Cyclophosphamide (CPA) was purchased from J&K Scientific (Beijing, China). O-methylhydroxylamine (OMHA) was obtained from Steraloids (Wilton, NH). Nicotinamide adenine dinucleotide phosphate (NADPH), uridine diphosphoglucuronic acid (UDPGA), alamethicin, and 3′-phosphoadenosine-5′-phosphosulfate (PAPS) were purchased from Sigma-Aldrich (St. Louis, MO). Propofol and propofol glucuronide were purchased from TargetMol (Shanghai, China). Galangin was obtained from Weikeqi Biotech (Sichuan, China). Assay kits for ALT (alanine aminotransferase), AST (aspartate aminotransferase) and GSH (glutathione) were purchased from Jiancheng Bioengineering Institute (Nanjing, China). Anti-PER2 (ab180655) and anti-GAPDH antibodies (ab9485) were purchased from Abcam (Cambridge, MA). Anti-CYP2B10 (TA504328) and anti-SULT1A1 (TA501951) antibodies were obtained from OriGene (Rockville, MD). Anti-UGT1A9 (bs-4224R) antibody was purchased from Bioss (Beijing, China). Anti-REV-ERBα antibody (AB10130) was purchased from Sigma-Aldrich (St Louis, MO). All primary antibodies were diluted with 5% BSA at a ratio of 1:1,000. The secondary antibody was purchased from Huaan Biotechnology (Hangzhou, China) and diluted with 5% skim milk at a ratio of 1:5,000. siPer2 (siRNA targeting Per2) and siNC (a negative control for siPer2) were obtained from TranSheep Biotech (Shanghai, China). pRL-TK vector was purchased from Promega (Madison, WI).

Per2 heterozygotes (on a C57BL/6 background) were obtained from Cyagen Biosciences (Guangzhou, China) (Supplementary Material S1). Homozygotes with exons 4-6 deleted (named Per2 ^Del4-6 mice) were generated by inter-crossing heterozygous mice. Per2-deficient mice and their wild-type littermates were housed and maintained under a 12 h light/12 h dark cycle (lights on at 6:00 AM (= Zeitgeber time 0/ZT0) and lights off at 6:00 PM (= ZT12)), with free access to food and water. Animal experimental procedures were approved by Institutional Animal Care and Use Committee of Guangzhou University of Chinese Medicine (Appr. date: 2020-11-19; IACUC Issue No: ZYD-2020-111) and were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals. Efforts were made to minimize suffering and the number of mice used in the experiments.

DNA was extracted from mouse tail (0.5–1 mm). PCR reactions were performed with 400 ng DNA template. Amplification program consisted of an initial denaturation at 94°C for 3 min, 35 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 35 s, and extension at 72°C for 35 s. The products were subjected to 2% agarose gel electrophoresis, and bands were imaged using the Omega Lum G imaging system (Aplegen). The primers are listed in Table 1.

Per2 ^Del4-6 and control wild-type mice (8–12 weeks, male) were treated with a single dose of CPA (100 mg/kg) by intraperitoneal injection at ZT14. At predetermined time points (15, 30, 60 and 120 min), mice (n = 5 per time point) were rendered unconscious with isoflurane, and plasma samples were collected by retro-orbital bleeding. Plasma samples were processed for UPLC-QTOF/MS analysis as previously described (Sadagopan et al., 2001; Zhao et al., 2019). Of note, the CPA metabolite, 4-hydroxycyclophosphamide (4-OH-CPA) is unstable in the plasma (Sadagopan et al., 2001). To quantify 4-OH-CPA, plasma samples were treated with O-methylhydroxylamine (OMHA) to transform 4-OH-CPA into a stable product OH-CPA O-methyloxime (Sadagopan et al., 2001). Additionally, liver samples were collected at two time points (i.e., 30 and 120 min), and processed to measure the drug/metabolite concentrations in the livers as previously described (Zhang et al., 2018a).

Per2 ^Del4-6 and control wild-type mice (8–12 weeks, male, n = 5 per group) were injected (i.p.) with CPA at 100 mg/kg or with vehicle at ZT14. Four hours after drug administration, mice were sacrificed to collect plasma and liver samples, followed by biochemical analyses.

ALT, AST, and GSH were measured with their assay kits according to manufacturer’s instructions (Jiancheng Bioengineering Institute, Nanjing, China).

Livers were collected from Per2 ^Del4-6 and control wild-type mice. Liver microsomes were prepared by sequential ultracentrifugation, first at 9,000 g for 10 min and then at 100,000 g for 1 h as previously described (Van et al., 1992). CPA was used as a specific substrate to determine the microsomal activity of CYP2B10 (Pass et al., 2005; Zhao et al., 2019). In brief, liver microsomes (4 mg/ml) was incubated with CPA (20 μM), NADPH (1.5 mM), and MgCl₂ (5 mM) in potassium phosphate (50 mM, pH 7.4) at 37°C for 2 h. The reaction was terminated by adding ice-cold acetonitrile (containing an internal standard). To quantify 4-OH-CPA, the resultant mixture was treated with OMHA (1.5 mg/ml) to transform 4-OH-CPA into a stable product OH-CPA O-methyloxime. The reaction mixture was centrifuged at 13,000 g for 15 min and the supernatant was subjected to UPLC-QTOF/MS analysis.

Glucuronidation assays were performed by incubating liver microsomes with propofol. The incubation mixture contained liver microsomes (4 mg/ml), propofol (100 μM), MgCl₂ (0.88 mM), saccharolactone (4.4 mM), alamethicin (11.2 μM) and UDPGA (3.5 mM) in 50 mM potassium phosphate (pH 7.4). After 30 min, the metabolic reaction was terminated by adding 200 μl ice-cold water/acetonitrile (50:50, v/v) (containing an internal standard). The resulting mixture was centrifuged at 2,000 g for 10 min, and the supernatant was subjected to UPLC-QTOF/MS analysis.

Livers were collected from Per2 ^Del4-6 and control wild-type mice. Liver S9 fraction was prepared by centrifugation at 9,000 g for 10 min (Richardson et al., 2016). Sulfation activity was determined using an incubation method with liver S9 fraction as previously described (Guo et al., 2018). Briefly, liver S9 fraction (4 mg/ml), PAPS (50 μM) and galangin (6 μM) were incubated in potassium phosphate buffer (50 mM, pH 7.4) at 37°C for 3 h. The reactions were terminated by adding 200 μl ice-cold acetonitrile (containing an internal standard), followed by vortex and centrifugation at 13,000 g for 15 min. The supernatant was subjected to UPCL-QTOF/MS analysis.

Drugs and metabolites were quantified using an UPLC-QTOF/MS system (AB SCIEX, CA) and a Phenomenex C18 column (2 × 100 mm, 1.6 μm; Phenomenex, Torrance). The mobile phases were 0.1% formic acid in acetonitrile (mobile phase A) and 0.1% formic acid in water (mobile phase B). For determination of CPA and OH-CPA O-methyloxime, the gradient elution program was as follows: 0–2 min, 90% B; 2–6 min, 90-10% B; and 6–8 min, 10–90% B. The mass spectrometer was operated at the positive ion full scan mode. For determination of propofol glucuronide, the gradient elution program was as follows: 0–2 min, 80% B; 2–4 min, 80-30% B; and 4–5 min, 30-80% B. For determination of galangin sulfate, the gradient elution program was as follows: 0–1 min, 90%B; 1–3.5 min; 90-10% B; 3.5–4.5 min, 10% B; and 4.5–5 min, 10–90% B. The mass spectrometer was operated at the negative ion full scan mode for propofol glucuronide and galangin sulfate. The flow rate was set at 0.3 ml/min. Peak areas were recorded with extract masses as follows: m/z 261.03 ± 0.05 Da for CPA, 306.09 ± 0.05 Da for OH-CPA O-methyloxime; m/z 353.16 ± 0.05 Da for propofol glucuronide; and m/z 348.9 ± 0.05 Da for galangin sulfate.

Hepa-1c1c7 cells were cultured in DMEM containing 10% fetal bovine serum (FBS). AML-12 cells were cultured in DMEM/F12 supplemented with 0.1% dexamethasone, 1% ITS (i.e., insulin, transferrin and selenium) and 10% FBS. Cells were seeded into 6-well plates. Once growing to a density of 60–70%, cells were co-transfected with Per2 expression plasmid (2 μg) or siPer2 (50 nM) or control using jetPRIME transfection reagent (Polyplus Transfection, Illkirch, France) according to the manufacturer’s protocol. After 24 or 48 h, cells were collected for mRNA and protein quantification.

Total RNA was extracted from mouse liver and cell samples with TRIzol reagent (Accurate Biotech, Hunan, China) following the manufacturer’s instructions, and used as a template for cDNA synthesis. qPCR reactions were performed with SYBR Green Master Mix (Vazyme, Nanjing, China) using a Biometra Toptical Thermocycler (Analytik Jena, Goettingen, Germany) as previously described (Yamamura et al., 2010). Amplification reactions consisted of an initial denaturation at 95°C for 5 min, followed by 40 cycles of denaturation at 95°C for 15 s, annealing at 60°C for 30 s, and extension at 72°C for 30 s. Peptidylprolyl isomerase B (Ppib) was used as an internal control. Relative mRNA expression was calculated using the 2^-∆∆Ct method and normalized to the control group. Primer sequences are shown in Table 1.

Mouse tissues and cell samples were lysed in RIPA buffer supplemented with a protease inhibitor cocktail (HY-K0010, Monmouth Junction, NJ). Protein concentration was detected by a BCA assay kit (Beyotime, Shanghai, China). After denaturing at 95°C for 10 min, protein samples (40 μg) were subjected to SDS-polyacrylamide gel electrophoresis (10% gel) and transferred to polyvinylidene fluoride membranes (Millipore, Bedford, MA). The membranes were sequentially incubated with primary antibody and HRP-conjugated secondary antibody. Protein bands were visualized using the Omega Lum G imaging system (Aplegen), and band density was analyzed by using FlourChem software (Bio-Rad). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. Protein expression was normalized to GAPDH.

HEK293T cells were cultured in DMEM supplemented with 10% FBS and seeded into 48-well plates. Cells were transfected with Cyp2b10 luciferase (firefly) reporter plasmid (50 ng), pRL-TK (15 ng) and overexpression plasmid (Per2 or Rev-erbα, 200 ng) or blank pcDNA3.1 (control) using jetPRIME transfection reagent (PolyplusTransfection, Illkirch, France). After 24 h, cells were lysed in 45 μl passive lysis buffer (Promega, Madison, WI), and cell lysate was collected to determine the luciferase activities using the Dual-Luciferase Reporter Assay system on a GloMax 20/20 luminometer (Promega). Firefly luciferase activities were normalized to renilla luciferase values, and expressed as relative luciferase units.

Data are presented as mean ± SD (standard deviation). Sample sizes are provided in figure legends. Student’s t-test was used to analyze a statistical difference between two groups. One-way ANOVA followed by Bonferroni post hoc test was used for multiple group comparisons. Statistical analyses were performed with GraphPad Prism 7 (GraphPad Software Inc., San Diego, CA). Pharmacokinetic analysis was performed using WinNonlin software (Pharsight Corp, St. Louis, MO). The level of significance was set at p < 0.05 (*).

Per2-deficient mice with exons 4-6 deleted (named Per2 ^Del4-6 mice) were obtained from Cyagen Biosciences (Guangzhou, China) (Figure 1A). PCR genotyping of mice was performed using tail biopsies with the primers F1, F2 and R1/2, resulting in a 707 bp fragment for Per2 ^Del4-6 mice and a 514 bp fragment for wild-type allele (Figure 1B). qPCR assay with primer set 1 (targeting the knockout sequence) confirmed the absence of wild-type Per2 transcript in mouse liver (Figure 1C). Western blotting showed that hepatic PER2 protein was markedly reduced in Per2 ^Del4-6 mice, however, it was not completely lost (Figure 1D). qPCR assay with primer set 2 targeting non-knockout sequence suggested the presence of a new version of Per2 transcript in Per2 ^Del4-6 mice albeit at a reduced level (Figure 1C). This new transcript may be translated to a PER2-related protein that retains a reactivity with the commercial PER2 antibody, thereby explaining why PER2 protein was detected in Per2 ^Del4-6 mice (Figure 1D). Taken together, Per2 ^Del4-6 mice were characterized with loss of wild-type Per2 transcript and markedly reduced PER2 protein.

Per2 expression displays a robust diurnal oscillation with the nadir at ZT2 and the zenith at ZT14 (Yamamura et al., 2010; Yoo et al., 2017). Therefore, the impact of Per2 on expression of drug-metabolizing genes was assessed at two diurnal time points (i.e., ZT2 and ZT14) using Per2 ^Del4-6 mice. According to qPCR analyses, hepatic expression of several drug-metabolizing genes (including Cyp2a4/2a5, Cyp2b10, Ugt1a1, Ugt1a9, Ugt2b36, Sult1a1, and Sult1e1) was altered (and majority were down-regulated) at both ZT2 and ZT14 in Per2 ^Del4-6 mice (Figure 2A). Of note, Cyp2b10, Ugt1a9, and Sult1a1 were three genes considerably affected, independent of diurnal times (Figure 2B). Consistent with the mRNA changes, the proteins of CYP2B10, UGT1A9 and SULT1A1 in the liver were significantly reduced in Per2 ^Del4-6 mice (Figure 2C). In addition, Per2 ^Del4-6 mice showed decreased enzymatic activities of CYP2B10, UGT1A9 and SULT1A1 against their specific substrates (CPA for CYP2B10, propofol for UGT1A9, and galangin for SULT1A1) based on in vitro incubation assays (Figure 3). Taken together, these data revealed a role of Per2 in regulation of drug-metabolizing enzymes.

Since CYP2B10 expression was altered in the liver of Per2 ^Del4-6 mice, we next tried to test whether this affects the metabolism and pharmacokinetics of substrate drugs. To this end, we performed pharmacokinetic experiments of CPA with Per2 ^Del4-6 and control mice. Per2 ^Del4-6 mice showed increases in plasma CPA concentrations (Figure 4A). Accordingly, AUC (area under the curve, a measure of systemic exposure) of CPA was significantly increased in Per2 ^Del4-6 mice (Table 2). In the meantime, liver CPA concentrations were increased in Per2 ^Del4-6 mice (Figure 4B). By contrast, the plasma and hepatic levels of the metabolite 4-hydroxycyclophosphamide (4-OH-CPA) were lower in Per2 ^Del4-6 mice, suggesting reduced metabolism of CPA (Figures 4C,D). Therefore, Per2 ablation altered the pharmacokinetics of CPA in mice by down-regulating drug metabolism.

CPA is an anticancer prodrug whose cytotoxic effects depend on metabolic activation to its metabolites such as 4-OH-CPA (Richardson et al., 2016). Our previous study shows that CPA treatment induces hepatotoxicity in mice, which is positively related to formation of 4-OH-CPA (Zhao et al., 2019). We thus examined whether CPA-induced hepatotoxicity is affected in Per2 ^Del4-6 mice. We found lower levels of ALT and AST activities, but a higher GSH level in the liver of Per2 ^Del4-6 mice, indicating reduced hepatotoxicity in the genetically modified mice (Figure 5). This was consistent with reduced metabolism of CPA and lowered formation of 4-OH-CPA (Figure 4). Altogether, CPA hepatotoxicity is reduced in Per2 ^Del4-6 mice due to decreased metabolism.

Next, we assessed the regulatory effects of Per2 on CYP2B10 expression in mouse Hepa-1c1c7 hepatoma cells by performing overexpression and knockdown assays. We confirmed that overexpression plasmid significantly elevated the levels of PER2 mRNA and protein, whereas siRNA decreased PER2 expression, in Hepa-1c1c7 cells (Figure 6A). Per2 overexpression significantly increased cellular levels of CYP2B10 mRNA and protein (Figure 6B). By contrast, Per2 knockdown reduced cellular expression of CYP2B10 (Figure 6B). In an identical manner, we examined the effects of Per2 on CYP2B10 expression in mouse AML-12 hepatocytes (Figure 7). Per2 overexpression significantly increased CYP2B10 expression, whereas Per2 knockdown deceased CYP2B10 expression, in AML-12 cells (Figures 7A,B). Taken together, Per2 positively regulates CYP2B10 expression in both Hepa-1c1c7 and AML-12 cells, consistent with our in vivo findings (Figure 2).

PER2 is an integral component of circadian clock, and functions as a co-repressor to repress the transcriptional activity of the BMAL1/CLOCK heterodimer, a well-known activator of many clock-controlled genes (Langmesser et al., 2008). Thus, it was speculated that PER2 may regulate CYP2B10 expression through an indirect mechanism that involves BMAL1/CLOCK and an intermediate regulator. This intermediate regulator should be a target of BMAL1/CLOCK and a repressor of CYP2B10. A survey of the literature suggested REV-ERBα as a candidate for such intermediate regulator because it transcriptionally inhibits CYP2B10 and the expression itself is directly driven by BMAL1/CLOCK (Dunlap, 1999; Crumbley and Burris, 2011; Zhang et al., 2018b). We therefore tested whether REV-ERBα mediates PER2 regulation of CYP2B10. We found that REV-ERBα mRNA and protein were significantly increased in the liver of Per2 ^Del4-6 mice (Figure 8A). This was expected because PER2 can repress the expression of REV-ERBα by inhibiting the transcriptional activity of BMAL1/CLOCK (Liu et al., 2008). We confirmed that REV-ERBα is a transcriptional repressor of Cyp2b10 in HEK293T cells based on luciferase reporter assays (Figure 8B). Intriguingly, PER2 dose-dependently induced the transcription of Cyp2b10 (Figure 8C). However, the induction effect of PER2 was attenuated in Rev-erbα silenced cells (Figure 8D). Moreover, overexpressing Rev-erbα reversed the effect of PER2 on Cyp2b10 transcription (Figure 8E). Taken together, our findings suggest that PER2 regulates CYP2B10 expression by down-regulating REV-ERBα.