Pharmacokinetic Alteration of Paclitaxel by Ferulic Acid Derivative

P-glycoprotein (P-gp) is known to be involved in multidrug resistance (MDR) and modulation of pharmacokinetic (PK) profiles of substrate drugs. Here, we studied the effects of synthesized ferulic acid (FA) derivatives on P-gp function in vitro and examined PK alteration of paclitaxel (PTX), a well-known P-gp substrate drug by the derivative. Compound 5c, the FA derivative chosen as a significant P-gp inhibitor among eight FA candidates by in vitro results, increased PTX AUCinf as much as twofold versus the control by reducing PTX elimination in rats. These results suggest that FA derivative can increase PTX bioavailability by inhibiting P-gp existing in eliminating organs.


Introduction
P-glycoprotein (P-gp) is a membrane efflux transporter that is a member of the ATP-binding cassette (ABC) superfamily. This transporter is widely distributed in normal organs including the brain [1], liver [2], intestine [3] and kidney [4] to protect our body from exogenous compounds including toxins by extruding them out of the body organs. P-gp is associated with a multidrug resistance (MDR) phenomenon which makes many potent anticancer drugs ineffective and is one of the reasons responsible for the failure of cancer chemotherapy [5][6][7]. Therefore, P-gp inhibitors can be used to overcome MDR. Moreover, this concept is extended to improve the absorption of orally delivered medications by inhibiting P-gp located in the gastrointestinal (GI) tract [8][9][10].
Ferulic acid (FA) is one of the most ubiquitous phenolic compounds in plants, and the organic compound and its derivatives possess multiple physiological and pharmaceutical functions such as antioxidant, antimicrobial, anti-inflammatory, antithrombosis and anticancer activities [11]. In the anticancer effect, it has been reported that FA and its derivatives induce cell cycle arrest, autophagy, apoptosis and cytotoxicity in several cancer cells such as leukemia, breast cancer, cervical cancer, etc. [12][13][14]. However, the effect of FA and its derivatives on P-gp modulation have not been investigated well. Only a few studies have assessed recently the effect of FA on the reversal of the P-gp-mediated MDR in cancer [15,16]. Therefore, we synthesized several FA derivatives in the study and evaluated their function on P-gp in vitro and in vivo systems.

Methylation of Ferulic Acid
Ferulic acid 2 g (10.3 mmol) was dissolved in 40 mL of methanol and 1.2 mL of concentrated sulfuric acid (H 2 SO 4 ) was added, and heated under reflux overnight. Then, 5% NaHCO 3 was added to make alkaline solution (pH 10) and extracted with diethyl ether. The solution was dried over MgSO 4 and the solvent was removed in vacuo.

General Procedure for the Synthesis of Esters
To the methyl ferulate (0.5 g, 2.4 mmol) in dry acetone (6 mL) solution, pottassium carbonate (K 2 CO 3 ) (0.4 g, 2.9 mmol) was added; 1-Iodopropan (0.26 mL, 2.9 mmol) was added and the reaction mixture was heated at 60 • C for 3 h. After cooling, the reaction mixture was extracted with diethyl ether, dried over MgSO 4 and the solvent was removed in vacuo. Column chromatography (hexane: ethylacetate = 1:1) afforded the desired product.

General Procedure for Hydrolysis
The ester obtained from the procedure 2.2.2 (0.5 mmol) was dissolved in acetone (6 mL), and 2 N NaOH (6 mL) was added and the reaction mixture was heated to 90 • C for 2 h. The reaction mixture was extracted with diethyl ether. The water layer was acidified with 10% HCl solution (pH 2) and extracted with diethyl ether. The organic solution was dried over MgSO 4 and the solvent was removed in vacuo to obtain the desired product.

Cytotoxicity Studies in P-gp Overexpressed Cells
The effect of eight FA derivatives on cytotoxicity was studied in P-gp overexpressed human breast cancer cells (MCF-7/ADR) using the SRB assay [18]. The details of the cell culture condition and the assay method were presented in our previous reports [10,19]. Verapamil (VER, 100 µM), one of the P-gp inhibitors, was used as a positive control. The half maximal inhibitory concentration (IC 50 ) values were calculated with Table Curve2D ® version 5.01 software (Systat Software Inc., San Jose, CA, USA). The assay was performed in triplicate.

Human P-glycoprotein ATPase Activity Assay
The effects of compounds 5c, 5f, 5g and 5h on P-gp ATPase activity at different concentrations (20, 50 and 100 µM) were examined in human P-gp membranes using an ATPase assay kit according to the method reported previously [10,19]. VER was used as a P-gp inhibitor and an ATPase stimulator. The ATPase activities were expressed as the rate of phosphate release per milligram of membrane protein and converted to the relative ratio versus the control. This assay was performed in duplicate.

Pharmacokinetic Study
The pharmacokinetic (PK) study was performed using male Sprague-Dawley rats (6 weeks old and 200 g-235 g) commercially available from Orient Bio (Seongnam, Korea) [10,19]. All animal procedures were approved by the Institutional Animal Care and Use Committee of Ewha Womans University (No. 2012-01-019, approved on 3 April 2012), Republic of Korea.

Pharmacokinetic Analysis
The following PK parameters of PTX after oral single dosing to rats were estimated by non-compartmental analysis using WinNonlin ® Professional version 5.2 software (Pharsight Corporation, Mountain View, CA, USA): The area under the plasma concentration-time curve from 0 h to infinity (AUC inf ), elimination half-life (t 1/2 ), apparent volume of distribution after oral single dosing (Vz/F), oral clearance (Cl/F), maximum plasma concentration (C max ) and the time required to reach C max (T max ). The relative bioavailability (RB, %) of PTX was calculated with the following formula: Pharmaceutics 2019, 11, 593 5 of 11 AUC inf po control is the AUC inf obtained from oral single administration of PTX alone, and AUC inf po co-administration is the AUC inf obtained from oral single co-administration of PTX and compound 5c.

Data Analysis
Statistical analysis was conducted using Tukey and Dunnett T3 s tests in conjunction with a one-way analysis of variance (ANOVA) for accumulation, efflux, and PK studies, respectively. Free GraphPad Prism was used for statistical analyses (Version 8.00, La Jolla, San Diego, CA, USA). Mean data were presented with standard deviations (SD). A p value < 0.05 was considered statistically significant.

Synthesis of Ferulic Amides, FA Derivatives
FA derivatives were synthesized as shown in Scheme 1. Ferulic acid (1) was protected with methyl alcohol to give methyl ester (2). The ester was then reacted with potassium carbonate and different alkyl halides such as propyl iodide, isobutyl iodide or benzyl bromide to afford corresponding esters (3) and purified by column chromatography. Then, the protected methyl ester was hydrolyzed with 2 N sodium hydroxide to give the acids (4). The coupling of acid group with various amines using PyBOP as coupling agent resulted in the desired amides (5).
oral clearance (Cl/F), maximum plasma concentration (Cmax) and the time required to reach Cmax (Tmax). The relative bioavailability (RB, %) of PTX was calculated with the following formula:

% 100
AUCinf po control is the AUCinf obtained from oral single administration of PTX alone, and AUCinf po co-administration is the AUCinf obtained from oral single co-administration of PTX and compound 5c.

Data Analysis
Statistical analysis was conducted using Tukey and Dunnett T3′s tests in conjunction with a oneway analysis of variance (ANOVA) for accumulation, efflux, and PK studies, respectively. Free GraphPad Prism was used for statistical analyses (Version 8.00, La Jolla, San Diego, CA, USA). Mean data were presented with standard deviations (SD). A p value < 0.05 was considered statistically significant.

Synthesis of Ferulic Amides, FA Derivatives
FA derivatives were synthesized as shown in Scheme 1. Ferulic acid (1) was protected with methyl alcohol to give methyl ester (2). The ester was then reacted with potassium carbonate and different alkyl halides such as propyl iodide, isobutyl iodide or benzyl bromide to afford corresponding esters (3) and purified by column chromatography. Then, the protected methyl ester was hydrolyzed with 2 N sodium hydroxide to give the acids (4). The coupling of acid group with various amines using PyBOP as coupling agent resulted in the desired amides (5).

P-gp Inhibitory Effect of FA Derivatives in Vitro
VER (a positive control) and FA derivatives were shown to be not toxic to MCF-7/ADR cells at a concentration of 100 μM (Table S1). When eight FA derivatives (Figure 1, 100 μM) were treated with DNM in MCF-7/ADR cells, compounds 5c, 5f, 5g and 5h decreased the IC50 values of DNM to about 1/5 of negative control (DNM alone) value (Table 1). Even compounds 5c and 5h dramatically reduced the IC50 values of DNM (2.2 and 2.6 μM, respectively) which was lower than that of VERtreated group (3.1 μM).

BA Enhancing Effect of Compound 5c in Vivo
The in vitro results suggested that compound 5c was the most potent P-gp inhibitor. As shown in Table 3, co-administration of the lowest dose of compound 5c (0.5 mg/kg) did not alter any PK parameters of PTX, as compared to PO control (PTX alone). However, with higher doses of compound 5c (2 and 5 mg/kg), the AUC inf of PTX was increased (control; 773 ± 169 ng·h/mL vs. 5 mg/kg; 1456 ± 367 ng·h/mL, p < 0.05) and Cl/F of PTX was decreased approximately to one half (control; 8.22 ± 1.84 L/h vs. 2 mg/kg; 4.10 ± 1.53 L/h (p < 0.05) and 5 mg/kg; 4.53 ± 1.04 L/h (p < 0.05), respectively). The t 1/2 was extended over two-fold (control; 2.8 ± 0.86 h vs. 5 mg/kg; 6.8± 1.3 h, p < 0.01) ( Table 3). The drug exposure in the body was enhanced due to the reduction of oral clearance when it was co-administered with compound 5c, except for a dose of 0.5 mg/kg (Figure 3). On the other hand, the C max of PTX and Vz/F were not significantly changed in the presence of the compound. Table 3. Pharmacokinetic (PK) parameters of paclitaxel (PTX) after oral administration of PTX (25 mg/kg) alone, or oral co-administration of PTX (25 mg/kg) and compound 5c (0.5, 2 and 5 mg/kg, respectively) to rats.

Discussion
FA enhanced the cytotoxicity of PTX and doxorubicin in human nasopharyngeal cancer cells by inhibiting the P-gp function [15,16]. FA enhanced cellular accumulation of calcein-acetoxymethyl ester and rhodamine123, well-known P-gp substrates, in P-gp overexpressed cells in a dosedependent manner [15]. In silico study, FA showed hydrogen bonding and hydrophobic interaction with transmembrane domain of P-gp [15]. In our study, newly synthesized FA derivatives including compound 5c increased the cytotoxicity and cellular accumulation of DNM in P-gp overexpressed breast cancer cells. Moreover, some parts of structure in the FA and its derivatives showed a similarity with VER, a well-known P-gp inhibitor. Because of the similar parts of the structure, FA derivatives might show the effect on the reversal of the P-gp-mediated MDR in cancer cells.
In addition, the dose-dependent decrease in ATPase activity of VER was also observed in our previous study [10]. However, all the values were over 2 (as a ratio to blank), suggesting ATPase stimulating effect of VER. In our previous studies, the lower dose (50 μΜ) of VER showed less effects on DNM accumulation and efflux [8] than the higher dose (100 μΜ) of VER [10,19]. Therefore, we used 100 μΜ VER and FA derivatives to observe P-gp inhibitory activity as much as possible in in vitro studies. According to the results of ATPase activity assay, four FA derivatives may inhibit P-gp activity by depleting ATP as an ATPase stimulator because a ratio to blank was over 2.
FA derivative, compound 5c chosen as the most potent P-gp inhibitor among the eight synthesized FA derivatives from in vitro studies increased PTX BA by reducing oral PTX clearance  Table 3. Pharmacokinetic (PK) parameters of paclitaxel (PTX) after oral administration of PTX (25 mg/kg) alone, or oral co-administration of PTX (25 mg/kg) and compound 5c (0.5, 2 and 5 mg/kg, respectively) to rats.

Discussion
FA enhanced the cytotoxicity of PTX and doxorubicin in human nasopharyngeal cancer cells by inhibiting the P-gp function [15,16]. FA enhanced cellular accumulation of calcein-acetoxymethyl ester and rhodamine123, well-known P-gp substrates, in P-gp overexpressed cells in a dosedependent manner [15]. In silico study, FA showed hydrogen bonding and hydrophobic interaction with transmembrane domain of P-gp [15]. In our study, newly synthesized FA derivatives including compound 5c increased the cytotoxicity and cellular accumulation of DNM in P-gp overexpressed breast cancer cells. Moreover, some parts of structure in the FA and its derivatives showed a similarity with VER, a well-known P-gp inhibitor. Because of the similar parts of the structure, FA derivatives might show the effect on the reversal of the P-gp-mediated MDR in cancer cells.
In addition, the dose-dependent decrease in ATPase activity of VER was also observed in our previous study [10]. However, all the values were over 2 (as a ratio to blank), suggesting ATPase stimulating effect of VER. In our previous studies, the lower dose (50 μΜ) of VER showed less effects on DNM accumulation and efflux [8] than the higher dose (100 μΜ) of VER [10,19]. Therefore, we used 100 μΜ VER and FA derivatives to observe P-gp inhibitory activity as much as possible in in vitro studies. According to the results of ATPase activity assay, four FA derivatives may inhibit P-gp activity by depleting ATP as an ATPase stimulator because a ratio to blank was over 2.
FA derivative, compound 5c chosen as the most potent P-gp inhibitor among the eight synthesized FA derivatives from in vitro studies increased PTX BA by reducing oral PTX clearance ) 0.5, ( ) 2 or (X) 5 mg/kg.

Discussion
FA enhanced the cytotoxicity of PTX and doxorubicin in human nasopharyngeal cancer cells by inhibiting the P-gp function [15,16]. FA enhanced cellular accumulation of calcein-acetoxymethyl ester and rhodamine123, well-known P-gp substrates, in P-gp overexpressed cells in a dose-dependent manner [15]. In silico study, FA showed hydrogen bonding and hydrophobic interaction with transmembrane domain of P-gp [15]. In our study, newly synthesized FA derivatives including compound 5c increased the cytotoxicity and cellular accumulation of DNM in P-gp overexpressed breast cancer cells. Moreover, some parts of structure in the FA and its derivatives showed a similarity with VER, a well-known P-gp inhibitor. Because of the similar parts of the structure, FA derivatives might show the effect on the reversal of the P-gp-mediated MDR in cancer cells.
In addition, the dose-dependent decrease in ATPase activity of VER was also observed in our previous study [10]. However, all the values were over 2 (as a ratio to blank), suggesting ATPase stimulating effect of VER. In our previous studies, the lower dose (50 µM) of VER showed less effects on DNM accumulation and efflux [8] than the higher dose (100 µM) of VER [10,19]. Therefore, we used 100 µM VER and FA derivatives to observe P-gp inhibitory activity as much as possible in in vitro studies. According to the results of ATPase activity assay, four FA derivatives may inhibit P-gp activity by depleting ATP as an ATPase stimulator because a ratio to blank was over 2.
FA derivative, compound 5c chosen as the most potent P-gp inhibitor among the eight synthesized FA derivatives from in vitro studies increased PTX BA by reducing oral PTX clearance in vivo. While C max of PTX was not significantly altered in the presence of compound 5c, the t 1/2 of PTX was significantly extended when PTX was co-administered with 2 or 5 mg/kg of the FA derivative ( Table 3). The increased AUC inf of the drug observed in the higher dose groups (2 and 5 mg/kg) was a result of the decreased PTX elimination. In addition, the reduced variation was monitored in the plasma PTX concentration-time profiles of the higher dose groups (2 and 5 mg/kg), as compared with control or low dose group (0.5 mg/kg).
Previously we reported other novel P-gp inhibitors such as coumarin derivative [8], phenylbutenoid dimer [9] and xanthone analogue [10]. They and VER [8] enhanced BA (9.6-fold, 1.8-fold and 2.5-fold and 3.0-fold at 5 mg/kg, respectively) and C max of PTX when they were co-administered with P-gp substrate anticancer drug, PTX. It means that these inhibitors worked at drug absorption phase by modulating the intestinal P-gp function. On the other hand, the FA derivative in the present study appeared to modulate P-gp function in the elimination phase, unlike other P-gp inhibitors reported previously [8][9][10].
Our study showed the potential P-gp modulating effect of FA derivative, compound 5c, on PTX disposition. Indeed, P-gp efflux transporters are expressed in the hepatic canalicular membrane for biliary excretion and in the proximal tubule of the kidney for renal excretion [2,4]. In addition, PTX has been reported to be eliminated by hepatic metabolism by mainly CYP2C8 [21], renal excretion [22] and biliary excretion [23]. It has been reported that FA is quickly absorbed in the intestine within 5-15 min and 30 min in rats and humans, respectively [24], and is mainly excreted through kidney [25]. Therefore, the interaction between FA derivative, absorbed at GI tract, and PTX on P-gp efflux transporters might occur during the elimination process. There is a lack of knowledge about the effect of FA derivatives and/or their metabolites on the hepatic metabolism and biliary excretion, and it needs to be clarified.
According to the reports about the relationship between FA and renal function in rodents, FA gives the positive effect on the kidney injury [26][27][28]. In lipopolysaccharide (LPS)-induced acute kidney injured mice, the phenolic compound (50 mg/kg and 100 mg/kg, intraperitoneal (IP) injection, 1 h before and 2 h after a single IP injection of LPS) reduces apoptosis and inflammation but increases adenosine generation [27], and in diabetes-induced rats, FA (50 mg/kg, orally for 8-week) protects hyperglycemia-induced kidney damage by antioxidation, anti-inflammation and autophagy [28]. In the present study, a single dose of FA derivative (0.5, 2 or 5 mg/kg) may not affect renal function in healthy rats to reduce PTX elimination. However, further studies are required to examine the effect of FA derivatives on renal function and their toxicities in the body following single or multiple dosing.

Conclusions
Four FA derivatives exhibited P-gp inhibitory function among eight synthesized compounds, and compound 5c was selected as the most potent P-gp inhibitor candidate in in vitro investigation. This compound significantly improved the BA of PTX by approximately twofold at the dose of 5 mg/kg, when it was co-administered with PTX in rats. Therefore, the co-administration of compound 5c could offer a therapeutic benefit in oral administration of P-gp substrate drugs by reducing drug elimination. Further studies are required to examine the effect of compound 5c on PTX metabolism to clarify the mechanism for the decreased PTX elimination caused by the FA derivative and to investigate the effect of the FA derivative, compound 5c on renal function.