Design, Synthesis, and Evaluation of Bis‐Chalcones and Bis‐Flavones as Selective COX‐2 Inhibitors

ABSTRACT

Over the past two decades, the search for safe and effective COX‐2 inhibitors has spurred extensive research on flavonoids. Within this context, synthetic flavonoid dimers have emerged as a promising subclass with potential anti‐inflammatory activity. To investigate whether dimerization enhances their potency and selectivity, novel A‐fused bis‐chalcones and A‐fused bis‐flavones were synthesized and evaluated for their inhibitory activity against isolated human COX‐1 and COX‐2 enzymes, as well as their effects on prostaglandin E₂ production in whole human blood. Interestingly, the most active compound identified was a monomeric chalcone sharing the same substitution pattern as the top‐performing bis‐chalcone, suggesting that key structural features drive activity regardless of dimerization. While the dimeric nature of bis‐chalcones did not enhance COX‐2 inhibition or selectivity in this study, these results provide valuable insights into structure–activity relationships. Furthermore, all active compounds against the isolated enzyme showed reduced potency in whole blood, possibly due to plasma protein binding limiting bioavailability. This study highlights the importance of rational design for further development of dimeric flavonoids, in particular strategies aimed at optimizing bioavailability.

Novel bis‐chalcones were successfully synthesized and demonstrated COX‐2 selective inhibition and the ability to inhibit PGE2 production in whole human blood. Notably, a monomeric chalcone with a similar substitution pattern exhibited even greater activity, providing valuable insights to guide the future design and optimization of flavonoid‐based COX‐2 inhibitors.

Abbreviations

AA

arachidonic acid

COXs

cyclooxygenases

DCM

dichloromethane

DIPEIA

N,N‐diisopropylethylamine

DMSO

dimethyl sulfoxide

LPS

lipopolysaccharide

MEM

methoxyethoxymethyl ether

MOM

methoxymethyl ether

NSAIDs

nonsteroidal anti‐inflammatory drugs

PGs

prostaglandins

PI

propidium iodide

TFA

trifluoroacetic acid

TLC

thin‐layer chromatography

TMS

tetramethylsilane

TXBSI

thromboxane synthase inhibitor

1. Introduction

Cyclooxygenase (COX), also known as prostaglandin‐endoperoxide synthase, is a well‐established, extensively studied pharmacodynamic target for the development of new anti‐inflammatory drugs (Rouzer and Marnett 2009). This enzyme catalyzes the conversion of arachidonic acid into prostaglandins (PGs), key inflammatory mediators responsible for classic symptoms such as fever, swelling, and pain (Ribeiro, Freitas, Lima, and Fernandes 2015). Specifically, COX catalyzes the multistep conversion of AA into PGH₂, the precursor of all other PGs synthesized in specific tissues (Ribeiro, Freitas, Lima, and Fernandes 2015).

COX has several homodimeric isoforms (COX‐1, −2, and −3) located in the endoplasmic reticulum and nuclear envelopes of various cells types. These isoforms share similar biological functions, exhibit up to 65% sequence homology, and have comparable molecular weights of approximately 72 kDa (Rouzer and Marnett 2009). COX‐1, in particular, is a constitutive isoform expressed consistently at low levels in most tissues, where it supports essential homeostatic functions such as gastric mucosal protection and vasoconstriction, showing only a modest increase during inflammation (Ribeiro, Freitas, Tomé, et al. 2015). In turn, COX‐2 is an inducible isoform, triggered by cytokines, endotoxins, or mitogens, and can be upregulated up to 20‐fold, leading to elevated levels of PGs during inflammation. (Ribeiro, Freitas, Tomé, et al. 2015) A key structural difference between COX‐1 and COX‐2 lies in their binding pocket size, which is almost 17% greater in the latter isoform due to a valine replacing an isoleucine amino acid at position 523. (Rouzer and Marnett 2009; Ahmadi et al. 2022) This characteristic has led to the design of structurally larger inhibitors that are sterically incompatible with the narrower COX‐1 substrate‐binding site.

COX‐3 is a variant of COX‐1, predominantly expressed in the brain. Its physiological role remains less defined than COX‐1 and COX‐2, yet its involvement in central pain and fever regulation suggests its potential as a therapeutic target for certain analgesic and antipyretic drugs (Biswas et al. 2023).

Inhibition of COX is a well‐established mechanism of action for nonsteroidal anti‐inflammatory drugs (NSAIDs). By blocking COX activity, NSAIDs reduce PG levels, alleviating inflammation‐related symptoms (Ribeiro, Freitas, Lima, and Fernandes 2015). However, most NSAIDs inhibit COX nonselectively, and secondary effects like gastrointestinal ulcers and bleeding, particularly with prolonged use. Selective inhibitors of COX‐2 (coxibs) were designed to mitigate these side effects (Arora et al. 2020). However, for some of them, such as rofecoxib and valdecoxib, an increased risk of cardiovascular diseases, including strokes, heart attacks, and blood clots, was evidenced and eventually led to their withdrawal from the market (Arora et al. 2020). Consequently, the development of safer but effective COX‐2 inhibitors with reduced long‐term side effects remains critical.

Flavonoids are natural polyphenolic compounds with well‐established bioactivities (Ribeiro, Freitas, Lima, and Fernandes 2015). They are characterized by their common rigid backbone, a three‐ring system of 15 carbons, two aryl rings, A and B, connected by an O‐heterocyclic ring called C (Ribeiro, Freitas, Lima, and Fernandes 2015). Many flavonoids and their chalcone precursors are widely recognized for their remarkable antioxidant activities (Ana et al. 2008; Gomes et al. 2012). They have been further tested as COX inhibitors (Ribeiro, Freitas, Tomé, et al. 2015; Sogawa et al. 1993) and COX expression modulators (Ribeiro, Freitas, Lima, and Fernandes 2015; López‐Posadas et al. 2010). Flavonoids, such as luteolin, apigenin, kaempferol, among others, have been reported to act as effective modulators of COX‐2 expression and to directly inhibit COX‐2 enzymatic activity (Ribeiro, Freitas, Lima, and Fernandes 2015; Ribeiro, Freitas, Tomé, et al. 2015).

Still, a significant number of flavonoid subcategories are understudied. Dimeric flavonoids, such as biflavonoids or proanthocyanins, have in the past decades attracted scientific interest, primarily because many of them were originally identified in medicinal plants. One newly proposed subclass, A‐fused bis‐flavonoids, represents fully synthetic dimers, in which two flavonoid units are fused through their A rings. Cyclization of bis‐chalcones is one of the preferred routes to obtain these compounds. Converting chalcones into flavonoids remains one of the most widely used synthetic strategies, as it mimics the natural biosynthetic pathways of these compounds (Pereira et al. 2023).

Given their unexplored structural features and potential, this research aims to assess whether dimeric A‐fused bis‐chalcones and A‐fused bis‐flavonoids constitute novel scaffolds for anti‐inflammatory drug development. For comparison, a monomeric chalcone was also evaluated as a potential COX‐2‐selective inhibitor. Assessment of the anti‐inflammatory activity of the tested compounds was performed using both isolated enzyme assays and a whole human blood model.

2. Results and Discussion

2.1. Synthesis

In this section, the synthesis of the tested compounds will be described, which were fully characterized by ¹H and ¹³C NMR (Figures S1–S27), as well as by mass spectrometry (MS) and high‐resolution mass spectrometry (HRMS) (Figures S28–S35). Among the various methods for the synthesis of flavones described in the literature, the cyclodehydrogenation reaction of bis‐chalcones, the precursors of bis‐flavones, was the selected method to obtain the A‐fused bis‐flavones described in the present work. This method was chosen because it is the only one that was previously implemented for the synthesis of bis‐flavones (Pereira et al. 2023). Bis‐chalcones were initially synthesized through Claisen–Schmidt condensation of compounds 4a,b with arylaldehydes 5–7. One methyl group, intended to remain throughout the synthesis, plus two methoxyethoxymethyl (MEM) or methoxymethyl (MOM) groups (for easy removal) were added to protect the hydroxy groups of the initial moiety.

Scheme 1 illustrates the synthetic pathway used to make the bis‐chalcone precursor. Phloroglucinol 1 undergoes Friedel‐Crafts acylation with acetic anhydride, catalyzed by BF₃‐Et₂O, yielding compound 2 (68%). Compound 2 was doubly protected by reaction with either MOM‐Cl or MEM‐Cl in basic medium in the presence of N,N‐diisopropylethylamine (DIPEIA) at room temperature. A mixture of mono‐ and di‐protected compounds was always formed, requiring rigorous purification by column chromatography using hexane/ethyl acetate (3:2) as eluent. The presence of two MOM or MEM moieties was confirmed by ¹H NMR. The MOM groups are composed of two symmetric CH₂ with their four protons showing chemical shifts corresponding to a singlet at δ _H = 5.22 ppm and another singlet at δ _H = 5.16 ppm for the two protons of the last methylene group (Figure S3). The resonance of the terminal methyl groups exhibited the same pattern, corresponding to two singlets at δ _H = 3.40 ppm and δ _H = 3.24 ppm, corresponding to three protons each signal. The resonances of the protons of the two acetyl groups were identical, appearing as one singlet, integrating to six protons at δ _H = 3.47 ppm.

SCHEME 1.

Synthesis of compound 4, the bis‐chalcone central unit/core.

The MEM groups enabled singlets at δ _H = 5.27 ppm with four protons corresponding to the resonance of the two symmetric CH₂ and δ _H = 5.01 ppm with two protons due to the resonance of the remaining CH₂ group at the 2‐position. The terminal methyl groups appear at similar chemical shifts to those observed for the MOM terminal methyl groups. However, additional multiplet signals were detected at δ _H = 3.79 ppm (four protons from two symmetric CH₂), δ _H = 3.73 ppm (two protons from the CH₂ of the MEM group at 2‐position), and δ _H = 3.55 ppm for the remaining six protons of the methyl groups.

Integration of the NMR signals showed complete symmetry of the molecule and confirmed the protection of the two hydroxy groups of 2. The remaining hydroxy group of 3 was alkylated with dimethyl sulfate (Me₂SO₄) in dichloromethane (DCM) using potassium carbonate as the base. This reaction produced compound 4 in moderate yield (63%).

Afterward, Claisen–Schmidt condensation of compound 4 with benzaldehydes 5–9 in strong alkaline conditions at room temperature yielded bis‐chalcones 10–12 (Scheme 2). Their structure was confirmed by NMR, displaying the typical signals of chalcones, but with double integration due to their symmetry (Figure S6). More specifically, two signals corresponding to symmetric H‐α and H‐β protons of the two unsaturated carbonyl systems, and the signal due to the resonance of H‐5 of the central A‐ring were observed. All bis‐chalcones 8–13 showed two doublets due to the resonance of H‐α at δ _H = 7.35–7.44 ppm and H‐β at δ _H = 6.87–7.09 ppm, both with a coupling constant J _Hα,Hβ = 16.1–16.2 Hz characteristic of double bonds with an E‐configuration. These bis‐chalcones underwent MOM or MEM deprotection in DCM/trifluoroacetic acid (TFA) at room temperature, and the polyhydroxylated bis‐chalcones 11–13 were obtained, as shown in Scheme 2 (Singh et al. 2014; Gula'csi et al. 1998; Durgapal et al. 2020). These mild conditions were strong enough for MOM or MEM removal, but incapable of removing the remaining methyl group. To synthesize A‐ring fused bis‐flavones, the cyclodehydrogenation step required unprotected hydroxyls at positions 4 and 6 of the core ring; the methyl group was also essential for the success of the reaction.

SCHEME 2.

Synthesis of bis‐chalcones 11–13.

Bis‐flavones 14 and 15 were synthesized in 30%–35% yield by cyclodehydrogenation of bis‐chalcones 11 and 12 using catalytic I₂ in DMSO at reflux for 30 min (Scheme 3).

SCHEME 3.

Synthesis of bis‐flavones 14–15.

The formation of bis‐flavones 14–15 was confirmed by both NMR and MS analysis, as well as by the appearance of fluorescence in the reaction product. ¹H NMR analysis confirmed the disappearance of the doublet signals of the symmetrical H‐α and H‐β protons, accompanied by the appearance of a new singlet at δ _H = 6.63 ppm, corresponding to the two symmetric unsaturated protons at the 3‐position of the C rings. Additionally, the reaction induced an upfield change of the chemical shift of proton H‐5, while the remaining protons retained their original characteristics (Figure S25). The new A‐fused bis‐flavones, rich in methoxy groups, represent a scarcely explored flavonoid dimer class; their further derivatization is ongoing.

2.2. In Vitro COX‐2 and COX‐1 Inhibition

The synthesized bis‐chalcones (11–13) and bis‐flavones (14–15) were tested in vitro as inhibitors of human recombinant COX‐2 and COX‐1 enzymes. The study continued by exploring structure–activity relationships, specifically the influence of hydroxy groups on potency and the significance of the bis‐chalcone dimer. Therefore, the mono derivative eriodictyol chalcone 16, purchased from Enamine, was tested and compared with two previously synthesized bis‐chalcones 17 and 18 (data not shown). Bis‐chalcone 17 features two catechol moieties and acts as a selective inhibitor of COX‐2, and the chlorinated bis‐chalcone 18 is a non‐selective inhibitor of both COX‐2 and COX‐1 (Figure 1). Celecoxib and SC‐560, two established positive controls, allowed for scaling based on inhibitory potency.

FIGURE 1.

Chemical structures of the tested compounds.

To prevent bias from spectral overlap and methodological interference, UV–Vis absorption and fluorescence spectra of all compounds were checked before the assay (data not shown).

Table 1 depicts the results obtained for the two A‐fused bis‐flavones, all the bis‐chalcones and monomeric chalcone tested, and the positive controls used, celecoxib and SC‐560.

TABLE 1.

Chemical structure and in vitro inhibitory activity of the compounds under study on COX‐2 and COX‐1.

No.	Structure	COX‐1 inhibitory activity (%) or IC50 value (mean ± SEM)	COX‐2 inhibitory activity (%) or IC50 value (mean ± SEM)	Selectivity index
11		< 30%100 μM	< 30%100 μM	—
12		< 30%100 μM	< 30%100 μM	—
13		19.2 ± 0.1 μMbc	2.0 ± 0.3 μM ab	9.6
14		< 30%100 μM	< 30%100 μM	—
15		< 30%100 μM	< 30%100 μM	—
16		21 ± 2 μMc	0.58 ± 0.03 μMa	36.2
17		17.0 ± 0.6 μMbc	1.5 ± 0.3 μMab	11.3
18		10.6 ± 0.6 μMb	8.6 ± 0.7 μMc	1.2
Celecoxib		50 ± 1 μMd	0.20 ± 0.04 μMa	250
SC‐560		0.0114 ± 0.0007 μMa	4.0 ± 0.2 μMb	0.003

Note: Concentrations shown in superscript indicate the highest concentrations that could be tested without interference with the methodology. The IC₅₀ with different lowercase superscript letters are significantly different from each other (p < 0.05). Statistical analysis was only applied within the same enzyme.

Bis‐chalcones 11 and 12 showed no inhibitory activity against the studied COX isoforms, whereas all other bis‐chalcones exhibited measurable activity. Notably, the monomeric chalcone 16 exhibited the lowest IC₅₀ at 0.58 ± 0.03 μM and demonstrated the highest selectivity, being 36.2 times more potent against COX‐2 than COX‐1. Compound 13 displayed similar potency to bis‐chalcone 17, presenting an IC₅₀ of 2.0 ± 0.3 μM for COX‐2, with a selectivity index (SI) of 9.6. The two A‐fused bis‐flavones (compounds 14 and 15) also showed no inhibitory activity against either COX isoform.

Bis‐chalcones 11–13, synthesized and tested in this study, are meant to be compared to fully unprotected bis‐chalcones, represented by compound 17. In contrast to compound 17, they possess only a single methylated hydroxy group on ring A. This structural difference may help elucidate the significance of that hydroxy group for the activity. Compound 17 is, in theory, thermodynamically more stable in a V shape (Figure 2), where two hydroxy groups establish internal hydrogen bonds with the carbonyl groups, leaving one free hydroxy group unpaired. This remaining hydroxy group could be important for the activity of the molecule.

FIGURE 2.

Structure of bis‐chalcone 17 optimized by DFT calculations (B3LYP/6‐31G).

By protecting this hydroxy group of the central ring (17–13), the bis‐chalcone showed no statistically relevant difference (IC₅₀ value of 2.0 ± 0.3 μM for COX‐2, SI = 9.6). This result highlights that the side rings possess a stronger influence over the activity of the bis‐chalcones, especially the catechol moiety, compared to that specific hydroxy group. Yet, this group might have a small role to play in the interaction with COX‐1.

The other two bis‐chalcones, 11 and 12, showed no activity toward any of the enzymes. This result highlights the importance of substituent groups at the side rings on the observed inhibition effect. A stronger influence over the activity of the bis‐chalcones is especially noticed with the catechol moiety.

Bis‐chalcone 18 is the only unspecific COX inhibitor in this study, with an SI of 1.2, a property given by its two chlorine groups in the B rings and fewer hydroxy groups present.

Bis‐chalcones are an underexplored class of compounds regarding their potential as COX‐2 inhibitors. El‐Sabbagh et al. (2013) synthesized a series of bis‐chalcone analogs containing N‐arylpyrazoles as central core and different halogens and tested them in vitro against isolated ovine COX‐2 and ovine COX‐1. These compounds were designed to be closely related to the chemical structure of coxib‐type drugs. Their most potent compound, with methoxy groups on ring B and p‐chloroaryl substituent in the A‐ring, had an IC₅₀ for COX‐2 of 0.29 μM and an exceeding selectivity index of 330. Unlike the bis‐chalcones evaluated in this study, their most active compound lacked any hydroxy group or other hydrogen donor group. Despite lacking this feature, these compounds were very active. Given the different enzyme sources, direct comparisons to the results reported in this work need careful consideration. Ovine ortholog versions of these enzymes are commonly used due to being cheaper than human isoforms while sharing around 61% sequence homology and similar active pockets (Rouzer and Marnett 2009; Filizola et al. 1997). However, even minor variations in enzymes between animals and humans significantly affect the translatability of findings (Rocha et al. 2025).

One of the most active compounds was 16, the monomeric chalcone. This chalcone has the same substitution pattern as synthetic bis‐chalcone 17, a catechol moiety on ring B, and three hydroxy groups at ring A. It has been hypothesized and evidenced that the dimeric property of bis‐chalcone favors some bioactivities over monomeric chalcones (Pereira et al. 2023). However, it was also made clear that the substitution pattern of bis‐chalcones has a greater net effect regarding potency than its dimeric feature. A direct comparison of compounds 16 and 17 reveals that the added chalcone and catechol moieties did not enhance, but slightly worsened, COX‐2 selectivity. While both compounds do not vary greatly from each other, it is an important conclusion regarding the role of the dimeric nature of bis‐chalcones for COX‐2 inhibition. Another key discovery is the improved selectivity toward COX‐2, an unexpected result. Rational drug design for selectivity toward COX‐2 over COX‐1 comes from the analysis of the differences between the active pocket of both enzymes. From COX‐2 to COX‐1, two key valine residues are changed (Val434 and Val 523) to isoleucine (Ile434 and Ile 523). The bulkier isoleucine residues in COX‐1 reduce the available space and remove a side pocket characteristic of COX‐2. Therefore, the pocket of COX‐1 is much smaller and more rigid compared to COX‐2 (Rouzer and Marnett 2009). Also, the entrance channel for the active site of COX‐2 is larger than COX‐1. Given this information, one might expect larger molecules, such as bis‐chalcones, to more readily access the active pocket of COX‐2 than that of COX‐1. However, this generalization is not universally valid. Compound 18, a bis‐chalcone with a similar size to 17, was found to be an unspecific inhibitor of both COX isoforms, highlighting the superior role of the chlorine substitution pattern over the bis‐chalcone skeleton. Compound 16 was also one of the most selective inhibitors found, meaning that despite entering both pockets due to its reduced size, its interactions are specific and effective with the COX‐2 pocket over COX‐1.

Chalcone 16 has previously been investigated for its anti‐inflammatory properties through alternative mechanisms of action. Sogawa et al. (1993) tested several hydroxychalcones, including chalcone 16, against different enzymes, including in vitro 5‐lipoxygenase from RBL‐1 cells, in vitro cyclooxygenase from sheep vesicles, and in vitro lipid peroxidation from rat liver microsomes. It showed good activity against 5‐lipoxygenase and lipid peroxidation, but low activity against COX. The differences between the results reported by Sogawa et al. (1993) and those obtained in this work pinpoint to variations in the experimental protocols, particularly the source of the COX enzyme and the lack of isoform specificity. For instance, the COX used was the sheep seminal vesicle COX, whereas this work employed the human enzyme. Sousa et al. (2020) explored the capability of several chalcones to modulate the oxidative burst in neutrophils and the release of neutrophil extracellular traps (NETs). Chalcone 16 was amongst the most active chalcones in inhibiting the production of reactive species, and it was the only compound capable of modulating NETs release in a concentration‐dependent manner (Sousa et al. 2020). The potential of this chalcone as a promising anti‐inflammatory agent with multiple mechanisms of action should hence be considered. This result contributes to the current literature by presenting a study that uses isolated human recombinant COX‐1 and COX‐2 enzymes, employing a method that is easily translated to human applications.

The lack of activity of A‐fused bis‐flavones 14 and 15 can probably be attributed to the absence of free hydroxy groups, a critical feature in all active compounds present in this study and commonly found in flavonoid‐based COX inhibitors described in the literature (Md Idris et al. 2022). Previous reports indicate that flavones lacking free hydroxy groups failed to inhibit COX‐2 activity in isolated enzyme assays and whole blood tests (Ribeiro, Freitas, Tomé, et al. 2015; Md Idris et al. 2022). Within the flavonoid dimer subclass, several naturally occurring biflavonoids have already shown the capability to inhibit COX‐2, but all these compounds contain several free hydroxy groups. Park et al. (2006) synthesized a new family of synthetic flavonoid dimers without any hydroxy groups that inhibited the COX‐2‐mediated PGE₂ production in Lipopolysaccharide (LPS)‐treated RAW cells. It should be noted that a direct comparison with the results of this work is not feasible, as the authors evaluated an indirect measure of COX activity in LPS‐treated RAW cells rather than testing the isolated enzyme. Lim et al. (2009) later synthesized another synthetic bis‐flavonoid with potent COX‐2 inhibitory activity, but this derivative contained two hydroxy groups. This study represents the first evaluation of A‐fused bis‐flavonoids against isolated human COX‐1 and COX‐2 enzymes, highlighting the need for further research to fully explore the potential of this new subclass.

2.3. Inhibition Kinetic Analysis

To further understand the mechanism of COX‐2 inhibition by this derivative, inhibition‐type kinetic studies were performed. Exhibiting one of the most potent and selective in vitro COX‐2 inhibitory profiles, compound 16 was selected for further studies to determine its mode of inhibition. These studies considered both the inspection of Lineweaver–Burk plots and statistical analysis of experimental data using nonlinear regression. Microsoft Office Excel and its add‐in Solver were used for these calculations (Figures S36–S39; Bezerra et al. 2013). A total of three experiments were conducted, considering three inhibitor concentrations in the range of 0–0.63 μM and three substrate concentrations in the range of 6.25–100 μM. The experimental data were obtained from the slope of the kinetic reaction from the first 2 min after substrate addition, corresponding to the initial velocity of the linear steady state phase. This data was sequentially fitted to various inhibition models (without inhibition, competitive, noncompetitive, uncompetitive, or mixed inhibition). From this fitting, the corresponding kinetic parameters and the sum of the square errors' values for each model were obtained and compared to determine the best model. The extra sum‐of‐squares F test and the Akaike information criterion (AIC) test were further considered for model validation. The most appropriate model was chosen by comparing the extra sum of the square error value F test (f: 0.95). If the –f value is superior to 0, the more complex model should be applied. Additionally, the AIC was employed with the model with the lowest corrected AIC score deemed the most plausible. Based on the experimental data, chalcone 16 demonstrated a mixed‐type inhibition mechanism for COX‐2 (Figure 3). The corresponding kinetic constant values (V _max, K _m, K _ic, and K _iu) are presented in Table 2.

FIGURE 3.

Lineweaver–Burk plot of COX‐2 inhibition by compound 16. The mean values are represented by the different symbols. The fitted lines correspond to the best‐fit mixed inhibition model.

TABLE 2.

Type of inhibition (using Solver supplement) of compound 16 against COX‐2 activity and respective kinetic parameters values: V _max, K _m, K _ic, and K _iu (mean ± SEM).

Type of COX‐2 inhibition	V max (μmol/min)	K m (μM)	K ic (μM)	K iu (μM)
Mixed	2800 ± 91	6.0 ± 0.9	0.20 ± 0.03	0.34 ± 0.03

A mixed‐type inhibitor can bind to both the free enzyme and the enzyme–substrate complex, affecting both K _m and V _max. This can be confirmed by visual analysis of Figure 3. With the increase of inhibitor concentration, the V _max (intersection with y‐axis equals 1/V _max) reduces and K _m (intersection with x‐axis equals −1/[K _m]) increases, a pattern consistent with a mixed‐type inhibition. Still, Lineweaver–Burk plots are not the most optimal method to assess experimental error, so statistical fitting with nonlinear regression using the Solver add‐on was necessary to complement this result (Bezerra et al. 2013).

The presence of multiple hydroxy groups in 16 granted exceptional binding capabilities with many acceptors through hydrogen bonds. Therefore, it is expected that this molecule is not confined solely to the active site of the enzyme but may also interact with other regions.

2.4. Viability of Erythrocytes and Leukocytes

To conduct human blood assays, viability tests with erythrocytes and leukocytes were performed using the active compounds against in vitro isolated COX‐2 at 100 μM. In erythrocyte viability, a statistically relevant loss was detected with compounds 13 and 17 (Figure 4) at 100 μM, while after a reduction to 75 μM, no differences were observed compared to the control. The decrease in viability caused by compound 18 was minimal and not statistically different from the control. Compounds 16 and 18 showed no effect on erythrocyte viability at 100 μM.

FIGURE 4.

Erythrocyte viability (%) based on hemoglobin absorbance, after incubation with bis‐chalcones. Each value represents mean ± SEM of at least three experiments. **p < 0.01, ***p < 0.001, compared to the control (without compound under study, 100% viability).

Following these results, leukocyte viability was assessed using Annexin V/PI (propidium iodide), with compounds 13, 17, and 18 at 75 μM and compound 16 at 100 μM (Figure 5). In the assays, the nonionic surfactant Triton X‐100 served as the positive control.

FIGURE 5.

Flow cytometry‐based evaluation of leukocyte viability using Annexin V and PI after 5 h of incubation. (a) Annexin V+/PI−; (b) Annexin V+/PI+. Each value represents mean ± SEM of at least three experiments. (c) Representative flow cytometry plots of Annexin V ([Annexin V; x‐axis]/PI [y‐axis]) binding assay for untreated control; (d) treated with chalcone 16 (100 μM).

No apoptotic behavior was detected under the conditions tested for any compound; therefore, these concentrations were used in subsequent assays.

2.5. Determination of PGE₂ Production in Human Whole Blood

The impact of bis‐chalcones 13, 17–18, and chalcone 16 on PGE₂ levels was evaluated using whole human blood. Whole blood PGE₂ quantification is an in vitro assay in which human blood cells are stimulated with LPS to induce a pro‐inflammatory state, leading to COX‐2 overexpression and subsequent PGE₂ overproduction. PGE₂ is the most prevalent prostaglandin, exerting several physiological and pathological effects associated with most inflammatory diseases (Legler et al. 2010). Therefore, targeting the production of this prostaglandin can be directly associated with several physiological benefits in inflammatory disease treatment.

This method was chosen because it employs a complex biological matrix, including erythrocytes, leukocytes, platelets, plasma proteins, and their active cell–cell interactions that more accurately reflects in vivo conditions than assays with isolated enzymes, thereby offering a more reliable model for drug development. Therefore, whole blood assays allow compounds to be tested at higher concentrations ex vivo while better replicating the physiological environment found in living humans.

Each compound was tested in at least three different concentrations. Celecoxib was used as a positive control and exhibited a PGE₂ production inhibitory activity of 79% ± 5% at 5 μM. All the tested compounds showed the capability to effectively reduce the PGE₂ levels at the highest concentration tested (Figure 6).

FIGURE 6.

Impact of compounds 16, 17, 18, and 13 on PGE₂ production by blood inflammatory cells as determined by enzyme immunoassay. Each value represents mean ± SEM of at least three experiments; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.001 significance thresholds after two tail t‐student comparison with control sample results (stimulated with LPS).

Chalcone 16 inhibited 54% ± 4% PGE₂ production at 75 μM, the highest concentration tested. Bis‐chalcone 17 achieved an inhibition of 44% ± 7% at the same concentration. Statistically, all bis‐chalcones showed similar potency at the maximum concentration of 75 μM. At 50 μM, chalcone 16 preserved most of its potency, reaching 46% ± 5% inhibition, whereas the bis‐chalcones experienced a marked reduction in activity, with inhibition levels not exceeding 20%. This makes chalcone 16 the most active compound against isolated COX‐2 since it showed significant activity at 50 μM, near 50% inhibition, but still less active than the positive control, celecoxib.

Compared to isolated enzyme in vitro assays, all compounds showed lower potency in whole blood assays. These differences can be ascribed to the different nature of the methods used. The assay with isolated enzyme only follows the reaction of arachidonic acid with COX‐2 in a simplified environment, while the whole blood assay encompasses a complex matrix where COX‐2 activity is not specifically isolated.

Hence, by measuring PGE₂ levels, we can assess the inhibition of PGE₂ production but not account for additional pharmacodynamic and pharmacokinetic variables. To our knowledge, no prior research has investigated the inhibitory effects of bis‐chalcones on PGE₂ production within whole human blood or comparable complex systems. Although only compounds active against isolated COX‐2 were tested, they may also affect other steps in the PGE₂ biosynthesis, such as modulation of COX‐2 expression or inhibition of other enzymes involved in the pathway, such as phospholipase A₂. To reach COX‐2, overexpressed in different leukocytes (mainly monocytes and macrophages) after LPS stimulation, the compounds must traverse the plasma, cross cell membranes, and access intracellular compartments such as the endoplasmic reticulum and nuclear envelope (Morita et al. 1995). This path will reduce the concentration of inhibitors that effectively reaches the enzyme, thereby diminishing its observed effect. Studying the pharmacokinetics of these compounds could be done experimentally or computationally. Given the screening level of these assays, we opted for a computational approach.

Some computational software packages exist to predict the administration, distribution, metabolism, excretion, and toxicity (ADMET) properties of drug candidates in earlier phases of drug development. In this work, the plasma protein binding (PPB) potential is the most applicable. The software PreADMET, available online, was used to theoretically predict this data for the novel bis‐chalcones alongside the positive control, celecoxib, and chalcone 16, based on 2500 molecular descriptors: constitutional, topological, electrostatic, physico‐chemical, and geometrical. (Lee et al. 2004) According to Table 3, all tested compounds had superior plasma protein binding compared to the positive control, which corroborates the experimental results. Although compound 16 displayed an IC₅₀ for COX‐2 inhibition in the in vitro assay similar to that of celecoxib, in the whole blood assay, celecoxib proved to be 10‐fold more potent.

TABLE 3.

Theoretical PPB values for compounds 13, 16–18, and the positive control celecoxib obtained with PreADMET web‐based software.

Compound	PPB value (%)
17	100
13	100
16	100
18	100
Celecoxib	91

Nevertheless, all synthetic bis‐chalcones tested were able to reduce PGE₂ levels. However, their activity profiles became more homogeneous in the whole blood assay. For example, bis‐chalcones 17 and 18 exhibited markedly different inhibitory potentials in the in vitro assay yet both achieved similar effectiveness in the whole blood assay. The superior potency of compound 17 observed with the isolated COX‐2 enzyme was not maintained in whole blood, likely due to its limited ability to effectively reach the enzyme. Although the presence of seven hydroxy groups, including two catechol moieties and three central hydroxy groups, provided optimal binding potential to COX‐2, it also enhanced plasma protein binding, thereby reducing its bioavailability.

To the best of our knowledge, bis‐chalcones have never been tested before as PGE₂ production inhibitors in whole human blood or any other complex system.

3. Conclusion

This study represents the first step toward expanding the library of A‐fused bis‐flavones. Two new derivatives featuring several methoxy groups were successfully synthesized. Further deprotection and derivatization of these compounds are necessary to fully assess the potential of this novel class of flavonoid dimers. Several new bis‐chalcone precursors, along with a monomeric chalcone, were evaluated against isolated COX‐1 and COX‐2, as well as for their ability to inhibit PGE₂ production in whole human blood. Notably, while potent and selective COX‐2 inhibitors were identified in vitro, the monomeric chalcone 16 emerged as the most promising compound. This study demonstrated that the dimeric nature of bis‐chalcone did not enhance COX‐2 inhibition or selectivity. Moreover, the polyhydroxylated bis‐chalcones fail to translate their in vitro potency into whole human blood activity, likely due to extensive plasma protein binding. Enhancing bioavailability may unlock the full therapeutic potential of these compounds, laying the foundation for novel, effective anti‐inflammatory treatments.

4. Materials and Methods

4.1. General

Reagents and solvents for synthesis were purchased as reagent‐grade and used without further purification unless otherwise stated. The following reagents were purchased from Sigma Chemical Co. (St. Louis, MO, USA): chloromethyl methyl ether chloride, 2‐methoxyethoxymethyl chloride, substituted benzaldehydes (4‐methoxybenzaldehyde, 3,4‐dimethoxybenzaldehyde, 3,4‐dihydroxybenzaldehyde), dimethyl sulfoxide (DMSO), acetylsalicylic acid, gentamicin sulfate, glycine, Cremophor El, l‐glutamine, lipopolysaccharides from Escherichia coli 026:B6 (LPS), hemin, Ampliflu red, human recombinant COX‐1 and COX‐2, Trizma hydrochloride. Celecoxib and SC‐560 were acquired from Biosynth Ltd. (Compton, UK). The thromboxane synthase inhibitor (TXBSI) and eriodictyol chalcone 16 were acquired from Enamine (Kyiv, Ukraine). Absolute ethanol was obtained from Fischer Chemical (New Hampshire, USA). Arachidonic acid (AA) peroxide‐free was obtained from Cayman Chemicals (Ann Arbor, MI, USA). Phloroglucinol was purchased from BDH Laboratory Reagents. The ethylenediamine tetraacetic acid K₃ (EDTA) and heparin tubes to collect the blood were purchased from Vacuette S.A. (Porto, Portugal).

Reactions were controlled by thin‐layer chromatography (TLC) using silica gel 60 HF254 plates. Column chromatography was performed using flash silica gel (40–60 μm, 60A). Melting points were determined with a Büchi melting point B‐540 apparatus and are uncorrected. NMR spectra were recorded with 300 or 500 MHz (300 MHz (¹H), 75 MHz (¹³C), or 500 MHz (¹H), 125 MHz (¹³C)) Bruker Avance III NMR spectrometers, with tetramethylsilane (TMS) as the internal reference (Figures S1–S27). Unequivocal ¹H assignments (δ, ppm) and ¹³C assignments (δ, ppm) were made based on the analysis of 2D HSQC, HMBC, and NOESY experiments, if necessary. Atom numbering was made according to the IUPAC name of the molecule or with custom attribution with visual representation in bis‐chalcones and bis‐flavones at Figures 7 and 8. Peak positions are given in parts per million (ppm) using the residual non‐deuterated solvent as the internal standard. Data are reported as follows: chemical shift (ppm), integrated intensity, multiplicity (indicated as: s, singlet; br s, broad singlet; d, doublet; t, triplet; q, quartet; m, multiplet and combination thereof), coupling constants (J) values in Hertz (Hz). Positive‐ion electrospray (ESI⁺) mass spectra were performed using a linear ion trap mass spectrometer LXQ (ThermoFinnigan, San Jose, CA). Data acquisition and analysis were performed using the Xcalibur Data System (version 2.0; ThermoFinnigan, San Jose, CA). High‐mass‐resolving ESI‐MS were conducted in a Q‐Exactive hybrid quadrupole Orbitrap mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) (Figures S28–S35). Untested intermediary compounds containing the protection groups MOM or MEM might not have their mass spectra analysis included, due to the instability of these groups under the ionization conditions. Purity of the tested compounds was confirmed by analysis of their ¹H and ¹³C NMR spectra.

FIGURE 7.

Structure and numbering of bis‐chalcone skeleton.

FIGURE 8.

Structure and numbering of bis‐flavone skeleton.

4.2. Synthesis

4.2.1. General Procedure for the Synthesis of 1,1′‐(2,4,6‐Trihydroxy‐1,3‐Phenylene)Bis(Ethan‐1‐One) (2)

Acetic anhydride (1.50 mL, 15.86 mmol) was added to a solution of boron trifluoride diethyl etherate (8 mL, 63.43 mmol). After 15 min, phloroglucinol 1 (1.0 g, 7.920 mL) was added slowly in portions. The mixture was stirred at 90°C for 14 h under nitrogen, poured into ice, quenched with potassium acetate (4.1 M, 10 mL), and extracted with diethyl ether (3 × 50 mL). The organic layers were washed with a saturated solution of potassium bicarbonate and water, dried over anhydrous sodium sulfate, and concentrated under reduced pressure. The residue was purified by silica gel column chromatography with hexane/ethyl acetate (4:1) as eluent.

1,1′‐(2,4,6‐Trihydroxy‐1,3‐phenylene)bis(ethan‐1‐one) (2), white solid, 80% yield, m.p. 161.6°C–163.4°C. ¹ H NMR (300 MHz, Acetone‐d ₆ ) δ 5.94 (s, 1H, H‐5), 2.66 (s, 6H, 1‐COCH ₃ , 3‐COCH ₃ ). ¹³ C NMR (75 MHz, Acetone‐d ₆ ) δ 203.8 (1‐COCH₃, 3‐COCH₃), 171.9 (C‐2, C‐4, C‐6), 103.8 (C‐1, C‐3), 94.7 (C‐5), 32.0 (1‐COCH₃, 3‐COCH₃). MS (ESI⁺) m/z (%): 211.0 [M + H]⁺ (50) HRMS (ESI⁺) m/z calcd for C₁₀H₁₁O₅[M⁺H]⁺ = 211.0601 (aprox.); found: 211.0600.

4.2.2. General Procedure for the Synthesis of 1,1′‐[2‐Hydroxy‐4,6‐Bis(Methoxymethoxy)‐1,3‐Phenylene]Bis(Ethan‐1‐One) (3a)

To a solution of 2 (0.30 g, 1.43 mmol) in DCM (5 mL) at 0°C, N,N‐diisopropylethylamine (DIPEIA) (5 mL, 28.74 mmol) was added, followed by slow addition of methoxymethyl chloride (MOM‐Cl) (0.23 mL, 3.00 mmol). The mixture was stirred overnight at room temperature, quenched with water, and the product was extracted with DCM. The organic layer was washed with saturated NaHCO₃ solution and brine, dried with anhydrous sodium sulfate, and concentrated under reduced pressure. The resulting residue was purified by silica gel column chromatography using hexane/ethyl acetate (3:2) as eluent to yield compound 3a.

1,1′‐[2‐Hydroxy‐4,6‐bis(methoxymethoxy)‐1,3‐phenylene]bis(ethan‐1‐one) (3a), colorless oil, 57% yield. ¹ H NMR (300 MHz, CDCl₃) δ 13.97 (s, 1H, 2‐OH), 6.39 (s, 1H, H‐5), 5.22 (s, 4H, 4‐OCH ₂ , 6‐OCH ₂ ), 3.47 (s, 6H, 4‐OCH₂OCH ₃ , 6‐OCH₂OCH ₃ ), 2.55 (s, 6H, 1‐COCH ₃ , 3‐COCH ₃ ). ¹³ C NMR (75 MHz, CDCl₃) δ 202.0 (1‐COCH₃, 3‐COCH₃), 160.7 (C‐4, C‐6), 121.4 (C‐2), 97.8 (C‐5), 94.4 (4‐OCH₂OCH₃, 6‐OCH₂OCH₃), 56.8 (4‐OCH₂OCH₃, 6‐OCH₂OCH₃), 32.8 (1‐COCH₃, 3‐COCH₃).

4.2.3. General Procedure for the Synthesis of 1,1′‐{2‐Hydroxy‐4‐[(Methoxyethoxy) Methoxy]‐6‐[(3‐Methoxypropoxy)Methoxy]‐1,3 Phenylene}Bis(Ethan‐1‐One) (3b)

2‐Methoxyethoxymethyl chloride (MEM‐Cl) (0.72 mL, 9.52 mmol) was added slowly to a solution of 2 (0.50 g, 2.38 mmol) in DCM (20 mL) with N,N‐diisopropylethylamine (DIPEIA) (4.14 mL, 23.80 mmol) at 0°C. Then, the mixture was left stirring overnight at room temperature. The mixture was diluted with water. The organic layer was separated and washed with saturated aqueous NaHCO₃ solution and then with brine, dried with anhydrous sodium sulfate, and concentrated under reduced pressure. The resulting residue was purified by silica gel column chromatography using hexane/ethyl acetate (1:1) as eluent to yield compound 3b.

1,1′‐{2‐Hydroxy‐4‐[(methoxyethoxy)methoxy]‐6‐[(3‐methoxypropoxy)methoxy]‐1,3 phenylene}bis(ethan‐1‐one) (3b), colorless oil, 65% yield. ¹ H NMR (300 MHz, CDCl₃) δ 13.99 (s, 1H, 2‐OH), 6.47 (s, 1H, H‐5), 5.35 (s, 4H, 4‐OCH ₂ O‐, 6‐OCH ₂ O‐), 3.87–3.78 (m, 4H, 4‐OCH₂OCH ₂ ‐, 6‐OCH₂OCH ₂ ‐), 3.61–3.52 (m, 4H, 4‐OCH₂OCH₂CH ₂ O‐, 6‐OCH₂OCH₂CH ₂ O‐), 3.38 (s, 6H, 2× ‐CH₂OCH ₃ ), 2.58 (s, 6H, 1‐COCH ₃ , 3‐COCH ₃ ). ¹³ C NMR (75 MHz, CDCl₃) δ 202.1 (1‐COCH₃, 3‐COCH₃), 163.2 (C‐4, C‐6), 160.6 (C‐2), 110.6 (C‐3, C‐1), 93.4 (4‐OCH₂O‐, 6‐OCH₂O‐), 91.8 (C‐5), 71.4 (4‐OCH₂OCH₂ CH₂O‐, 6‐OCH₂OCH₂ CH₂O‐), 68.5 (4‐OCH₂OCH₂‐, 6‐OCH₂OCH₂‐) 59.1 (2× ‐CH₂OCH₃), 32.8 (1‐COCH₃, 3‐COCH₃).

4.2.4. General Procedure for the Synthesis of 1,1′‐[2‐Methoxy‐4,6‐Bis(Methoxymethoxy)‐1,3‐Phenylene]Bis(Ethan‐1‐One) (4a)/1,1'‐{2‐methoxy‐4‐[(2‐methoxyethoxy)methoxy]‐6‐[(3‐methoxypropoxy)methoxy]‐1,3‐phenylene)bis(ethan‐1‐one) (4b)

Dimethyl sulfate (0.28 mL, 2.95 mmol) was added to a suspension of 3a/b (0.20 g, 0.95 mmol) and potassium carbonate (0.66 g, 4.76 mmol) in acetone (30 mL). The mixture was stirred at reflux overnight, quenched with NaOH solution (1 M), and the product was extracted with DCM (30 mL). The organic phase was dried over anhydrous sodium sulfate, evaporated to dryness, and the residue was purified by silica gel column chromatography using DCM as eluent to yield compound 4a/b.

1,1′‐[2‐Methoxy‐4,6‐bis(methoxymethoxy)‐1,3‐phenylene]bis(ethan‐1‐one) (4a), colorless oil, 63% yield. ¹ H NMR (300 MHz, CDCl₃) δ 6.73 (s, 1H, H‐5), 5.18 (s, 4H, 4‐OCH ₂ , 6‐OCH ₂ ), 3.74 (s, 3H, 2‐OCH ₃ ), 3.47 (s, 6H, 4‐OCH₂OCH ₃ , 6‐OCH₂OCH ₃ ), 2.52 (s, 6H, 1‐COCH ₃ , 3‐COCH ₃ ). ¹³ C NMR (75 MHz, CDCl₃) δ 201.0 (1‐COCH₃, 3‐COCH₃), 155.8 (C‐4, C‐6), 120.6 (C‐2), 97.6 (C‐5), 94.7 (4‐OCH₂OCH₃, 6‐OCH₂OCH₃), 64.3 (2‐OCH₃), 56.5 (4‐OCH₂OCH₃, 6‐OCH₂OCH₃), 32.5 (1‐COCH₃, 3‐COCH₃).

1,1′‐{2‐Methoxy‐4‐[(2‐methoxyethoxy)methoxy]‐6‐[(3‐methoxypropoxy)methoxy]‐1,3‐phenylene)bis(ethan‐1‐one) (4b), colorless oil, 63% yield. ¹ H NMR (300 MHz, CDCl₃) δ 6.77 (s, 1H, H‐5), 5.25 (s, 4H, 4‐OCH ₂ O‐, 6‐OCH ₂ O‐), 3.82–3.73 (m, 4H, 4‐OCH₂OCH ₂ ‐, 6‐OCH₂OCH ₂ ‐), 3.71 (s, 3H, 2‐OCH₃), 3.56–3.51 (m, 4H, 4‐OCH₂OCH₂CH ₂ O‐, 6‐OCH₂OCH₂CH ₂ O‐), 3.36 (s, 6H, 2× ‐CH₂OCH₃), 2.49 (s, 6H, 1‐COCH ₃ , 3‐COCH ₃ ). ¹³ C NMR (75 MHz, CDCl₃) δ 201.0 (1‐COCH₃, 3‐COCH₃), 155.8 (C‐4, C‐6), 155.1 (C‐2), 120.7 (C‐3, C‐1), 98.0 (4‐OCH₂O, 6‐OCH₂O), 93.8 (C‐5), 71.4 (4‐OCH₂OCH₂ CH₂O‐, 6‐OCH₂OCH₂ CH₂O‐), 68.2 (4‐OCH₂OCH₂‐, 6‐OCH₂OCH₂‐), 64.3 (2‐OCH₃), 59.0 (2× ‐CH₂OCH₃), 32.5 (COCH₃).

4.2.5. General Procedure for the Synthesis of 3,4‐Bis[(2‐Methoxyethoxy)Methoxy]Benzaldehyde (7)

2‐Methoxyethoxymethyl chloride (MEM‐Cl) (0.72 mL, 9.52 mmol) was added slowly to a solution of 3,4‐dihydroxybenzaldehyde (0.50 g, 3.62 mmol) in DCM (20 mL) with N,N‐diisopropylethylamine (DIPEIA) (4.14 mL, 23.80 mmol) at 0°C. Then, the mixture was left stirring overnight at room temperature. The mixture was diluted with water. The organic layer was separated and washed with saturated aqueous NaHCO₃ solution and then with brine, dried with anhydrous sodium sulfate, and concentrated under reduced pressure. The resulting residue was purified by silica gel column chromatography using hexane/ethyl acetate (1:1) as eluent to yield compound 7.

3,4‐Bis[(2‐methoxyethoxy)methoxy]benzaldehyde (7), colorless oil, 75% yield. ¹ H NMR (300 MHz, CDCl₃) δ 9.87 (s, 1H, CHO), 7.72 (d, J = 2.0 Hz, 1H, H‐2), 7.53 (dd, J = 8.4, 2.0 Hz, 1H, H‐6), 7.34 (d, J = 8.4 Hz, 1H, H‐5), 5.42 (s, 2H, 3‐OCH ₂ O‐), 5.38 (s, 2H, 4‐OCH ₂ O‐), 3.92–3.84 (m, 4H, 3‐OCH₂OCH ₂ ‐, 4‐OCH₂OCH ₂ ‐), 3.61–3.51 (m, 4H, 3‐OCH₂OCH₂CH ₂ O‐, 4‐OCH₂OCH₂CH ₂ O‐), 3.38 (s, 3H, 3‐OCH₂OCH₂CH₂OCH₃), 3.37 (s, 3H, 3‐OCH₂OCH₂CH₂OCH ₃ ).

4.2.6. General Procedure for the Synthesis of Bis‐Chalcones 8–10

Compound 4a–b (0.5 mmol) was dissolved in methanol (40 mL), and NaOH (4.0 g) was added. After 30 min of stirring at room temperature, the corresponding aldehyde (2.1 equiv., 1.05 mmol) was added to the mixture. The mixture was stirred at room temperature for 24–48 h, poured into ice and quenched with diluted hydrochloric acid to adjust the pH to 4. The product was extracted with DCM (3 × 20 mL). The organic layers were dried over anhydrous sodium sulfate, evaporated to dryness, and the residue was purified by silica gel column chromatography.

(2E,2′ E)‐1,1′‐[2‐Methoxy‐4,6‐bis(methoxymethoxy)‐1,3‐phenylene]bis[3‐(3,4‐dimethoxyphenyl)prop‐2‐en‐1‐one] (8) yellow solid, 63% yield, m.p. 225.6–227.3°C. ¹ H NMR (300 MHz, CDCl₃) δ 7.38 (d, J = 16.1 Hz, 2H, H‐β, H‐β′), 7.14 (dd, J = 8.5, 2.0 Hz, 2H, H‐6′, H‐6″), 7.08 (d, J = 2.0 Hz, 1H, H‐2′, H‐2″), 6.91 (d, J = 16.1 Hz, 2H, H‐α, H‐α′), 6.87 (d, J = 8.5 Hz, 2H, H‐5′, H‐5″), 6.81 (s, 1H, H‐5), 5.18 (s, 4H, 4‐OCH ₂ OCH₃, 6‐OCH ₂ OCH₃), 3.92 (s, 6H, 3′‐OCH ₃ , 3″‐OCH ₃ ), 3.92 (s, 6H, 4′‐OCH ₃ , 4″‐OCH ₃ ), 3.70 (s, 3H, 2‐OCH ₃ ), 3.43 (s, 6H, 4‐OCH₂OCH ₃ , 6‐OCH₂OCH ₃ ). ¹³ C NMR (75 MHz, CDCl₃) δ 193.9 (1‐COCH₃‐, 3‐COCH₃‐), 156.4 (C‐4, C‐6), 151.6 (C‐4′, C‐4″), 149.3 (C‐3′, C‐3″), 146.2 (C‐β, C‐β′), 127.4 (C‐1′, C‐1″), 126.6 (C‐α, C‐α′), 123.3 (C‐6′, C‐6″), 118.3 (C‐1, C‐3), 111.1 (C‐5′, C‐5″), 110.0 (C‐2′, C‐2″), 97.6 (C‐5), 94.7 (4‐OCH₂‐, 6‐OCH₂‐), 63.2 (2‐OCH₃), 56.5 (4‐OCH₂OCH₃, 6‐OCH₂OCH₃ ), 56.0 (3′‐OCH₃, 3″‐OCH₃), 55.95 (4′‐OCH₃, 4″‐OCH₃). MS (ESI⁺) m/z (%): 608.9 [M + H]⁺ (100). HRMS (ESI⁺) m/z calcd for C₃₃H₃₆O₁₁: 609.2300 [M + H]⁺; found: 609.2310.

(2E,2′ E)‐1,1′‐[2‐Methoxy‐4,6‐bis(methoxymethoxy)‐1,3‐phenylene]bis[3‐(4‐methoxyphenyl)prop‐2‐en‐1‐one] (9), orange solid, 79% yield, m.p. 90.0°C–91.3°C. ¹ H NMR (300 MHz, CDCl₃) δ 7.49 (d, J = 9.0 Hz, 4H, H‐2′,6′, H‐2″, 6″), 7.38 (d, J = 16.0 Hz, 2H, H‐β, H‐β′), 6.90 (d, J = 16.0 Hz, 2H, H‐α, H‐α′), 6.89 (d, J = 9.0 Hz, 4H, H‐3′, 5′, H‐3″, 5″), 6.80 (s, 1H, H‐5), 5.16 (s, 4H, 4‐OCH ₂ , 6‐OCH ₂ ), 3.82 (s, 6H, 4′‐OCH ₃ , 4″‐OCH ₃ ), 3.68 (s, 3H, 2‐OCH ₃ ), 3.40 (s, 6H, 4‐OCH₂OCH ₃ , 6‐OCH₂OCH ₃ ). ¹³ C NMR (75 MHz, CDCl₃) δ 193.8 (1‐COCH₃, 3‐COCH₃), 161.8 (C‐4′, C‐4″), 156.4 (C‐2), 145.8 (C‐β, C‐β′), 130.3 (C‐2′,6′, C‐2″,6″) 127.2 (C‐1′, C‐1″), 126.4 (C‐α, C‐α′) 118.4 (C‐1, C‐3), 114.4 (C‐3′,5′, C‐3″,5″), 97.5 (C‐5), 94.7 (4‐OCH₂OCH₃, 6‐OCH₂OCH₃), 63.3 (2‐OCH₃), 56.5 (4′‐OCH₃, 4″‐OCH₃), 55.4 (4‐OCH₂OCH₃, 6‐OCH₂OCH₃). MS (ESI⁺) m/z (%): 549.0 (100). HRMS (ESI⁺) m/z calcd for C₃₁H₃₂O₉: 549.2046 [M + H]⁺; found: 549.2104.

(2E,2′ E)‐1,1′‐{2‐Methoxy‐4,6‐bis[(2‐methoxyethoxy)methoxy]‐1,3‐phenylene}bis(3‐{3,4‐bis[(2‐methoxyethoxy)methoxy]phenyl}prop‐2‐en‐1‐one) (10), colorless oil, 55% yield. ¹ H NMR (300 MHz, CDCl₃) δ 7.42 (d, J = 1.7 Hz, 1H, H‐2′, H‐2″), 7.34 (d, J = 15.9 Hz, 1H, H‐β, H‐β′), 7.25–7.11 (m, 4H, H‐5′,‐6′, H‐5″, −6″), 6.91 (d, J = 15.9 Hz, 2H, H‐α, H‐α′), 6.88 (s, 1H, H‐5), 5.35 (s, 4H, 3′‐OCH ₂ O‐, 3″‐OCH ₂ O‐), 5.32 (s, 4H, 4′‐OCH ₂ O‐, 4″‐OCH ₂ O‐), 5.27 (s, 4H, 4‐OCH ₂ O‐, 6‐OCH ₂ O‐), 3.92–3.81 (m, 8H, 3′‐OCH₂OCH ₂ ‐, 4′‐OCH₂OCH ₂ ‐, 3″‐OCH₂OCH ₂ ‐, 4″‐OCH₂OCH ₂ ‐), 3.80–3.71 (m, 4H, 4‐OCH₂OCH ₂ ‐, 6‐OCH₂OCH ₂ ‐), 3.68 (s, 3H, 2‐OCH ₃ ), 3.62–3.47 (m, 14H, 3′‐OCH₂OCH₂CH ₂ O‐, 4′‐OCH₂OCH₂CH ₂ O‐, 3″‐OCH₂OCH₂CH ₂ O‐, 4″‐OCH₂OCH₂CH ₂ O‐, 4‐OCH₂OCH₂CH ₂ O‐, 6‐OCH₂OCH₂CH ₂ O‐, 2‐OCH₂OCH₂CH ₂ O‐), 3.38 (s, 6H, 3′‐OCH₂OCH₂CH₂OCH ₃ , 3″‐OCH₂OCH₂CH₂OCH ₃ ), 3.37 (s, 6H, 4′‐OCH₂OCH₂CH₂OCH ₃ , 4″‐OCH₂OCH₂CH₂OCH ₃ ), 3.35 (s, 6H, 4‐OCH₂OCH₂CH₂OCH ₃ , 6‐OCH₂OCH₂CH₂OCH ₃ ). ¹³ C NMR (75 MHz, CDCl₃) δ 193.6 (1‐COCH₃, 3‐COCH₃), 156.5 (C‐4, C‐6), 156.4 (C‐2), 149.6 (C‐4′, C‐4″), 147.3 (C‐3′, C‐3″), 145.5 (C‐β, C‐β′), 129.0 (C‐1′, C‐1″), 127.4 (C‐α, C‐α′), 124.2 (C‐6′, C‐6″), 118.4 (C‐1, C‐3), 116.4 (C‐2′, C‐2″), 116.1 (C‐5′, C‐5″), 98.0 (C‐5), 94.6 (4′‐OCH₂O‐, 4″‐OCH₂O‐), 94.1 (3′‐OCH₂O‐, 3″‐OCH₂O‐), 93.7 (4‐OCH₂O‐, 6‐OCH₂O‐), 71.5 (4′‐OCH₂OCH₂ CH₂O‐, 4″‐OCH₂OCH₂ CH₂O‐, 3′‐OCH₂OCH₂ CH₂O‐, 3″‐OCH₂OCH₂ CH₂O‐) 71.4 (4‐OCH₂OCH₂ CH₂O‐, 6‐OCH₂OCH₂ CH₂O‐), 68.1 (4‐OCH₂OCH₂‐, 6‐OCH₂OCH₂‐), 68.0 (3′‐OCH₂OCH₂‐, 4′‐OCH₂OCH₂‐, 3″‐OCH₂OCH₂‐, 4″‐OCH₂OCH₂‐), 63.3 (2‐OCH₃), 59.1 (3′‐OCH₂OCH₂CH₂OCH₃, 3″‐OCH₂OCH₂CH₂OCH₃, 4′‐OCH₂OCH₂CH₂OCH₃, 4″‐OCH₂OCH₂CH₂OCH₃), 58.98 (4‐OCH₂OCH₂CH₂OCH₃, 6‐OCH₂OCH₂CH₂OCH₃).

4.2.7. General Procedure for the Synthesis of Unprotected Bis‐Chalcones 11–13

Trifluoroacetic acid (TFA) (1 mL) was added to a solution of 8–10 (0.5 mmol) in DCM (10 mL) at 0°C. The mixture was stirred for 30–60 min. until total consumption of the starting material, controlled by TLC. The mixture was treated with methanol, followed by evaporation until the total removal of TFA.

(2E,2′ E)‐1,1′‐(4,6‐Dihydroxy‐2‐methoxy‐1,3‐phenylene)bis[3‐(3,4‐dimethoxyphenyl)prop‐2‐en‐1‐one] (11) orange solid, 99% yield, m.p. 134.6°C–141.7°C. ¹ H NMR (300 MHz, CDCl₃) δ 13.66 (s, 2H, 4‐OH, 6‐OH), 7.89 (d, J = 15.5 Hz, 2H, H‐β, H‐β′), 7.72 (d, J = 15.5 Hz, 2H, H‐α, H‐α′), 7.30 (dd, J = 8.4, 2.0 Hz, 2H, H‐6′, H‐6″), 7.15 (d, J = 2.0 Hz, 2H, H‐2′, H‐2″), 6.93 (d, J = 8.4 Hz, 2H, H‐5′, H‐5″), 6.35 (s, 1H, H‐5), 3.95 (s, 12H, 3′‐OCH ₃, 3″‐OCH ₃, 4′‐OCH ₃, 4″‐OCH ₃), 3.82 (s, 3H, 2‐OCH ₃). ¹³ C NMR (75 MHz, CDCl₃) δ 193.0 (1‐COCH₃, 3‐COCH₃), 169.4 (C‐4, C‐6), 167.4 (C‐2), 151.7 (C‐4′, C‐4″), 149.3 (C‐3′, C‐3″), 144.8 (C‐β, C‐β′), 128.0 (C‐1′, C‐1″), 124.0 (C‐α, C‐α′), 122.9 (C‐6′, C‐6″), 111.3 (C‐1, C‐3), 111.0 (C‐5′, C‐5″), 109.9 (C‐2′, C‐2″), 101.7 (C‐5), 64.9 (2‐OCH₃), 56.1 (3′‐OCH₃, 3″‐OCH₃), 56.0 (4′‐OCH₃, 4″‐OCH₃). MS (ESI⁺) m/z (%): 521.2 [M + H]⁺(100). HRMS (ESI⁺) m/z calcd for C₂₉H₂₈O₉: 521.1733 [M + H]⁺; found: 521.1795.

(2E,2′ E)‐1,1′‐(4,6‐Dihydroxy‐2‐methoxy‐1,3‐phenylene)bis[3‐(4‐methoxyphenyl)prop‐2‐en‐1‐one] (12) orange solid, 99% yield, m.p. 184.4°C–185.6°C. ¹ H NMR (300 MHz, DMSO‐d ₆ ) δ 11.87 (s, 2H, 4‐OH, 6‐OH), 7.70 (d, J = 8.8 Hz, 4H, H‐2′, H‐6′, H‐2″, H‐6″), 7.54 (d, J = 15.9 Hz, 2H, H‐β, H‐β′), 7.25 (d, J = 15.9 Hz, 2H, H‐α, H‐α′), 7.01–6.92 (m, 4H, H‐3′, H‐5′, H‐3″, H‐5″), 6.27 (s, 1H, H‐5), 3.81 (s, 6H, 4′‐OCH ₃ , 4″‐OCH ₃ ), 3.63 (s, 3H, 2‐OCH ₃ ). ¹³ C NMR (75 MHz, DMSO‐d ₆ ) δ 193.06 (1‐COCH₃‐, 3‐COCH₃‐), 162.8 (C‐4, C‐6), 161.87 (C‐2), 144.7 (C‐β, C‐β′), 131.01 (C‐2′, 6′, C‐2″, 6″), 127.46 (C‐4′, C‐4″), 125.7 (C‐α, C‐α′), 115.0 (C‐3′, 5′, C‐3″, 5″), 112.8 (C‐3, C‐1), 99.6 (C‐5), 64.1 (2‐ OCH ₃ ), 55.9 (4′‐OCH₃, 4″‐OCH₃). MS (ESI⁺) m/z (%): 461.2 [M + H]⁺ (100), HRMS (ESI⁺) m/z calcd for C₃₁H₃₂O₉: 461.1522 [M + H]⁺; found: 461.1583.

(2E,2′ E)‐1,1′‐(4,6‐Dihydroxy‐2‐methoxy‐1,3‐phenylene)bis[3‐(3,4‐dihydroxyphenyl)prop‐2‐en‐1‐one] (13), red solid, 99% yield, m.p. (≥ 230°C decomposition). ¹ H NMR (300 MHz, DMSO‐d ₆ ) δ 17.23 (s, 1H, 2‐OH), 13.47 (s, 2H, 4‐OH, 6‐OH), 9.73 (s, 2H, 4′‐OH, 4″‐OH), 9.33 (s, 2H, 3′‐OH, 3″‐OH), 7.88 (d, J = 15.5 Hz, 2H, H‐β, H‐β′), 7.68 (d, J = 15.5 Hz, 2H, H‐α, H‐α′), 7.14 (d, J = 2.1 Hz, 2H, H‐2′, H‐2″), 7.05 (dd, J = 8.2, 2.1 Hz, 2H, H‐6′, H‐6″), 6.81 (d, J = 8.1 Hz, 2H, H‐5′, H‐5″), 5.98 (s, 1H, H‐5), 3.44 (s, 3H, 2‐OCH ₃). ¹³ C NMR (126 MHz, Acetone) δ 193.1 (1‐COCH₃‐, 3‐COCH₃‐), 168.4 (C‐4, C‐6), 148.8 (C‐4′, C‐4″) 145.5 (C‐β, C‐β′), 145.2 (C‐3′, C‐3″), 127.2 (C‐α, C‐α′), 123.0 (C‐6′, C‐6″), 123.0 (C‐6′, C‐6″), 118.4 (C‐1, C‐3), 115.7 (C‐5′, C‐5″), 114.5 (C‐2′, C‐2″), 100.34 (C‐5), 64.5 (2‐OCH₃). MS (ESI⁺) m/z (%): 465.3 [M + H]⁺.

4.2.8. General Procedure for the Synthesis of A‐Fused Bis Flavones 14–15

To a solution of 11–12 (0.1 mmol) in DMSO (1 mL), I₂ (0.1 equiv., cat) was added, and the mixture was left stirring at 190°C in a sand bath for 30 min and then quenched with Na₂SO₃ (aq.) saturated solution. The mixture was filtered under vacuum, and the remaining precipitate was washed with cold water, dissolved in chloroform/methanol, and purified by silica gel column chromatography using DCM/MeOH 10:1 as eluent.

2,8‐Bis(3,4‐dimethoxyphenyl)‐5‐methoxy‐4H,6H‐pyrano[3,2‐g]chromene‐4,6‐dione (14), white solid, 33% yield, m.p. 291.6°C–293.0°C. ¹ H NMR (300 MHz, CDCl₃) δ 7.69 (dd, J = 8.5, 2.1 Hz, 2H, H‐6′, H‐6″), 7.63 (d, J = 2.1 Hz, 2H, H‐2′, H‐2″), 7.03 (d, J = 8.5 Hz, 2H, H‐5′, H‐5″), 6.70 (s, 2H, H‐3, H‐7), 4.13 (s, 3H, 2‐OCH ₃), 4.03 (s, 6H, 3′‐OCH ₃, 3″‐OCH ₃), 3.99 (s, 6H, 4′‐OCH ₃, 4″‐OCH ₃). ¹³ C NMR (75 MHz, CDCl₃) δ 176.4 (C‐4, C‐6), 163.0 (C‐5), 161.6 (C‐2, C‐8), 158.6 (C‐9a, C‐10a), 152.4 (C‐4′, C‐4″), 149.4 (C‐3′, C‐3″), 122.9 (C‐1′, C‐1″), 120.4 (C‐6′, C‐6″), 116.5 (C‐4a, C‐5a), 111.3 (C‐5′, C‐5″), 109.2 (C‐2′, C‐2″), 107.1 (C‐3, C‐7), 63.5 (5‐OCH₃), 56.2 (3′‐OCH₃, 3″‐OCH₃, 4′‐OCH₃, 4″‐OCH₃). MS (ESI⁺) m/z (%): 517.2 [M + H]⁺ (35). HRMS (ESI⁺) m/z calcd for C₂₉H₂₄O₉: 517.1420 [M + H]⁺; found: 517.1487.

5‐Methoxy‐2,8‐bis(4‐methoxyphenyl)‐4H,6H‐pyrano[3,2‐g]chromene‐4,6‐dione (15), white solid, 35% yield, m.p. 295.6°C–297.0°C. ¹ H NMR (300 MHz, CDCl₃) δ 7.42 (d, J = 7.8 Hz, 4H, H‐2′,6′, H‐2″,6″), 6.84 (d, J = 7.8 Hz, 5H, H‐10, H‐3′,5′, H‐3″,5″), 6.50 (s, 2H, H‐3, H‐7), 4.00 (s, 3H, 2‐OCH ₃), 3.83 (s, 6H, 4′‐OCH ₃, 4″‐OCH ₃). ¹³ C NMR (75 MHz, CDCl₃) δ 176.7 (C‐4, C‐6), 162.4 (C‐5), 161.9 (C‐2, C‐8), 160.7 (C‐9a, C‐10a), 157.4 (C‐4′, C‐4″), 127.6 (C‐3′,5′, C‐3″,5″), 123.1 (C‐1′, C‐1″), 116.0 (C‐4a, C‐5a), 114.6 (C‐2′,6′, C‐2″,6″), 112.0 (C‐10), 108.3 (C‐3, C‐7), 63.4 (5‐OCH₃), 55.6 (4′‐OCH₃, 4″‐OCH₃). MS (ESI⁺) m/z (%): 457.1 [M + H]⁺. HRMS (ESI⁺) m/z calcd for C₂₇H₂₀O₇: 457.1209 [M + H]⁺; found: 457.1274.

4.3. Computational Methods

To study the optimized structure of bis‐chalcone 17, density functional theory (DFT) calculations were used. B3LYP was the functional used, and it represents a generalized hybrid gradient approximation functional, incorporating a 20% mixture of exact exchange along with additional semiempirical adjustments to both semi‐local exchange and correlation components (Bursch et al. 2022). The structure was optimized using the B3LYP functional with a 6‐31G basis in the gas phase. The optimized structure was absent of any imaginary frequencies, proving the minimum energy state for the system. All calculations were performed with the Gaussian 09 program (G09), version A.02 (Frisch et al. 2009) and visualized with GaussView, version 6 (Dennington et al. 2016).

4.4. Evaluation of the In Vitro Inhibition of COX‐1 and COX‐2 Activity

Human recombinant COX‐1 or COX‐2 enzymes oxidize the substrate arachidonic acid (AA), yielding PGG₂, which is subsequently converted at the peroxidase site of COX in the presence of the fluorescence probe (Ampliflu red). This reaction produces PGH₂ and resorufin, with the latter responsible for the fluorescence signal detected at λ _Ex/Em = 53/587 nm. Before the experiments, absorption spectra were traced for all compounds to guarantee that none of them interfered with the assays at the wavelengths used.

Briefly, human recombinant COX‐1 or COX‐2 (2.5 ng/μL) was combined with diluted hemin (1 μM), COX cofactor, and Ampliflu red (100 μM) in assay buffer (Trizma hydrochloride pH = 8.0 and EDTA 0.1 μM). Subsequently, 85 μL of this reaction mixture was added to a 96‐well black plate containing 5 μL of compound (0–100 μM) or DMSO and incubated for 5 min. Following this, 10 μL of AA solution (25 μM) was added, and fluorescence was immediately monitored for 15 min at 25°C using a microplate reader (Synergy HT, BIO‐TEK, Winooski, VT, USA) with excitation at 530 ± 25 nm and emission detection at 590 ± 35 nm. Celecoxib (0–1 μM) was used as the positive control for COX‐2, and SC‐560 (0–40 nM) was the positive control used for COX‐1. The amount of DMSO used [5% (v/v)] did not interfere with the assay, and none of the tested compounds showed fluorescence emission in the specified wavelength ranges of the assay. The percentage of inhibition of COX activity was calculated using a single point of RFU after 5 min of reading. These results were obtained from a minimum of three distinct and independent experiments.

The selective indexes (SI) are a calculated parameter used to quickly evaluate the selectivity of a given compound to COX‐2 over COX‐1. These values were calculated according to the following formula (Equation 1). Hence, for compounds presenting IC₅₀, the calculation was done using Equation (1):

$$\mathrm{SI}=\frac{{\mathrm{IC}}_{50}^{\mathrm{COX}‐1}}{{\mathrm{IC}}_{50}^{\mathrm{COX}‐2}}$$ (1)

A high SI value means higher selectivity for COX‐2.

4.5. Inhibition Kinetic Analysis

The inhibition mechanism of the COX‐2 enzyme was determined for the most active compound tested, 16 (IC₅₀ = 0.5 ± 0.03 μM). A protocol similar to the in vitro COX inhibition assay was employed. In a 96‐well plate, compound 16 was incubated at different concentrations (0.16, 0.31, and 0.63 μM) with COX‐2 for 5 min, followed by the addition of AA (6.3, 25.0, and 100.0 μM). The resulting fluorescence was then measured using the previously established protocol.

The kinetics of the system were described using the generalized Michaelis–Menten equation (Equation 2), with various simplifications corresponding to different types of inhibition:

$$v_{\text{inic}}=\frac{V_{\max }\left(S\right)}{K_{\mathrm{m}}\left(1+\frac{\left[I\right]}{K_{\mathrm{ic}}}\right)+\left(S\right)\left(1+\frac{\left[I\right]}{K_{\mathrm{iu}}}\right)}$$ (2)

where v _inic = initial velocity of formation of resorufin in ΔRFU per minute, V _max = maximum achievable velocity when for the 2.5 ng/μL of enzyme is fully saturated with substrate, S = AA concentration (expressed in mM), K _m = Michaelis–Menten constant (expressed in μM), K _ic = inhibitor dissociation constant for the enzyme–inhibitor complex (expressed in μM), K _iu = inhibitor dissociation constant of enzyme–substrate–inhibitor complex (expressed in μM).

Each nonlinear regression of the results was conducted based on the data from a minimum of three independent experiments. Data fitting was performed in an Excel Microsoft Office spreadsheet using the Solver supplement, according to the guidelines described by Bezerra et al. (2013) and Dias et al. (2014).

In each case, the parameters of the equation were estimated using Solver, with initial guesses based on the parameter values obtained from the simplest model (absence of inhibition). These initial estimates were then used for the models corresponding to competitive, noncompetitive, uncompetitive, and mixed inhibition.

To determine the mechanism of inhibition, the models were compared using the extra sum‐of‐squares F test and the Akaike information criterion (AIC) test. The error associated with the kinetic constant values was assessed using the Jackknife procedure, which involved calculating the standard deviation of parameter estimates obtained from Solver when each experimental data point was omitted iteratively.

Additionally, Lineweaver–Burk plots were analyzed for each concentration of inhibitor and substrate to provide further insights into the kinetics of the system.

4.6. Blood Collection

Following the declaration of Helsinki, and with the approval of patient‐related procedures and protocols by the Ethics Committee of Centro Hospitalar do Porto, Portugal, venous blood samples were collected from healthy human donors after their informed consent. The blood collection was performed via peripheral venipuncture using lithium heparin tubes.

4.7. Viability of Erythrocytes and Leukocytes

The erythrocyte viability was assessed via hemolysis analysis based on a previously described methodology with modifications (Rocha et al. 2024). Briefly, 800 μL of human blood was collected in heparinized vacuum tubes and transferred to a microtube containing 110 μL of DPBS‐gentamicin and 100 μL of the tested compounds (0–100 μM), dissolved in a mixture of DMSO with cremophor‐ethanol 1% (1:10). Afterwards, 600 μL of this mixture was added to each well of a 12‐well plate, followed by 200 μL of DPBS. The mixture was homogenized and incubated for 5 h at 37°C. After incubation, each mixture was transferred to a microtube and centrifuged at 1500 g for 5 min at 4°C. A 100 μL aliquot of the supernatant was transferred to a transparent 96‐well plate, and absorbance was measured using a microplate reader at 540 nm (maximum absorbance of hemoglobin) and 630 nm (maximum absorbance of methemoglobin). Triton X‐100 at two different concentrations (0.3% and 0.6%) was used as a positive control.

Leukocyte viability was assessed using the staining protocol provided by BD's FITC Annexin V apoptosis detection Kit I. The previously described blood sample mixture was distributed into a 12‐well plate and incubated for 5 h at 37°C. After incubation, the content of each well was resuspended, and 100 μL of solution was transferred to 15 mL centrifuge tubes. Subsequently, 6 mL of lysis buffer (1:10 ratio from the BD Pharm Lyse lysing solution) was added. Following 10 min of incubation, the samples were centrifuged at 400 g for 5 min at 20°C. The resulting pellet was resuspended in 2 mL of PBS and subsequently centrifuged at 400 g for 5 min at 20°C. The supernatant was discarded, and 100 μL of the solution Annexin V/PI was added. After a 15‐min incubation period, the samples were diluted with 400 μL of the binding buffer and analyzed using a BD Accuri C6 flow cytometer. To isolate the leukocyte population, a polygon gate was set based on the light scattering properties (forward vs. side scatter plot), excluding debris and other blood cells. Fluorescence signals from at least 10,000 events were collected in logarithmic mode and analyzed using BD Accuri C6 software. PI incorporation was monitored in channel 3 (FL3) while Annexin V incorporation was followed in channel 1 (FL1). The results were expressed as the percentage of Annexin V and/or PI‐positive cells.

4.8. Determination of PGE₂ Production in Human Whole Blood

The whole human blood assay to assess the PGE₂ production was conducted according to the previously described methodology (Rocha et al. 2024). Human blood (800 μL) was collected in heparinized vacuum tubes, placed in a 12‐well plate and incubated at 37°C for 15 min, with 10 μL TBXSI (1 μM), 50 μL acetylsalicylic acid (10 μg/mL), and 100 μL of compounds under study (0–100 μM) dissolved in DMSO/cremophor/ethanol 1% (1:10). Subsequently, 50 μL of LPS (10 μg/mL) was added, and the mixture was incubated for 5 h. The reaction was terminated by adding 1 mL of cold DPBS‐gentamicin buffer to each sample, followed by incubation on ice for 10 min. The samples were centrifuged at 1000 g for 15 min at 4°C. The supernatant was collected and stored at −20°C until further use. The PGE₂ levels in the samples were measured using a PGE₂ ELISA kit, following the manufacturer's instructions. Celecoxib (5 μM) was used as the positive control. The results were expressed as the mean percentage inhibition of PGE₂ production relative to the control. Each mean represents at least three independent experiments.

4.9. Binding to Plasma Proteins

Plasma protein binding (PDB) values were estimated for compounds 13, 16–18 and celecoxib using PreADMET (Lee et al. 2004), an online tool that predicts drug‐likeness as well as absorption, distribution, metabolism, excretion (ADME), and toxicological data.

4.10. Statistical Analysis

The results of the in vitro inhibitory activities of the tested compounds are expressed as half maximal inhibitory concentration (IC₅₀) ± SEM (n ≥ 3). The results for PGE₂ production inhibition are expressed as mean percentage inhibition ± standard error of mean (SEM). A one‐way analysis of variance (ANOVA) was performed to assess the statistical significance of differences among groups, followed by Dunnett's multiple comparison test. Statistical significance was established at p < 0.05. The IC₅₀ analysis were conducted using GraphPad Prism (version 9.0.0; GraphPad Software).

Author Contributions

Rui Pereira: conceptualization, methodology, investigation, writing – review and editing. Marisa Freitas: supervision, methodology, writing – review and editing. Alberto N. Araújo: writing – review and editing, methodology. Vera L. M. Silva: conceptualization, supervision, resources, writing – review and editing. Eduarda Fernandes: conceptualization, supervision, funding acquisition, writing – review and editing.

Funding

This work was supported by Fundação para a Ciência e a Tecnologia (UID/50006, UI/BD/151269/2021).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: ¹H NMR spectrum of compound 2 (300 MHz, Acetone‐d ₆ ).

Figure S2: ¹³C NMR spectrum of compound 2 (75 MHz, Acetone‐d ₆ ).

Figure S3: ¹H NMR spectrum of compound 3a (300 MHz, CDCl₃).

Figure S4: ¹³C NMR spectrum of compound 3a (75 MHz, CDCl₃).

Figure S5: ¹H NMR spectrum of compound 3b (300 MHz, CDCl₃).

Figure S6: ¹³C NMR spectrum of compound 3b (75 MHz, CDCl₃).

Figure S7: ¹H NMR spectrum of compound 4a (300 MHz, CDCl₃).

Figure S8: ¹³C NMR spectrum of compound 4a (75 MHz, CDCl₃).

Figure S9: ¹H NMR spectrum of compound 4b (300 MHz, CDCl₃).

Figure S10: ¹³C NMR spectrum of compound 4b (75 MHz, CDCl₃).

Figure S11: ¹H NMR spectrum of compound 7 (300 MHz, CDCl₃).

Figure S12: ¹H NMR spectrum of compound 8 (300 MHz, CDCl₃).

Figure S13: ¹³C NMR spectrum of compound 8 (75 MHz, CDCl₃).

Figure S14: ¹H NMR spectrum of compound 9 (300 MHz, CDCl₃).

Figure S15: ¹³C NMR spectrum of compound 9 (75 MHz, CDCl₃).

Figure S16: ¹H NMR spectrum of compound 10 (300 MHz, CDCl₃).

Figure S17: ¹³C NMR spectrum of compound 10 (75 MHz, CDCl₃).

Figure S18: ¹H NMR spectrum of compound 11 (300 MHz, CDCl₃).

Figure S19: ¹³C NMR spectrum of compound 11 (75 MHz, CDCl₃).

Figure S20: ¹H NMR spectrum of compound 12 (300 MHz, DMSO‐d ₆ ).

Figure S21: ¹³C NMR spectrum of compound 12 (75 MHz, DMSO‐d ₆ ).

Figure S22: ¹H NMR spectrum of compound 13 (300 MHz, DMSO‐d ₆ ).

Figure S23: ¹³C NMR spectrum of compound 13 (126 MHz, Acetone‐d ₆ ).

Figure S24: ¹H NMR spectrum of compound 14 (300 MHz, CDCl₃).

Figure S25: ¹³C NMR spectrum of compound 14 (75 MHz, CDCl₃).

Figure S26: ¹H NMR spectrum of compound 15 (300 MHz, CDCl₃).

Figure S27: ¹³C NMR spectrum of compound 15 (75 MHz, CDCl₃).

Figure S28: Mass spectrum of compound 2.

Figure S29: Mass spectrum of compound 8.

Figure S30: Mass spectrum of compound 9.

Figure S31: Mass spectrum of compound 11.

Figure S32: Mass spectrum of compound 12.

Figure S33: Mass spectrum of compound 13.

Figure S34: Mass spectrum of compound 14.

Figure S35: Mass spectrum of compound 15.

Figure S36: Mean values of the slopes (y values) and respective standard deviations as results of the in vitro inhibition of COX‐2 by chalcone 16 (0–1 μM) using three substrate concentrations (x values: 6,25, 25 and 100 μM).

Figure S37: Sum of the squares (sum) of the different models (without inhibition, competitive inhibition, noncompetitive inhibition, uncompetitive inhibition, and mixed inhibition) determined from the results obtained from COX‐2 inhibition by chalcone 16.

Figure S38: Comparison of the different models (without inhibition, competitive inhibition, noncompetitive inhibition, uncompetitive inhibition, and mixed inhibition), based on the COX‐2 inhibition by chalcone 16.

Figure S39: Error parameters determination (V _max, K _m, K _ic, and K _iu) for mixed inhibition model of COX‐2 by chalcone 16, through “Jackknife” procedure.

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.