Structure–Activity Relationship and Molecular Docking of Quinazolinones Inhibiting Expression of COX-2, IL-1β, iNOS, and TNF-α through NF-κB Pathways

Abstract

The quinazolinone scaffold is found in natural products and biologically active compounds, including inflammatory inhibitors. Major proteins or enzymes involved in the inflammation process are regulated by the amount of gene expression. Quinazolinone derivatives were investigated and developed against the inflammatory genes cyclooxygenase-2 (COX-2), interleukin-1β (IL-1β), tumor necrosis factor-α (TNF-α), and inducible nitric oxide synthase (iNOS) in the lipopolysaccharide (LPS)-stimulated RAW 264.7 macrophage cell line. The mRNA expressions were measured using a real-time quantitative polymerase chain reaction (RT-qPCR). Quinazolinone compounds at 62.5 μM demonstrated anti-COX-2 and anti-IL-1β mRNA expressions down to 0.50% and 3.10% gene expression, respectively, via inhibition of nuclear factor κB (NF-κB). Molecular docking was performed to explain the interaction between the binding site and the developed compounds as well as the structure–activity relationship of the quinazolinone moiety.

Inflammation is a biological response to protect organs or tissues from harmful stimuli such as infections, radioactivity, toxins, or injury. As the body heals or eliminates stimuli, inflammation is reduced; otherwise, it can worsen and manifest as pain, swelling, redness, and heat. Inflammation is a symptom of many diseases, including migraine, rheumatoid arthritis, osteoporosis, and cancer.¹ Cytokines such as lipopolysaccharide (LPS), interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α) stimulate cytokine receptors and release arachidonic acid (AA) from the cell membrane using phospholipase A2 (PLA2). AA is oxidized by the cyclooxygenase (COX) enzyme and furnishes prostaglandin H2 (PGH2) as a precursor for bioactive prostaglandins such as PGE2 and PGI2 as well as prostanoid thromboxane A2 (TXA2). Prostaglandin and prostanoid levels are related to the inflammatory response.² Moreover, the translation process of cytokines and the COX enzyme is controlled by gene expression affecting cytokines, enzyme production, and inflammation, respectively. NF-κB and activating factor (AP-1) are crucial transcriptional factors that regulate many genes.³

The quinazolinone core structure is found in several biologically active compounds, both natural and synthetic. For example, bouchardatine isolated from Bouchardatia neurococca has adipogenesis/lipogenesis effects,⁴ and febrifugine has been used for the treatment of malarial infections.⁵ Synthetic compounds 3–8 exhibited anti-inflammatory properties,⁶⁻¹¹ as shown in Figure 1. Various substituents have been investigated to modify the core structure of quinazolinone and improve its anti-inflammatory activity. For example, compound 3 having an indole moiety exhibited 82.4% edema,¹² while 3-(p-substituted phenyl)-6-bromo-4(3H)-quinazolinone (4) and methylsulfonyl-substituted 2,3-diarylquinazolinones (5) improved inhibition of COX-2 activity, with quinazolinone-bearing parazole (6) reducing acute inflammation by up to 53%. Some quinazolinone derivatives (7, 8) displayed transcription factor and activating factor inhibition that impacted gene transcription and gene expression, respectively.³ Therefore, NF-κB and AP-1 are interesting targets to inhibit inflammatory processes. However, data on the structure–activity relationship (SAR) of quinazolinone derivatives to NF-κB or AP-1 inhibition are scarce and disorganized.³

Figure 1.

Bioactive compounds containing a quinazolinone core structure.⁶⁻¹¹

Our recent study focused on green chemical methods to synthesize quinazolinones and investigate their biological properties, especially their anti-inflammatory effects. This study investigated the SAR of substituents on the quinazolinone core to reduce the expression of inflammatory genes, such as COX-2, IL-1β, TNF-α, and inducible nitric oxide synthase (iNOS) via NF-κB inhibition. Real-time quantitative polymerase chain reaction (RT-qPCR) was used to assess the inhibition of certain gene expressions, and computational docking and a Western blot were performed to describe the inhibitory mechanism.

Quinazolinone derivatives were synthesized in two ways for the SAR study (Scheme 1). Quinazolinones 9–18 were synthesized using Kerdphon’s method.¹³ The 2-aminobenzamides and alcohol derivatives were catalyzed using Cu(OAc)₂·H₂O under oxygen gas at 110 °C for 6–24 h and furnished the desired products in up to 99% yield (Scheme 1, Route A). Compounds 15 and 16 were prepared from methyl 2-aminobenzoate in two steps via amide formation to obtain the corresponding amides in 87% and 75% yield, respectively. This was followed by cyclization as described by Kerdphon.¹³

Scheme 1. Synthetic Routes for Quinazolinones Used in the SAR.

Quinazolinone compounds 19–27 containing an aliphatic or aromatic substituent at position R³ were prepared from 2-aminobenzamide and various aldehydes under reflux conditions using iodine in ethanol as the catalyst, giving the products in moderate to high isolated yields (Scheme 1, Route B).¹⁴

The cell viability of quinazolinones was determined using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay to investigate the highest concentrations of all synthesized compounds that were not harmful to RAW 264.7 cells. Compounds 21, 24, 25, and 27 were not soluble in Dulbecco’s modified Eagle’s medium (DMEM) at 125 and 250 μM. A concentration of 62.5 μM was the highest at which all compounds were not toxic to the cells (% cell viability greater than 80%) except for compound 27 (Figure 2). Thus, this concentration of quinazolinones was chosen for further study.

Figure 2.

Cell viability percentages of 31.25–250.00 μM quinazolinones on RAW 264.7 cells.

The anti-inflammatory activities of different types, numbers, and substituent positions on quinazolinones 9–26 were examined in RAW 264.7 cells using GAPDH as a housekeeping gene. The TNF-α, IL-1β, iNOS, and COX-2 genes were used to measure the amount of gene expression involved in the inflammatory processes (Figure 3).

Figure 3.

Effect of quinazolinones on TNF-α, IL-1β, iNOS, and COX-2 RNA expressions. Cells were treated with LPS (100 ng/mL) and quinazolinones (62.5 μM) or blank. Means with an asterisk indicate significant 95% difference from cells treated with only LPS.

To investigate the effects of quinazolinones on TNF-α gene expression, RAW 264.7 cells were treated for 16 h with LPS (100 ng/mL) either alone or in combination with one quinazolinone compound. RNA was then extracted using the RNAspin Mini RNA Isolation Kit, while cDNA was synthesized using the iScript cDNA Synthesis Kit. TNF-α mRNA was measured by using RT-PCR and normalized by using GAPDG as the housekeeping gene. Quinazolinones 10–12 containing the halogen atom (chlorine or fluorine) at position R¹ decreased LPS-induced TNF-α mRNA expression to less than 50%. When treated with compounds bearing an aliphatic substituent at position R³ (compounds 19, 21, and 22), LPS-induced TNF-α mRNA expression was in the range of 44–50%.

The inhibition of iNOS gene expression was also investigated. All synthetic quinazolinones reduced the expression of iNOS mRNAs except for compound 26, which contained a methoxy group at the para position R³ on the aromatic. The pyridine ring showed no improvement in the inhibition of iNOS mRNA expression (62.82% gene expression). The fluorine substituent at position R¹ (compounds 11 and 12) slightly reduced gene expression. Compared with compound 9, the R² substituent affected the mRNA expression of LPS-treated cells. A slight increase in the level of induced gene expression was observed when R² was an aliphatic group. The presence of an aromatic substituent at position R² only minimally reduced iNOS mRNA expression to less than 22% (compounds 14–16). The effect of varying the substituent at position R³ was also investigated. Compounds 17, 20, and 21 bearing a longer aliphatic chain at position R³ and 18, 19 and 22 with a bulky group at R³ decreased gene expression of iNOS mRNA. When R³ was an aromatic substituent, the electron-donating group on the aromatic ring did not have any effect on inhibition of gene expression, while the electron-withdrawing group −CF₃ showed excellent inhibition of iNOS mRNA expression at 13.98% (compound 25).

The IL-1β gene, which is crucial for controlling the production of IL-1β during inflammatory processes, was investigated. Quinazolinone and its derivatives were found to inhibit the IL-1β gene formation. When the R¹ substituent on the aromatic ring was a halogen such as chlorine or fluorine (compounds 10–12), IL-1β mRNA expression decreased to 10.85% compared to cells treated with only LPS. The pyridine ring (13) did not significantly improve the inhibitory effect, while aliphatic substituents installed at position R³ (18, 19, and 21) displayed good to excellent inhibition of gene expression for IL-1β formation compared to dexamethasone. When R³ was a methyl, hexyl or octyl group, IL-1β mRNA was expressed at 58.79%, 3.10%, and 6.92%, respectively (compounds 17, 20, and 21). Furthermore, bulky R³ compounds 18 and 19 were also used as inflammatory inhibitors and achieved gene expression of IL-1β at 14.02% and 4.02%, respectively. The cyclohexyl and phenyl groups at position R³ were also examined for gene expression activity, and IL-1β mRNA expression decreased to 75.72% and 47.82%, respectively. These results suggested that bulky and long chains of aliphatic substituents increased the level of inhibition of IL-1β gene expression. To improve the inhibition of gene expression on the aromatic R³ substituent, the electron-donating and electron-withdrawing groups (compounds 24–26) were used as substituents on the aromatic ring and achieved IL-1β mRNA expression of less than 35%.

The quinazolinones effectively blocked LPS-induced COX-2 gene formation, with halogen substituents such as chlorine or fluorine at aromatic position R¹ (compounds 10–12) exhibiting up to 9-fold COX-2 mRNA expression inhibition compared to compound 9. However, the pyridine ring of compound 13 was unable to control the COX-2 gene formation. The substituents on R² were also examined. When methyl or phenyl was the substituent, the COX-2 mRNA expression was reduced to 28.85% and 28.46%, respectively (compounds 14 and 15). A bulky group at position R² increased inhibition activity of COX-2 gene formation (compound 16), while compared to compound 9 the substituent at position R³ effectively reduced LPS-induced COX-2 gene production. These compounds displayed better inhibition of COX-2 gene expression, similar to commercial dexamethasone, with the longer-chain aliphatic substituents at position R³ reducing the level of COX-2 gene production. When methyl was the substituent at position R³, the level of COX-2 gene expression was 60.96%. Replacing the methyl group with a longer side chain such as isopropyl or sec-butyl led to better inhibition of COX-2 mRNA formation at 20.47% and 0.50%, respectively. To increase the inhibitory property, n-hexyl was used at position R³, giving 95% inhibition of the COX-2 gene expression (compound 20). A longer-side-chain substituent (eight carbons) was also tested for gene expression, and the COX-2 mRNA formation increased slightly to 11.14% (compound 21). Having cyclohexyl or phenyl at position R³ gave a good reduction of the level of COX-2 gene formation at more than 80%. Various other substituents at the aromatic position R³ were also investigated for inhibition of gene expression. Surprisingly, methyl at the meta position (compound 24) displayed excellent inhibition of LPS-induced COX-2 gene production with only 3.79% COX-2 gene expression, whereas the trifluoromethyl and methoxy groups had no inhibitory effect.

The concentration-dependent manner was investigated using compounds 21 and 22 with concentrations ranging from 0.1 to 62.5 μM. The results showed that when the concentrations decreased, either COX-2 or IL-1β gene expression displayed more production, as shown in Figure 4. Quinazolinones acted in a dose-dependent manner on inhibition of COX-2 and IL-1β expression in LPS-induced RAW 264.7 cells.

Figure 4.

Effect of quinazolinone concentrations on IL-1β and COX-2 RNA expressions. Cells were treated with LPS (100 ng/mL) and quinazolinones (0.1–62.5 μM) or blank.

In the gene expression experiments, the quinazolinones had the highest inhibition of COX-2 mRNA synthesis, so this inhibitory pathway was further examined. According to the literature, quinazolines and quinazolinones enabled downregulation of COX-2 in both gene expression and protein production. Many studies on these derivatives suggested that it involves suppression of the NF-κB pathway.^3,15,16 Here, NF-κB p65 was determined by Western blot. NF-κB and IκB complexes were degraded by phosphorylation using IκB kinase (IKK). Phosphorylated NF-κB p65 (pNF-κB) is transported into the nucleus, contributing to upregulation of inflammation-related genes. The ratios of pNF-κB/β-actin and NF-κB/β-actin obtained from LPS-induced RAW 264.7 cells which were treated with compounds 20 and 24 were measured. Our findings demonstrated that compound 20 resulted in greater phosphorylation inhibition on NF-κB than compound 24. These results correspond to the COX-2 and IL-1β mRNA expression (compound 20 inhibited the gene expression better than compound 24), as shown in the Supporting Information.¹⁷

Therefore, NF-κB was selected for a molecular docking study to clarify the mechanism of action of our synthetic quinazolinones. All potential anti-inflammatory candidates were docked into the DNA binding site of NF-κB, with cocrystallized DNA used for specifying the grid box for docking. The surfaces of the receptors are depicted in Figure 5. Residues that play important roles in DNA binding are Arg54, Arg56, Tyr57, Cys59, Glu60, His64, Gly65, Gly66, Lys144, Lys145, Lys241, Lys272, Gln274, Lys275, Arg305, and Arg306.¹⁸ These key amino acid residues were used to select the binding pose of the candidates for visualization.

Figure 5.

Mechanism of COX-2 gene expression and its potent biomolecular target NF-κB receptor to achieve in silico screening of anti-inflammatory agents.

All of the selected compounds, except for compounds 25 and 26, were able to accommodate the DNA binding site, with binding energies ranging from −4.8 to −6.6 kcal mol^–1. The binding poses were chosen based on molecular interactions and binding energy. Binding energies and molecular interactions are listed in Table 1.

Table 1. Binding Energies and Molecular Interactions of Candidate Anti-inflammatory Agents on NF-κB Receptor.

		molecular interactions
compound	binding energy (kcal mol–1)	H-bonding	hydrophobic	electrostatic
10	–4.8	Ser240, Lys241, Asn247	Ser246, Asp271, Lys272, Arg305, Phe307	–
11	–5.2	Ser240, Lys241, Arg305, Asp271	Ser246, Asn247, Lys272, Phe307	–
12	–5.0	Ser240, Lys241, Asn247, Arg305, Asp271	Ser246, Lys272, Phe307	–
17	–5.0	His141, Lys241	Tyr57, Leu207, Ser208, Ala242, Pro243	–
18	–5.1	Asp239, Lys241	Leu207, Ala242	His141
19	–5.1	Asp239, Lys241	Tyr57, Leu207, Ala242	His141
20	–5.3	Asp239	Tyr57, Lys241, Ala242, Pro243	His141
21	–5.0	Lys241, Ser208	Cys59, Leu207, Ala242	His141
22	–6.0	Asp239	Leu207, Lys241, Ala242	His141
23	–6.4	Ser208, Asp239	Tyr57, Lys241, Ala242, Pro243	His141
24	–6.6	Ser208, Asp239	Tyr57, Lys241, Ala242	His141
dexamethasone (+)	–6.0	Ser240, Asn247, Asp271, Arg305	Lys241, Ser246, Lys272, Gln306, Phe307	–

Our finding suggested that two quinazolinone binding regions can be classified based on substrate specificity, referred to as active sites I and II in this study (Figure 5). Active site I was constructed of the side chains of Arg54, Arg56, Cys59, Glu60, Lys144, Lys145, Lys241, and Lys272 of chain A and Arg305 of chain B, while active site II was built of amino acids Arg54, Arg56, Tyr57, Cys59, Lys144, and Lys241 of chain B. Site I was larger than site II. Quinazolinones with substituents at C2, including compounds 17–24, preferred to bind to this pocket due to decreased steric repulsion. Thus, active site II was the smaller binding site. Only small quinazolinone derivatives, namely, compounds 10, 11, and 12, favored this binding site.

The interaction profiles (Figure 6) also confirmed that smaller quinazolinone derivatives, especially compounds 10, 11, and 12, bound at site II. The interactions were predominantly hydrogen bonding with Ser240, Lys241, Asn247, Asp271, and Arg305 residues. The carbonyl group in the quinazolines formed hydrogen bonds with the Lys241 and Asn247 residues. The amino group interacted with amino acid Ser240 and contributed to the interaction between the imine moiety of the pyrimidine ring with Lys241.

Figure 6.

2D plots of protein–ligand interactions between anti-inflammatory candidates and NF-κB receptor.

The effect of a halide atom on C6 and C7 was considered. The presence of a fluorine atom played a crucial role in determining the binding ability. The fluorine atom enabled hydrogen bonding with the Arg305 and Asp271 residues. These bonds were absent when a chlorine atom replaced the fluorine atom. Overall, a fluorine atom substituent was found to significantly enhance the binding ability compared to a chlorine atom, with the presence of a fluorine atom on C7 resulting in a stronger binding affinity to NF-κB than a fluorine at C6.

The C2-substituted quinazolinone derivatives 17–24 displayed binding to a different site than the smaller quinazolinones, which were bound to active site I. The binding of these compounds to the receptor was facilitated by hydrogen-bonding, hydrophobic, and electrostatic interactions as well as π–π and amide−π interactions. Notably, the amino group of these derivatives displayed hydrogen bonding with the receptor via interaction with Asp239 and Lys241 residues, similar to the smaller quinazolinones. Moreover, the aromatic and pyrimidine rings of these compounds interacted with His41 through π–cation interactions.

Compounds 18, 19, 20, and 21 with alkyl substituents at C2 had stronger interactions between protein–ligand complexes than compound 17 because an increase in the number of carbon atoms in the alkyl chain led to an increase in hydrophobic interactions with the receptor. This finding was consistent with the in vitro experiment on the expression of the COX-2 gene. The alkyl side chain interacted mainly with Tyr57, Lys241, and Cys59, with maximum binding ability observed in compound 20 with a six-carbon-atom side chain. However, as the length of the alkyl side chain increased, steric effects on the binding ability were observed, as seen in compounds 21 and 22, indicating an unfavorable interaction with the Tyr57 residue. Compound 20 inhibited the COX-2 gene expression the most, consistent with in silico molecular modeling, giving the lowest binding energy among these alkylated quinazolinone derivatives.

When C2-substituted aromatic quinazolinones 23, 24, 25, and 26 were compared to alkylated quinazolinones, these compounds bound to the same site. However, the aromatic ring substituent affected the binding abilities of these compounds. Alkylated quinazolinones and compounds 23 and 24 showed similar binding profiles, while the imine moiety of the pyrimidine ring interacted with Asp239 via hydrogen bonding and the aromatic and pyrimidine rings interacted with the His141 residue via electrostatic interactions. The aromatic moiety also contributed to π interactions with Lys241 and Tyr57 residues via π–alkyl and π–π interactions, which were stronger than the hydrophobic interactions observed in alkylated quinazolinones. Compound 24, with a methyl group on the aromatic ring, had a hydrophobic interaction with the receptor that was not present in compound 23. Compounds 25 and 26 did not bind to NF-κB receptors, attributed to an unfavorable interaction of the trifluoromethyl group with the receptor and the shape of the p-methoxyphenyl moiety, which did not fit into the receptor pocket. Additionally, we also compared these observed interactions with those of the standard anti-inflammatory drug dexamethasone. The results suggested that dexamethasone interacted with key amino acids similar to those quinazolinones with a binding energy of −6.0 kcal mol^–1. Dexamethasone interacted with NF-κB via conventional hydrogen bonds with residues Ser240, Asn247, Asp271, and Arg305 along with van der Waals interactions with residues Lys241, Ser246, Lys272, Gln306, and Phe307. This suggested that quinazolinone derivatives inhibited COX-2 gene expression via the NF-κB signaling pathway.

In summary, quinazolinones were synthesized to investigate the structure–activity relationship of inflammatory gene expression by using GAPDH as a housekeeping gene. Quinazolinones having an aliphatic substituent at position R³ showed good to excellent inhibition of inflammatory gene expression, including COX-2, iNOS, and IL-1β mRNA. However, to decrease COX-2 and iNOS mRNA expression, the aliphatic chain at the R³ position should not be longer than eight carbons because it may have steric hindrance to target proteins that impact COX-2 gene expression such as NF-κB. The bulky aliphatic substituent at position R³ displayed higher inhibition of IL-1β mRNA expression than the long aliphatic chain substituent.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.3c00098.

• Compound synthesis, characterization, biological activity protocols, and molecular docking (PDF)

Author Contributions

All of the authors approved the final version of the manuscript.

The authors declare no competing financial interest.