Does In Vitro Cytochrome P450 Downregulation Translate to In Vivo Drug-Drug Interactions? Preclinical and Clinical Studies With 13-cis-Retinoic Acid

Abstract

All-trans-retinoic acid (atRA) downregulates cytochrome P450 (CYP)2D6 in several model systems. The aim of this study was to determine whether all active retinoids downregulate CYP2D6 and whether in vitro downregulation translates to in vivo drug-drug interactions (DDIs). The retinoids atRA, 13cisRA, and 4-oxo-13cisRA all decreased CYP2D6 mRNA in human hepatocytes in a concentration-dependent manner. The in vitro data predicted ~ 50% decrease in CYP2D6 activity in humans after dosing with 13cisRA. However, the geometric mean area under plasma concentration-time curve (AUC) ratio for dextromethorphan between treatment and control was 0.822, indicating a weak induction of dextromethorphan clearance following 13cisRA treatment. Similarly, in mice treatment with 4-oxo-13cisRA-induced mRNA expression of multiple mouse Cyp2d genes. In comparison, a weak induction of CYP3A4 in human hepatocytes translated to a weak in vivo induction of CYP3A4. These data suggest that in vitro CYP downregulation may not translate to in vivo DDIs, and better understanding of the mechanisms of CYP downregulation is needed.

Figure 1

Dose‐dependent downregulation of CYP2D6 mRNA by all‐trans‐retinoic acid (at RA), 13‐cis‐retinoic acid (13cis RA), and 4‐oxo‐13cis RA in three human hepatocyte donors. Panels (a,b,c) show donor 1, panels (d,e,f) show donor 2, and panels (g,h,i) show donor 3. The treatments shown are 13cis RA in panels a, d, and g, at RA in panels b, e, and h, and 4‐oxo‐13cis RA in panels c, f and i. The data are presented as mean and range of independent mRNA measurements (n = 3) in comparison to vehicle control. For donor 3, three separate experiments done on independent days are also shown. The minimum effect (E_min) and half‐maximal effective concentration (EC₅₀; μM) shown were fit as described in the methods and are presented as mean and 90% confidence interval for donor 1 and donor 2. For donor 3, the median values from three replicate experiments are shown and the E_min and EC₅₀ for each replicate experiment are provided in Table S1 . Inconsistent changes in CYP2D6 mRNA were observed in response to the small heterodimer partner inducer control, GW4064, at 1 μM (average 1.0‐fold, 1.2‐fold, and 0.3‐fold change in donors 1, 2, and 3, respectively).

Figure 2

Simulations of the predicted time course of CYP2D6 mRNA (a) and protein (b) in human hepatocytes assuming a zero‐order mRNA synthesis rate of 0.04 pmol/hour and mRNA and protein elimination rate constants of 0.04 hour⁻¹ and 0.015 hour⁻¹, respectively. Simulations were conducted as described in the Methods section with 100% (solid black line), 75% (dashed line), 50% (dotted line), or 25% (solid gray line) inhibition of mRNA synthesis rate.

Figure 3

The effects of 13‐cis‐retinoic acid (13cis RA), all‐trans‐retinoic acid (at RA), and 4‐oxo‐13cis RA on CYP3A4 mRNA and activity. Data for CYP3A4 mRNA (white bars) and activity (hashed bars) in donor 1 (a,b,c), donor 2 (d,e,f), and one replicate experiment in donor 3 (g,h,i) hepatocytes are presented as mean and range of the measured effect (n = 3 replicates per donor). The data from other two replicate experiments in donor 3 are shown in Figure S3 . The control rifampicin (25 μM) increased CYP3A4 mRNA by an average 38‐fold, 17‐fold, and 176‐fold and CYP3A4 activity by an average 3.0‐fold, 2.8‐fold, and 16‐fold in donors 1, 2, and 3, respectively. The concentrations listed on the x‐axis are the nominal concentrations added to the hepatocytes. The red boxes indicate the nominal concentrations that mimic clinically relevant circulating unbound concentrations.

Figure 4

Serum concentration‐time profiles of the study drugs and the pharmacokinetic measures of CYP2D6 and CYP3A4 activity markers. Panel (a) shows the serum concentration vs. time profiles for 13‐cis‐retinoic acid (13cis RA) (circles) and its metabolites 4‐oxo‐13cis RA (squares), and all‐trans‐retinoic acid (at RA) (triangles) following the final 40 mg dose of 13cis RA on study day 15. Panels (b,c) show the serum concentrations of dextromethorphan (b) and dextrorphan (c) on study days 1 (prior to 13cis RA dosing, circle with solid line) and 15 (after 13cis RA dosing, square with dashed line). Concentration data are presented as mean and range of individual values (n = 8). Insets are data presented on semilog scale. The calculated formation clearances (Cl_f) of dextromethorphan metabolites dextrorphan (d) and 3‐methoxymorphinan (e) on the two study days are shown for seven subjects. One subject was identified as an outlier and removed from all analyses involving urine data. Panel (f) shows 6β‐hydroxycortisol Cl_f on the two study days in six subjects. One subject had no measurable 6β‐hydroxycortisol in urine collected from study day 15 and was excluded.

Figure 5

Effects of 13‐cis‐retinoic acid (13cis RA), all‐trans‐retinoic acid (at RA), or 4‐oxo‐13cis RA on cytochrome P450 (CYP) mRNA expression in mouse livers following treatment with each study retinoid. The mice (n = 6 for each treatment) were treated for 4 days with each retinoid at 5 mg/kg i.p. as described in the Methods section and the fold change in small heterodimer partner (Shp) and Cyp mRNA expression (black bars) was evaluated in comparison to vehicle controls (white bars). The panels show mRNA data after treatment with 13cis RA (a), at RA (b), or 4‐oxo‐13cis RA (c), and the insets provide P values calculated for each gene compared to control as described in methods. *P < 0.05.