Evaluation of the usefulness of plasma 4β‐hydroxycholesterol concentration normalized by 4α‐hydroxycholesterol for accurate CYP3A phenotyping

Abstract

Plasma 4β‐hydroxycholesterol (OHC) has drawn attention as an endogenous substrate indicating CYP3A activity. Plasma 4β‐OHC is produced by hydroxylation by CYP3A4 and CYP3A5 and by cholesterol autoxidation. Plasma 4α‐OHC is produced by cholesterol autoxidation and not affected by CYP3A activity. This study aimed to evaluate the usefulness of plasma 4β‐OHC concentration minus plasma 4α‐OHC concentration (4β‐OHC–4α‐OHC) compared with plasma 4β‐OHC concentration and 4β‐OHC/total cholesterol (TC) ratio in cross‐sectional evaluation of CYP3A activity. Four hundred sixteen general adults were divided into 191 CYP3A5*1 carriers and 225 non‐carriers. Twenty‐six patients with chronic kidney disease (CKD) with CYP3A5*1 allele were divided into 14 with CKD stage 3 and 12 with stage 4–5D. Area under the receiver operating characteristic curve (AUC) for the three indices were evaluated for predicting presence or absence of CYP3A5*1 allele in general adults, and for predicting CKD stage 3 or stage 4–5D in patients with CKD. There was no significant difference between AUC of 4β‐OHC–4α‐OHC and AUC of plasma 4β‐OHC concentration in general adults and in patients with CKD. AUC of 4β‐OHC–4α‐OHC was significantly smaller than that of 4β‐OHC/TC ratio in general adults (p = 0.025), but the two indices did not differ in patients with CKD. In conclusion, in the present cross‐sectional evaluation of CYP3A activity in general adults and in patients with CKD with CYP3A5*1 allele, the usefulness of 4β‐OHC–4α‐OHC was not different from plasma 4β‐OHC concentration or 4β‐OHC/TC ratio. However, because of the limitations in study design and subject selection of this research, these findings require verification in further studies.

Study Highlights.

• WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

Plasma 4β‐hydroxycholesterol (4β‐OHC) concentration and 4β‐OHC/total cholesterol (TC) ratio are conventionally used in cross‐sectional evaluation of cytochrome P450(CYP)3A activity in study populations. Because 4β‐OHC is produced by hydroxylation by CYP3A4/5 and autoxidation of cholesterol, plasma 4β‐OHC concentration may increase unrelated to CYP3A activity. The 4α‐hydroxycholesterol (4α‐OHC) is produced by autoxidation of cholesterol, and not metabolized by CYP3A. Subtracting plasma 4α‐OHC concentration from plasma 4β‐OHC concentration (4β‐OHC–4α‐OHC) may provide more accurate assessment of CYP3A activity, but its usefulness has not been investigated in detail.

• WHAT QUESTION DID THIS STUDY ADDRESS?

Is 4β‐OHC–4α‐OHC useful for accurate CYP3A phenotyping in general adults or patients with chronic kidney disease (CKD) with CYP3A5*1 allele compared with plasma 4β‐OHC concentration or 4β‐OHC/TC ratio?

• WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

This study investigated the predictive value of the three CYP3A activity indices for the presence or absence of CYP3A5*1 allele in general adults and for CKD stage 3 or stage 4–5D in patients with CKD with CYP3A5*1 allele using receiver operating characteristic curve, used to compare the accuracy of diagnostic tests. In general adults and in patients with CKD with CYP3A5*1 allele, the usefulness of 4β‐OHC–4α‐OHC in cross‐sectional evaluation of CYP3A activity is not different from plasma 4β‐OHC concentration or 4β‐OHC/TC ratio.

• HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

The necessity of subtracting plasma 4α‐OHC concentration from plasma 4β‐OHC concentration is considered low in general adults and in patients with CKD with CYP3A5*1 allele.

INTRODUCTION

The cytochrome P450(CYP)3A subfamily is involved in the metabolism of diverse substrates of human P450 and plays an important role in the metabolism of ~40% of the drugs used in clinical setting. ¹ CYP3A activity varies widely among individuals due to not only environmental factors, including use of concomitant medications, but also genetic and physiological factors. ² Therefore, assessment of CYP3A activity in individual patients is important to ensure the efficacy of CYP3A substrate drugs and to reduce adverse effects.

For quantification of CYP3A activity, probe drugs such as midazolam, alprazolam, and triazolam are commonly used. Endogenous markers such as urinary 6β‐hydroxycortisol to cortisol ratio, metabolic clearance of cortisol, and 4β‐hydroxycholesterol (4β‐OHC) have also been proposed. ³ Of these biomarkers, 4β‐OHC is produced by hydroxylation at the 4β position of cholesterol by CYP3A4 and CYP3A5. ⁴ , ⁵ , ⁶ Because 4β‐OHC has a long half‐life and little circadian change, ⁷ single blood sampling can be used to evaluate CYP3A phenotypes. The 4β‐OHC is mainly metabolized slowly by CYP7A1, ⁵ and CYP7A1 activity is not affected by renal dysfunction. ⁸ Therefore, 4β‐OHC may be useful for the evaluation of CYP3A activity in infants, older individuals, and particularly in patients with renal dysfunction.

Plasma 4β‐OHC concentration and plasma 4β‐OHC concentration normalized by total cholesterol (4β‐OHC/TC ratio) are conventionally used to evaluate CYP3A activity using 4β‐OHC as an indicator. ⁹ , ¹⁰ , ¹¹ Plasma 4β‐OHC concentration has been shown to be reduced by inhibitor of CYP3A4, ¹² and elevated by inducers of CYP3A. ⁴ , ⁷ , ¹² In addition, a study using CYP3A‐humanized mice showed that plasma 4β‐OHC level and 4β‐OHC/TC ratio correlated positively with hepatic CYP3A4 protein levels. ⁶ These studies provided the rationale of using these endogenous biomarkers for assessing CYP3A activity. However, 4β‐OHC is produced not only by metabolism of cholesterol by CYP3A but also by autoxidation of cholesterol. ¹³ Hence, 4β‐OHC has been reported to increase in inappropriately stored samples. ¹³ Furthermore, Ikegami et al. ¹⁴ reported that serum 4β‐OHC concentrations in patients with chronic hepatitis C virus infection were significantly higher compared with healthy volunteers, and the elevated levels were significantly reduced by anti‐viral therapy compared to pre‐therapy levels. These results suggest that oxidative stress or inflammation in the liver promotes cholesterol autoxidation. Therefore, plasma 4β‐OHC concentration and 4β‐OHC/TC ratio when used as biomarkers may overestimate CYP3A activity in inappropriately stored samples or pathological conditions with increased oxidative stress or inflammation, because the 4β‐OHC produced by autoxidation is included.

On the other hand, 4α‐hydroxycholesterol (4α‐OHC) is a stereoisomer of 4β‐OHC produced by autoxidation of cholesterol ¹³ and is not affected by drugs that increase or decrease plasma 4β‐OHC concentration. ⁴ , ⁷ , ¹² The 4β‐OHC and 4α‐OHC are distributed abundantly in low‐density lipoprotein among plasma lipoproteins, ⁴ and in vitro oxidation of low‐density lipoprotein resulted in the formation of 4β‐OHC and 4α‐OHC at a ratio close to one. ¹⁵ Hence, we hypothesized that subtracting plasma 4α‐OHC concentration from plasma 4β‐OHC concentration (4β‐OHC–4α‐OHC) should allow more accurate CYP3A phenotyping by removing the effects of sample storage condition and oxidative stress or inflammation on plasma 4β‐OHC concentration. However, few studies have evaluated 4β‐OHC–4α‐OHC, and its usefulness has not been investigated in detail.

In this study, to reveal the usefulness of 4β‐OHC–4α‐OHC, plasma 4β‐OHC concentration, and 4β‐OHC/TC ratio in cross‐sectional evaluation of CYP3A activity, we conducted two evaluations. First, CYP3A5 shows gene polymorphisms, and carriers of at least one CYP3A5*1 (wild type) allele (CYP3A5*1/*1 or CYP3A5*1/*3) have higher CYP3A activity than non‐carriers (CYP3A5*3/*3). ¹⁶ , ¹⁷ , ¹⁸ Hence, we evaluated the accuracy of the three CYP3A activity indices in predicting the presence or absence of CYP3A5*1 allele in a sample of adults from the general population (general adults). Second, because elevated oxidative stress and inflammation are observed in patients with CKD, ¹⁹ , ²⁰ we considered that 4β‐OHC–4α‐OHC may provide more accurate CYP3A phenotyping than plasma 4β‐OHC concentration or 4β‐OHC/TC ratio especially in these patients. Studies have shown that CYP3A activity decreases as CKD pathology progresses. ²¹ , ²² , ²³ Thus, we evaluated the ability of the three indices of CYP3A activity in predicting CKD stage 3 or CKD stage 4–5D in patients with CKD. A previous study in patients with CKD revealed that CYP3A activity decreased with decrease in estimated glomerular filtration rate (eGFR) in CYP3A5*1 allele carriers, whereas the association between CYP3A activity and eGFR was not significant in non‐carriers. ²⁴ Therefore only carriers of CYP3A5*1 allele were analyzed among the patients with CKD.

METHODS

Subjects

In this observational study, we analyzed cross‐sectional data of participants in the Japan Multi‐Institutional Collaborative Cohort Study (J‐MICC study) ²⁵ that uses genetic and clinical data from general Japanese population to confirm and detect interactions of genes and environment associated with lifestyle‐related diseases. The participants in the J‐MICC study were aged 35–69 years, and were volunteers residing in the areas defined by local governmental administration, participants of health checkup examinations conducted by local governments, individuals attending health checkup centers, or individuals attending a cancer hospital. Those who could not complete the lifestyle questionnaire and those who did not accept follow‐up examinations were excluded. All participants in the J‐MICC study provided written informed consent, and the study protocol was approved by the ethics committees of institutions participating in the J‐MICC study. We randomly selected the data of 500 general adults who underwent a health examination at Kyoto Prefectural University of Medicine, and included 416 participants with body mass index (BMI) lower than 30 kg/m², eGFR higher than 60 mL/min/1.73 m², total bilirubin lower than 2.0 mg/dL, and alanine aminotransaminase (ALT) lower than 100 IU/L in the present study. The following participant characteristics and laboratory data were obtained from health examination records: sex, age, body weight, height, eGFR, total bilirubin, ALT, and TC. All plasma samples for quantifying plasma 4β‐OHC and 4α‐OHC concentrations were stored at −80°C until analysis.

We also analyzed 63 patients with CKD who attended the Department of Urology, Faculty of Medicine at Oita University, and met the following selection criteria: over 18 years of age; no use of concomitant drugs that strongly induce or inhibit CYP3A, ²⁶ such as rifampicin, carbamazepine, clarithromycin, and itraconazole; BMI lower than 30 kg/m², eGFR lower than 60 mL/min/1.73 m², total bilirubin lower than 2.0 mg/dL, and ALT lower than 100 IU/L. Blood sample was collected from a vein into blood sampling tube containing EDTA‐3 K as anticoagulant. Four hundred μL of whole blood was aliquoted for gene polymorphism measurement and the remaining blood sample was centrifuged at 2330 g for 5 min at 4°C. After centrifugation, the plasma was transferred into transparent polypropylene tube. Whole blood samples were stored at −40°C and plasma samples at −80°C until analysis. The following patient characteristics and laboratory data were obtained from electronic medical records: therapeutic drugs, sex, age, body weight, height, eGFR, total bilirubin, ALT, and TC. The scientific purpose and procedures of this study were explained to each patient, and each gave written informed consent.

BMI in general adults and patients with CKD were calculated as weight (kg) divided by square of height (m²). The eGFR was calculated using The Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) equation for Japanese. ²⁷

This study was conducted in compliance with the Declaration of Helsinki, and was approved by the ethics committee of Meiji Pharmaceutical University (approval numbers: 3023, 202026), the ethics committee of Kyoto Prefectural University of Medicine (approval number: ERB‐C‐1384), and the ethics committee of Oita University Hospital (approval number: 1925).

Genotyping procedures for CYP3A5

A total of 14,539 individuals from 12 regions in Japan participated in the J‐MICC study and were genotyped at RIKEN Center for Integrative Medicine using a Human OmniExpressExome‐8 version 1.2 BeadChip array (Illumina Inc.). ²⁸ Quality control filtering of samples and single nucleotide polymorphisms (SNPs) were conducted as reported previously. ²⁸ Briefly, the following samples were excluded: genotype call rate less than or equal to 0.99, contradictory sex information, one sample each of two pairs identified as closely related pair, and estimated to have ancestry other than Japanese. SNPs with genotype call rate less than or equal to 0.98 and/or Hardy–Weinberg equilibrium exact test p value less than 1.0 × 10⁻⁶ were also excluded. The quality control filtering extracted 14,091 participants and 570,162 SNPs. Among these, we obtained the genotype data of the above‐mentioned 416 participants. From the genotyped data that passed the quality control filtering, we determined the presence or absence of rs776746 (6986 A>G, CYP3A5*3). ¹⁷ When the CYP3A5*3 allele was not detected, the test allele was designated CYP3A5*1. Finally, participants were classified into CYP3A5*1 carriers (CYP3A5*1/*1 and CYP3A5*1/*3) and non‐carriers (CYP3A5*3/*3).

For patients with CKD, total DNA was extracted from 400 μL of whole blood sample using Maxwell 16 DNA Purification Kit (Promega, Tokyo, Japan). Each sample was analyzed for the SNP rs776746 (6986 A>G, CYP3A5*3). Allelic discrimination reaction was carried out using TaqMan genotyping assay (Thermo Fisher Scientific) with a LightCycler Nano System (Roche Applied Science, Penzberg, Germany). Based on the presence or absence of the CYP3A5*3 allele, the patients were classified into CYP3A5*1 carriers and non‐carriers.

Measurement of plasma 4β‐ and 4α‐hydroxycholesterol concentrations

Plasma 4β‐ and 4α‐OHC concentrations were measured according to our previous report. ²³ Briefly, 50 μL of plasma sample was saponified with 28% sodium methoxide methanol solution followed by liquid–liquid extraction, and the organic phase was evaporated to dryness under a N₂ gas stream. The residue was reconstituted with derivatization solution (250 mg of 2‐methyl‐6‐nitrobenzoic anhydride, 75 mg of 4‐dimethylamino‐pyridine, 200 mg of picolinic acid, 7.5 mL of pyridine, and 1 mL of triethylamine) followed by liquid–liquid extraction. The organic phase was evaporated to dryness under a N₂ gas stream, and the residue was reconstituted with acetonitrile. The internal standard was 4β‐OHC‐D₇, an isotope of 4β‐OHC. The pretreated sample was analyzed using ultra‐high performance liquid chromatography (LC) coupled to tandem mass spectrometry (MS/MS) system (Shimadzu, Kyoto, Japan) consisted of a Nexera X2 LC system and quadrupole mass spectrometer (LCMS‐8040). The 4β‐OHC, 4α‐OHC, and 4β‐OHC‐D₇ were measured in positive ion mode of electrospray ionization. The MS/MS transitions for the di‐picolinyl esters were m/z 613.3 → 490.5 for 4β‐OHC and 4α‐OHC, m/z 620.3 → 497.6 for 4β‐OHC‐D₇. The lower limit of quantification was 0.5 ng/mL for both 4β‐ and 4α‐OHC. The within‐batch accuracy was 92.1%–106.6% for 4β‐OHC and 89.9%–107.2% for 4α‐OHC. The within‐batch precision (% coefficient of variation) was below 9.9% for both 4β‐ and 4α‐OHC.

Data analyses

Comparison of three CYP3A activity indices and plasma 4α‐OHC concentration

For general adults, CYP3A5*1 carriers and non‐carriers were analyzed separately, and whether there were differences in the three CYP3A activity indices between CYP3A5*1 carriers and non‐carriers was examined. Because previous study in patients with CKD suggests that renal function affects CYP3A activity in CYP3A5*1 carriers but has little effect in non‐carriers, ²⁴ only patients with CKD with CYP3A5*1 allele were analyzed. In these patients, CKD stage 3 (30 ≤ eGFR < 60 mL/min/1.73 m²), and CKD stage 4–5D (eGFR < 30 mL/min/1.73 m²) were analyzed separately, and whether there were differences in the three CYP3A activity indices between CKD stage 3 and CKD stage 4–5D was examined. Continuous variables are expressed as mean ± standard deviation (SD) or median (interquartile range), depending on normality evaluated by Shapiro–Wilk test. Variables between groups were compared using Welch's t‐test or Mann–Whitney U test. Categorical variables are expressed as absolute number (%) and analyzed using Fisher's exact test. Statistical analyses were performed using Graph Pad Prism 8 (GraphPad Software) and p value less than 0.05 was considered statistically significant.

Comparison of three CYP3A activity indices by receiver operating characteristic curve analysis

Receiver operating characteristic (ROC) curve is used to compare the accuracy of diagnostic tests, and the larger the area under the ROC curve (AUC), the better is the test performance. ²⁹ ROC curves were created to evaluate the predictive value of plasma 4β‐OHC concentration, 4β‐OHC/TC ratio, and 4β‐OHC–4α‐OHC for the presence or absence CYP3A5*1 allele in general adults, or for CKD stage 3 or stage 4–5D in CKD patients with CYP3A5*1 allele. To evaluate the usefulness of 4β‐OHC–4α‐OHC compared with plasma 4β‐OHC concentration or 4β‐OHC/TC ratio, we assessed the AUC, cutoff value, sensitivity, and specificity obtained from the ROC curves. The AUC of 4β‐OHC–4α‐OHC was compared with the AUC of plasma 4β‐OHC concentration or 4β‐OHC/TC ratio using the DeLong test. ³⁰ Statistical analyses were performed using R software version 4.3.1 (http://www.r‐project.org) and p value less than 0.05 was considered statistically significant.

RESULTS

Subject characteristics

Table 1 shows the participant characteristics and laboratory data. Of the 416 general adults, 191 were CYP3A5*1 carriers and 225 were non‐carriers. Except for total bilirubin, there were no significant differences in participant characteristics and laboratory data between the CYP3A5*1 carriers and non‐carriers. Of the 63 patients with CKD, 26 were CYP3A5*1 carrier, 14 of whom were CKD stage 3 and 12 were CKD stage 4–5D. Participant characteristics and laboratory data except for eGFR were not significantly different between CKD stage 3 and stage 4–5D.

TABLE 1.

Demographic and laboratory data of subjects analyzed in the study.

	General adults		CKD patients with CYP3A5*1 allele
	CYP3A5*1 carrier	Non‐carrier	CKD stage 3	CKD stage 4‐5D
Number of subjects; n	191	225	14	12
Sex (male/female); n (%)	56 (29.3)/135 (70.7)	66 (29.3)/159 (70.7)	6 (42.9)/8 (57.1)	10 (83.3)/2 (16.7)
Age (years)	57.0 (46.5–64.0)	55.0 (48.0–66.0)	49.2 ± 13.1	52.4 ± 15.9
Body mass index (kg/m2)	21.8 (19.7–23.7)	21.6 (19.5–23.6)	20.2 ± 2.6	22.5 ± 3.7
Estimated glomerular filtration rate (mL/min/1.73 m2)	77.9 (71.9–86.5)	78.8 (73.6–85.1)	40.1 ± 5.3	13.2 ± 9.2*
Total bilirubin (mg/dL)	0.70 (0.60–0.90)	0.80 † (0.60–0.90)	0.65 ± 0.16	0.62 ± 0.27
Alanine aminotransferase (IU/L)	15.0 (12.0–21.0)	16.0 (12.0–20.0)	13.3 ± 4.2	12.9 ± 8.2
Total cholesterol (mg/dL)	210.0 (188.0–234.0)	218.0 (189.0–238.0)	223.9 ± 32.6	193.4 ± 55.7

Note: Data are expressed as numbers (%) for categorical variables, and mean ± standard deviation or median (interquartile range) for continuous variables. Sex was compared by Fisher's exact test. Normality of the data were tested by Shapiro–Wilk test. Parametric data were compared by Welch's t‐test. Nonparametric data were compared by Mann–Whitney U test.

Abbreviation: CKD, chronic kidney disease.

^* p < 0.05 versus CKD stage 3.

^† p < 0.05 versus CYP3A5*1 carrier.

CYP3A activity indices and plasma 4α‐OHC concentration in general adults and patients with CKD with CYP3A5*1 allele

Figure 1 shows plasma 4β‐OHC concentration, 4β‐OHC/TC ratio, 4β‐OHC–4α‐OHC, and plasma 4α‐OHC concentration in general adults divided into CYP3A5*1 carriers and non‐carriers. Plasma 4β‐OHC concentrations (median [interquartile range]) in CYP3A5*1 carriers and non‐carriers were, respectively, 36.6 (IQR: 28.0–46.9) and 31.2 (IQR: 22.1–39.8) ng/mL (p < 0.0001); 4β‐OHC/TC ratios were 0.18 (IQR: 0.14–0.22) and 0.14 (IQR: 0.11–0.18, p < 0.0001); and 4β‐OHC–4α‐OHC were 31.2 (IQR: 22.0–40.8) and 24.4 (IQR: 16.4–34.4) ng/mL (p < 0.0001). All three indices were significantly higher in CYP3A5*1 carriers than in non‐carriers. Plasma 4α‐OHC concentrations (median [interquartile range]) in CYP3A5*1 carriers and non‐carriers were 5.9 (IQR: 4.8–7.0) and 6.1 (IQR: 4.9–7.0) ng/mL, respectively, and no significant difference was detected (p = 0.32).

FIGURE 1.

Comparison of (a) plasma 4β‐OHC concentration (ng/mL), (b) 4β‐OHC/TC ratio, (c) 4β‐OHC–4α‐OHC (ng/mL), and (d) plasma 4α‐OHC concentration (ng/mL) between CYP3A5*1 carriers and non‐carriers in general adults. Data normality was analyzed by Shapiro–Wilk test, and all parameters were analyzed by Mann–Whitney U test. The horizontal line represents the median. 4α‐OHC, 4α‐hydroxycholesterol; 4β‐OHC, 4β‐hydroxycholesterol; CYP, cytochrome P450; TC, total cholesterol.

Figure 2 shows plasma 4β‐OHC concentration, 4β‐OHC/TC ratio, 4β‐OHC–4α‐OHC, and plasma 4α‐OHC concentration according to CKD stage in patients with CKD with CYP3A5*1 allele. Plasma 4β‐OHC concentrations (mean ± SD) in CKD stage 3 and stage 4–5D were, respectively, 49.3 ± 16.0 and 30.2 ± 15.6 ng/mL (p = 0.0053); 4β‐OHC/TC ratios were 0.22 ± 0.065 and 0.15 ± 0.047 (p = 0.0067); and 4β‐OHC–4α‐OHC were 42.1 ± 16.3 and 23.9 ± 14.8 ng/mL (p = 0.0064). All three indices were significantly lower in CKD stage 4–5D than in stage 3. Plasma 4α‐OHC concentrations (median [interquartile range]) in CKD stage 3 and stage 4–5D were 7.4 (IQR: 6.0–8.0) and 5.3 (IQR: 4.7–7.5) ng/mL, respectively, with no significant difference between the two groups (p = 0.12).

FIGURE 2.

Comparison of (a) plasma 4β‐OHC concentration (ng/mL), (b) 4β‐OHC/TC ratio, (c) 4β‐OHC–4α‐OHC (ng/mL), (d) plasma 4α‐OHC concentration (ng/mL) between CKD stage 3 and stage 4–5D in patients with CKD with CYP3A5*1 allele. Data normality was analyzed by Shapiro–Wilk test. Plasma 4α‐OHC concentration (ng/mL) was analyzed by Mann–Whitney U test, and the horizontal line represents the median. The other parameters were analyzed by Welch's t‐test, and the horizontal line represents the mean. 4α‐OHC, 4α‐hydroxycholesterol; 4β‐OHC, 4β‐hydroxycholesterol; CKD, chronic kidney disease; TC, total cholesterol.

Comparison of three CYP3A activity indices by receiver operating characteristic curve analysis

Figure 3 shows the ROC curves of the three CYP3A activity indices for the prediction of CYP3A5*1 allele status in general adults. Figure 4 shows the ROC curves of the three CYP3A activity indices for the prediction of CKD stage (3 or 4–5D) in patients with CKD with CYP3A5*1 allele. Table 2 shows the AUC (95% confidence interval [CI]), optimal cutoff values, sensitivity, and specificity of the ROC curves for the three CYP3A activity indices in general adults and in patients with CKD with CYP3A5*1 allele.

FIGURE 3.

Receiver operating characteristic curves of CYP3A activity indices for predicting presence or absence of CYP3A5*1 allele in general adults. 4α‐OHC, 4α‐hydroxycholesterol; 4β‐OHC, 4β‐hydroxycholesterol; TC, total cholesterol.

FIGURE 4.

Receiver operating characteristic curves of CYP3A activity indices for predicting CKD stage 3 or stage 4–5D in patients with CKD with CYP3A5*1 allele. 4α‐OHC, 4α‐hydroxycholesterol; 4β‐OHC, 4β‐hydroxycholesterol; CKD, chronic kidney disease; TC, total cholesterol.

TABLE 2.

Validity of three CYP3A activity indices for predicting presence of CYP3A5*1 allele in general adults and CKD stage 3 or stage 4–5D in patients with CKD with CYP3A5*1 allele.

	AUC (95% CI)	Cut‐off value	Sensitivity	Specificity	p value
General adults
Plasma 4β‐OHC concentrations (ng/mL)	0.63 (0.58–0.68)	36.5	0.52	0.69	0.63
4β‐OHC/TC	0.66 (0.60–0.71)	0.17	0.57	0.69	0.025
4β‐OHC – 4α‐OHC (ng/mL)	0.63 (0.58–0.69)	23.6	0.73	0.48	–
CKD patients with CYP3A5*1 allele
Plasma 4β‐OHC concentrations (ng/mL)	0.80 (0.62–0.99)	43.9	0.71	0.91	0.33
4β‐OHC/TC	0.78 (0.59–0.97)	0.19	0.71	0.83	0.96
4β‐OHC–4α‐OHC (ng/mL)	0.79 (0.59–0.98)	31.9	0.79	0.83	–

Abbreviations: 4α‐OHC, 4α‐hydroxycholesterol; 4β‐OHC, 4β‐hydroxycholesterol; AUC, area under the curve; CI, confidence interval; CKD, chronic kidney disease; TC, total cholesterol.

p value: versus 4β‐OHC–4α‐OHC.

In general adults, the AUC of the ROC curve for plasma 4β‐OHC concentration was 0.63 (95% CI: 0.58–0.68), that for 4β‐OHC/TC ratio was 0.66 (95% CI: 0.60–0.71), and that for 4β‐OHC–4α‐OHC was 0.63 (95% CI: 0.58–0.69). There was no significant difference in AUC between 4β‐OHC–4α‐OHC and plasma 4β‐OHC concentration (p = 0.63). On the other hand, the AUC of 4β‐OHC–4α‐OHC was significantly smaller than that of 4β‐OHC/TC ratio (p = 0.025). Cutoff value, sensitivity, specificity for plasma 4β‐OHC concentration were 36.5 ng/mL, 0.52, and 0.69; those for 4β‐OHC/TC ratio were 0.17, 0.57, and 0.69; those for 4β‐OHC–4α‐OHC were 23.6 ng/mL, 0.73, and 0.48.

In patients with CKD with CYP3A5*1 allele, the AUC of the ROC curve for plasma 4β‐OHC concentration was 0.80 (95% CI: 0.62–0.99), that for 4β‐OHC/TC ratio was 0.78 (95% CI: 0.59–0.97), and that for 4β‐OHC–4α‐OHC was 0.79 (95% CI: 0.59–0.98). There was no significant difference in AUC between 4β‐OHC–4α‐OHC and plasma 4β‐OHC concentration (p = 0.33) or 4β‐OHC/TC ratio (p = 0.96). Cutoff value, sensitivity, and specificity for plasma 4β‐OHC concentration were 43.9 ng/mL, 0.71, and 0.91; those for 4β‐OHC/TC ratio were 0.19, 0.71, and 0.83; those for 4β‐OHC–4α‐OHC were 31.9 ng/mL, 0.79, and 0.83.

DISCUSSION

In this study, to determine whether correction of plasma 4β‐OHC concentration by subtracting plasma 4α‐OHC concentration is useful for more accurate assessment of CYP3A activity, we compared 4β‐OHC–4α‐OHC with plasma 4β‐OHC concentration and 4β‐OHC/TC ratio that are conventionally used to assess CYP3A activity. We investigated the predictive value of the three CYP3A activity indices for the presence or absence of CYP3A5*1 allele in general adults and for CKD stage 3 or stage 4–5D in patients with CKD with CYP3A5*1 allele using ROC curve analysis. To the best of our knowledge, this is the first report comparing 4β‐OHC–4α‐OHC with plasma 4β‐OHC concentration and 4β‐OHC/TC ratio in general adults and patients with CKD with CYP3A5*1 allele.

Plasma 4β‐OHC concentration and 4β‐OHC/TC ratio have been used to explore the inducibility of CYP3A activity by novel therapeutics, ⁹ , ¹⁰ , ¹¹ but changes in these endogenous indices have been reported to be small compared to change in the AUC of midazolam, ⁹ , ¹⁰ a standard probe for CYP3A phenotyping. Plasma 4β‐OHC concentration has been reported to increase independent of CYP3A activity under conditions of inappropriate sample storage ¹³ and increased inflammation or oxidative stress due to hepatitis. ¹⁴ Thus, plasma 4β‐OHC concentration and 4β‐OHC/TC ratio when used as biomarkers may overestimate CYP3A activity. On the other hand, 4α‐OHC, a stereoisomer of 4β‐OHC, is produced from cholesterol nonenzymatically, ¹³ and is not metabolized by CYP3A. ⁴ , ⁷ , ¹² Although the ratio of 4β‐OHC to 4α‐OHC produced in vivo is unknown, oxidation of low‐density lipoprotein in vitro produces both 4β‐OHC and 4α‐OHC at a ratio close to one. ¹⁵ Hence, we hypothesized that subtracting plasma 4α‐OHC concentration from plasma 4β‐OHC concentration should provide more accurate assessment of CYP3A activity by controlling for the effect of 4β‐OHC autoxidation.

In general adults, the three CYP3A activity indices were significantly higher in CYP3A5*1 carriers compared with non‐carriers (Figure 1a–c). This result is consistent with previous studies using CYP3A substrate drug and endogenous marker. ¹⁸ , ³¹ Therefore, the three CYP3A activity indices may be useful for assessing the effect of CYP3A5 gene polymorphisms on CYP3A activity. Next, using ROC curves to predict the presence or absence CYP3A5*1 allele in general adults, we compared the predictive value of 4β‐OHC–4α‐OHC with that of plasma 4β‐OHC concentration or 4β‐OHC/TC ratio. There was no significant difference in the AUC between 4β‐OHC–4α‐OHC and plasma 4β‐OHC concentration and the AUC of 4β‐OHC/TC ratio was significantly larger than that of 4β‐OHC–4α‐OHC (Table 2, Figure 3). In addition, sensitivity of 4β‐OHC–4α‐OHC was higher than that of plasma 4β‐OHC concentration and 4β‐OHC/TC ratio, but specificity of 4β‐OHC–4α‐OHC was not high (Table 2). Hence, in general adults, the usefulness of 4β‐OHC–4α‐OHC in evaluating CYP3A activity may not be higher than plasma 4β‐OHC concentration and 4β‐OHC/TC ratio. Because the collection, storage, and pretreatment of samples were performed in a similar manner in CYP3A5*1 carriers and non‐carriers, the degree of autoxidation due to sample management was unlikely to differ markedly between the two groups. In addition, there was no significant difference in plasma 4α‐OHC concentration between CYP3A5*1 carriers and non‐carriers (Figure 1d). These results thus indicate that there is no difference between the two groups in the degree of increase in apparent plasma 4β‐OHC concentration due to nonenzymatically production in general adults.

In patients with CKD with CYP3A5*1 allele, the three CYP3A activity indices were significantly lower in CKD stage 4–5D compared with CKD stage 3 (Figure 2a–c). This result is consistent with our previous study that CYP3A activity decreased with decrease in eGFR in patients with CKD carrying the CYP3A5*1 allele, ²⁴ and suggests that all three indices are useful for the assessment of CYP3A activity during renal dysfunction. Next, using ROC curves to predict CKD stage 3 or stage 4–5D in patients with CKD with CYP3A5*1 allele, we compared the predictive value of 4β‐OHC–4α‐OHC with that of plasma 4β‐OHC concentration or, 4β‐OHC/TC ratio. There was no significant difference in the AUC between 4β‐OHC–4α‐OHC and plasma 4β‐OHC concentration or 4β‐OHC/TC ratio (Table 2, Figure 4). The sensitivity of 4β‐OHC–4α‐OHC was higher than that of plasma 4β‐OHC concentration and 4β‐OHC/TC ratio, but specificity of 4β‐OHC–4α‐OHC was the same or lower (Table 2). Therefore, in patients with CKD with CYP3A5*1 allele, 4β‐OHC–4α‐OHC would not be superior to plasma 4β‐OHC concentration and 4β‐OHC/TC ratio in evaluating CYP3A activity. Because the process of collection, storage, and pretreatment of samples is similar in patients with CKD stage 3 and those with CKD stage 4–5D, the degree of autoxidation due to sample management is considered to be equivalent in the two groups. Although, it has been reported that oxidative stress increase with deterioration of renal function, ²⁰ there was no significant difference in plasma 4α‐OHC concentration between CKD stage 3 and stage 4–5D (Figure 2, Table 1). Thus, these results suggest that there are no differences between two groups in the degree of increase in apparent plasma 4β‐OHC concentration due to nonenzymatically production in patients with CKD with CYP3A5*1 allele.

This study has some limitations. First, although eGFR, BMI, total bilirubin, and ALT were used as selection criteria, general adults included individuals with various diseases, such as hypertension, diabetes mellitus, and dyslipidemia. The effect of disease on CYP3A activity was not considered in this study. Further, because the medications taken by individual subjects were unknown, the effect of drugs that induce or inhibit CYP3A activity ²⁶ cannot be ruled out. The AUC of the three ROC curves for predicting the presence or absence of CYP3A5*1 allele in general adults were ~0.6 (Table 2), indicating low prediction accuracy. ²⁹ It is possible that the prediction accuracy of the presence or absence of CYP3A5*1 allele may improve by considering the effects of disease and medications taken. Second, the sample size of patients with CKD with CYP3A5*1 allele may be too small to detect differences between indices of CYP3A activity. Hence, the present results should be interpreted with caution, and large‐scale clinical studies are needed to assess the usefulness of 4β‐OHC–4α‐OHC in patients with CKD. Third, there was no significant difference in plasma 4α‐OHC concentration between CKD stage 3 and stage 4–5D (p = 0.12; Figure 2d). Thus, the usefulness of 4β‐OHC–4α‐OHC in individuals with significantly different plasma 4α‐OHC concentrations could not be evaluated. Note that CYP3A activity indices were not compared between general adults and patients with CKD, because general adults included patients with various diseases. To evaluate the usefulness of 4β‐OHC–4α‐OHC among subjects with a wide range of plasma 4α‐OHC concentrations, further studies need be conducted in healthy volunteers and patients with various CKD stages or pathological conditions in which oxidative stress is elevated by inflammation.

This study indicated that the usefulness of 4β‐OHC–4α‐OHC is not different from plasma 4β‐OHC concentration or 4β‐OHC/TC ratio when used in cross‐sectional evaluation of CYP3A activity in general adults or in patients with CKD with CYP3A5*1 allele. However, as mentioned above, the present results should be interpreted with caution due to the lack of information on diseases and concomitant medications in general adult subjects, and the small sample size of patients with CKD with CYP3A5*1 allele. Furthermore, the usefulness of 4β‐OHC–4α‐OHC in groups with different plasma 4α‐OHC concentrations requires further studies in healthy volunteers and individuals with various pathological conditions causing increased oxidative stress.

In conclusion, in general adults and patients with CKD with CYP3A5*1 allele, the usefulness of 4β‐OHC–4α‐OHC in cross‐sectional evaluation of CYP3A activity is not higher than plasma 4β‐OHC concentration or 4β‐OHC/TC ratio, and the necessity of correcting plasma 4β‐OHC concentration by plasma 4α‐OHC concentration is considered low. Further large‐scale studies are required to evaluate the usefulness of 4α‐OHC in evaluating CYP3A activity.

AUTHOR CONTRIBUTIONS

A.O., Y.S., T.K., R.Tan., T.S., K.W., K.M., H.I., and K.O. designed the research. A.O., Y.S., H.S., T.K., M.N., Y.M., R.Tan., H.O., R.Tat., and T.A. performed the research. A.O., Y.S., T.K., R.Tan., and K.O. analyzed the data. A.O., Y.S., T.K., R.Tan., T.S., and K.O. wrote the manuscript.

FUNDING INFORMATION

This study was supported by Grants‐in‐Aid for Scientific Research in Priority Areas of Cancer (No. 17015018) and Innovative Areas (No. 221S0001), and by the Japan Society for the Promotion of Science (JSPS) KAKENHI Grant (Nos. 16H06277 and 22H04923 [CoBiA]) from the Japanese Ministry of Education, Culture, Sports, Science and Technology. This work was also supported in part by funding for the BioBank Japan Project from the Ministry of Education, Culture, Sports, Science and Technology from April 2003 to March 2015, and the Japan Agency for Medical Research and Development since April 2015.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

Oda A, Suzuki Y, Sato H, et al. Evaluation of the usefulness of plasma 4β‐hydroxycholesterol concentration normalized by 4α‐hydroxycholesterol for accurate CYP3A phenotyping. Clin Transl Sci. 2024;17:e13768. doi: 10.1111/cts.13768