Effects of meal composition and meal timing on the expression of genes involved in hepatic drug metabolism in rats

The field of chronotherapy, the interdisciplinary science that aims to synchronize drug administration and biological rhythms, is receiving increasing attention. Drug clearance [1], toxicity [2] and pharmacokinetics all display clear changes over the light-dark cycle (reviewed in [3]). In addition, daily patterns are present in the manifestations and severity of many diseases, further highlighting the necessity for chronotherapy.

Daily rhythms in the mammalian body are mastered by the endogenous circadian (meaning approximately 24 h) clock in the suprachiasmatic nucleus of the hypothalamus (SCN) [4, 5] that is mainly entrained to the external environment by light. In addition to the central SCN clock, clocks in peripheral organs such as the liver are both entrained by the SCN via neural and endocrine pathways and by external stimuli such as nutrition [6]. Desynchronization between these central and peripheral clocks can alter the metabolic phenotype of an organism [7].

The first step of drug metabolism by the liver is carried out by phase I enzymes, the cytochrome p450 enzymes (P450). Of all the different isoforms of P450 enzymes present in the liver, the Cyp1-3 families are contributing to 70–80% of the metabolism of clinically prescribed drugs [8]. Transcription of P450 is regulated by the nuclear receptors constitutive androstane receptor (CAR) and the pregnane X receptor (PXR) [9]. In mouse liver, Car expression shows strong oscillating circadian patterns, whereas this is absent for Pxr [10]. In rodents, some but not all, P450 enzymes reveal daily variations in activity [11]. CYP3A4, the most important drug metabolizing P450 enzyme in humans, shows circadian oscillation in vitro [12]. In addition to the nuclear receptors, aminolevulinic acid synthase 1 (ALAS1), a rate-limiting enzyme in the production of heme, and cytochrome P450 oxireductase (POR) are key enzymes in P450 function. Both enzymes are regulated by the molecular clock [13–15].

The molecular clock consists of a sophisticated autoregulatory feedback loop with a period of approximately 24 hours. In short, the proteins CLOCK (Circadian Locomotor Output Cycles Kaput) and BMAL (brain muscle ARNT-like 1) form a dimer that regulates the expression of Per (period) and Cry (cryptochrome). As a heterodimer, PER and CRY inhibit the nuclear transcriptional activity of CLOCK and BMAL. By doing so they inhibit their own transcription which activates CLOCK:BMAL activity again [6, 16–18]. In addition to Per and Cry, the CLOCK:BMAL complex regulates a vast variety of metabolic genes such as Pparα (peroxisome proliferator-activated receptor alpha) and Rev-erb that link the clock to physiological processes. CLOCK:BMAL also regulates the expression of PAR bZIP proteins such as DBP (albumin D-site binding protein) which in turn is responsible for the circadian oscillations in Car expression [19, 20].

Contrary to the central SCN clock, the peripheral liver clock is particularly sensitive to changes in meal timing. When rats are subjected to restricted feeding in the resting (light) phase, and thus eat at an inappropriate time of day, clock gene expression in the liver is reversed and uncoupled from the master clock in the SCN [21–23]. In addition, also meal composition may affect the molecular clock system (reviewed in [24]). Since both meal timing and composition may affect the hepatic molecular clock, and the molecular clock regulates expression of genes involved in drug metabolism, it is evident that changes in eating behavior due to shiftwork, sleep disturbance and the 24/7 availability of highly palatable foods in our western society could have severe consequences for the practice of chronotherapy.

Therefore, the aim of our study was to address the following questions: 1) Do nuclear receptors and P450 isoform homologues to human P450 enzymes involved in drug metabolism display day-night oscillations?; 2) Does the ingestion of a free choice high-fat-high-sugar (fcHFHS) diet affect the hepatic expression of nuclear receptors and P450 enzymes? 3) Does meal timing affect these day-night oscillations? To this end, rats were subjected to either a chow or a fcHFHS diet [25]. Food was either available ad libitum (ad lib) or rats were only given access to food in the light period (light fed, or LF rats) or in the dark period (dark fed, or DF rats). To differentiate between the effect of the hepatic molecular clock and the timing of food intake, we conducted an additional experiment in which rats were either subjected to an ad lib chow diet or to a regimen where access to food was given in 6 predictable bouts per day (6-meals-a-day (6M) experiment). We determined daily patterns of hepatic gene expression by qPCR.

Several aspects of hepatic drug metabolism display pronounced day-night variations, such as drug clearance and the associated drug toxicity [2]. In this paper, we describe the effects of meal timing and meal composition on the mRNA expression of nuclear receptors and P450 enzymes involved in hepatic drug metabolism in rats. We found that changes in meal timing but not meal composition induce profound phase shifts in daily rhythms of Car, Pxr, Alas1 and Por which are mediated by the hepatic molecular clock. These changes are likely to affect P450 mediated drug metabolism.

Phase I drug metabolism is carried out by cytochrome P450 enzymes. Circadian oscillations of a wide variety of P450 enzymes and nuclear receptors have been described in literature (reviewed in [11]). We show for the first time in rats that Pxr displays significant daily oscillations in ad lib chow fed conditions. None of the other nuclear receptors and P450 isoforms that we tested showed significant day-night rhythmicity. Surprisingly, although described as a highly oscillating gene, Car did not show a significant day-night rhythm when ad lib fed, although mRNA expression in the dark period tended to be higher than in the light period. The same was observed for Cyp2b2 expression, a classic CAR target gene. None of the P450 enzymes, we tested, displayed a robust daily rhythm. Since the methodology to assess whether a gene displays a significant day-night oscillation greatly differs between studies (from simply testing a two time point day/night difference to the more advanced algorithms such as JTK cycle that we used), comparisons between studies are difficult. In addition, species differences might play a role. Using a cosine-wave pattern algorithm, Yang and coworkers found that nuclear receptors Car, Shp and Rxr were rhythmically expressed in mouse liver, whereas Pxr was not [10]. However, another study reported clear diurnal variation in Pxr expression in mouse liver [31]. Also in our study Pxr showed significant rhythmicity when rats were fed chow ad lib.

Heme production by 5-aminolevulinate synthase 1 (ALAS1) and electron transport by P450 oxidoreductase (POR), are essential for P450 activity. Several studies have shown a strong reciprocal interaction between components of the molecular clock and both heme production and POR activity [13–15], offering a posttranslational mechanism for the regulation of P450 enzyme activity. We measured Alas1 and Por mRNA expression in ad lib chow fed rats and indeed found very strong day-night oscillations which suggests a functional day night difference in P450 mediated drug metabolism.

We investigated the effect of meal composition and meal timing on the cytochrome P450 enzymes. The effect of meal composition was investigated by feeding rats a fcHFHS diet, a diet intervention specifically designed to model the overconsumption of palatable food in our Western society, including the choice aspect [32, 33]. fcHFHS feeding did not blunt or induce the daily expression pattern of any of the genes tested. However, the fcHFHS diet did increase the overall expression of Car, Cyp2b2 and Cyp3a2. Rats on the fcHFHS diet had a higher daily caloric intake than the chow groups, while in the second experiment the 6M groups had a lower caloric intake than the ad lib and 6M groups. In the second experiment basal expression of Car, Cyp2b2 and Cyp3a1 was not affected by the six-meal feeding regimen, we thus assume that the difference in the first experiment in basal expression of Car, Cyp2b2 and Cyp3a2 is due to the ingestion of fat and/or sugar. Literature on this subject is contradictory; a high fat diet for 13 weeks increased Pxr and CYP3a2 in rats [34], whereas a high fat diet for 11 weeks decreased liver Car, Pxr and Cyp3a2 expression in mice [35]. Dietary lipids do not have a uniform effect on P450 expression; ingestion of lard or corn oil by rodents induced Cyp2e1, Cyp1a2, Cyp2b, Cyp3a and Cyp4a, whereas this did not affect Cyp2a1 and Cyp2c11 expression [36]. Even less is known about the effect of high sucrose intake on hepatic P450 expression. An early study showed that high sucrose feeding decreased total CYP content [37], whereas another study in rodents showed that a high-fat-high-sucrose diet increased Cyp1a2, Cyp1a1, Cyp2b10 and Cyp2c29 expression while decreasing Cyp3a11 expression [38]. Moreover, the mechanism(s) by which dietary components affect P450 expression are unclear at the moment, as well as the functional consequences for P450 function. Also, Car expression is also upregulated during fasting which is mediated via PPARα [28, 39] further complicating the interpretation of this data. We did not observe an effect of the fcHFHS diet on Alas1 and Por mRNA expression.

Meal timing had pronounced effects on the daily rhythms of the studied genes. The daily pattern in mRNA expression of nuclear receptors Car and Pxr, and P450 cofactors Alas1 and Por was completely shifted between rats that only had access to food in the light period compared to the dark period. Interestingly, this effect was essentially the same when rats were fed a fcHFHS diet except for Pxr which displayed a more moderate shift on a fcHFHS diet compared to the chow diet. These results reinforce the notion that timing of food intake rather than diet composition determines daily rhythms in gene expression. However, since hepatic clock gene expression also shifts with the restricted food intake in our timing experiment (Opperhuizen et al., 2016), it is difficult to confirm whether the shifts in gene expression observed in the different feeding schedules are caused by shifts in clock gene expression or feeding activity. To separate the influence of clock genes from food intake, we designed a second experiment in which we measured gene expression in the livers of ad lib fed rats and rats that received 6 meals a day, equally distributed over the 24 h. This paradigm removes the strong day-night rhythm in feeding activity, without abolishing or shifting the day-night rhythms in hepatic clock gene expression [40]. Interestingly, feeding rats according to a 6-meals-a-day schedule in this study only affected the rhythmic expression of Pxr and Cyp2c6, while Car, P450 enzymes, Alas1 and Por expression did not differ between the ad lib and 6M fed rats. Expression of clock genes Cry1 and Dbp was not affected in the 6M experiment (Eggink et al 2017, to appear in Chronobiology International). We can thus conclude that while some aspects of hepatic drug metabolism, such as Pxr and Cyp2c6 expression seem to be sensitive for changes in meal composition and meal timing, other components are under the exclusive control of the hepatic molecular clock.

Taken together, the data from these experiments show that when food intake is restricted to an inappropriate time of day (for rodents that is eating during the light phase), the rhythmic expression of genes involved in hepatic drug metabolism is shifted, but the rhythmicity itself is not abolished. Surprisingly, when the influence of rhythmic meal timing is removed, as shown in the 6M experiment, daily gene expression of most aspects of P450-mediated drug metabolism still follows the expression pattern of ad lib fed animals, indicating that a factor, different from meal timing, affects circadian gene expression of these genes. The functional consequences of these meal timing-induced phase shifts for hepatic drug metabolism need to be further investigated, although the profound effect on Alas1 and Por expression indicates that P450 mediated drug metabolism will be affected. Johnson et al recently showed that food intake during the light phase leads to shifts in acetaminophen toxicity [13], highlighting the role of food intake in chronotoxicity of drugs. Although direct extrapolation to clinical practice is challenging, especially since the nocturnal and diurnal difference between rodents and humans, these data show that disturbed meal timing can profoundly affect the diurnal pattern in gene expression of nuclear receptors and enzymes that are heavily involved in xeno- and endobiotic metabolism and might thus also have consequences for the practice of chronotherapy.

All relevant data are within the paper and its supporting information files.

This work was supported by the Technology Foundation STW grant “Feeding OnTime” (#12195) for SEF and AK and an AMC PhD Scholarship for JOE.