Crocetin Halts Tumorigenesis in Metastatic Breast Cancer via Disrupting CYP2J2/EET Axis

Abstract

CYP2J2 catalyzes the conversion of arachidonic acid into epoxyeicosatrienoic acids (EET), which have gained significant attention due to their oncogenic properties. Differential overexpression of CYP2J2 in tumors results in increased EET production. The role of EET in tumorigenesis is quite clear, but inhibition of CYP2J2 on EET formation with its molecular target is poorly understood. Research is ongoing to identify CYP2J2 inhibitors and the role of the CYP2J2/EET axis in halting tumorigenesis. In these contexts, we aimed to address the same through in silico, in vitro, and in vivo approaches using crocetin, a phyto-based druggable molecule. In vitro mechanistic investigations using human liver microsomes (HLM) indicate that crocetin could act as a specific and reversible CYP2J2 inhibitor. In silico molecular docking analysis and molecular dynamics (MD) simulation explicate strong and stable interactions between crocetin and active site of human CYP2J2. In vivo investigations in BALB/c mice reveal that crocetin could enhance the plasma exposure of rivaroxaban (CYP2J2 substrate) by delaying metabolism and attenuating CYP2J2 protein expression in the liver tissues. Additionally, in vitro studies using HLM and mouse liver microsomes (MLM) suggest that crocetin could substantially hinder EET formation. Thereafter, crocetin showcases its antitumor effect in the mouse model of breast cancer involving downregulation of CYP2J2 expression and impediment of EET formation in the tumor tissues. Further, proteomics data from crocetin-treated tumor tissues with gene ontology analysis (KEGG database) reflect upregulation/downregulation of key proteins and biological processes associated with attenuating tumorigenesis. It is a footstep toward understanding the CYP2J2/EET axis for targeted breast cancer therapy.

Introduction

Cytochrome P450 (CYP) enzymes are vital for the biotransformation of xenobiotics and are responsible for the phase-I oxidative metabolism of approximately 80% of widely prescribed medications. Among these CYP enzymes, CYP2J2 is the sole one belonging to the human CYP2J subfamily. Unlike other CYP isozymes, CYP2J2 is expressed mostly in extrahepatic tissues. , As it constitutes only 1 to 2% of the total CYP content in the liver, it has limited contribution in metabolizing drugs of clinical practice. − CYP2J2 has garnered substantial attention due to its epoxygenase activity to transform arachidonic acid into regioselective epoxyeicosatrienoic acids (EET), which include 14,15-EET, 11,12-EET, 8,9-EET, and 5,6-EET. All these EET are important lipid mediators that may actively facilitate the proliferation, migration, and adhesion of cancerous cells through phosphorylating EGFR and activating downstream PI3K-AKT and MAPK signaling pathways shown in overexpressing CYP2J2 carcinoma cells. − Furthermore, CYP2J2 also exhibits significant expression in various tumor types, where overexpression of CYP2J2 and elevated levels of EET are shown to trigger malignancy. − From that perspective, elevated EET's level driven by CYP2J2 overexpression in tumors mediated by EET represents that the CYP2J2/EET axis can be a promising target for therapeutic intervention in treating human malignancies including breast cancer. In this direction, various research groups are exploring CYP2J2 inhibitors, which showed significant antitumor effects in vitro and in vivo by reducing EET biosynthesis. − Although a tentative signaling pathway is not quite clear to date and research works are ongoing to elucidate that these inhibitors have been shown to slow down cancer cell growth and metastasis, promote cancer cell death, decrease its cell adhesion, migration, and invasion, and increase the survival time of tumor-bearing mice. −

Natural products and their derivatives have shown significant potential for advancing chemotherapeutics, exhibiting a broad range of structural diversity and favorable pharmacological and molecular traits conducive to drug development. Between 1981 and 2014, it is worth noting that of the 131 anticancer drugs that emerged, 49% were derived from natural products and their derivatives. Paclitaxel, docetaxel, vincristine, and vinblastine are only a few examples from the overwhelming list of plant-derived anticancer therapeutics dominating the pharma market. In this context, we aim to explore crocetin, an aglycone of crocin naturally occurring in the stigma of Crocus sativus Linne and the fruits of Gardenia jasminoides Ellis, because of its noteworthy pharmacological effects against several cancers via obstructing the growth factor signaling pathway, arresting the cell cycle, and triggering apoptosis. − Moreover, crocin extract containing crocetin is highly safe and has favorable drug-like properties. , However, as of now, there is no evidence available in the literature for its effect on the CYP2J2/EET axis in any cancer including its CYP2J2 inhibitory action.

Hence, our current study aimed to explore the CYP2J2 inhibitory potential of crocetin along with its underlying mechanism using three different approaches: in vitro (human liver microsomes/HLM and mouse liver microsomes/MLM), in silico (molecular docking), and in vivo (normal mouse model for validation of pharmacokinetic interaction and mouse model of breast cancer for validation of antitumor efficacy).

Results and Discussion

Crocetin Showed CYP2J2 Inhibition

To investigate the inhibitory effects of crocetin on CYP2J2, our first objective was to conduct in vitro CYP2J2 inhibition studies. The liver serves as the primary organ responsible for drug metabolism in the body. Liver microsomes are, therefore, valuable in vitro models for investigating CYP-mediated metabolism. Thus, we employed HLM for these inhibition studies. Unlike other CYP isoforms, the United States Food and Drug Administration (USFDA) has not yet recommended any specific in vitro marker reaction to evaluate the CYP2J2-mediated metabolism. Hence, numerous substrates are being investigated by different research groups to examine the metabolism mediated by CYP2J2. Astemizole was selected as our experimental CYP2J2 substrate because in the case of other substrates such as ebastine, terfenadine, and amiodarone, their metabolites (such as hydroxy-ebastine and terfenadine alcohol) or the substrate (such as amiodarone) themselves have been found to suppress the activity of CYP2J2; , however, astemizole does not exhibit any such liability, which makes it more suitable to study the metabolic role of CYP2J2. Therefore, we chose ’astemizole O-demethylation’ as an index reaction for CYP2J2 inhibition studies. First of all, kinetic parameters pertaining to this reaction were assessed to validate the reaction conditions optimized for evaluating the CYP2J2 inhibition. O-desmethyl astemizole as a metabolite was quantified in the reaction mixture using liquid chromatography-tandem mass spectrometry (LC-MS/MS). The determined values for the maximum velocity of the uninhibited reaction (V _max) and the Michaelis constant (K _m) of the index reaction were 163.40 ± 4.21 pmol/min/mg of protein and 1.13 ± 0.08 μM, respectively (Figure A). The results of this study are consistent with the reported values. ,

1.

Michaelis–Menten plot for the formation of O-desmethyl astemizole in HLM (A); IC₅₀ curve of terfenadone (B); danazol (C); and crocetin (D) for CYP2J2-catalyzed astemizole O-demethylation in HLM. IC₅₀ shift of ritonavir (E) and crocetin (F) for CYP2J2-catalyzed astemizole O-demethylation in HLM. Data are represented as mean ± SEM (n = 3).

Following the determination of the K _m value, the concentration of astemizole was then fixed at 1 μM for further studies to assess the inhibitory effect of crocetin on CYP2J2. For this, the half-maximal inhibitory concentration (IC₅₀) of crocetin for CYP2J2 was estimated in HLM. The IC₅₀ value of an inhibitor is directly related to its potency, which determines the minimum amount of inhibitor needed to achieve the desired inhibitory effect. IC₅₀ of terfenadone and danazol as the standard CYP2J2 inhibitors was also evaluated alongside the test candidate and was determined to be 3.63 ± 0.23 and 13.16 ± 1.17 μM, respectively (Figure B,C). Earlier reports for terfenadone showed a relatively lower IC₅₀ than our experimental value, which can be attributed to the differences in the enzyme system and hydrogen source reducing agent as the inhibitory effect of terfenadone against CYP2J2 activity has mainly been assessed using the recombinant CYP2J2 enzyme and nicotinamide adenine dinucleotide phosphate (NADPH) generating system. , IC₅₀ of danazol, another standard CYP2J2 inhibitor, aligns with the previously reported values. , We used these two standard CYP2J2 inhibitors with different inhibitory potential (terfenadone with IC₅₀ < 10 μM & danazol with IC₅₀ > 10 μM) in all of our present in vitro assays. The experimental IC₅₀ for crocetin was found to be 13.48 ± 1.02 μM (Figure D). These findings indicate that crocetin could inhibit CYP2J2 in HLM. Several phytochemicals are reported for their CYP2J2 inhibitory activity, like bilobetin and broussochalcone A, but those have a shortfall of nonselectivity toward CYP2J2 inhibition. ,− Our previous investigation reveals the CYP inhibitory activity of crocetin for the USFDA-recommended panel of CYP isoforms for drug interaction assessment with IC₅₀ values in the following order: CYP2D6 (IC₅₀: 20.35 μM) > CYP2C8 (IC₅₀: 35.32 μM) > CYP1A2/CYP2B6/CYP2C9/CYP2C19/CYP3A4 (IC₅₀ > 50 μM) representing the selective nature of crocetin for CYP2J2 inhibition compared to other CYP isoforms.

Thereafter, to examine the underlying mechanism behind crocetin-mediated CYP2J2 inhibition, an IC₅₀ shift assay was conducted, in which the enzymatic activity of CYP2J2 was assessed following the preincubation of 0 and 30 min with crocetin in the presence of NADPH. This assay reflects the time-dependent inhibition potential of the test compound. A standard time-dependent CYP2J2 inhibitor, ritonavir, was evaluated simultaneously and exhibited a 2.4-fold-shift in the IC₅₀ value (Figure E). This finding aligns with the outcomes published in the existing literature. The present investigation reveals that the IC₅₀ shift value for crocetin is <1.5-fold. Thus, crocetin did not exhibit any time-dependent inhibitory effect on CYP2J2 (Figure F). Dronedarone and amiodarone are the other well-known drugs having time-dependent CYP2J2 inhibitory action.

The deficit in irreversible inhibition of crocetin for CYP2J2 led us to further examine the reversible mode of inhibition and determination of the inhibition constant (K _i) value of crocetin for CYP2J2 in HLM, in which the metabolite formation rate was investigated by varying the concentration of astemizole (substrate) and crocetin (test candidate). To perform this study, the obtained K _m value of astemizole (∼1 μM) and IC₅₀ value of crocetin (∼13 μM) were employed to choose the concentration range of astemizole from 0.5 to 4 μM and crocetin from 6.75 to 54 μM. The K _i values of terfenadone and danazol as standard inhibitors were also assessed and were found to be 1.60 ± 0.05 and 20.44 ± 0.53 μM, respectively (Figure A–D). Crocetin displayed a mixed-type of inhibition, and the K _i value was determined to be 42.72 ± 3.08 μM with corresponding R ² value & α-value of 0.97 and 8.92, respectively. These findings were illustrated during visual inspection of the x- and y-intercepts, as well as the slopes for different experimental data sets in the Lineweaver–Burk plots and then further ascertained using the graphical analysis of Dixon plots (Figure E,F).

2.

Lineweaver–Burk and Dixon plots for the effect of terfenadone (A, B); danazol (C, D); and crocetin (E, F) on the kinetics of CYP2J2-catalyzed astemizole O-demethylation in HLM. Data are represented as mean ± SEM (n = 3).

This phenomenon of mixed inhibition is possible when the CYP enzyme metabolizes substrate through distinct active sites or when the inhibitor obstructs several active sites of the CYP enzyme, interfering with its enzymatic functions. The mixed inhibitor can disrupt the function of an enzyme by attaching itself to the enzyme–substrate complex or the unbound enzyme. it has varying levels of attraction to binding affinities for either. The affinity between an enzyme and substrate is modified by the presence of an inhibitor, and the α-value quantifies this change. The current data indicate that the observed α-value was 8.92, i.e., >1, which indicates its competitive nature. These findings indicate that crocetin has a strong propensity for association with the free CYP2J2 enzyme compared to the enzyme–substrate complex. Consequently, crocetin prevents the binding of the substrate with the enzyme. Danazol was also reported for its mixed-type inhibition pattern for CYP2J2-catalyzed astemizole O-demethylation with an inclination toward competitive inhibition (α-value: 18.3).

Crocetin Inhibited the Metabolic Depletion of CYP2J2 Substrates

The in vitro CYP2J2 inhibition studies suggest that crocetin could significantly inhibit the enzymatic activity of CYP2J2 in HLM. However, before moving further with the in vivo investigations using the mouse model, we want to ascertain the CYP2J2 inhibitory capability of crocetin in MLM. For this, we planned to explore the metabolic depletion of astemizole as a CYP2J2 substrate in the presence of crocetin by employing species-specific microsomes (MLM & HLM). Here, the use of species-specific microsomes also enables us to understand interspecies differences in the metabolism of the substrate. Following a 30 min of incubation period, the percentage of astemizole remaining in HLM and MLM was determined to be 27 ± 5 and 25 ± 4%, respectively. Thus, astemizole exhibits a similar kind of metabolic behavior in both the microsomes tested. In addition to the test compound, the standard inhibitors terfenadone (Figure A,B) and danazol (Figure C,D) were also evaluated simultaneously. Here, with increasing concentration of terfenadone (i.e., from 0 to 100 μM), the percentage of astemizole remaining was increased from 37 to 88% in HLM and 32 to 84% in MLM. In the case of danazol (from 0 to 100 μM), the percentage of astemizole remaining was increased from 20 to 56% in HLM and 20 to 98% in MLM. In a similar manner, the increasing concentration of crocetin (from 0 to 100 μM) increased the percentage of astemizole remaining from 23 to 58% in HLM and 22 to 54% in MLM. Results indicate a positive correlation between the concentration of crocetin and the percentage of astemizole remaining in both the microsomes studied (Figure E,F).

3.

Metabolic depletion of astemizole in the presence of terfenadone (A, B), danazol (C, D), and crocetin (E, F) using HLM and MLM. Data are represented as mean ± SEM (n = 3). Data are compared: terfenadone/danazol/crocetin (0 μM) at 30 min vs terfenadone/danazol/crocetin (1 to 100 μM) at 30 min. p < 0.05/0.01/0.001 (*/**/***) denotes statistically significant; ns denotes not significant.

Another CYP2J2 substrate, rivaroxaban, which we were going to use in upcoming in vivo studies, was similarly studied for its metabolic depletion in the presence of crocetin. Following a 30 min incubation period, it was seen that 81 ± 2% of rivaroxaban remained in HLM, whereas the remaining percentage in MLM was determined to be 70 ± 1%, indicating the relatively rapid metabolic clearance of rivaroxaban in MLM compared to that in HLM. Terfenadone (Figure A,B) and danazol (Figure C,D) as the standard inhibitors were also assessed. In this study, as the concentration of terfenadone increased from 0 to 100 μM, the percentage of rivaroxaban remaining also increased from 84 to 100% in HLM and 71 to 100% in MLM. Similarly, as the concentration of danazol increased from 0 to 100 μM, the percentage of rivaroxaban remaining increased from 77 to 98% in HLM and 71 to 97% in MLM. In a similar manner, the increase in crocetin concentration (ranging from 0 to 100 μM) resulted in a corresponding rise in the percentage of rivaroxaban remaining, with values increasing from 83 to 100% in HLM and 68 to 91% in MLM. The results clearly reveal that the percentage of rivaroxaban remaining increased with increasing crocetin concentration in both tested microsomes (Figure E,F).

4.

Metabolic depletion of rivaroxaban in the presence of terfenadone (A & B), danazol (C & D), and crocetin (E & F) using HLM (A) and MLM (B). Data are represented as mean ± SEM (n = 3). Data are compared: terfenadone/danazol/crocetin (0 μM) at 30 min vs terfenadone/danazol/crocetin (1 to 100 μM) at 30 min. p < 0.05/0.01/0.001 (*/**/***) denotes statistically significant; ns denotes not significant.

Crocetin Interacted with the Active Site of Human CYP2J2

Before evaluating crocetin’s CYP2J2 inhibitory potential in the in vivo model, a molecular docking investigation between the human CYP2J2 enzyme and crocetin was conducted. By examining the involved amino acid residues, the preferred binding orientation, the major interactions, interaction energy, binding affinity, and the involved forces, we can gain a deeper knowledge of the crocetin-mediated CYP2J2 interaction (Figure A,B). During the molecular docking analysis, the best complex structure found as the best pose among all the conformers has been selected based on the lowest CDOCKER interaction energy (−47.653 kcal/mol). More negative CDOCKER interaction energy indicates more significant binding between crocetin and the human CYP2J2 enzyme. Molecular docking investigation reveals that crocetin interacted with the active site pocket of the enzyme through hydrogen bonding interaction with the amino acid residues such as GLN A: 228, LEU A: 449, and ARG A: 138 in the distance of 2.95, 2.06, and 2.75 Å, respectively, π-alkyl bond interaction with amino acid residue PHE A: 310 in the distance of 4.37 Å, alkyl bond interaction with amino acid residues LEU A: 378, LEU A: 402, ILE A: 487, PRO A: 381, ILE A: 376, VAL A: 380, and ILE A: 127 in the distance of 4.70, 4.85, 5.29, 4.93, (5.46 & 5.08), 4.16, and 5.07 Å, respectively, and attractive charge interaction with amino acid residue ARG A: 138 in the distance of 5.07 Å. Based on the aforementioned analysis, we have determined that the molecular docking and in vitro investigation findings align regarding crocetin’s CYP2J2 interacting potential. Furthermore, we performed an in silico alanine scanning mutagenesis analysis to strengthen our findings. The results showed (Figure S1) a notable reduction in binding affinity (CDOCKER interaction energy of −41.518 kcal/mol) and loss of key hydrogen bonding interactions in the mutant model, supporting the critical role of these residues in ligand binding.

5.

3D stereo image (A) and 2D stereo image (B) of molecular docking analysis for the interaction of crocetin with the active site of the human CYP2J2 enzyme. During MD simulation, RMSD profile plotted against time in nanosecond (ns) (C) and root-mean-square fluctuation (RMSF) of protein backbone atoms (D).

To gain comprehensive insights into the dynamic stability, conformational behavior, and binding interactions of crocetin, a 100 ns molecular dynamics (MD) simulation was carried out for the protein–ligand complex. Analysis of the simulation trajectories revealed an average root-mean-square deviation (RMSD) of 0.116 nm (Figure C), indicating that the complex reached equilibrium quickly and remained structurally stable throughout the simulation period. The consistently low RMSD values, along with the limited fluctuations observed, suggest that crocetin binding did not disrupt the structural integrity of the protein, highlighting the stability of the complex under physiological conditions. Furthermore, the average RMSF was found to be 0.131 nm (Figure D), indicating an overall acceptable level of flexibility across the protein residues. The relatively low RMSF values, particularly at the ligand binding regions, suggest that these sites remained structurally stable, an important characteristic for sustaining high-affinity interactions and ensuring functional specificity. The compactness and folding behavior of the protein were evaluated using the radius of gyration (Rg) analysis (Figure S2A). The average Rg value of 2.132 nm remained consistent throughout the simulation, indicating that the protein maintained a stable tertiary structure without noticeable unfolding or expansion. This structural stability upon ligand binding suggests that crocetin does not compromise the protein’s integrity, which is an encouraging characteristic for its potential as a drug candidate. Analysis of hydrogen bond formation further highlighted the strength and stability of the protein–ligand interactions (Figure S2B). The average number of hydrogen bonds observed was 2.211 for the complex with crocetin. Hydrogen bonding plays a crucial role in enhancing the specificity and affinity of ligand binding, and the observed interactions indicate that the ligands consistently maintained contact with key residues in the binding pocket during the simulation. Importantly, the total protein–ligand interaction energy was calculated to be −71.726 kcal/mol, indicating a highly favorable and stable binding interaction. This strongly negative energy value highlights crocetin’s potential as a potent agonist of the target protein. Together with the regular hydrogen bonding patterns and structural stability seen during the simulation, these results imply that crocetin might have a substantial biological function. The results of the MD simulations not only confirm the structural stability and strong binding affinity of the test candidate but also emphasize its potential as a promising lead for further experimental validation and optimization. The combination of low RMSD and RMSF values, stable Rg, consistent hydrogen bonding patterns, and highly favorable interaction energy collectively indicates that the experimental compound exhibits a favorable pharmacological profile and warrants further studies.

Crocetin Significantly Altered the Pharmacokinetics of Rivaroxaban in a Mouse Model

The positive in vitro and in silico findings made it abundantly evident that crocetin could significantly alter human CYP2J2 enzymatic activity. Knowing that CYP2J2 accounts for only 1 to 2% of the total CYP content in the liver, we were interested in determining the extent of pharmacokinetic interaction crocetin can cause at the in vivo level. We believed that encouraging in vivo findings at the preclinical level could be helpful prior to expensive and risky clinical exploration. However, variations in the extent of the pharmacokinetic interaction may exist in different species. We used the mouse model because this model exhibit CYP2J2 in the form of CYP2J5 (mouse homologue). So, we explored the pharmacokinetic interaction between the CYP2J2 substrate and crocetin (CYP2J2 inhibitor) using the BALB/c mice. Rivaroxaban was employed as a CYP2J2 substrate in this investigation, as it undergoes metabolism primarily by CYP2J2 (41.1%), with a lesser contribution from CYP3A4 (27.3%). Furthermore, there are reported instances in the literature wherein rivaroxaban was exclusively utilized as a substrate for CYP2J2 in preclinical models to investigate CYP2J2-mediated pharmacokinetic interactions. , Thus, we decided to study the pharmacokinetics of rivaroxaban in the presence of crocetin.

The impact of crocetin on the pharmacokinetics of rivaroxaban in the mouse model is represented in Figure A,B. Here, crocetin was pretreated orally for seven consecutive days prior to rivaroxaban administration. We chose to administer crocetin for multiple days in order to assess the sustained effect on the protein expression alterations. Administration of crocetin resulted in a significant increase in the C _max of rivaroxaban by 2.0-fold. Additionally, the AUC_0–t and AUC_0–∞ of rivaroxaban were both enhanced by 1.8-fold. In parallel to the aforementioned pharmacokinetic parameters, crocetin treatment notably delayed the rivaroxaban clearance by 45%. Although the T _max of rivaroxaban was reduced notably upon concomitant administration with crocetin, there was no statistically significant effect on its volume of distribution. In parallel to rivaroxaban, we also evaluated the pharmacokinetic profile of crocetin (Figure S3). The combination group achieved a C _max of 37.8 ± 9.6 μM crocetin in the mouse model. The results indicate that a sufficient level of crocetin has reached the systemic circulation to inhibit CYP2J2, which could be linked to our experimental IC₅₀ of crocetin for CYP2J2 in HLM (∼13 μM).

6.

Mean plasma concentration vs time profile (A) and main pharmacokinetic parameters (B) of rivaroxaban (4 mg/kg) in mice after oral administration as alone and in combination with crocetin (50 mg/kg). Data are represented as mean ± SEM (n = 5). C _max, highest plasma concentration; T _max, time to reach highest plasma concentration; AUC_0–t , area under the curve for plasma concentration from zero to last measurable plasma sample time; AUC_0–∞, area under the curve for plasma concentration from zero to infinity; V _d/F, volume of distribution after oral administration; and Cl/F, clearance after oral administration. Data are compared: rivaroxaban alone vs rivaroxaban with crocetin. p < 0.05/0.01/0.001 (*/**/***) denotes statistically significant; WB analysis and densitometry data for CYP2J2 protein expression in the liver tissue of the mice (C). β-actin expression: endogenous loading control.

After studying the pharmacokinetic effect of crocetin treatment, we aimed to investigate the impact of crocetin on the protein expression of CYP2J2, as currently there is limited information on the protein expression level of CYP2J2. For this, we investigated the protein expression of CYP2J2 in liver tissue. The results indicate that administration of crocetin markedly down-regulated the CYP2J2 expression to 0.80-fold compared to the rivaroxaban alone treatment (Figure C). Therefore, results indicate that crocetin treatment-mediated CYP2J2 inhibition occurred at the protein level, which modulates the pharmacokinetics of the CYP2J2 substrate in the mouse model.

Crocetin Remarkably Blocked EET Formation

After successfully establishing crocetin as a CYP2J2 inhibitor using in vitro, in silico, and in vivo approaches, we further checked the ability of crocetin to impede EET generation by employing the CYP2J2 inhibitor. CYP2J2 enzyme is known to catalyze the metabolism of arachidonic acid into all the four regioisomeric EET, which are well-known for their oncogenic properties. The epoxidation process is most likely to occur at position 14,15, resulting in 37% of the total EET products. This is followed by 18% for 11,12-EET, 24% for 8,9-EET, and 21% for 5,6-EET. Thus, to validate the role of the CYP2J2 inhibitor in blocking EET production, we evaluated the generation of EET from arachidonic acid in the presence of crocetin using HLM at the in vitro level first. In addition to the test compound, the standard inhibitors, terfenadone and danazol, were also evaluated simultaneously (Figure A,B).

7.

Effect of terfenadone (A); danazol (B); and crocetin (C) on the formation of EET from arachidonic acid in HLM. Data are represented as mean ± SEM (n = 3). Data are compared: terfenadone/danazol/crocetin (0 μM) vs terfenadone/danazol/crocetin (1 to 100 μM). p < 0.05/0.01/0.001 (*/**/***) denotes statistically significant.

As the concentration of terfenadone increased from 0 to 100 μM, the percentage of EET generation decreased from 100 to 13% (14,15-EET), 100 to 17% (11,12-EET), 100 to 13% (8,9-EET), and 100 to 15% (5,6-EET). In the case of danazol (0 to 100 μM), the percentage of EET generation decreased from 100 to 35% (14,15-EET), 100 to 39% (11,12-EET), 100 to 34% (8,9-EET), and 100 to 53% (5,6-EET). Similarly, the increase in crocetin concentration (ranging from 0 to 100 μM) resulted in a corresponding decline in the percentage of EET generation, with values decreasing from 100 to 64% (14,15-EET), 100 to 62% (11,12-EET), 100 to 62% (8,9-EET), and 100 to 63% (5,6-EET). Results indicate that the conversion of arachidonic acid to EET was impeded in the presence of crocetin and this effect is more pronounced with its increasing concentration (Figure C). Consequently, results infer that crocetin as a CYP2J2 inhibitor can hinder EET formation from arachidonic acid at the in vitro level.

Crocetin Lowered the Tumor Burden in Mouse Model of Breast Cancer by Suppressing CYP2J2 Expression and EET Level

After investigating the potential of crocetin in preventing the generation of EET at the in vitro level, we were curious to observe its action in the in vivo setting. To accomplish this, we used a mouse mammary carcinoma model employing 4T1 cells since the rapid proliferation capabilities of these cells in BALB/c mice mimic the progression and spread of human breast cancer, thus conferring an extensively employed rodent research model to study the spontaneous metastasis of breast cancer. Figure A displays the tumor images from the tumor control group and the test group subjected to crocetin treatment (50 mg/kg) for seven consecutive days. A significant reduction (76%) in the tumor size (Figure B) and tumor volume (83%) (Figure C) was observed upon the treatment with crocetin in comparison to the tumor control group.

8.

Representative tumor images (A), changes in tumor weight (B), and changes in tumor volume (C) of various study groups; WB analysis and densitometry data for CYP2J2 protein expression in the tumor tissue (D) and in the liver tissue (E) of the various study groups; and relative EET level in the tumor tissue of the various study groups (F). Data are represented as the mean ± SEM (n = 5). β-actin expression: endogenous loading control. Statistical significance level: p < 0.05/0.01/0.001 (*/**/***).

To ascertain the involvement of CYP2J2/EET axis toward its antitumor efficacy, we further explored the expression of CYP2J2 in the tumor tissue using Western blot analysis. CYP2J2 has been shown to overexpress in breast cancer and is known to promote tumor cell growth as well as cell proliferation by abrogating apoptosis. , Our findings from the present study suggest that administration of crocetin notably down-regulated the intratumoral CYP2J2 protein expression (0.21-fold) in comparison to the tumor control group (Figure D). In addition, the protein expression of hepatic CYP2J2 was also assessed, and downregulation (0.23-fold) was observed compared to that of the tumor control group (Figure E).

Following the assessment of CYP2J2 protein expression, our objective was to measure the level of EET in the tumor tissue to correlate with the CYP2J2 expression, as the role of CYP2J2-linked EET in breast cancer progression is not well understood to date. Furthermore, extensive research is currently being conducted to monitor the levels of EET to establish prognostic biomarkers for the diagnosis of breast cancer. Gratifyingly, crocetin administration significantly decreased the 14,15-EET formation to 0.69-fold, 11,12-EET formation to 0.74-fold, 8,9-EET formation to 0.74-fold, and 5,6-EET formation to 0.73-fold (Figure F). These results unequivocally indicate that the antitumor efficacy of crocetin could be attributed to its inhibition potential toward CYP2J2 and consequent EET generation. As EET is involved explicitly in tumorigenesis, this particular investigation contributes to advancing knowledge toward identifying targeted breast cancer therapeutics.

The role of EET is quite clear, but the tentative signaling pathway involved in targeting the CYP2J2/EET axis is relatively unexplored to date. Thus, we conducted untargeted proteomics analysis to identify the significantly dysregulated proteins in the crocetin-treated tumor tissues compared to those in untreated tumor tissues. Among the 1275 proteins that were evaluated, 174 proteins were up-regulated and 121 proteins were down-regulated considering the change of 2.0-fold (Log₂ fold change of ≥ +1.0 for upregulation & ≤ −1.0 for downregulation) and p-value <0.05 (−Log₁₀ p-value >1.3) (Figure A, Table S1).

9.

Volcano plot representing the dysregulated proteins in the tumor tissue upon treatment of crocetin (crocetin treated vs untreated) to compute the fold change in the protein expressions along with its statistical significance. Log₂ fold change of ≥ +1.0 for upregulation & ≤ −1.0 for downregulation and p-value <0.05 (−Log₁₀ p-value >1.3). Functional profiling of the up-regulated genes (B) and down-regulated genes (C) through gene ontology enrichment analysis.

The functional analysis was performed using the g:Profiler tool. After gene ontology (GO) enrichment analysis, the significant pathways/processes were associated with the up-regulated and down-regulated proteins are presented (Figures B,C and S4–S6). The detected significant molecular pathways were related to molecular functions (MF), biological process (BP) & cellular component (CC). We explored the KEGG resource, which is a widely used database elucidating signaling pathways for biological systems (Table S2).

Hypoxia is known to promote breast cancer cell proliferation, and the observed antitumor activity of crocetin is closely related to the observed downregulation of glycolysis/gluconeogenesis, oxidative phosphorylation, and TCA cycle (↑Pgk2 Pgk-2). − It can be correlated to the ROS scavenging activity of crocetin, which is reported in the literature using cancer cell lines. Additionally, upregulation of PPAR signaling pathway by crocetin treatment (↑Alox15 Alox12l) shows a similar line of action as PPARγ agonists are under investigation against breast cancer. , Induction of apoptosis is a key phenomenon of an anticancer agent, and data suggest upregulation of apoptosis upon crocetin treatment (↓Serpina1b Aat2 Dom2 Spi1–2, ↓Serpina1c, ↓Rps18 ps6 Gm10260 Gm17352, ↓Ttn, and ↓Rpl8). A similar line of crocetin’s effect on apoptosis is previously reported using in vitro/in vivo platform. , Fatty acids are known vital mediators for cancer metastasis due to their role in the remodeling of the tumor microenvironment. Current data indicate that treatment with crocetin caused upregulation of fatty acid metabolism and fatty acid degradation linked to alleviating metastasis. To be considered as a therapeutic strategy against cancer, research is ongoing by targeting fatty acid metabolism. , Based on the present finding, targeted proteomics with due validation of identified proteins by Western blotting/immunohistochemistry and additional metabolomics data can be useful toward successfully validating the molecular signaling pathway of crocetin. In this context, literature reported crocetin’s mode of action, such as apoptosis induction, EGFR inhibition, etc., should be prioritized, and additional mechanisms like ROS scavenging should be kept in mind in establishing a putative pathway for anticancer action of crocetin. ,− In vivo (preclinical/clinical) data for in vitro-identified CYP2J2 inhibitors is scarce. Still, data availability will be highly beneficial in prompting the establishment of tentative pathways involving CYP2J2 inhibition. Apart from its immense potential for druggability and therapeutic action, its toxicological aspects should be scrutinized before moving forward in developing such a promising anticancer candidate based on CYP2J2 inhibition.

Conclusions

In conclusion, the following inference can be derived based on the present investigations: (a) crocetin is a reversible inhibitor with a mixed-type of inhibition toward human CYP2J2; (b) it can boost the plasma exposure of rivaroxaban (CYP2J2 substrate); (c) it has substantial potential to hinder the formation of CYP2J2-mediated EET from arachidonic acid; (d) it showcases its antitumor efficacy in mouse mammary carcinoma model involving inhibition of CYP2J2 expression and impediment of EET level; and (e) it can regulate various key BPs to attenuate tumorigenesis. Crocetin is found to be a specific CYP2J2 inhibitor and it is highly useful not only as a therapeutic candidate but also can act as a probe to unravel signaling pathways. Considering the significant potential of crocetin against the CYP2J2/EET axis, identification of the significantly dysregulated proteins and their involvement in the database-based driven signaling pathway linked to angiogenesis, apoptosis, and metastasis could reinforce elucidating the molecular mechanism of CYP2J2/EET axis for targeted breast cancer therapy.

Experimental Section

Materials

Astemizole (purity ≥98%), rivaroxaban-d₄ (purity ≥99%), danazol (purity ≥98%), (±)­14,15-EET (purity ≥98%), (±)­11,12-EET (purity ≥98%), (±)­8,9-EET (purity ≥98%), (±)­5,6-EET (purity ≥98%), (±)­14,15-EET-d ₁₁ (purity ≥99%), and arachidonic acid (peroxide free) (purity ≥98%) were purchased from Cayman Chemical. Terfenadone and O-desmethyl astemizole were procured from Toronto Research Chemicals. Rivaroxaban was received as a gift sample from Mylan Laboratories Ltd. (Bollaram, India). Mebendazole (purity ≥98%) and ritonavir (purity ≥99%) were purchased from TCI and Indian Pharmacopoeia Commission, respectively. NADPH tetrasodium salt (purity ≥98%), MgCl₂ hexahydrate, phosphate buffer solution, ammonium formate (MS-grade), and ammonium acetate (MS-grade) were procured from Sigma-Aldrich. HLM (lot no. #PL050F–C) pooled from 50 human donors and MLM (lot no. #2513990-A) pooled from 100 BALB/c mice donors were obtained from Gibco. CYP2J2 pAb (lot no. CP1D10A) was purchased from MyBioSource. Antirabbit IgG HRP-linked Ab (catalog no. 7074S) was obtained from Cell Signaling Technology. Acetonitrile, methanol, and formic acid of MS-grade were purchased from Thermo Fisher Scientific. All other experimental materials were of bioreagent grade or above. Ultrapure and distilled water (make: Merck-Millipore; model: Direct-8) were used for analysis and animals, respectively.

Test Article

In order to use in the current experiments, crocetin was synthesized from commercially available crocin (TCI, Product no. C1527), , which was then purified and characterized using ¹H NMR, ¹³C NMR, and HRMS. The purity of crocetin (>96%) was determined by HPLC (Figure S7).

CYP2J2 Inhibition Potential of Crocetin in HLM (In Vitro)

Estimation of Kinetic Parameters

CYP2J2-catalyzed astemizole O-demethylation was used as an index reaction in this investigation. , For this reaction, the stock solution and further dilutions of astemizole were prepared by using DMSO and phosphate buffer (50 mM, pH 7.4), respectively. The stock solution of the metabolite, O-desmethyl astemizole, was prepared in DMSO and diluted in methanol to prepare the standard solutions. The reaction mixture (100 μL) was comprised of phosphate buffer (50 mM, pH 7.4), MgCl₂ (3.3 mM), astemizole (0.1 to 100 μM), microsomal protein, i.e., HLM (0.1 mg/mL), and NADPH (1.2 mM). The reaction commenced with the addition of NADPH and thereafter was incubated for a duration of 15 min at a temperature of 37 °C in a shaking water bath (make: New Brunswich Scientific; model: CLASSIC C76). Following incubation, the reaction was quenched by the addition of acetonitrile (100 μL) containing mebendazole (500 ng/mL), which served as an internal standard (IS). Samples were vortex-mixed, centrifuged (14,000 rpm, 10 min), decanted into inner vials, and quantified by LC–MS/MS (make: Thermo Fisher Scientific; model: TSQ Endura). The metabolite was measured in the reaction mixture using a matrix-match protein standard by LC-MS/MS in the present and subsequent in vitro experiments (Table S3). Reactions were performed in triplicate. The concentration of organic content in the reaction was maintained at ≤1% (v/v).

Determination of IC₅₀

The CYP2J2 inhibitory activity of crocetin was evaluated using the above-mentioned optimized reaction conditions, except for the concentration of astemizole, which was now employed at around the experimental K _m value. Terfenadone and danazol, known as CYP2J2 inhibitors, were also examined for their inhibitory activity as a standard inhibitor in addition to crocetin. The stock solution of crocetin, test compound, and standard inhibitors was prepared in DMSO before being diluted in methanol. The experiment was conducted in triplicate, and the reaction without the standard/test inhibitor was served as a control.

Assessment of the IC₅₀ Shift

The time-dependent inhibitory potential of crocetin for CYP2J2 was evaluated. The composition and the incubation conditions of the reaction mixture were consistent with the aforementioned description. Concisely, the test candidate was subjected to preincubation with HLM in the presence of NADPH for 0 and 30 min. Afterward, the CYP2J2 substrate was added to the preincubated reaction mixture and further incubated, processed, and analyzed, as mentioned above. Ritonavir was included as a standard inhibitor in this study. The reaction was carried out in triplicate.

Evaluation of the Mechanism of Inhibition

In order to assess the inhibitory mechanism of crocetin for CYP2J2 and determine its K _i, the selected concentration levels were 0.5 times of K _m, K _m value itself, 2 times of K _m, and 4 times of K _m for probe substrate (0.5 to 4 μM) and 0, 0.5 times the IC₅₀, the IC₅₀ value, 2 times the IC₅₀, and 4 times the IC₅₀ for test candidate (0 to 54 μM). Reaction conditions and sample analysis were the same as those mentioned above. Terfenadone and danazol were employed as standard inhibitors. The study was carried out in triplicate. Ultimately, acquired data were utilized to generate the Lineweaver–Burk and Dixon plots.

Data Analysis

The GraphPad Prism 5.0 software was employed to evaluate each parameter. The V _max and K _m values were determined by using the data of the metabolite generated during the reaction. These parameters were calculated by fitting the rate of metabolite formation at varying substrate concentrations to the Michaelis–Menten equation by using nonlinear regression analysis. The IC₅₀ of standard/test inhibitor was estimated using data on the % control activity exhibited by CYP2J2 upon inhibition at a log of varied concentrations of the inhibitor. The % control activity was calculated based on the metabolite generated in the control samples compared to the metabolite generated in the presence of a standard/test inhibitor. The fold-shift in IC₅₀ values was computed as the ratio between IC₅₀ obtained in the preincubation group for 0 min and IC₅₀ obtained in the preincubation group for 30 min. To determine the K _i and mechanism of CYP inhibition by standard/test inhibitor, data pertaining to the rate of metabolite generation at the various concentrations of substrate and inhibitor were fitted to the various equations of enzyme inhibition models.

Metabolic Depletion of CYP2J2 Substrates in the Presence of Crocetin Using Species-Specific Microsomes (In Vitro)

Astemizole as a CYP2J2 Substrate

Metabolic depletion of astemizole in the presence of crocetin was conducted using a substrate depletion approach in species-specific microsomes like HLM and MLM. Briefly, 100 μL of reaction mixture consisted of phosphate buffer (100 mM, pH 7.4), MgCl₂ (3.3 mM), astemizole (1 μM), microsomal protein, i.e., HLM/MLM (0.5 mg/mL), and crocetin (0, 1, 5, 10, 50, and 100 μM). The reaction was commenced by the addition of NADPH (1.2 mM) and terminated by the addition of acetonitrile after an incubation time of 0 and 30 min, followed by vortex mixing (2 min) and centrifugation (14,000 rpm, 10 min). Then, the supernatant was taken out and transferred into vials to analyze astemizole. Samples were analyzed by LC-MS/MS (Table S4). Study was carried out in triplicate. A reaction mixture deprived of NADPH was considered a negative control. Astemizole quantified in the samples incubated for 0 min was considered 100%. Samples after 30 min of incubation in the presence and absence of crocetin were compared with samples after 0 min to assess the percent remaining of astemizole and determine the effect of crocetin on astemizole depletion in various microsomes.

Rivaroxaban as a CYP2J2 Substrate

Metabolic depletion of rivaroxaban in the presence of crocetin was conducted using the same methodology described above for astemizole. Samples were analyzed by LC-MS/MS (Table S5).

Molecular Interactions between Crocetin and Human CYP2J2 (In Silico)

To further understand the interaction pattern of crocetin with the CYP2J2 enzyme, we performed molecular docking studies. There were no experimental structures available in the Protein Data Bank, so we retrieved the predicted protein structure from the AlphaFold protein structure database with the UniProt: P51589, source organism: Homo sapiens (human) and AlphaFold id: AF–P51589-F1-model_v2. , AlphaFold is an artificial intelligence/AI system established by DeepMind that predicts a protein’s 3D structure from its amino acid sequence. , To confirm the reliability of the predicted structure, we performed comprehensive validation of the structure. As part of this, we generated a Ramachandran plot using MolProbity online server (Available from http://molprobity.biochem.duke.edu/), which showed that 98.2% (491/500) of all residues were in favored (98%) regions and 100.0% (500/500) of all residues were in allowed (>99.8%) regions, with no significant outliers in the active site region (Figure S8). This indicates a high-quality backbone geometry and supports the structural plausibility of the model. The molecular docking study was implemented using the Biovia Discovery Studio client 4.1 platform following the protocol discussed by Robertson and co-workers and Kumar et al. − We have predicted the multiple active sites at the surface of the enzyme using the Biovia discovery studio 4.1 client platform from the “define and edit binding site” using the module “generate active site from receptor cavities” and docked the ligand in each site to identify the favorable binding site (identified most favorable active site coordinate X: 5.499, Y: 0.591, Z: −3.317, the radius of the sphere: 22.900). To prepare ligands, the molecules were run through the Discovery Studio platform’s “small-molecule module”, where several ligand conformers were generated. Each of these generated conformers was subsequently employed in the CDOCKER module for molecular docking using a CHARMm-based molecular dynamic scheme. , The CDOCKER interaction energy parameter (kcal/mol) was examined for all receptor–ligand complexes, and the highest-scoring (more negative; hence favorable to binding) poses with only noncovalent interactions (hydrogen bonds (conventional and carbon), ionic bonds, hydrophobic interactions, etc.) were kept for future investigation. Moreover, to strengthen our findings, we have performed an in silico mutagenesis analysis in which GLN228 and ARG138 were individually mutated to alanine (GLN228 → ALA228 and ARG138 → ALA138), followed by redocking and comparative binding energy estimation. This is a standard method used in alanine scanning mutagenesis, where alanine is substituted to test the importance of a residue’s side chain in molecular interactions since alanine is small and uncharged.

Thereafter, MD simulation was conducted using GROMACS version 2024.4 to evaluate the dynamic stability and interaction profile of crocetin in the complex with the target protein. The CHARMM36-jul2022 all-atom force field was employed for the protein, while ligand topologies compatible with the CHARMM force field were generated via the SwissParam online tool (http://www.swissparam.ch/). , The protein–ligand complex was solvated in a triclinic box filled with the CHARMM-modified TIP3P water model, maintaining a minimum distance of 1.0 nm between the solute and the box boundaries. Appropriate counterions (Na⁺ and Cl^–) were added to neutralize the system. Energy minimization was performed using the steepest descent algorithm to remove steric clashes. Subsequently, two equilibration steps were carried out: a 100 ps NVT (constant number of particles, volume, and temperature) phase using the V-rescale (modified Berendsen) thermostat at 300 K, followed by a 100 ps NPT (constant number of particles, pressure, and temperature) phase using the Parrinello–Rahman barostat at 1 bar. During both equilibration phases, positional restraints were applied to the heavy atoms of the protein–ligand complex to maintain the structural integrity. The production MD simulation was then executed for 100 ns under periodic boundary conditions using the leapfrog integrator with a 2 fs time step. Long-range electrostatic interactions were treated using the particle mesh Ewald (PME) method 27, and all bond lengths were constrained using the LINCS algorithm. Trajectories were saved every 10 ps for subsequent analysis. Postsimulation evaluation of the structural and interaction stability was performed using several metrics, including RMSD, RMSF, Rg, and hydrogen bond analysis.

Pharmacokinetic Interaction of Crocetin with CYP2J2 Substrate (In Vivo)

Animal Model and Ethical Approval

The experiment was conducted using healthy female BALB/c mice (25 to 30 g of body weight). Animals were housed as groups in an IVC system and placed in a pathogen-free and regulated housing environment. The animals were provided with unrestricted access to a standardized pellet diet and water. The studies adhered to the protocols established by the ‘Committee for the Control and Supervision of Experiments on Animals/CCSEA’, a governmental body based in New Delhi, India. The experimental protocols were authorized by our ‘Institutional Animal Ethics Committee/IAEC’ under the approval number 316/82/2/2023.

Test Article Dose and Dose Formulation

The present investigation was executed using rivaroxaban as a CYP2J2 substrate. The dose of rivaroxaban was 4 mg/kg, which was selected by converting the prescribed maximum daily human dose to the equivalent mouse dose. Crocetin was used at a dose of 50 mg/kg based on the existing literature on its efficacy. − The composition of the solution used to prepare the dose of rivaroxaban was 5% DMSO + 40% PEG-400 + 55% water (v/v). An aqueous suspension containing sodium carboxymethylcellulose (0.1%, w/v) was employed as a vehicle to prepare the dose of crocetin. The dose was prepared freshly and administered at a rate of 10 mL/kg.

Study Design

To investigate the impact of crocetin on the in vivo pharmacokinetics of a CYP2J2 substrate, rivaroxaban was employed. The study was conducted on a total of 50 animals. Overnight-fasted (∼12 h) animals were randomly assigned to two study groups. Each study group was additionally divided into five subgroups for sparse sampling, with each subgroup consisting of five animals. Study groups were as follows: Group 1 received rivaroxaban alone, whereas Group 2 received rivaroxaban plus crocetin. Both rivaroxaban and crocetin were administered orally. In Group 2, crocetin was administered for six consecutive days, and on the seventh day, rivaroxaban was given 0.5 h after crocetin administration. Blood samples (∼110 μL) were collected at various time points (0.083, 0.25, 0.5, 0.75, 1, 2, 3, 4, 6, and 8 h) postdosing of rivaroxaban. The samples were obtained using the retro-orbital venipuncture and transferred into microcentrifuge tubes containing an anticoagulant. The samples were centrifuged (14,000 rpm, 10 min) to separate plasma (50 μL each) and kept in a deep freezer (−80 °C) until analysis. In parallel, the animals were euthanized by using carbon dioxide asphyxiation and sacrificed by cervical dislocation. After this, their livers were extracted, cleaned with normal saline solution, dried using tissue paper, and stored in a deep freezer.

Bioanalysis

The preparation of individual stock solutions (1 mg/mL) of rivaroxaban and crocetin was carried out in DMSO, followed by further dilutions in methanol to obtain the desired concentrations. Each plasma sample (50 μL) was processed with the addition of acetonitrile (200 μL) containing IS (rivaroxaban-d₄, 250 ng/mL), vortex-mixed, centrifuged (14,000 rpm, 10 min), decanted into the vial, and analyzed by LC-MS/MS (Table S5). The same plasma samples were then used for the quantification of crocetin (Table S6). The estimation of rivaroxaban and crocetin in mice plasma was done using the matrix-match calibration curve.

Pharmacokinetic Data Evaluation

The concentration data obtained at different time points were analyzed using PK Solutions software (Summit Research Services, USA) by employing a noncompartmental model to estimate the essential pharmacokinetic parameters.

Western Blotting for Protein Expression of CYP2J2

The protein expression of CYP2J2 in the liver tissue of the experimental animals was evaluated by using Western blot analysis. The standard immunoblotting methodology was employed for this purpose. The liver tissue homogenate was prepared by using an equal proportion of tumor tissue from each experimental animal in RIPA buffer containing a cocktail of protease and phosphatase inhibitors. Subsequently, the protein content was quantified using the Bradford method. The protein sample underwent electrophoresis using SDS-PAGE and then transferred onto PVDF membranes. Subsequently, the membranes were treated with a blocking solution containing bovine serum albumin (5% w/v). Finally, the membranes were exposed to specific primary antibodies for a duration of 4 h at room temperature. Following the washing step using a solution of Tris-buffered saline and Tween 20, the blots were subjected to probing using a secondary antibody that is species-specific and is coupled with HRP. Following a comprehensive washing procedure, the detection of immunoreactive protein bands was accomplished using the Western Bright ECL HRP substrate (Bio Rad, Hercules, USA) and subsequently detected with a Chemi-doc detection system (Syngene, Frederick, USA). Then, densitometry analysis was conducted using the ImageJ software.

Generation of EET from Arachidonic Acid in the Presence of Crocetin (In Vitro)

To investigate the suppressive impact of crocetin on the generation of EETs, arachidonic acid was used as a substrate, and metabolites [(±)­14,15-EET, (±)­11,12-EET, (±)­8,9-EET, and (±)­5,6-EET] were analyzed in the presence of crocetin using HLM following earlier reports. Arachidonic acid was supplied in an ethanol solution (250 mg/mL), and further dilutions were performed using phosphate buffer (100 mM, pH 7.4). EET was also supplied as an ethanol solution (100 μg/mL), and subsequent dilutions were done with methanol to prepare the standard solutions. In this study, the reaction mixture (100 μL) consisted of arachidonic acid (100 μM), microsomal protein, i.e., HLM (0.5 mg/mL), MgCl₂ (3.3 mM), phosphate buffer (100 mM, pH 7.4), and crocetin (0, 1, 5, 10, 50, and 100 μM). The reaction was initiated by the addition of NADPH (1.2 mM). After 30 min of incubation, the reaction was terminated by adding ethyl acetate containing IS [(±)­14,15-EET-d ₁₁, 300 ng/mL], followed by vortex-mixing (2 min) and centrifuged (14,000 rpm, 10 min). Then, the organic layer was separated, dried in a vacuum dryer, reconstituted with acetonitrile, and transferred into vials for the analysis of EET by LC-MS/MS. The study was carried out in triplicate. The samples in the absence of crocetin were considered a control sample with 100% formation of EET. Data from the samples in the presence of crocetin were compared with control samples in order to assess the inhibitory effect of crocetin on the generation of EET from arachidonic acid via CYP2J2.

Effect of Crocetin on CYP2J2 Expression & EET Level in Breast Cancer Mouse Model (In Vivo)

Cell Culture

The present study was executed by employing a 4T1 mouse mammary carcinoma cell line, which was graciously gifted by Dr. Avinash Bajaj from Regional Centre for Biotechnology (Faridabad, India). These cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum, 1% streptomycin, and penicillin-G. These were grown in culture flasks under controlled conditions of 37 °C in a 5% CO₂ incubator.

Animal Model and Ethical Approval

The experiment was performed in adult female BALB/c mice (5 to 7 weeks old; 25 to 30 g body weight). Animal maintenance and ethical guidelines (IAEC approval number: 339/83/8/2023) are as previously mentioned.

Test Article Dose and Dose Formulation

Crocetin was used at a dose of 50 mg/kg. The dose formulation and route of dose administration were the same as those in the aforementioned pharmacokinetic interaction studies.

Study Arm and Treatment Protocol

The experiment was conducted following the previously published protocol. , At first, mouse mammary cancer cells (1.5 × 10⁶ 4T1 in 200 μL of serum-free RPMI media) were injected subcutaneously into the mammary fat pad of each mouse, specifically around the right mammary gland, to develop a tumor. When the palpable tumor developed around 1 week of tumor initiation, animals were randomly divided into two groups containing five animals each. Group 1 served as the tumor control group, while Group 2 received a daily dose of crocetin (50 mg/kg) for 7 days. Animal’s body weight was documented daily. After completion of the treatment period, animals were humanely euthanized using carbon dioxide and subjected to cervical dislocation. Tumor and liver tissue were then extracted, rinsed with normal saline, dried with tissue paper, and weighed. Afterward, the tissues were rapidly frozen using liquid nitrogen and stored at −80 °C in a deep freezer.

Measurement of Tumor Weight and Tumor Volume

Tumor tissues were weighed using an analytical balance (make: Mettler Toledo; model: XS205). Tumor volume was calculated using the formula: 0.5 × length × width².

Assessment of CYP2J2 Protein Expression in the Tumor and Liver Tissue

Western blotting was performed to assess the protein expression of CYP2J2 in the tumor and liver tissue homogenate using the standard immunoblotting protocol described above.

Determination of EET Concentration in the Tumor Tissue

To assess the EET level, tumor tissue homogenate (350 mg/mL) was prepared at first using 0.2% formic acid (v/v) in methanol and water in the ratio of 80:20, processed with acetonitrile containing IS, vortex-mixed (2 min), centrifuged (14,000 rpm, 10 min), decanted into the inner vial, and analyzed by LC-MS/MS.

SWATH-Based Untargeted Proteomics

The proteomics study was carried out following earlier reported protocols. − First of all, the proteins from the tumor tissue were extracted after tissue homogenization using RIPA buffer solution and were subsequently precipitated by chilled acetone overnight (12 h) at −20 °C. The samples were then centrifuged (15,000g, 15 min). The protein pellets were resuspended in Tris-HCl with urea at pH 8.5. Protein level was estimated by the Bradford assay. After that, 10 μg of protein from each sample was taken and gradually diluted to bring the urea concentration below 1 M before proceeding with reduction, alkylation, and trypsin digestion. Reduction of the sulfide bond was performed by 25 mM dithiothreitol/DTT at 56 °C, followed by alkylation using 55 mM iodoacetamide/IAA at room temperature (in dark). These samples were then subjected to trypsin digestion for 16 h at 37 °C, and subsequently, the peptides were desalted by a Ziptip C₁₈ cartridge, and the peptides were dried in a vacuum concentrator. The tryptic peptides were then analyzed on a quadrupole-TOF hybrid mass spectrometer (make: Sciex, model: TripleTOF 6600) coupled to an Eksigent-LC system. Peptides were loaded on a trap column to perform online desalting. After that, peptides were separated on a reverse-phase C₁₈ analytical column in a 57 min long gradient using buffer A (water with 0.1% formic acid) and buffer B (acetonitrile with 0.1% formic acid). Data were acquired using Analyst version 1.7.1 software. Mass spectrometry data acquired via data-independent acquisition (DIA) in SWATH-MS mode, comprising.wiff and.wiff.scan files, were processed using Spectronaut software (Biognosys AG, Switzerland) utilizing the direct DIA pipeline. Protein identification was facilitated by searching against the Mus musculus reference proteome obtained from UniProtKB, which served as the spectral library. To ensure high-confidence identifications, protein and precursor identifications were filtered at a stringent 1% false discovery rate (FDR) threshold, determined using a target-decoy approach integrated within Spectronaut. The direct DIA analysis was conducted with default parameters incorporating advanced spectral prediction and adaptive extraction windows. To account for technical variability, cross-run normalization was applied globally across samples, employing the ‘Automatic’ normalization strategy recommended by Spectronaut. After identification and relative quantification, results were exported in the ‘Run Pivot Report’ format, yielding normalized protein abundance values across all samples.

Statistical Analysis

The statistical analysis for all other experiments, except proteomics study, was performed using an unpaired t test through an online t-test calculator (GraphPad Prism). A p-value below the thresholds of 0.05, 0.01, and 0.001 was deemed to indicate statistical significance. To account for intrinsic variability in proteomics data, we employed median-based normalization followed by log10 transformation to mitigate the effects of skewed intensity distributions. The following thresholds were applied to identify differentially expressed proteins in cases relative to controls: upregulation was defined as log2­(fold change) ≥1 and −log10 (p-value) >1.3, while downregulation was defined as log2­(fold change) ≤ −1 and −log10 (p-value) >1.3. Statistical analysis was performed using Microsoft Excel & R (version 4.3.1). Functional enrichment analysis was conducted using g: Profiler (version 2023), leveraging GO annotations to elucidate BPs associated with dysregulated proteins.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c03187.

• Molecular docking analysis and MD simulation data; pharmacokinetic data of crocetin in mouse model; GO enrichment analysis for dysregulated proteins in proteomics study; Ramachandran plot of selected CYP2J2 protein structure; chromatographic purity of crocetin by HPLC; proteomics-based up-regulated and down-regulated proteins/genes data; GO analysis using KEGG database; and LC-MS/MS conditions for quantification of O-desmethyl astemizole, astemizole, rivaroxaban, and crocetin (PDF)

D.M.: investigation, methodology, formal analysis, and writing-original draft; A.J.: investigation and methodology; G.K.: investigation; MD.Q.A.: investigation; N.A.: investigation; V.K.: investigation; P.K.O.: investigation; S.D.S.: supervision; S.B.: supervision and data analysis; A.G.: supervision; U.N.: conceptualization, validation, supervision, and writing-review and editing.

This research was supported by the Council of Scientific and Industrial Research, India, under the internal budget heads of CSIR-IIIM (MLP21006). UN is thankful to the Department of Science and Technology, India, for the required support from Bose Institute toward publication.

The authors declare no competing financial interest.