In vitro and in situ effects of atorvastatin and ezetimibe on P-glycoprotein expression and function

P-glycoprotein (P-gp) is a membrane transporter responsible for the active efflux from the cell. Inhibition of the activity may lead to clinically significant drug-drug interactions. This study was performed to investigate the effects of atorvastatin and ezetimibe on the function and expression of P-gp. The in vitro rhodamine-123 (Rho123) efflux assay and Western blot in Caco-2 cells, and the in situ rat single-pass intestinal permeability model followed by high performance liquid chromatography were developed. Rho123 intracellular accumulation in 100 µM of atorvastatin- and ezetimibe-treated cells was significantly higher than that in control cells (p<0.05). P-gp expression was decreased by 100 µM atorvastatin and ezetimibe. Intestinal effective permeability of digoxin in the presence of atorvastatin (3 and 100 µM), ezetimibe (10 and 100 µM) was significantly increased (p<0.05). Both drugs  inhibited P-gp activity in vitro and in situ . Atorvastatin and ezetimibe down-regulated the expression of P-gp in vitro .  Video Clip of Methodology : Single-pass intestinal permeability: 17 min 26 sec Full Screen Alternative


Introduction
In recent years, the potential for significant clinical drug -drug interactions (DDIs) has emerged owing to the growing use of drugs and herbal medicines. Most drugs are administered orally and must, therefore, be adequately and consistently absorbed to achieve successful therapy. Membrane transporter is an important risk factor for developing DDI. The active efflux of drugs from enterocytes back into the intestinal lumen is one of the major barriers to oral drug delivery. P-glycoprotein (P-gp), an ATP binding cassette protein, is a membrane transporter responsible for the active efflux and the secretion of passively diffused drugs and xenobiotics from the cell. It is expressed on several barrier epithelia, including the small intestine, bloodbrain barrier, liver, kidneys, and bile canaliculi. This multidrug transporter has broad substrate specificity, with a tendency towards lipophilic, cationic com-pounds. Thus, P-gp plays critical role in the absorption and excretion of drugs, and the distribution of drugs to various physiological spaces and target tissues.
The transport activity of P-gp can be inhibited in vivo by a variety of drugs, such as verapamil, ketoconazole, and cyclosporine, which may affect the absorption of the drugs themselves and that of concomitantly used drugs.
Competitive or non-competitive inhibition of P-gp by drugs or other xenobiotics may lead to pronounced inter-individual variability of its transport activity, resulting in clinically significant DDIs. DDIs affect the drugs' bioavailability and effectiveness and may cause drug toxicity and unwanted adverse effects. Therefore, owing to the clinical importance of P-gp related DDIs, it is critical to identify drugs that are inhibitors of P-gp so that adverse DDIs might be minimized or avoided (Ando et al., 2015;Li et al., 2014;O'Connor et al., 2015).
Atorvastatin, an effective hypolipidemic synthetic drug, is one of the most widely prescribed drugs and the most widely prescribed statin worldwide. Atorvastatin is an inhibitor of 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMG CoA), which catalyzes the conversion of HMG CoA to mevalonate, an early rate limiting step in the cholesterol biosynthesis pathway. By inhibiting this enzyme, atorvastatin reduces hepatocyte cholesterol levels, resulting in up-regulation of lowdensity lipoprotein (LDL) receptors and, consequently, increased clearance of LDL-cholesterol (LDL-C) from the blood plasma. It reduces total cholesterol, LDL, and triglyceride (TG) while elevating the blood level of high -density lipoprotein (HDL). It shows poor oral bioavaila -bility (12% in humans and 6% in rats), poor solubility, and extensive hepatic first-pass biotransformation, preventing its effectiveness in most chemotherapeutic applications (Adams et al., 2012;Horwich and Mc-Lellan, 2007;Rodde et al., 2014;Vatsa et al., 2012).
is widely used as a lipid lowering agent and, unlike statins, inhibits intestinal and biliary cholesterol and phyto-sterol absorption by selectively blocking the Niemann-Pick C 1-like 1 protein (NPC1L1) in the brush border of jejunum and, to a lesser extent, in the duodenum and the ileum. It can significantly lower LDL-C and non-high-density lipoprotein cholesterol (non-HDL-C) when used alone or in combination with statin therapy. Ezetimibe prevents the NPC1L1/sterol complex from interacting with AP2 in clathrin-coated vesicles. It may change the shape of NPC1L1 to render it incapable of binding to sterols or interfere with the binding of free cholesterol to the cell membrane. After oral administration, it is rapidly absorbed from the gut lumen and conjugated by UDP-glucuronosyl-transferases (Bandyopadhyay et al., 2012;Lioudaki et al., 2011;Oswald et al., 2006a;Oswald et al., 2006b;Oswald et al., 2007;Phan et al., 2012;Suchy et al., 2011).
The purpose of this study was to investigate the in vitro and in situ effects of atorvastatin and ezetimibe on the membrane transporter function of P-gp. The effects of the drugs on P-gp expression in Caco-2 cells were also investigated.
Since atorvastatin and ezetimibe are only slightly soluble in water, atorvastatin and ezetimibe were primarily dissolved in methanol and ethanol, respectively, for the initial stock solution, which was diluted to produce the needed working solutions (Oswald et al., 2006a;Rodde et al., 2014). Vehicle controls were performed with methanol and ethanol (without drugs) to control for their potential effects.

In vitro study
Cell culture The Caco-2 cells were purchased from Pasteur Institute of Iran (Iran) and were cultured in DMEM containing 10% fetal calve serum and 1% penicillin-streptomycin. The culture flasks were incubated in a CO2-incubator (Germany) at 37°C with a 5% CO2 atmosphere. The culture medium was changed 2-3 times per week. After 2-3 weeks culture and differentiation of Caco-2 into intestinal-like cells, the cells were detached from the culture flask by addition of 0.25% trypsin-EDTA. The cells were then washed with phosphate buffer solution (pH = 7.4) and seeded at the needed density for the following experiments.

Cytotoxicity study (MTT assay)
The cytotoxicity of different concentrations of atorvastatin and ezetimibe were investigated using the MTT assay. Caco-2 cells were seeded into 96-well plates at a density of 15 × 10 3 cells per well. After 24 hours, the medium was replaced with 200 µL per well of atorvastatin and ezetimibe at different concentrations (diluted with complete culture medium) or complete culture medium as control. Three wells for each concentration or control were considered. After 24 hours, the MTT solution was added to each well at final concentration of 2 mg/mL and incubated for 4 hours at 37ºC in the CO2 incubator. The MTT solution was discarded and the produced formazan crystals were solubilized with 200 µL/well of DMSO and 25 µL/well Sorensen buffer. The optical densities (ODs) were measured with an enzyme linked immunosorbent assay (ELISA) micro-plate reader (Statfax-2100, Awareness, USA) at 570 nm with background subtraction at 630 nm. The percentage of cell viability was calculated using the following formula: The mean ± standard deviation (SD) was calculated for 0.1, 3, and 100 µM concentrations of atorvastatin and 0.2, 10, and 100 µM concentrations of ezetimibe.

Rho123 uptake assay
Rho123 is a well-known P-gp substrate and is secreted by Caco-2 cells via P-gp efflux. Treatment with an inhibitor of P-gp (verapamil) will promote the accumulation of Rho123 in Caco-2 cells. In this study, Caco-2 cells were seeded in 24-well plates and allowed to attach for 24 hours. The old medium was removed, and cells were washed with PBS. Then, new culture media containing 100 µM atorvastatin, 100 µM ezetimibe, or 300 µM verapamil without FCS or fetal bovine serum (FBS) were added to separate wells and incubated. Because of the protein binding properties of atorvastatin (>98%) (Lennernäs, 2003) and ezetimibe (99.5-99.8%) (Oswald et al., 2007), which may interfere with the results, we did not add FCS or FBS to the culture medium during this treatment step. After 24 hours, the old medium was discarded and cells were washed with PBS. Rho123 solution [DMEM containing 10 mM HEPES (pH = 7.4) and 5 μM Rho123] was added and incubated for 3 hours at 37°C, followed by three washes with ice-cold PBS (pH = 7.4). Cells were lysed in 1% Triton X-100 and centrifuged (3-18k, Sigma) at 1500 g for 10 min. The cellular accumulation of Rho123 in each sample (Rh-123 contents of supernatant) was measured quantitatively by fluorescence spectrophotometry (λex = 485 nm, λem = 532 nm) (FP-750, Jasco, Japan) and normalized by the protein content of each sample (supernatant), as determined by a protein assay kit (Pars Azmoon, Iran) (Mohammadzadeh et al., 2014;Zastrea et al., 2002;Zrieki et al., 2010).

Western blotting analysis of P-gp expression
Expression of P-gp in Caco-2 cells was assayed by Western blot after cells were treated with atorvastatin, ezetimibe, and verapamil. Caco-2 cells were seeded in a 6-well plates at 10 6 cells per well and treated for 48 hours with 100 µM atorvastatin, 100 µM ezetimibe, or 300 µM verapamil in culture medium in separate wells (n = 4). Culture medium alone was considered the control (n = 4). Solutions were then discarded, and cells were washed with PBS and incubated for 5 min with 0.25% trypsin/EDTA in 37°C. The detached cells were washed twice with PBS. Lysis buffer (Triton X-100 50 mM, NaCl 150 mM, EDTA 5 mM, 1% protease inhibitor cocktail, Tris-HCl, pH=7.4) was added and cell suspen-sions were centrifuged at 1,500 xg for 5 min. Proteins were resolved by 12.5% sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) (80 V, 120 min) and electrophoretically transferred to polyvinylidene di-fluoride (PVDF) membrane using a semidry Western blotting system (Bio-Rad, Hercules, CA, USA). All membranes were blocked in 3% non-fat-dried milk for 1 hour at room temperature and then washed three times with Tris buffered saline (TBS) with 0.1% Tween 20). Membranes were subsequently incubated overnight with mouse monoclonal anti-P-gp antibody (1/1000 in TBS). Membranes were washed with TBS and then incubated with HRP-conjugated rabbit antimouse secondary antibody for 2 hours. Membranes were washed, and the proteins were then detected using an ECL kit. The proteins were visualized by exposing the membrane to medical X-Ray film for 5 min. β-Actin was the internal standard and was detected using rabbit polyclonal anti-β-actin as primary antibody and HRP-conjugated goat anti-rabbit IgG as secondary antibody. P-gp expression is presented as the ratio of P-gp band intensity to β-actin band intensity in the same blot run (P-gp/β-actin) (Wanga et al., 2009).

In situ study
The rat single-pass intestinal permeability experiment with P-gp inhibition study was used to investigate whether intestinal P-gp-mediated transport of digoxin, a known substrate of P-gp, was affected by atorvastatin and ezetimibe. The effective permeability of digoxin was calculated and compared in the presence and absence of atorvastatin, ezetimibe and verapamil.

Animals
Male Sprague-Dawley rats (200-250 g) were purchased from Pasteur Institute of Iran. The rats were housed in standard cages under standard husbandry conditions (22 ± 2ºC, 12/12 hours light/dark cycle and 55-60% relative humidity). Standard rodent chow and clean water were available ad libitum. Prior to experimentation, rats were fasted for 12-16 hours (water ad libitum).

P-gp inhibition study by single-pass intestinal perfusion test (Video Clip)
The rats were anesthetized with an intraperitoneal injection of thiopantal sodium (60 mg/kg body weight). A midline abdominal incision of 3-4 cm was made and approximately 10 cm of the jejunum segment of the intestine was isolated and cannulated at both ends with polypropylene tubes. Body tempered saline was used for moisturizing the exposed segment. The segment was rinsed with 37ºC normal saline to wash and clear the segment, then PBS (pH=7.4) containing 50 mg/L phenol red without drug (blank solution) was pumped through the segment at a constant flow rate of 0.2 mL/ min (Qin). Blank perfused solution was collected at the outlet and used to prepare digoxin high performance liquid chromatography (HPLC) calibrator solutions and for stability studies (Patel and Barve, 2012;Zakeri-Milani et al., 2007).
The SPIP method was carried out in anesthetized rats as described above, and the blank solution containing the substance of interest was perfused at a constant flow rate of 0.2 mL/min using a volumetric infusion pump (Argus Medical AG, Switzerland). After reaching the steady state, each perfusion experiment lasted for 90 min, and perfusates were quantitatively collected at 10 min intervals (2 mL). Some changes, such as water absorption and secretion, during perfusion may cause errors in the calculated permeability values. Therefore, 50 mg/L phenol red was added as a non-absorbable marker to correct the results. Digoxin (20 µM, a typical substrate of P-gp), verapamil (300 µM, a typical P-gp inhibitor), and different concentrations of atorvastatin (0.1, 3 and 100 µM) and ezetimibe (0.2, 10 and 100 µM), each one in a blank solution with digoxin 20 µM, were administrated, as described above (Patel and Barve, 2012;Zakeri-Milani et al., 2007). The SPIP method was performed in three rats for each concentration of each drug and the blank solution. At the end of procedure, the length of the segment was measured (cm) and the animal was euthanatized. Samples were stored in an ultra-low temperature freezer (Jal Tajhiz Production, Iran) at -70°C until HPLC analysis. The concentrations of phenol red in the perfused (outlet) samples were measured at 560 nm using an ultraviolet-visible (UV-VIS) spectrophotometer (Ultra-spec 2000, Pharmacia, Pfizer, USA) (Valizadeh et al., 2012). The amount of digoxin in the perfused samples was detected by HPLC.

HPLC analysis of digoxin in perfused samples
HPLC system (Smartline manager 5000, Smartline UV detector 2600 and Smartline pump 1000, Knauer Advanced Scientific Instruments, Germany) was used to detect the amount of digoxin in the perfused samples. The mobile phase was 35% v/v of acetonitrile in water, which was filtered through a sintered glass filter P5 (1.0-1.6 μm, Winteg, Germany) and degassed in a sonicator. The mobile phase was pumped in isocratic mode at a flow rate of 2 mL/min at ambient temperature. UV detection was performed at 218 nm, and the injection volume of samples was 20 μL. The HPLC column was Knauer-15VE081ESJ-150X 4.6 mm with precolumn-Eurospher 100-5 C8 (Wanga et al., 2009).

Calculation and analysis of data
Effective permeability (Peff) values were calculated from the steady state concentrations of digoxin in the perfusate samples. Steady state was confirmed by the ratio of the outlet to inlet concentrations (corrected for water transport) versus time (Patel and Barve, 2012;  Figure 1 illustrates the steady state charts.

Phenol red correction
Corrected Cout (outlet concentration of the drug) was calculated from the following equation (Prasad and Bhasker, 2012). Cout (corr) = Corrected outlet concentration of the drug; Cout = Outlet concentration of the drug Calculations were based on outlet perfusate steady state concentrations achieved after the selected time points (Figure 1). The steady state intestinal effective permeability (Peff) was calculated according to following equation: Peff represents effective permeability (cm/s), Qin represents the perfusion rate (0.2 mL/min), Cin and Cout represent the concentrations of the test drug entering and leaving the segment, respectively, r is the radius of the intestinal segment (0.18 cm), and L is the length of the intestinal segment (cm)

Statistical analysis
Data are presented as the mean ± SD. Statistical analyses were performed using one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test, and the differences between two groups were determined using the unpaired t-test. Statistical tests were performed with SPSS13.0 (SPSS Inc., Chicago, IL, USA), where p<0.05 and p<0.01 were considered to be statistically significant.

Cytotoxicity
The MTT cytotoxicity assay on Caco-2 cells treated with 0.1, 3, and 100 µM atorvastatin, 0.2, 10, and 100 µM ezetimibe, 300 µM verapamil and 20 µM digoxin solutions showed no cytotoxicity (Figure 2). Statistical analyses showed that there was no significant difference in cell viability among the different treatment groups, and the concentrations used were considered for the following in vitro and in situ tests.

Rho123 accumulation in Caco-2 cells
To investigate the effects of atorvastatin and ezetimibe on the function of P-gp as an efflux pump in vitro, we measured the accumulation of Rho123 in Caco-2 cells treated with atorvastatin, ezetimibe and verapamil. As shown in Figure 3, treatment with 100 µM atorvastatin and 100 µM ezetimibe significantly increased the intracellular accumulation of Rho123 in Caco-2 cells relative to control (p<0.05). The mean intracellular concentration of Rho123 in control cells was 424.4 ± 89.3, while in atorvastatin, ezetimibe and verapamil treated cells, it was 646.0 ± 13.1, 681.7 ± 135.0, and 749.4 ± 32.1 pg/mg protein respectively.

Immunoblotting
After treating Caco-2 cells with atorvastatin, ezetimibe (100 µM), or verapamil (300 µM), immunoblotting of cell extracts was performed using a monoclonal antibody against P-gp to measure P-gp expression. Protein expression of β-actin was measured as an internal immunoblotting control. P-gp expression is presented as the ratio of P-gp band intensity to β-actin band intensity (P-gp/β-actin) and was compared with verapamil control bands in the same blot run (Mohammadzadeh et al., 2014). Low density of immunoblotting bands of atorvastatin and ezetimibe treatment decreased P-gp protein levels relative to control (Figure 4).

Adsorption experiments
Preliminary adsorption experiments showed that there was no considerable adsorption of compounds on the tubing or syringe and that the stock and working standard solutions of digoxin and phenol red were stable during the experiments.

Single-pass intestinal permeability
The effects of the drugs on P-gp function were investigated in situ using the rat SPIP method. Inhibition of P-gp activity with atorvastatin, ezetimibe, and verapamil treatment was investigated using digoxin as the typical P-gp substrate. The Peff values of digoxin (20 µM) in the absence and presence of verapamil were 3.4 ± 0.8 and 8.9 ± 0.7 × 10 -5 cm/s, respectively, whereas they were 4.4 ± 0.6, 6.0 ± 1.1, and 7.3 ± 0.7 × 10 -5 cm/s with 0.1, 3 and 100 µM atorvastatin, respectively. The Peff values of digoxin in the presence of 0.2, 10 and 100 µM ezetimibe were 4.8 ± 0.4, 6.5 ± 0.4, and 6.5 ± 1.0 × 10 -5 cm/s, respectively. As shown in Figure 5, 0.1 µM atorvastatin did not significantly increase the Peff of digoxin compared with the control group (digoxin alone) (p>0.05), but Peff at high concentrations, 10 and 100 µM, was significantly different (p<0.05 and p<0.01, respectively). Ezetimibe (0.2 µM) treatment did not significantly increase the Peff of digoxin relative to the control group (p>0.05), whereas 10 and 100 µM did significantly increase Peff relative to control (p<0.05).

Discussion
Based on data from the in vitro Rho123 uptake study, treatment of Caco-2 cells with atorvastatin increased the accumulation of Rho123 compared with control, suggesting that atorvastatin inhibits P-gp in vitro. We also found in situ that atorvastatin significantly increased digoxin Peff. The results of the in situ study showed a dose-dependent effect of atorvastatin, in the range of 0.1-100 µM, on P-gp. Protein expression of Pgp in atorvastatin-treated Caco-2 cells was low relative to control cells, confirming inhibition of P-gp by atorvastatin.
Although there are a few studies about the effect of atorvastatin on P-gp function, we did not find any evidence that is in contrast to our results. Vatsa et al. reported that atorvastatin is a moderate substrate for CYP P450 and is extensively metabolized via CYP P450mediated oxidation and phase II glucuronidation both in humans and rats (Vatsa et al., 2012). The drug is a CYP3A4 substrate, and most of CYP3A4 substrates are P-gp substrates as well (Boyd et al., 2000). Vatsa et al. stated that atorvastatin is a P-gp substrate at lower dose and an inhibitor at higher doses (Vatsa et al., 2012). The P-gp substrate activity of atorvastatin was also suggested by some other studies (Williams and Feely, 2002;Wu et al., 2000). Boyd et al. found that 100 µM atorvastatin inhibited digoxin secretion in Caco-2 cells (transport from the basolateral to apical aspect of the monolayer) by 58%, equivalent to the inhibition observed with verapamil (Boyd et al., 2000). Another study showed that atorvastatin increased the bioavailability of digoxin (Lennernäs, 2003), which may be explained by the inhibition of P-gp by atorvastatin.
Ezetimibe-treated Caco-2 cells exhibited more accumulation of Rho123 than cells treated with control. Rho123 is a P-gp substrate, and its accumulation may be due to the inhibitory effect of ezetimibe on P-gp. In contrast, in the rat SPIP test, 10 and 100 µM ezetimibe increased the Peff of digoxin by 91% relative to digoxin alone treated cells. Although 0.2 µM ezetimibe increased the Peff of digoxin, the effect was not statistically significant. Digoxin is a well-known substrate of P-gp and enhancement of its permeability may be due to inhibition of Pgp efflux by ezetimibe. The inhibition by ezetimibe in situ on P-gp was not dose dependent, and the Peff of digoxin was approximately the same at ezetimibe concentrations greater than 10 µM. P-gp expression in Caco-2 cells treated with 100 µM ezetimibe was less dense than that in non-treated cells, illustrating the inhibitory effects of ezetimibe on P-gp.
Ezetimibe is a biopharmaceutics classification system (BCS) class II drug with high permeability, and it undergoes rapid first-pass metabolism and P-gp efflux (Bandyopadhyay et al., 2012). To our knowledge, after oral administration of ezetimibe, it is released from the dosage form in the upper small intestine, where it is taken up into enterocytes based on the regional expression and activity of P-gp and multidrug resistant protein 2 (MRP2) (Oswald et al., 2006a (Suchy et al., 2011). They also showed that up-regulation of intestinal P-gp in healthy subjects was associated with markedly decreased serum concentrations and a sterol-lowering effect of ezetimibe (Suchy et al., 2011;Bandyopadhyay et al., 2012). In another study, the intestinal expression of ATP-binding cassette, subfamily B, member 1 transporter (ABCB1) was inversely correlated with the effects of ezetimibe on plant sterol serum concentrations. The disposition and the cholesterol-lowering effect of ezetimibe were markedly influenced by the activity of intestinal efflux transport proteins, P-gp, and MDR2 (Oswald et al., 2006b). Oswald et al. concluded that the extent of oral absorption of ezetimibe is controlled by intestinal P-gp and that this efflux transport protein is involved, at least in part, in the long-lasting recirculation of the drug to the pharmacodynamic sites of action (Oswald et al., 2006a). Co-administration of ezetimibe and cyclosporine resulted in enhancement of the blood plasma levels of cyclosporine (Oswald et al., 2007). Cyclosporine is a known substrate of P-gp (Wessler et al., 2013), and the increase in its levels may be due to the Pgp inhibitory effect of ezetimibe.
The clinical dosage ranges for atorvastatin and ezetimibe are 10-80 mg/day and up to 10 mg/day (Lennernäs, 2003), respectively, which are within the concentration ranges used in this study. Therefore, atorvastatin and ezetimibe may inhibit P-gp at doses used clinically. Co-administration and interaction of atorvastatin and ezetimibe with other drugs that are Pgp substrates may result in the occurrence of adverse drug reactions owing to increases in drug absorption and the concentration in blood plasma or tissues. The DDIs become more important when the therapeutic index (the margin between effectiveness and toxicity) of the drug is narrow (e.g., digoxin).

Conclusion
Atorvastatin and ezetimibe inhibited P-gp efflux activity in vitro and in situ. The inhibitory effect of atorvastatin was dose-dependent (0.1-100 µM), whereas the effect of ezetimibe (0.2-100 µM) was not dosedependent. Both drugs down-regulated the expression of P-gp in vitro. The inhibitory effects of atorvastatin and ezetimibe on P-gp should be considered when predicting potential DDIs when drugs are coadministered with drugs that are P-gp substrates.

Financial Support
Drug Applied Research Center of Tabriz University of Medical Sciences (Grant No. 91-123)

Ethical Issue
All experiments were conducted in accordance with the humanity and animal ethic protocols. The animal study protocols were registered by the research ethical committee of Tabriz University of Medical Sciences (Ref No: TBZMED.REC.1394.378).

Conflict of Interest
Authors declare no conflict of interest