Effect of isotretinoin on CYP2D6 and CYP3A activity in patients with severe acne

Abstract

Aims:

Previously, retinoids have decreased CYP2D6 mRNA expression in vitro and induced CYP3A4 in vitro and in vivo. This study aimed to determine whether isotretinoin administration changes CYP2D6 and CYP3A activities in patients with severe acne.

Methods:

Thirty-three patients (22 females and 11 males, 23.5 ± 6.0 years old) expected to receive isotretinoin treatment completed the study. All participants were genotyped for CYP2D6 and CYP3A5. Participants received dextromethorphan (DM) 30 mg orally as a dual-probe substrate of CYP2D6 and CYP3A activity at two study timepoints: pre-isotretinoin treatment and with isotretinoin for at least 1 week. The concentrations of isotretinoin, DM and their metabolites were measured in 2-h postdose plasma samples and in cumulative 0–4-h urine collections using liquid chromatography-mass spectrometry.

Results:

In CYP2D6 extensive metabolizers, the urinary dextrorphan (DX)/DM metabolic ratio (MR) (CYP2D6 activity marker) was numerically, but not significantly, lower with isotretinoin administration compared to pre-isotretinoin (geometric mean ratio [GMR] [90% confidence interval (CI)] 0.78 [0.55, 1.11]). The urinary 3-hydroxymorphinan (3HM)/DX MR (CYP3A activity marker) was increased (GMR 1.18 [1.03, 1.35]) and the urinary DX-O-glucuronide/DX MR (proposed UGT2B marker) was increased (GMR 1.22 [1.06, 1.39]) with isotretinoin administration compared to pre-isotretinoin.

Conclusions:

Administration of isotretinoin did not significantly reduce CYP2D6 activity in extensive metabolizers, suggesting that the predicted downregulation of CYP2D6 based on in vitro data does not translate into humans. We observed a modest increase in CYP3A activity (predominantly CYP3A4) with isotretinoin treatment. The data also suggest that DX glucuronidation is increased following isotretinoin administration.

1 |. INTRODUCTION

Cytochrome P450 2D6 (CYP2D6) is responsible for the metabolism of approximately 20–25% of commonly used drugs, although it only accounts for 2–4% of total liver CYP content.^1–4 The disposition of CYP2D6 substrate drugs shows high inter-individual variability due to genetic polymorphisms.⁵ In addition, CYP2D6-mediated metabolism varies due to differences in protein expression.⁶ A positive correlation exists between CYP2D6 copy number and mRNA levels in human liver tissue.⁷ CYP2D6 mRNA expression in the liver also varies between individuals with the same genotype, suggesting transcriptional regulation of CYP2D6.^8,9 The mechanisms of CYP2D6 transcriptional regulation have not been well defined. Previous studies have shown that CYP2D6 is generally not inducible by common xenobiotic inducers.^10,11 In human hepatocytes and CYP2D6 humanized mice, CYP2D6 activity is regulated by hepatocyte nuclear factor (HNF) 4α.^12,13 Small heterodimer partner (SHP) acts as a transcriptional repressor for CYP2D6 expression by interacting with HNF4α and repressing CYP2D6 transcription.^7,14

Isotretinoin (13-cis-retinoic acid [13-cisRA]) is a US Food and Drug Administration (FDA)-approved treatment for severe, recalcitrant, nodulocystic acne.¹⁵ Isotretinoin, its isomer all-trans-retinoic acid (atRA), and its active metabolite 4-oxo-13-cis-retinoic acid (4-oxo-13-cisRA) increase SHP mRNA expression and decrease CYP2D6 mRNA expression in human hepatocytes.¹⁶ However, in mice, treatment with 13-cisRA and atRA had no impact on Shp mRNA expression, while 4-oxo-13-cisRA increased the mRNA expression of Shp and mouse Cyp2d.¹⁶ In contrast, in CYP2D6 humanized mice, treatment with atRA induced SHP and downregulated CYP2D6 expression.¹⁴ The increase in SHP mRNA expression likely resulted in reduced CYP2D6 expression via repression of HNF4α activity.¹⁴ Based on human hepatocyte studies, a 35–77% decrease in CYP2D6 expression was predicted with combined effects of 13-cisRA, atRA and 4-oxo-13-cisRA with isotretinoin 20 mg twice daily dosing for 20 weeks.¹⁶ However, using dextromethorphan (DM) as a CYP2D6 probe substrate, no decrease in CYP2D6 activity was observed in healthy male CYP2D6 extensive metabolizers (EM) (n = 8) receiving isotretinoin 40 mg twice-daily dosing for 2 weeks.¹⁶ The mean dextrorphan/DM (DX/DM) under the concentration-time curve (AUC) ratios and formation clearance of DX were instead increased, indicating CYP2D6 was potentially weakly induced in vivo.¹⁶ This study only recruited healthy adult males and may not have fully captured the role of retinoids in regulating CYP2D6 activity.¹⁶ For example, females may have different CYP2D6 regulation as Cyp2d22 wild-type female mice have higher Cyp2d22 hepatic mRNA expression at the oestrous phase of the oestrous cycle compared to male mice and ovariectomy decreased Cyp2d22 mRNA expression in female mice.¹⁷ Previous studies have reported that the DM/DX urinary metabolic ratio (MRs) was significantly lower in healthy female CYP2D6 EMs compared to males.^18,19

Retinoids have also been shown to be in vitro inducers of cytochrome P450 3A4 (CYP3A4) via activation of pregnane X receptor (PXR)¹⁶ and/or constitutive androstane receptor (CAR).²⁰ Previously, isotretinoin administration resulted in a weak CYP3A4 induction in vivo in humans based on the formation clearance of 3-methoxymorphinan (3MM) from DM and 6βOH-cortisol formation as an endogenous hepatic CYP3A4 biomarker.¹⁶ However, this finding requires further confirmation. Our study aimed to test whether isotretinoin regulates CYP2D6 and CYP3A activity in male and female patients with severe acne using DM as a dual-probe substrate in vivo.

2 |. METHODS

2.1 |. Study participants

The study was registered on clinicaltrials.gov (NCT03076021) and approved by the University of Washington Investigational Review Board. All participants or parents of minors provided written informed consent. This study was conducted at the University of Washington. The study was designed to detect a 25% change in DX/DM MR with α = 0.05 and 80% power. Thirty-three patients (22 females and 11 males) with severe acne and expected to be treated with isotretinoin were recruited. The participants were ≥12 years old. Exclusion criteria included body weight <80 lbs or body mass index ≥30 kg/m²; history of allergic or adverse reactions to isotretinoin, vitamin A, or DM; pregnancy; inability to follow isotretinoin risk evaluation and mitigation strategy (iPLEDGE) requirements; had chronic or persistent cough accompanying asthma, smoking or COPD; had a productive cough, fever, kidney disease, liver disease, diabetes, mental illness requiring treatment; taking medications or supplements known to interact with DM or CYP2D6 activity or taking drugs known to increase the risk of adverse effects from DM within 14 days.

2.2 |. Study design and drug administration

Each participant received DM 30 mg orally on each of the two study timepoints. The first study timepoint was the control, which occurred prior to the administration of isotretinoin. Participants then took an average of 1.01 ± 0.20 mg/kg/day of isotretinoin per clinical care, with a median of 66 days of treatment (range 10–194 days), before study timepoint 2. Patients self-administered and self-monitored isotretinoin throughout the treatment period. In the 3 days leading up to study timepoint 2, participants documented the precise times they took isotretinoin. On both study timepoints, blood samples were collected (into foil-wrapped EDTA tubes and immediately placed on wet ice) 2 h post-DM administration and approximately 1 h after the start of a nonstandardized meal. Plasma was separated by centrifugation (10 min at 3000g) and stored at −80°C until analysis. Laboratory lights were covered with UV fluorescent light filter sleeves. Plasma was used for quantification of DM, DX, 3-hydroxymorphinan (3HM), 3MM, dextrorphan-O-glucuronide (DX-O-glucuronide), atRA, 13-cisRA, 4-oxo-13-cisRA and retinol (ROH) concentrations. Cumulative urine was collected 0–4 h post-DM administration on both study timepoints. The urine was kept refrigerated until the collection was complete, then all urine was combined, volume measured and aliquots stored at −80°C until analysis. Urine samples were used to quantify DM, DX, 3HM, 3MM and DX-O-glucuronide.

2.3 |. CYP2D6 and CYP3A5 genotyping

Buccal swabs were collected for genomic DNA extraction using Qiagen QIAamp DNA Mini Kit (Germantown, MD, USA). CYP2D6 and CYP3A5 assays were performed on a StepOnePlus instrument following the recommended protocols from the manufacturer (Applied Biosystems, Waltham, MA, USA). CYP2D6 copy number was determined and samples were genotyped for nine core single nucleotide polymorphisms (SNPs), CYP2D6*2, *3, *4, *6, *9, *10, *17, *35, and *41, using Taqman allelic discrimination assays (Invitrogen, Carlsbad, CA, USA). CYP2D6 genotypes and activity scores (ASs) were determined using previously published genotyping strategy and methods.^21–23 The CYP2D6 ASs were calculated by the summation of values that were assigned to each allele. CYP2D6 metabolizer phenotypes were assigned with the corresponding AS values: poor (AS = 0), intermediate (AS = 0.5), extensive (AS = 1–2) and ultrarapid (AS > 2). Samples were also genotyped for three non-functional SNPs of CYP3A5, CYP3A5*3, *6, and *7, using TaqMan allelic discrimination assays. Combined CYP3A5 allelic status for each of these three SNPs (a CYP3A5*3/*6/*7 metabolic composite) was used for characterizing extensive, intermediate and poor metabolizers of CYP3A5 as described previously.^24,25 The wild-type functional allele is CYP3A5*1. CYP3A5 extensive, intermediate and poor metabolizers were identified with two, one and zero functional alleles, respectively.

2.4 |. Quantification of retinoids, DM and metabolites in plasma and urine

All-trans retinol and atRA were purchased from Millipore Sigma (Burlington, MA, USA). The 13-cisRA was purchased from Toronto Research Chemicals (Toronto, Canada). The atRA-d₅ and 13-cisRA-d₅ were purchased from Cayman Chemical (Ann Arbor, MI, USA) and retinol-d₆ was purchased from Cambridge Isotopes Laboratories (Tewksbury, MA, USA). The 4-oxo-13-cisRA and 4-oxo-13-cisRA-d₃ were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Mass-spectrometry grade formic acid, water, methanol and acetonitrile were purchased from Fisher Scientific (Pittsburgh, PA, USA). DM, DX, 3-HM, 3-MM, dextromethorphan-d₃ and dextrorphan-d₃ were purchased from Sigma Aldrich (St. Louis, MO, USA).

Plasma retinoids were analysed and quantified using an AB Sciex 6500 QTRAP mass spectrometer (Framingham, MA, USA) coupled with Agilent 1290 Infinity II liquid chromatograph (Santa Clara, CA, USA) operated on atmospheric pressure chemical ionization positive ion mode using previously validated mAgilentethods.^26–28 Plasma DM and its metabolites (DX, 3MM and 3HM) were quantified as previously described.²¹ To quantify urine DM, DX, 3MM and 3HM, urine samples along with calibration curves and quality controls were diluted 1:10 with water and protein precipitated with an equal amount of ice-cold acetonitrile containing internal standards of DM-d₃ and DX-d₃. To quantify DX-O-glucuronide both in plasma and urine, samples were diluted by 1:1000 with water and protein precipitated with an equal amount of ice-cold acetonitrile containing internal standards of DX-d₃. The processed samples were analysed using an AB Sciex 6500 QTRAP mass spectrometer coupled with an Agilent 1290 Infinity II liquid chromatograph and equipped with a Kinetex EVO C18 column (2.1 × 100 mm, 2.6 μm) (Torrence, CA, USA). The acquisition method was as previously described^16,21 with minor modifications. Mass spectrometry (MS)/MS transitions monitored for the analytes were DM, m/z 272 > 215; DX, m/z 258 > 157; DX-O-glucuronide, m/z 434 > 258; 3HM, m/z 244 > 157; 3MM, m/z 258 > 171; DM-d₃, m/z 275 > 215; DX-d₃, m/z 261 > 157. The lower limits of quantification (LLOQ) for DM, DX, DX-O-glucuronide, 3HM and 3MM were 1, 1, 90, 3 and 0.33 nM in plasma and 1, 150, 9000, 45 and 6 nM in urine, respectively.

The liquid chromatography-MS/MS data were quantified using MultiQuant 3.0.3. All analyte concentrations were calculated by peak area ratios of the analytes to internal standards. The calibration curves for each analyte were fitted by linear regression and weighed by 1/x, with the exception of DX-O-glucuronide, which was weighted by 1/x². FDA Bioanalytical Guidance was used to refer to bioanalytical acceptance criteria.²⁹

2.5 |. Data and statistical analysis

The participants’ demographics are expressed as mean and standard deviation (SD). Retinoid concentrations in plasma are reported as geometric means and 90% confidence intervals (90% CIs). The geometric mean ratios (GMRs) and their 90% CIs were used to compare the changes in retinoid concentrations with isotretinoin treatment (study timepoint 2) to endogenous concentrations (study timepoint 1). The GMR of treatment/control is the geometric mean of each participant’s study timepoint 2 retinoid concentration over the study timepoint 1 retinoid concentration.

DX/DM and DX-O-glucuronide/DX MRs were calculated by dividing the molar concentration (plasma) or molar quantity (urine) of the metabolite by that of the parent. DX plus DX-O-glucuronide/DM urinary MR was calculated by dividing the sum of the molar quantity of metabolites by that of the parent. The 3MM/DM and 3HM/DX MRs in urine were calculated by molar quantity of metabolite by that of parent. Urine samples with concentrations above or within 5% of the LLOQ were accepted, and concentrations lower than that were excluded from data analysis. DX/DM and DX plus DX-O-glucuronide/DM MRs were analysed only in CYP2D6 EMs (AS = 1–2). Participants with all ASs were included in the analysis of 3MM/DM, 3HM/DX and DX-O-glucuronide/DX MRs. The GMR and 90% CI were used to compare plasma or urinary MRs between study timepoint 1 (pre-treatment control) and study timepoint 2 (with isotretinoin treatment). The GMR of treatment/control MR is the geometric mean of each participant’s study timepoint 2 MR over study timepoint 1 MR. The 90% CI of the GMR for the MR with isotretinoin treatment over pre-treatment control was compared to the bioequivalence range of 0.8 and 1.25 according to FDA Guidance of Clinical Drug Interaction Studies.³⁰ If the 90% CI does not include 1, there is an indication of a potential drug interaction. A non-parametric Mann-Whitney test was performed to evaluate sex differences in CYP2D6 activity pre-isotretinoin and with isotretinoin treatment among CYP2D6 EMs. The non-parametric Mann-Whitney test was also used to assess the differences in urinary 3HM/DX MRs between CYP3A5 genotypes pre-isotretinoin and with isotretinoin treatment. A P value <.05 was considered statistically significant. Statistical analyses were all conducted on GraphPad Prism 9.5.1 (GraphPad Software, Inc., La Jolla, CA, USA).

To evaluate the relationship between plasma retinoid concentrations (atRA, 13-cisRA and 4-oxo-13-cisRA) and urinary DX/DM, 3MM/DM and 3HM/DX MRs in CYP2D6 EMs, a linear mixed-effect regression analysis was conducted by treating two repeated measurements (pre-treatment control and with isotretinoin treatment) from each participant as a cluster. To conduct the analyses, both plasma retinoid concentrations and urinary MRs were log-transformed. The mixed effect modelling was only applied to the intercept, allowing the baseline urinary MRs to differ across participants, while the slope estimate (ie, the relationship between urinary MRs and plasma retinoid concentrations) was derived using fixed effect only as the population means. All analyses were performed using the NLME package in R Studio (R version 4.1.2) (R Studio, Boston, MA, USA).³¹

3 |. RESULTS

3.1 |. Demographics

Thirty-six participants were enrolled and 33 (22 females and 11 males; 25 White, seven Asian and one Black) completed the study (Supporting Information Table S1). The mean (± standard deviation [SD]) age was 23.5 ± 6.0 years, weight 64.6 ± 15.4 kg and height 168.5 ± 10.9 cm. Two participants had a CYP2D6 AS of 0, one had an AS of 0.5, five had an AS of 1.0, five had an AS of 1.5, 17 had an AS of 2.0, one had an AS of 2.0–2.5 and two had an AS of 3.0. Twenty-six participants were CYP3A5 poor metabolizers with CYP3A5 *3/*3 genotype. Six participants were CYP3A5 intermediate metabolizers with CYP3A5 *1/*3 genotype. One participant was a CYP3A5 EM with CYP3A5 *1/*1 genotype.

3.2 |. Retinoid concentrations

The geometric mean and 90% CI of retinoid concentrations on both study timepoints are reported in Table 1. Administration of isotretinoin increased plasma retinoid (except retinol) concentrations on study timepoint 2 compared to study timepoint 1. The geometric mean of the steady-state concentration values for atRA, 13-cisRA and 4-oxo-13-cisRA measured at study timepoint 2 (2-h post-dose) were 41.1, 1085 and 2385 nM, respectively. The average increases in atRA, 13-cisRA and 4-oxo-13-cisRA concentrations were 17.7-, 535.3- and 558.5-fold at study timepoint 2 compared to study timepoint 1. Retinol concentrations decreased on average by 17% at study timepoint 2 compared to study timepoint 1.

TABLE 1.

Retinoid concentrations prior to treatment at study timepoint 1 (control) and with isotretinoin treatment at study timepoint 2 (treatment) for all participants (n = 33)

Retinoids	Study timepoint 1 Controla	Study timepoint 2 Treatmenta	Treatment/control GMRb (90% CI)
Retinol (μM)	1.5 (1.4–1.6)	1.2 (1.1–1.3)	0.8 (0.7–0.9)
atRA (nM)	2.8 (2.5–3.0)	41.1 (34.1–49.6)	14.9 (12.5–17.7)
13-cisRA(nM)	2.7 (2.3–3.2)	1085 (904–1301)	397.3 (311.4–506.8)
4-Oxo-13-cisRA (nM)	6.7 (5.2–8.7)	2385 (1912–2974)	356.2 (254.7–498.3)

Note: Retinoid concentrations were measured 2 h post dextromethorphan dose and approximately 1 h after a non-standardized meal.

Abbreviations: 4-oxo-13cisRA, 4-oxo-13-cis-retinoic acid; atRA, 13cisRA, 13-cis-retinoic acid; all-trans-retinoic acid; CI, confidence interval.

^aStudy timepoints 1 and 2 data were reported as geometric mean (90th percentile CI).

^bTreatment/control geometric mean ratio with 90th percentile CI were reported.

3.3 |. DM MRs

The 2-h post-dose plasma and 0–4-h post-dose cumulative urine MRs of DX/DM were used as in vivo markers of CYP2D6 activity (Figure 1). The 0–4-h post-dose cumulative urine 3HM/DX MR was used as an in vivo marker of CYP3A activity. The geometric mean and 90% CI of plasma and urine MRs on both study timepoints, as well as the GMR of treatment/control with 90% CI, are shown in Table 2.

FIGURE 1.

Dextromethorphan (DM) metabolic pathway. Surrogate markers for CYP2D6 (dextrorphan [DX]/DM metabolic ratio [MR]) and CYP3A (3-methoxymorphinan [3MM]/DM and 3-hydroxymorphinan [3HM]/DX MRs) are shown. CYP2D6, cytochrome P450 2D6; CYP3A, cytochrome P450 3A; UGT2B, UDP-glucuronosyltransferase (UGT) 2B.

TABLE 2.

Dextromethorphan and metabolite metabolic ratios pre and with isotretinoin treatment

	Number of participants	Study timepoint 1 Controlh	Study timepoint 2 Treatmenth	Treatment/control GMRi (90% CI)
DX/DM plasma MR	27a	3.0 (2.2–4.3)	3.1 (2.3–4.1)	1.01 (0.80–1.28)
DX/DM urinary MR	24b	16.0 (10.2–25.0)	12.5 (8.9–17.6)	0.78 (0.55–1.11)
(DX + DX-O-glucuronide)/DM urinary MR	21c	540.5 (329.9–885.7)	492.2 (324.4–746.8)	0.91 (0.59–1.40)
DX-O-glucuronide/DX plasma MR	32d	62.5 (54.6–71.6)	63.5 (53.7–75.2)	1.02 (0.90–1.15)
DX-O-glucuronide/DX urinary MR	25e	30.9 (26.2–36.6)	37.6 (31.9–44.3)	1.22 (1.06–1.39)
3MM/DM urinary MR	12f	0.11 (0.07–0.16)	0.07 (0.05–0.10)	0.65 (0.40–1.06)
3HM/DX urinary MR	29g	0.27 (0.24–0.31)	0.32 (0.28–0.38)	1.18 (1.03–1.35)

Abbreviations: 3HM, 3-hydroxymorphinan; 3MM, 3-methoxymorphinan; CI, confidence interval; DM, dextromethorphan; DX, dextrorphan; MR, metabolic ratio; Study timepoint 1, pre-isotretinoin; Study timepoint 2, with isotretinoin.

^aOnly participants who were extensive metabolizers of CYP2D6 (activity scores 1–2) were included.

^bOnly participants who were extensive metabolizers of CYP2D6 (activity scores 1–2) were included (three participants had concentrations below limit of quantification and were excluded from the analysis).

^cOnly participants who were extensive metabolizers of CYP2D6 (activity scores 1–2) were included (six participants had concentrations below limit of quantification and were excluded from the analysis).

^dOne participant had concentrations below limit of quantification and was excluded.

^eEight participants had concentrations below limit of quantification and were excluded.

^ftwenty-one participants had concentrations below limit of quantification and were excluded.

^gFour participants had concentrations below limit of quantification and were excluded.

^hStudy timepoints 1 and 2 data were reported as geometric mean (90th percentile CI).

ⁱTreatment/control geometric mean ratio (GMR) with 90th percentile CI was reported.

For participants who were CYP2D6 EMs (AS = 1–2), no statistically significant differences were observed between study timepoints 1 and 2 plasma DX/DM MRs (GMR 1.01 treatment/control) (Table 2 and Figure 2A). Both the urinary DX/DM and DX plus DX-O-glucuronide/DM MRs in CYP2D6 EMs (AS = 1–2) were numerically lower with isotretinoin treatment (GMR = 0.78 and 0.91 treatment/control), but the differences were not statistically significant (Table 2 and Figure 2B, C). The impact of the administration of isotretinoin on plasma and urine DX/DM MRs was also analysed in individual AS subgroups (Figure 2D–E and Supporting Information S1). No statistically significant differences were observed in plasma DX/DM MR in the AS = 2 group (GMR 0.86 treatment/control) (Figure 2D). The urine DX/DM MR in AS = 2 group was statistically significantly lower at study timepoint 2 than study timepoint 1 (GMR 0.58 treatment/control, 90% CI 0.41–0.82) (Figure 2E). No statistically significant differences in DX/DM 0–4-h urinary MR between female and male participants were observed at either study timepoint in CYP2D6 EMs (P = .2 and .8) (Supporting Information Figure S2).

FIGURE 2.

Cytochrome P450 2D6 (CYP2D6) activity prior to (ST1) and with isotretinoin treatment (ST2) based on CYP2D6 activity scores (AS). (A)-(C) include participants with activity scores = 1–2 and (D) and (E) only include participants with activity score = 2. (A) Plasma dextrorphan/dextromethorphan (DX/DM) metabolic ratios (MR) (n = 27), (B) urinary DX/DM MR (n = 24), (C) urinary DX plus DX-O-glucuronide/DM MR (n = 21), (D) plasma DX/DM MR (n = 17) and (E) urinary DX/DM MR (n = 15). 90% CI, 90th percentile confidence interval; GMR, geometric mean ratio.

The urinary 3MM/DM MR was not statistically significantly different with isotretinoin treatment compared to control (GMR 0.65 treatment/control) (Table 2 and Figure 3A). Urinary 3HM/DX MR increased with isotretinoin treatment compared to control (GMR 1.18 treatment/control, 90% CI 1.03–1.35) (Table 2 and Figure 3B). There was no statistically significant difference in urinary 3HM/DX MR between CYP3A5 expressors (CYP3A5 *1/*1 and *1/*3 genotypes) and CYP3A5 non-expressors (CYP3A5 *3/*3 genotype) at either study timepoint (P = .2 and .4) (Figure S3A,B). Additionally, there was no statistically significant difference observed in urinary 3HM/DX MR with isotretinoin treatment (study timepoint 2, 3HM/DX MR to study timepoint 1, 3HM/DX MR ratio) between CYP3A5 expressors and non-expressors (P = .5) (Supporting Information Figure S3C).

FIGURE 3.

CYP3A activity prior to (ST1) and with isotretinoin treatment (ST2). (A) Urinary 3-methoxymorphinan/dextromethorphan (3MM/DM) (n = 12) and (B) 3-hydroxymorphinan/dextrorphan (3HM/DX) (n = 29) metabolic ratios. 90% CI, 90th percentile confidence interval; GMR, geometric mean ratio.

The plasma DX-O-glucuronide/DX ratio was unchanged between the two study timepoints (GMR 1.02 treatment/control) (Table 2 and Figure 4A). Conversely, urinary DX-O-glucuronide/DX MR increased with isotretinoin treatment compared to control (GMR 1.22 treatment/control, 90% CI 1.06–1.39) (Table 2 and Figure 4B).

FIGURE 4.

UGT2B activity prior to (ST1) and with isotretinoin treatment (ST2). (A) Plasma dextrorphan-O-glucuronide/dextrorphan (DX-O-glucuronide/DX) (n = 32) and (B) urinary DX-O-glucuronide/DX (n = 25) metabolic ratios. 90% CI, 90th percentile confidence interval; GMR, geometric mean ratio.

3.4 |. Correlation between plasma retinoid concentrations and urinary DM MRs

Based on linear mixed-effect regression analysis, there were no significant correlations between changes in log plasma retinoid concentrations (atRA, 13-cisRA and 4-oxo-13-cisRA) and log urinary DX/DM MR (P = .2, .3, and .3, respectively), 3HM/DX MR (P = .5, .3, and .3, respectively) and 3MM/DM MR (P = .4, .5, and .4, respectively) with isotretinoin treatment in CYP2D6 EMs (Supporting Information Figure S4).

4 |. DISCUSSION

Our study assessed the effect of isotretinoin on CYP2D6 and CYP3A activities in patients with severe acne. Our results show no statistically significant downregulation of CYP2D6 activity in CYP2D6 EMs treated with isotretinoin compared to pre-treatment control. A potential weak induction of CYP3A activity with isotretinoin was observed. The 0–4-h urinary DX-O-glucuronide/DX MR increased with isotretinoin, preliminarily suggesting increased UDP-glucuronosyltransferase 2B (UGT2B) activity.

Previously, a concentration-dependent induction of SHP mRNA and downregulation of CYP2D6 mRNA were observed in human hepatocytes with retinoid (atRA, 13-cisRA and 4-oxo-13-cisRA) treatment and in vivo CYP2D6 downregulation was predicted due to the HNF4α-SHP regulation mechanism.¹⁶ Our findings suggest that the in vitro downregulation of CYP2D6 due to retinoids does not translate into in vivo effects in humans. This lack of translation has been suggested to be due to the lack of endogenous regulators such as bile acids, stellate cells and Kupffer cell-derived factors in in vitro hepatocyte systems and their potential role in SHP regulation.¹⁶ In support of this, treatment of rat primary hepatocytes with bile acids was shown to interfere with the activation of Ntcp gene expression by retinoids mediated through Fxr activation and Shp induction.³² In addition, estrogens and bile acids have been suggested to contribute to the regulation of CYP2D6 expression in CYP2D6 humanized mice.^33,34 To improve the prediction of retinoid effects on CYP2D6 regulation in humans in vivo, it may be necessary to include these endogenous regulators in in vitro systems.

In male mice, following atRA, 13-cisRA and 4-oxo-13-cisRA treatment, no correlation between Shp induction and mouse Cyp2d downregulation was observed.¹⁶ In male mice, treatment with 4-oxo-13cis-RA resulted in increased mRNA expression in Cyp2d isoforms and Shp.¹⁶ Likewise, elevated liver mRNA expression and activity of Cyp2d11, Cyp2d22, Cyp2d26 and Cyp2d40 was observed in pregnant mice along with increased RA signalling.³⁵ Notably, a retinoic acid response element was identified in mice Cyp2d40 promotor sequence.³⁵ However, whether or not involvement of RA signalling in Cyp2d regulation in mice shares a similar mechanism in humans remains unclear.

Previously, Stevison et al¹⁶ demonstrated a weak increase in DX/DM AUC ratio and DX formation clearance with isotretinoin, suggesting that 13-cisRA is a CYP2D6 inducer. In contrast, we found no statistically significant change in CYP2D6 activity with isotretinoin. Differences in study sizes and populations may explain differences in results. Our study was much larger (n = 33) and included more diverse CYP2D6 genotypes, than that of Stevison et al¹⁶ (n = 8), who excluded CYP2D6 *3 and CYP2D6*4 participants. In our study, the functional allele frequency among Caucasian participants (70%) aligns with the previously reported median value of 71% in Caucasian population.³⁶ The effect of isotretinoin on CYP2D6 activity in different CYP2D6 genotypes might be different. In our study, the CYP2D6 AS 2 participants showed a significant decrease in 0–4-h urinary DX/DM MR with isotretinoin treatment, suggesting downregulation of CYP2D6 activity (Figure 2E). This was not observed in participants with CYP2D6 ASs of 1 or 1.5 (Supporting Information Figure S1). Stevison et al¹⁶ recruited healthy males, whereas we had a predominantly female population with severe acne. However, the male/female difference did not appear to impact the outcome. Although some studies have reported that DM/DX (0–8 h¹⁸, 0–10 h¹⁹) urinary MRs were significantly lower in healthy, Caucasian, female CYP2D6 EMs compared to male, we found no significant difference in DX/DM 0–4-h urinary MR between females and males who were CYP2D6 EMs, possibly due to the smaller number of participants. This result is similar to a previous study reporting no significant difference in DM/DX 0–24-h urinary MRs between 11 female and seven male healthy volunteers who were CYP2D6 EMs.³⁷

Previous studies have suggested that retinoids induce CYP3A4 via PXR activation in vivo and in vitro.¹⁶ Our results are consistent with a weak induction of CYP3A activity with isotretinoin treatment based on the 0–4-h urinary 3HM/DX MR (GMR 1.18 treatment/control, n = 29) (Figure 3B). However, no statistically significant change was observed in the 0–4-h urinary 3MM/DM MR (n = 12), likely due to the smaller amount of available data (Figure 3A). Most of the participants had 3MM urinary concentrations below the LLOQ and therefore were not included in the data analysis. Hence, 3HM/DX MR was used as a CYP3A in vivo activity indicator as previously described.³⁸

In our study, we did not observe any significant differences in urinary 3HM/DX MR between various CYP3A5 genotypes both prior to and during isotretinoin treatment (Supporting Information Figure S3). A previous study reported that DM primarily undergoes conversion into 3MM primarily catalysed by CYP3A4, with a lesser contribution from CYP3A5 in human liver.³⁹ This suggests CYP3A5 enzymatic activity may have a relatively minor role in the overall catalytic activity of CYP3A from DM to 3MM in humans in vivo as well. The majority of participants in our study were identified as CYP3A5 poor metabolizers. Consequently, the urinary 3HM/DX MR we observed predominantly reflects the activity of CYP3A4, and induced CYP3A activity following isotretinoin administration might be primarily contributed by increased CYP3A4 activity.

An interesting finding in our study was the significant increase in urinary DX-O-glucuronide/DX MR following isotretinoin treatment (GMR 1.22 treatment/control) (Table 2 and Figure 4B). A previous study showed that UGT2B isoforms, including UGT2B4, UGT2B7, UGT2B15 and UGT2B17, are primarily responsible for the glucuronidation of DX to dextrorphan-O-glucuronide, suggesting the change in urinary MR may be due to induction of these UGTs.⁴⁰ UGT1A1 was found to be induced by approximately 2-fold with treatment with retinoids (30 μM atRA, 13-cisRA or 4-oxo-13-cisRA) in human hepatocytes.⁴¹ Another in vitro study indicated that vitamin A (100 μM) competitively inhibits UGT1A1, UGT2B4, UGT2B7 and UGT2B15 activity.⁴² Retinoids including atRA and its metabolite, 4-hydroxyretinoic acid (4-OH-RA), have been shown to be substrates of UGT2B7.⁴³ However, induction of UGT2B activities by retinoids in vitro or in vivo has not been previously shown. DX-O-glucuronide/DX MR has not been established or evaluated as an UGT2B in vivo marker. Hence, whether or not the observed increase in DX-O-glucuronide/DX MR with isotretinoin treatment indicates induction of UGT2B isoforms is inconclusive. Future in vitro studies would be helpful in understanding and predicting the mechanisms of retinoid impact on UGT2B isoform expression and activities in vivo.

This study employed the DM MR, a well-established surrogate biomarker, to investigate the alterations in the activities of CYP2D6 and CYP3A in a cohort of male and female participants with diverse CYP2D6 genotypes. This study has brought insights into varied responses of CYP2D6 activity to isotretinoin within EMs, stratified by activity score groups. Additionally, this study assessed changes in CYP3A activity associated with isotretinoin among participants possessing different CYP3A5 genotypes. Nevertheless, it is important to acknowledge a limitation in our study, which is the potential lack of sensitivity in detecting subtle alterations in CYP2D6 activity due to the use of a single timepoint assessment. The interpretation of the single timepoint MR might also be confounded by changes in the elimination rate of the metabolite.⁴⁰ For example, induction of DX glucuronidation might decrease the DX/DM MR. However, the trend towards decreased ratio was still observed when 0–4-h urinary DX plus DX-O-glucuronide/DM MR was analysed in the CYP2D6 EMs (Table 2 and Figure 2C). Similarly, 3HM/DX MR is also affected by the changes in the formation clearance and elimination of DX. The elimination pathway of DX (Figure 1) includes glucuronidation and metabolism to 3HM.⁴⁰ Based on larger DX-O-glucuronide in vitro formation clearance compared to 3HM, the major elimination pathway of DX is glucuronidation.⁴⁰

In summary, in vitro downregulation of CYP2D6 activity by isotretinoin does not translate into significant in vivo effects in CYP2D6 EMs. However, this study demonstrated a weak induction in CYP3A activity in vivo with isotretinoin administration. Future studies using other CYP3A probe substrates are necessary to determine if the CYP3A induction by retinoids in vivo is reproducible. This study has shown an increased dextrorphan-O-glucuronide to DX MR with isotretinoin treatment and reported a preliminary observation of increased apparent UGT2B activity. Future work is needed to demonstrate how isotretinoin treatment impacts the activity of UGT2B isoforms in vitro and in vivo.

What is already known about this subject

• Retinoids have been shown to decrease CYP2D6 mRNA expression in vitro.

• Only one small (n = 8) clinical study has evaluated whether CYP2D6 downregulation by retinoids in vitro could translate into humans in vivo.

• Retinoids have been shown to induce CYP3A4 expression in vitro and in vivo in humans.

What this study adds

• Isotretinoin did not significantly downregulate CYP2D6 activity in humans.

• Isotretinoin weakly induced CYP3A activity in humans.

• Preliminary data suggest that isotretinoin increases UGT2B activity.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author.