Induction of hepatic CYP3A4 expression by cholesterol and cholic acid: Alterations of gene expression, microsomal activity, and pharmacokinetics

Abstract

Human cytochrome P450 3A4 (CYP3A4) is a drug‐metabolizing enzyme that is abundantly expressed in the liver and intestine. It is an important issue whether compounds of interest affect the expression of CYP3A4 because more than 30% of commercially available drugs are metabolized by CYP3A4. In this study, we examined the effects of cholesterol and cholic acid on the expression level and activity of CYP3A4 in hCYP3A mice that have a human CYP3A gene cluster and show human‐like regulation of the coding genes. A normal diet (ND, CE‐2), CE‐2 with 1% cholesterol and 0.5% cholic acid (HCD) or CE‐2 with 0.5% cholic acid was given to the mice. The plasma concentrations of cholesterol, cholic acid and its metabolites in HCD mice were higher than those in ND mice. In this condition, the expression levels of hepatic CYP3A4 and the hydroxylation activities of triazolam, a typical CYP3A4 substrate, in liver microsomes of HCD mice were higher than those in liver microsomes of ND mice. Furthermore, plasma concentrations of triazolam in HCD mice were lower than those in ND mice. In conclusion, our study suggested that hepatic CYP3A4 expression and activity are influenced by the combination of cholesterol and cholic acid in vivo.

Plasma concentrations of triazolam (TRZ) in hCYP3A mice fed high cholesterol and cholic acid diet (HCD) were lower than those in hCYP3A mice fed normal diet (ND).

Abbreviations

4‐OH TRZ

4‐hydroxy triazolam

AUC

area under plasma concentration curve

AUCR

AUC ratio

CA

cholic acid

CYP

cytochrome P450

HCD

high cholesterol and cholic acid diet

hCYP3A

humanized CYP3A

HMGCR

3‐hydroxy‐3‐methylglutaryl‐coA reductase

HPLC

high performance liquid chromatography

IS

internal standard

KCZ

ketoconazole

KO

knockout

LC–MS/MS

liquid chromatography–tandem mass spectrometry

MAC

mouse artificial chromosome

PXR

pregnane X receptor

TRZ

triazolam

WT

wild type

α‐OH TRZ

α‐hydroxy triazolam

1. INTRODUCTION

Xenobiotics, such as drugs and exogenous toxicants, that are incorporated into the body are eliminated by various enzymes. Cytochrome P450 (CYP) is a large enzyme group of monooxygenases that catalyze dealkylation, hydroxylation, and epoxidation reactions of xenobiotics. ¹ CYP enzymes mediate the metabolism of 40% of small‐molecule drugs that were newly launched during the period from 2007 to 2016. ² The expression levels and activities of the enzymes are altered by various factors such as drugs, diseases, diets, smoking, alcohol, age, and race as well as genetic factors. ³

Human CYP3A4 is an important member of the CYP enzyme group that is abundantly expressed in the liver and small intestine and contributes to the metabolism of a large number of commercially available drugs. ⁴ In 15 patients receiving abdominal surgery, the contents of CYP3A4 protein in biopsies of the liver and duodenum varied by 51‐fold and ninefold, respectively. ⁵ Not only genetic polymorphism ⁶ but also exogenous compounds containing therapeutic drugs, environmental chemicals (e.g., pesticides, UV stabilizers, and plasticizers), and natural compounds in the diet (e.g., polyphenols and fat) ⁷ alter the expression level of CYP3A4. However, there is little information about what impact these factors have on the expression of CYP3A4 in vivo.

To clarify exogeneous factors that regulate the metabolism of CYP3A4, strictly controlled in vivo animal models expressing CYP3A4 are needed because many uncontrolled factors influence CYP3A4 in the human body. In our previous studies, some humanized mouse lines for pharmacokinetic analysis were developed using mouse artificial chromosome (MAC) vectors coding the genes of interest. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ Since an Mb‐sized genomic region can be loaded in a MAC vector, CYP3A‐humanized mice (hCYP3A mice) have a MAC vector that carries a human CYP3A gene cluster including the coding region and regulatory sequences. In a physiological condition, hCYP3A mice express CYP3A4 mRNA and protein in the liver and small intestine as in humans. ⁹ In addition, CYP3A4 was induced via ligand‐dependent activation of pregnane X receptor (PXR) in hCYP3A mice. ⁹ , ¹⁶ , ¹⁷ As for other human CYP3A enzymes, CYP3A5 and CYP3A7 proteins are barely expressed in the liver and intestine of adult hCYP3A mice because hCYP3A mice carry a CYP3A5 variant to cause a splicing defect and show infant‐specific expression of CYP3A7 as same as humans. ¹⁸ CYP3A43 is known to have little involvement in drug metabolism. ¹⁹ Therefore, this hCYP3A mouse model is thought to be a simplified model for assessing the alteration of CYP3A4 in the liver and intestine.

Bile acids are endogenous compounds that regulate the expression levels of CYP3A enzymes via activation of nuclear receptors. In human primary hepatocytes, exposure to bile acids increased CYP3A4 expression. ²⁰ When wild‐type (WT) mice were fed diets containing bile acids, mouse Cyp3a11 (mCyp3a11) expression was increased by cholic acid (CA) and ursodeoxycholic acid, but it was decreased by deoxycholic acid. ²¹ , ²² A diet containing cholesterol and CA (high cholesterol and CA diet, HCD) induced mCyp3a11 expression via PXR in the liver. ²² In contrast, a low cholesterol diet decreased mCyp3a11 expression. ²³ Thus, cholesterol is also assumed to be a regulator of mCyp3a11. However, it is still unknown whether CYP3A4 expression is induced by bile acids and cholesterol in vivo.

In this study, we examined the effects of an HCD on the expression levels of CYP3A4 in the liver and small intestine and the pharmacokinetics of a CYP3A substrate in hCYP3A mice. In addition, we examined whether hCYP3A mice fed the HCD could be used for drug–drug interaction (DDI) studies.

2. MATERIALS AND METHODS

2.1. Materials

Triazolam (TRZ) and cholesterol were purchased from Wako Pure Chemicals (Osaka, Japan). α‐Hydroxy (α‐OH) TRZ and 4‐hydroxy (4‐OH) TRZ were supplied by Nihon Upjohn Co. (Tokyo, Japan). 4β‐Hydroxy (4β‐OH) cholesterol was purchased from Cayman (Ann Arbor, MI). Ketoconazole (KCZ) was purchased from Tokyo Chemical Industry (Tokyo, Japan). Other reagents used in this study were purchased from commercial sources.

2.2. Animals

hCYP3A mice carrying a MAC vector containing the entire human CYP3A gene cluster and lacking the entire mouse Cyp3a gene cluster were generated and characterized previously. ⁹ Sex and developmental stage can influence the expression level of CYP3A4 in the liver of CYP3A‐humanized mice generated by chromosome engineering as in humans. ¹⁸ , ²⁴ Thus, all experiments were performed using adult (between 9 and 11 weeks of age) male hCYP3A mice with an ICR background. Male ICR mice purchased from Japan SLC (Shizuoka, Japan) were used as WT mice. All experimental animals were kept in an environment in which the temperature and light/dark cycle was controlled. The light cycle hours were between 7 a.m. and 7 p.m. Mice were administered combined anesthetic agents (medetomidine/midazolam/butorphanol = 0.3–0.75/4/5, mg/kg) before the surgical procedure. All animal experiments were approved by the Animal Care and Use Committee of Tottori University and Meiji Pharmaceutical University and were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals as adopted by the Committee on Animal Research of Tottori University. We conducted all animal experiments in accordance with ARRIVE guidelines.

2.3. Chows

CE‐2, containing approximately 0.1% cholesterol, was used as a normal diet (ND). HCD was CE‐2 supplemented with 1% cholesterol and 0.5% CA, a diet that was previously used in mouse experiments. ²⁵ , ²⁶ CE‐2 supplemented with only 0.5% CA (CA‐diet) was also used. These diets were purchased from CLEA Japan. hCYP3A mice were fed these diets for 1 week and then the livers and small intestines of the mice were collected. Since CA is required for efficient ingestion of cholesterol from the digestive tract, ²⁷ , ²⁸ a cholesterol‐supplemented diet without any CA was not prepared in this study. WT mice were fed ND or HCD for 1 week and then the livers and small intestines were collected.

2.4. Extraction of plasma cholesterol and bile acids

For determination of 4β‐OH cholesterol and cholesterol, 20 or 2 μL of plasma was mixed with 5 μL of methanol containing 4 ng/mL 4β‐OH cholesterol‐d7 (as an internal standard, IS). Then 5 μL of 1 mg/mL butylated hydroxytoluene (in ethanol) and 100 μL of 56 mg/mL potassium hydroxide (in ethanol) were added and the mixture was incubated (37°C, 1 h). After addition of 250 μL of water, 4β‐OH cholesterol and cholesterol were extracted with n‐hexane. The organic layer obtained by extraction twice for each sample was dried by a centrifugal evaporator and an organic solvent mixture (formic acid/ methanol/ acetonitrile = 1/500/500) was added into redissolve the samples. The supernatants were injected to a liquid chromatography‐tandem mass spectrometry (LC‐MS/MS) system. Sample preparation to determine bile acids in plasma was conducted in accordance with the protocol of Method Package of Bile Acids (Shimadzu, Kyoto, Japan).

2.5. Measurements of mRNA levels

The mRNA expression levels were determined by quantitative PCR performed on a CFX Connect Real‐time System (Bio‐Rad, Hercules, CA). The genes listed in Table S1 were measured by TaqMan Gene expression Assays (Thermo Fisher Scientific, Waltham, MA). mGapdh was used as an endogenous control.

2.6. In vitro metabolism of TRZ

The activities of TRZ α‐ and 4‐hydroxylation via CYP3A4 were determined using hCYP3A mouse liver or small intestine microsomes prepared as described previously. ¹⁸ Briefly, the basic incubation mixture of 0.2 mg/mL of liver microsomes or 0.1 mg/mL of small intestine microsomes and 100 μM TRZ supplemented with 0.1 mM EDTA and 100 mM potassium phosphate buffer (pH 7.4) was incubated (37°C, 30 min). The incubation was initiated by the addition of an NADPH‐generating system. To terminate the reaction, acetonitrile was added and then oxazepam (IS) in methanol was added. The centrifuged supernatant was injected into a high‐performance liquid chromatography (HPLC) system.

In an inhibition study, reaction mixtures containing 0.01, 0.1, and 1 μM KCZ and pooled liver microsomes were incubated as described above.

2.7. In vivo metabolism of TRZ

hCYP3A mice fed the ND were orally administered TRZ dissolved in 25% ethanol in saline. After 1 week of feeding the HCD without any drug administration, the same hCYP3A mice were administered TRZ orally again. In an inhibition study, 40 mg/kg KCZ, a CYP3A inhibitor, or a vehicle (0.375% methylcellulose in water/ethanol = 75/25, v/v) was administered orally to hCYP3A mice fed the ND or HCD. After 15 min of administration of KCZ, TRZ was administered orally. Blood samples were collected at 15, 30, 60, 120, 240, and 420 min after TRZ administration. Following protein precipitation, the levels of TRZ, α‐OH TRZ, and 4‐OH TRZ in plasma were determined by LC‐MS/MS as described previously. ²⁹

The trapezoidal rule was used to calculate the area under the plasma concentration‐time curve (AUC). The AUC ratio (AUCR), an established in vivo CYP3A4 activity marker in hCYP3A mice, ²⁹ was calculated as follows: AUC_{α‐OH TRZ}/AUC_TRZ = AUCR_α‐OH, AUC_{4‐OH TRZ}/AUC_TRZ = AUCR_4‐OH.

2.8. HPLC analysis

Metabolites of TRZ generated in microsomal mixtures were detected by a Prominence HPLC system (Shimadzu). Plasma concentrations of cholesterol, CA, tauro‐CA, TRZ, and metabolites of TRZ were determined by an LCMS8050 apparatus (Shimadzu) equipped with a Nexera XR HPLC system (Shimadzu). Detailed conditions of HPLC and other equipment are shown in Table S2.

2.9. Statistical analysis

A comparison of two groups was made with the Mann–Whitney U test. One‐way ANOVA with a post hoc test of Scheffé's F test was performed to compare multiple groups. p < .05 was considered statistically significant. Statcel‐4 (OMS, Tokyo, Japan) was used for these statistical analyses. The correlations between the hepatic expression level of CYP3A4 and the hepatic expression levels of other genes were determined by Pearson's correlation coefficients using JMP Pro software (version 16.2.0, SAS Institute Inc., Cary, NC), and the correlation coefficients (r‐values) were calculated.

3. RESULTS

3.1. Plasma concentrations of cholesterol and bile acids in hCYP3A mice fed different diets

Plasma concentrations of cholesterol, CA, and its major conjugate, tauro‐CA, were determined (Table 1). Plasma concentrations of these compounds in HCD and CA‐diet mice were higher than those in ND mice. Notably, plasma concentrations of cholesterol and tauro‐CA in HCD mice were significantly higher than those in ND mice.

TABLE 1.

Plasma concentrations of cholesterol and bile acids in hCYP3A mice fed the ND, HCD, or CA‐diet (n = 4/group).

Compounds	Diets
	ND	HCD	CA
Cholesterol (μg/mL)	1361 ± 152	2662 ± 856*	2250 ± 336
CA (ng/mL)	312 ± 243	1091 ± 1145	1350 ± 1869
Tauro CA (ng/mL)	56 ± 7	1601 ± 784*	1028 ± 664

Note: Data are shown as means with SD of four mice. Statistical analysis was performed by ANOVA with a post hoc test of Scheffé's F test.

^* p < .05 versus mice fed the ND.

3.2. Effects of diets containing bile acids and cholesterol on mRNA expression levels

The mRNA expression levels of various genes in hCYP3A mice fed different diets were compared. The mRNA expression levels of CYP3A4 in livers of HCD mice and CA‐diet mice were 7.4‐fold and 1.9‐fold higher, respectively, than those in ND mice (Figure 1A). A statistically significant increase of CYP3A4 mRNA expression was found in the livers of HCD mice. In the small intestine, no significant increases in the expression levels of CYP3A4 in HCD and CA‐diet mice compared with those in ND mice were found (Figure 1B). In addition, the effects of the HCD on mCyp3a11, a mouse homolog of CYP3A4, was compared in WT mice fed the ND and HCD (Figure S1). WT mice fed the HCD showed a significantly higher expression level of mCyp3a11 in the liver but not in the small intestine than that in WT mice fed the ND as shown in hCYP3A mice.

FIGURE 1.

Effects of the HCD and CA‐diet on expression levels of CYP3A4 in the liver and intestine. The mRNA expression levels of CYP3A4 in the liver (A) and intestine (B) of hCYP3A mice were determined by real‐time PCR. Data are shown as means with SD of four mice. *p < .05 versus mice fed the normal diet (ND). Statistical analysis was performed by ANOVA with a post hoc test of Scheffé's F test.

The expression levels of mCyp2c55, mMdr1a, and mCyp2b10 in the livers of hCYP3A mice fed supplemented diets were also higher than those in the livers of mice fed the ND (Figure 2). The expression levels of all of these genes were highest in HCD mice. On the other hand, hepatic mRNA expression levels of mCyp1a2 and mPxr were not different between mice fed different diets (Figure S2). Since the expression of hepatic CYP3A4 mRNA was most clearly increased in HCD mice, correlations between the mRNA expression level of CYP3A4 and mRNA expression levels of the other four genes were analyzed in individual ND and HCD mice (n = 4/group). Among these four genes, the expression levels of mCyp2c55 and mMdr1a showed strong positive correlations with those of CYP3A4 in ND and HCD mice (Table 2).

FIGURE 2.

Effects of the HCD and CA‐diet on the expression levels of mouse endogenous genes in the liver. The mRNA expression levels of (A) mouse Cyp2c55, (B) mouse mMdr1a, and (C) mouse Cyp2b10 in the livers of hCYP3A mice were determined by real‐time PCR. Data are shown as means with SD of four mice. *p < .05 versus mice fed the ND. Statistical analysis was performed by ANOVA with a post hoc test of Scheffé's F test.

TABLE 2.

Correlation coefficients of hepatic mRNA expression levels between CYP3A4 and other genes in hCYP3A mice fed the ND and HCD (n = 4/group).

Genes	Correlation coefficient (r)	p‐value
mPXR	.3327	.4207
mCyp2c55	.8382	.0093
mMdr1a	.7672	.0263
mCyp2b10	.6328	.0922
mCyp1a2	.4362	.2799

Note: The correlations between the hepatic expression level of CYP3A4 and these genes were determined by Pearson's correlation coefficient.

Since plasma levels of cholesterol and bile acids were increased in HCD and CA‐diet mice, the hepatic expression levels of marker genes affected by them were determined. The mRNA expression levels of mCyp7a1 and mCyp8b1, genes that are repressed by CA or other bile acids, ³⁰ in hCYP3A mice fed the HCD or CA‐diet were markedly lower than those in mice fed the ND (Figure S3A,B). In addition, the mRNA expression levels of 3‐hydroxy‐3‐methylglutaryl‐coA reductase (Hmgcr), a gene repressed under a high cholesterol condition, ³¹ in hCYP3A mice fed the HCD or CA‐diet were lower than those in mice fed the ND (Figure S3C).

3.3. Effect of HCD on CYP3A activity

Since the HCD induced a high level of hepatic CYP3A4 mRNA expression in hCYP3A mice, we examined the effect of the HCD on in vitro and in vivo CYP3A4 activity. Hydroxylation of TRZ, a typical CYP3A4 substrate, has been established as a CYP3A4 activity marker in hCYP3A mice. ¹¹ , ²⁹ Thus, the microsomal activity of TRZ hydroxylation was determined to reveal the effects of different diets on in vitro CYP3A4 activity in hCYP3A mice. In the liver, TRZ hydroxylation activities in HCD mice were significantly higher than those in ND mice (Figure 3A). On the other hand, TRZ hydroxylation activities in the small intestine showed small differences between HCD mice and ND mice (Figure 3B).

FIGURE 3.

Effects of the HCD on TRZ hydroxylation activities in liver and intestine microsomes of hCYP3A mice. TRZ (100 μM) was incubated with (A) individual liver microsomes (n = 4/each group) or (B) pooled intestine microsomes (n = 4/each group). TRZ α‐ and 4‐hydroxylation activities in the microsomes were determined. Data are shown as means with SD of four mice for the liver or of triplicate incubation for the intestine. *p < .05 versus mice fed the ND. Statistical analysis was performed by the Mann–Whitney U‐test.

Next, in vivo CYP3A4 activity in hCYP3A mice fed the HCD or ND was determined by pharmacokinetic analysis of TRZ and its metabolites. Plasma concentrations and AUC values of TRZ (AUC_TRZ), α‐OH TRZ (AUC_{α‐OH TRZ}), and 4‐OH TRZ (AUC_{4‐OH TRZ}) in HCD mice were lower than those in ND mice (Figure 4A–D). AUCR_α‐OH in HCD mice was significantly higher than that in ND mice (Figure 4E).

FIGURE 4.

Pharmacokinetics of TRZ and its metabolites in hCYP3A mice fed the ND or HCD. Plasma concentration versus time curves of (A) TRZ, (B) α‐OH TRZ, and (C) 4‐OH TRZ after oral TRZ administration (1 mg/kg) in hCYP3A mice fed the ND or HCD. (D) The values of AUC of TRZ and its metabolites were calculated by the trapezoidal rule. AUCR of α‐OH TRZ (E) and that of 4‐OH TRZ (F) were calculated by dividing the AUC value of each metabolite by the AUC value of TRZ. Data are shown as means with SD of four mice. *p < .05 versus mice fed the ND. Statistical analysis was performed by the Mann–Whitney U‐test.

4β‐Hydroxy of cholesterol is a biomarker for CYP3A4 activity in plasma of hCYP3A mice. ¹⁷ The plasma level of 4β‐OH cholesterol in hCYP3A mice fed the HCD was significantly higher than that in mice fed the ND. The plasma concentration ratio (4β‐OH cholesterol/cholesterol) was not significantly different in mice fed those two diets (Figure S4).

3.4. In vivo DDI study in hCYP3A mice fed the ND or HCD

Finally, the effect of a CYP3A inhibitor on in vivo CYP3A4 activity in hCYP3A mice was examined. KCZ was used as a CYP3A4 inhibitor because microsomal activity of TRZ hydroxylation in the liver of hCYP3A mice was inhibited by KCZ regardless of the diet (Figure S5). After pretreatment with KCZ or a vehicle, the mice were administered TRZ. Pretreatment with KCZ tended to increase plasma concentrations of TRZ and decrease those of α‐OH TRZ in both ND and HCD mice (Figure 5A,B). The percentages of decline in AUCR_α‐OH caused by KCZ were almost the same in the two groups (49% for ND mice and 46% for HCD mice) (Figure 5C).

FIGURE 5.

Effects of a CYP3A inhibitor on pharmacokinetics of TRZ and its metabolites in hCYP3A mice. Plasma concentration versus time curves of TRZ and α‐OH TRZ after oral TRZ administration (1 mg/kg) in hCYP3A‐MAC mice treated with KCZ or 0.5% methylcellulose (vehicle). The mice were fed (A) the ND or (B) the HCD for 1 week before administration. (C) AUCR of α‐OH TRZ (AUCR_α‐OH) was calculated by dividing the AUC value of α‐OH TRZ by that of TRZ. Data are shown as means with SD of four mice. *p < .05 versus mice treated with the vehicle for each of the diet groups. Statistical analysis was performed by the Mann–Whitney U‐test.

4. DISCUSSION

The results of this study showed that the HCD greatly increased hepatic CYP3A4 levels in hCYP3A mice (Figure 1). Since the effect of the CA‐diet on CYP3A4 expression in the liver was smaller than the effect of the HCD (Figure 1), we thought that the increment of hepatic CYP3A4 in mice fed the HCD is due to the supplementation of cholesterol to the CA‐diet. The effect of the HCD on CYP3A4 expression in the liver of hCYP3A mice was similar to a previous finding that an HCD increased mCyp3a11 expression in the liver of WT mice. ²² In addition, we showed an increase in the microsomal activity of CYP3A4, a decrease in the plasma concentration of a CYP3A4 substrate, and an increase of AUCR after oral administration (Figures 3 and 4). The decline in plasma concentrations of monohydroxy TRZ in mice fed the HCD would be due to enhancement of secondary metabolism to dihydroxy TRZ mediated by CYP3A4. ¹¹ , ³² The condition with high levels of CA and cholesterol may increase the expression of CYP3A4 in the liver and enhance the CYP3A4‐mediated metabolism in humans.

PXR is an important nuclear receptor that regulates the transcription of genes coding drug‐metabolizing enzymes including CYP3A4 and some transporters. ¹³ , ¹⁴ , ³³ , ³⁴ In the liver of hCYP3A mice, HCD increased not only CYP3A4 but also other PXR target genes including mCyp2c55 and mMdr1a (Figure 2). In addition, significant correlations were found between the mRNA expression levels of those genes and that the mRNA expression level of CYP3A4 (Table 2). Cholesterol metabolites induce CYP3A gene expression by activation of mouse PXR. ³⁵ , ³⁶ Several kinds of bile acids, including CA, deoxycholic acid, lithocholic acid, and chenodeoxycholic acid, also activate PXR. ³⁷ Since the expression of rate‐limiting enzymes in bile acid synthesis were suppressed by the HCD (Figure S3), the compensatory reaction to produce lithocholic acid and chenodeoxycholic acid would be preferred in mice fed the HCD, resulting in activation of PXR. ³⁸

We revealed that the HCD increased in vitro and in vivo metabolic activities of TRZ, a CYP3A probe drug (Figures 3 and 4). On the other hand, the 4β‐OH cholesterol/cholesterol ratio, a well‐established endogenous CYP3A biomarker in humans and hCYP3A mice, was not significantly changed by the HCD (Figure S4B). Since 4β‐OH cholesterol was increased by about twofold (Figure S4A), no significant change in the 4β‐OH cholesterol/cholesterol ratio was due to an increase of plasma cholesterol levels (Table 1). In addition to feeding supplemented cholesterol, PXR activation may also contribute to the high levels of plasma cholesterol. ³⁹ PXR ligands increased the plasma concentration of 4β‐OH cholesterol and the 4β‐OH cholesterol/cholesterol ratio dose‐dependently in hCYP3A mice ¹⁷ and healthy human subjects. ⁴⁰ , ⁴¹ , ⁴² In those studies, no significant change in plasma cholesterol levels was found. Thus, it was thought that the HCD increased the formation of 4β‐OH cholesterol but that the effect was too weak to increase the plasma 4β‐OH cholesterol/cholesterol ratio. Alternatively, cholesterol itself may inhibit 4β‐OH cholesterol formation as a CYP3A4 inhibitor. ⁴³ Therefore, attention should be paid to the use of 4β‐OH cholesterol/cholesterol ratio for assessing CYP3A activity in a condition showing abnormal profiles of cholesterol and bile acids.

It is expected that hCYP3A mice can be used for predicting CYP3A‐mediated DDIs because of the physiological expression of CYP3A4 in the liver and intestine. KCZ was used as a CYP3A inhibitor to evaluate CYP3A‐mediated DDIs in hCYP3A mice fed the ND or HCD. The effects of the HCD on plasma KCZ levels would be negligible because KCZ is metabolized by arylacetamide deacetylase and flavin‐containing monooxygenase, ⁴⁴ enzymes that would not be induced via PXR. ⁴⁵ α‐Hydroxylation of TRZ by CYP3A4 was inhibited by KCZ in hCYP3A mice fed the HCD or ND in vivo as shown in Figure 5C. Furthermore, the inhibitory effect of KCZ on TRZ metabolism was statistically significant in HCD mice. The results suggest that the HCD is not only a diet that enhances CYP3A4 activity in hCYP3A mice but also a useful diet for evaluating CYP3A4‐mediated DDIs by co‐treatment with inhibitors without interference.

In this study, CA and cholesterol levels were increased in hCYP3A mice by feeding a special diet. Although there is no human diet containing such a high level of CA, CA level may be increased by intake of pharmaceutics. Cholbam^® (CA capsule), which inhibits the synthesis of toxic intermediates of bile acids, is an approved drug for Zellweger spectrum disorders. ⁴⁶ Administration of Cholbam^® increases blood CA levels in patients. ⁴⁷ On the other hand, increased ingestion of dietary cholesterol fails to increase blood cholesterol levels because of the strict endogenous regulation of cholesterol biosynthesis and transport. ⁴⁸ , ⁴⁹ However, dietary fatty acids can affect blood cholesterol levels. ⁵⁰ Blood cholesterol levels are increased more by intake of saturated fatty acid than by intake of polyunsaturated fatty acid. In addition, some compensatory metabolites of squalene and cholesterol, which are produced in response to inhibition of cholesterol synthesis, are PXR ligands. ⁵¹ , ⁵² , ⁵³ Taken together, the results indicate that direct and indirect alterations of bile acids and cholesterol metabolism may modulate CYP3A activity in humans as shown in hCYP3A mice. In this study, treatment with 1% cholesterol diet (with 0.5% CA) for 1 week increased the expression of CYP3A4 in the liver of hCYP3A mice. It was shown in a previous study that long‐term treatment with 1% cholesterol diet (without CA) for 11 weeks repressed the expression levels of drug‐metabolizing enzymes including CYP3A and nuclear receptors in rats. ⁵⁴ Therefore, further study is needed to clarify the chronic effects of cholesterol and CA on CYP3A4 expression.

The effect of the HCD on mRNA expression level and microsomal activity of CYP3A4 was not significant in the small intestine (Figures 1B and 3B). This finding was consistent with results in WT mice (Figure S1B). In these experiments, mouse tissues were collected after fasting for 16 h because we intended to remove diets from the digestive tract to cleanup the small intestine. The fasting period without any effects by food might be sufficient to initialize HCD‐meditated changes in the expression levels of genes in intestinal epithelium cells.

The hCYP3A mice used in this study were entirely humanized for CYP3A but not PXR. The amino acid sequences of the DNA‐binding domain of PXR are very similar in humans and mice, but species differences in the ligand‐binding domain of PXR between humans and mice are well known. ⁵⁵ The selectivity and potency of various bile acids for PXR activation in hepatocytes are different in humans and mice. ³⁷ There is no detailed report showing species differences for PXR activation by various cholesterol derivatives. Since it was suggested that PXR is the factor most likely to increase CYP3A4 expression in the liver of hCYP3A mice fed an HCD, a double humanized model for PXR and CYP3A might help to reveal the involvement of human PXR in CYP3A4 induction by an HCD.

In conclusion, we revealed by using hCYP3A mice that a condition of high cholesterol and high CA increased CYP3A4 expression in the liver rather than in the intestine in vivo. In addition, we found the potency of diets to modulate pharmacokinetics of CYP3A4 substrates. Although factors responsible for metabolism of cholesterol and bile acids (e.g., nuclear receptors, enzymes, and intestinal flora) were not humanized in hCYP3A mice, this mouse model can provide valuable information for understanding the regulation of CYP3A4 and the effects on pharmacokinetics of CYP3A4 substrates by endogenous and exogeneous factors.

AUTHOR CONTRIBUTIONS

Participated in research design: Kobayashi K and Kazuki Y. Conducted experiments: Kobayashi K, Minegishi G, Kobayashi Y, Fujikura M., and Sano A. Contributed new reagents or analytic tools: Kazuki Y. Performed data analysis: Kobayashi K, Minegishi G. Wrote or contributed to the writing of the article: Kobayashi K, Minegishi G, and Kazuki Y.

FUNDING INFORMATION

This study was supported in part by Research Support Project for Life Science and Drug Discovery (BINDS) from AMED under Grant Number JP23ama121046 (Y. Kazuki and K. K.), JST, CREST Grant Number JPMJCR18S4, Japan (Y. Kazuki), and by JSPS KAKENHI Grant Number 23 K06242 (K. K.).

CONFLICT OF INTEREST STATEMENT

There are no conflicts of interest.

ETHICS STATEMENT

All animal experiments were approved by the Animal Care and Use Committee of Tottori University and Meiji Pharmaceutical University and were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals as adopted by the Committee on Animal Research of Tottori University. We conducted all animal experiments in accordance with ARRIVE guidelines.

Supporting information

Figure S1.

Figure S2.

Figure S3.

Figure S4.

Figure S5.

Table S1.

Table S2.

Minegishi G, Kobayashi Y, Fujikura M, Sano A, Kazuki Y, Kobayashi K. Induction of hepatic CYP3A4 expression by cholesterol and cholic acid: Alterations of gene expression, microsomal activity, and pharmacokinetics. Pharmacol Res Perspect. 2024;12:e1197. doi: 10.1002/prp2.1197

DATA AVAILABILITY STATEMENT

The authors declare that all of the data supporting the findings of this study are available within this article and its Supplemental Data.