Phenyl‐4,5‐Dihydro‐1H‐Pyrazole (2f)

Compound 2f (0.5 mmol) was prepared based on the general procedure of the reaction with chalcone 1f, using methanol. The resulting product yielded in 93%, with a purity of 99.31%, a pale yellow solid, and m.p. 117°C–118°C. ¹H NMR (500 MHz, CDCl₃) δ 7.18 (t, 2H), 7.02 (d, 8.1 Hz, 2H), 6.82 (m, 2H), 6.75 (d, 3,7 Hz, 1H), 6.50 (s, 2H), 5.11 (dd, 7.9; 12.4 Hz, 1H), 3.82 (s, 3H), 3.80 (s, 6H), 3.75 (dd, 12.4; 16.9 Hz, 1H), 3.07 (dd, 7.9; 16.9 Hz, 1H). ¹³C NMR (125 MHz, CDCl₃) δ 153.98, 144.93, 142.58, 138.05, 137.26, 135.35, 131.33, 129.03, 126.56, 125.13, 119.82, 113.72, 102.40, 65.53, 60.95, 56.26, 43.87. IR: νmax 1589, 1496, 1450, 1373, 1234, 1126, 748. MS (70 eV) m/z (%): 428 (M⁺, 80), 261 (80), 91 (100), 77 (30), 64 (20), 51 (10).

4.1.7. 5‐(Benzo[d][1,3]Dioxol‐5‐il)‐3‐(5‐Chlorothiophene‐2‐yl)‐1‐Phenyl‐4,5‐Dihydro‐1H‐Pyrazole (2g)

Compound 2g (0.5 mmol) was prepared by reacting chalcone 1g in ethanol according to the general procedure. The resulting product yielded in 88%, with a purity of 99.03%, a brown solid, and m.p. 139°C–140°C. ¹H NMR (500 MHz, CDCl₃) δ 7.17 (m, 2H), 7.01 (d, 8.1 Hz, 2H), 6.76 (m, 6H), 5.90 (s, 2H), 5.16 (dd, 7.2; 12.2 Hz, 1H), 3.71 (dd, 12.2; 16.8 Hz, 1H), 3.02 (dd, 7.2; 16.8 Hz, 1H). ¹³C NMR (125 MHz, CDCl₃) δ 148.52, 147.24, 144.39, 142.19, 136.13, 135.49, 131.18, 129.03, 126.51, 124.94, 119.57, 119.23, 113.56, 108.78, 106.25, 101.29, 64.57, 43.73. IR: νmax 1597, 1489, 1442, 1381, 1234, 1033, 748. MS (70 eV) m/z (%): 382 (M⁺, 50), 261 (20), 91 (100), 77 (30), 64 (20), 51 (10).

4.1.8. 4‐(3‐(5‐Chlorothiophene‐2‐yl)‐1‐Phenyl‐4,5‐Dihydro‐1H‐Pyrazole‐5‐il)‐N,N‐Dimethylaniline (2h)

Compound 2h (1 mmol) was prepared according to the general procedure by reaction of chalcone 1 h in ethanol. The resulting product yielded 88%, with a purity of 98.38%, a yellow solid, and m.p. 130°C–132°C. ¹H NMR (500 MHz, CDCl₃) δ 7.15 (m, 4H), 7.03 (m, 2H), 6.80 (d, 3.9 Hz, 1H), 6.76 (m, 1H), 6.72 (d, 3.9 Hz, 1H), 6.66 (m, 2H), 5.17 (dd, 7.2; 12.2 Hz, 1H), 3.70 (dd, 12.2; 16.7 Hz, 1H), 3.04 (dd, 7.2; 16.7 Hz, 1H), 2.91 (s, 6H). ¹³C NMR (125 MHz, CDCl₃) δ 150.12, 144.66, 142.23, 135.85, 130.85, 129.69, 128.45, 126.81, 126.48, 124.72, 119.28, 113.60, 112.99, 64.47, 43.76, 40.63. IR: νmax 1597, 1496, 1442, 1388, 1118, 732. MS (70 eV) m/z (%): 381 (M⁺, 70), 261 (10), 147 (100), 91 (50), 77 (20), 64 (10), 51 (5).

4.2. Anticancer Activity Assay

The cytotoxic activity of N‐phenylpyrazoline derivatives (2a–2h) was evaluated using the MTT assay against T47D (RRID:CVCL_0553), MCF‐7 (RRID:CVCL_0031), and WiDr (RRID:CVCL_2760) cancer cell lines, alongside normal Vero (RRID:CVCL_0059) cells to determine selectivity. Cells were seeded in 96‐well plates and incubated at 37°C in 5% CO₂ for 24 h. Test compounds and positive controls (doxorubicin and 5‐fluorouracil) were prepared as stock solutions in DMSO and serially diluted in culture medium to final concentrations of 3.13–400 µg/mL. We tested each concentration in triplicate (n = 3 technical replicates). After 24 h of treatment, cells were washed with PBS and incubated with MTT solution for 4 h. Formazan crystals were dissolved using 10% SDS, and absorbance was measured at 595 nm using a microplate reader. We calculated cell viability, determined IC₅₀ values as in Equation (1), and calculated the selectivity index (SI) by comparing cytotoxicity against cancer and normal cells, as shown in Equation (2).

$$\mathrm{Cell}\mathrm{viab}\mathrm{ility}=\frac{{\mathrm{A}}_{\mathrm{treated}\mathrm{cell}}-{\mathrm{A}}_{\mathrm{blank}}}{{\mathrm{A}}_{\mathrm{control}\mathrm{cell}}-{\mathrm{A}}_{\mathrm{blank}}}\times 100\%$$ (1)

$$\mathrm{Sele}\mathrm{ctivity}\mathrm{index}=\frac{{\mathrm{IC}}_{50(\mathrm{normal}\mathrm{cell})}}{{\mathrm{IC}}_{50(\mathrm{cancerous}\mathrm{cell})}}$$ (2)

4.3. Computational Methods

4.3.1. Molecular Docking

The structure of ligand 2g was initially constructed using GaussView 5.0 and subsequently optimized with Gaussian 09 W software [60], following our previous work by Wulan et al. [24]. Geometry optimization was performed using density functional theory (DFT) at the B3LYP/6‐31G level to obtain the energetically favored conformation. The crystal structure of COX‐2 complexed with celecoxib (PDB ID: 3LN1) was retrieved from the Protein Data Bank. Molecular docking simulations were carried out using AutoDock Vina [61]; AutoDock Tools was used for system preparation and grid generation. Docking protocol validation was performed by redocking the native ligand into the COX‐2 active site. We used a cubic grid box (15 × 15 × 15 Å3) with a grid spacing of 1.000 Å, centered at x = 77.873, y = −15.448, and z = −7.948. All docking simulations were conducted with an exhaustiveness value of 100, and identical parameters were applied to all ligands. Protein–ligand interactions were analyzed and visualized using Discovery Studio Visualizer 2017 (DSV 2017) [62].

4.3.2. Molecular Dynamics Simulation

Molecular dynamics (MD) simulations were performed using the YASARA software package [63]. Protein–ligand complexes were generated from the docking results obtained with AutoDock Vina. Before simulation, the complexes were energy‐minimized using the steepest descent algorithm with semi‐empirical quantum mechanical calculations based on the Austin Model 1 (AM1) method37. Periodic boundary conditions were applied to define the simulation cell. Following preparation, MD simulations were executed using a modified md_run.mcr script, with the temperature maintained at 310.15 K, pressure control set to manometer (1D) with pressure = 1 atm, and a total simulation time of 120 000 ps. Upon completion of the simulations, trajectory analyses were performed using the md_analyze.mcr script to obtain RMSD, RoG, SASA, RMSF, and secondary structure content (DSSP).

4.4. Pharmacokinetic Prediction (ADMET)

The ADMET (absorption, distribution, metabolism, excretion, and toxicity) properties of 2a–2h were predicted using the pkCSM web server [64]. We assessed absorption using predicted aqueous solubility and human intestinal absorption (HIA). Distribution parameters included predicted volume of distribution, blood–brain barrier permeability (BBB), and central nervous system (CNS) distribution. Metabolic behavior was evaluated by predicting cytochrome P450 enzyme inhibition to estimate metabolic stability. Excretion was analyzed using predicted total clearance and renal elimination. Toxicity was predicted using rat‐based models to provide an initial estimate of the compounds' safety profile.

Author Contributions

Nikma Farismatur Riza: writing – original draft, methodology, formal analysis, data curation, investigation. Fia Fathiana Wulan: writing – review and editing, formal analysis, investigation, visualization. Tegar Asanda Ghifari: writing – review and editing, formal analysis, investigation. Endang Astuti: writing – review and editing, validation, supervision, visualization. Tutik Dwi Wahyuningsih: writing – review and editing, conceptualization, validation, resources, supervision, visualization, project administration, funding acquisition.

Conflicts of Interest

The authors declare no conflicts of interest

Use of Generative AI and AI‐Assisted Technologies in the Writing Process

During the preparation of this work, the authors used Claude (Anthropic) to assist with language refinement. After using this tool, the author reviewed and edited the content as needed and take full responsibility for the content of the publication.

Supporting information

Supporting Information for this article is available under https://doi.org/10.1002/cbdv.71787.

Supporting File: cbdv71787‐sup‐0001‐SuppMat.docx

Data Availability Statement

The data that support the findings of this study are available in the Supplementary Material of this article.