Tissue Specific Modulation of cyp2c and cyp3a mRNA Levels and Activities by Diet-Induced Obesity in Mice: The Impact of Type 2 Diabetes on Drug Metabolizing Enzymes in Liver and Extra-Hepatic Tissues

Various diseases such as type 2 diabetes (T2D) may alter drug clearance. The objective of this study was to evaluate the effects of T2D on CYP450 expressions and activities using high-fat diet (HFD) as a model of obesity-dependent diabetes in C57BL6 mice. The cyp450 mRNA expression levels for 15 different isoforms were determined in the liver and extra-hepatic tissues (kidneys, lungs and heart) of HFD-treated animals (n = 45). Modulation of cyp450 metabolic activities by HFD was assessed using eight known substrates for specific human ortholog CYP450 isoforms: in vitro incubations were conducted with liver and extra-hepatic microsomes. Expression levels of cyp3a11 and cyp3a25 mRNA were decreased in the liver (>2–14-fold) and kidneys (>2-fold) of HFD groups which correlated with a significant reduction in midazolam metabolism (by 21- and 5-fold in hepatic and kidney microsomes, respectively, p < 0.001). HFD was associated with decreased activities of cyp2b and cyp2c subfamilies in all organs tested except in the kidneys (for tolbutamide). Other cyp450 hepatic activities were minimally or not affected by HFD. Taken together, our data suggest that substrate-dependent and tissue-dependent modulation of cyp450 metabolic capacities by early phases of T2D are observed, which could modulate drug disposition and pharmacological effects in various tissues.


Introduction
Type 2 diabetes (T2D) has become a worldwide public health concern as prevalence of the disease continues to rise [1]. In 2014, the American Diabetes Association reported that 29.1 million Americans, or 9.3% of the population, had T2D [2]. In addition to anti-diabetic drugs, T2D patients commonly require multiple drug therapies to treat a wide range of comorbidities such as hypertension, stroke, dyslipidemia, atherosclerosis and coronary artery disease [3]. Clinical practice reveals that T2D patients show highly variable pharmacokinetics and responses to several drugs used to treat T2D and its related comorbidities [4][5][6][7]. For instance, variable drug dosages and effects are observed for drugs such as clopidogrel, warfarin, cyclosporine and tacrolimus, as well as for anti-hypertensive and cholesterol

Animals
Five-week-old male C57BL/6 mice (n = 45) were purchased from Charles River Laboratories (Montreal, QC, Canada). Animals were housed in a temperature-, light-, and humidity-controlled environment, they were housed 2 per cage and were maintained at an ambient temperature of 21 • C on a 12-h light/dark cycle with free access to water and food ad libitum. The obese diabetic mice were developed according to the experimental protocol described previously [51]. Briefly, one week after their arrival, mice (mean weight 20.0 ± 1.0 g) were fed with a high-fat diet (HFD) (Bio-Serv Diet #F3282, Frenchtown, NJ, USA, 60% fat by energy) or the standard normal diet (ND) (Teklad Global 18% protein diet; Harlan Teklad, Madison, WI, USA, 15% fat by energy) for 8 weeks. Body weight was measured weekly while blood glucose and insulin were determined at Week 8. After 8 weeks of diet, HFD mice were stratified according to their body weight and two groups were formed as follows; low responders to HFD (LDR) (<39.9 g) and high responders to HFD (HDR) (39.9-45 g). Peyot et al. have demonstrated that the LDR are less obese, develop intermediate severity of insulin resistance and have mild impairment in glycemia, while the HDR are more obese, insulin resistant, hyperinsulinemic and hyperglycemic [51]. Animals were sacrificed by cervical dislocation, and organs including heart, lungs, kidneys and liver were quickly excised, washed with cold TRIS 100 mM buffer (pH 7.4) and immersed in liquid nitrogen (−80 • C). Experimental protocols were approved by the institutional committee of animal protection and were carried out in accordance with the Guide for the Care and Use of Experimental Animals of the Canadian Council on Animal Care.

Isolation of RNA and Preparation of cDNA
For each organ tested (using a pool of 3 mice/organ), about 100 mg of tissue was homogenized in 1 mL of Trizol and incubated for 5 min at room temperature. Chloroform (200 µL) was added, the mixture shaken for 15 s and then, centrifuged at 16,000× g for 30 min at 4 • C. The aqueous supernatant (500 µL) was transferred and ethanol 70% was added (1:1 v/v). RNA was extracted using the Qiagen kit (RNeasy Mini kit; Qiagen Sciences, MD, USA) according to the manufacturer's recommendations. RNA concentration and quality was assessed by spectrometry. Total RNA (2 µg) from each sample was used for reverse transcription. RNA, random primers (6 µg) and dNTP (25 mM) were preheated for 5 min at 65 • C. Then, 5X-first strand buffer, 80 units of RNAse inhibitor, DTT (0.01 M) and 400 units of Superscript II (Invitrogen, Carlsbad, CA, USA) were added to a final volume of 40 µL. Reverse transcription was carried out for 50 min at 42 • C and stopped by heating to 70 • C for 15 min (final RNA concentration 50 ng/µL). The resulting cDNA was frozen at −80 • C until analyzed.
The relative quantification of various gene expressions was calculated to the comparative CT method using the formula 2 −∆CT [52,53]. Only CT values ≤ 35 were included in the analyses. Since CT values > 35 were not reliable and considered below the detection level of the assay, a CT value of 35 to 38 was defined not quantifiable (NQ) while a value of CT > 38 as not detectable (ND). For their part, mRNA levels associated with the expression of each isoenzyme under a specific diet condition (ND, LDR, HDR) were determined using a calibrator and the following formula 2 −∆∆CT [52]. The calibrator was prepared at the same mRNA concentration using a pool of RNA obtained for each tissue (Clontech A Takara, Bio Company, Mountain View, CA, USA). Determination of mRNA levels was performed in triplicate for each sample, and three independent experiments were repeated to confirm results.

Preparation of Microsomes
Microsomes from liver, kidneys, lungs and heart were prepared according to our previously described methods with slight modifications [54]. Briefly, tissue (pools of organs from 3 mice) was homogenized in an ice-cold buffer consisting of 50 mM-150 mM-1 mM TRIS-KCL-EDTA buffer (liver and kidneys) or 100 mM-150 mM-1 mM PO 4 -KCL-EDTA buffer (lungs and heart) and both buffers containing protease inhibitors namely, PMFS (0.01 mM) and DTT (0.5 mM). Microsomal subcellular fraction was prepared by centrifugation (10,000 g × 20 min, at 4 • C) followed by ultra-centrifugation (100,000 g × 90 min, at 4 • C). The microsomal pellets were resuspended in the same buffer (without PMSF and DTT), and frozen at −80 • C until in vitro metabolism experiments were performed. The protein concentration of the microsomes was determined by the Bradford method using bovine serum albumin as the standard.

Effects of the HFD on Hepatic and Extra-Hepatic cyp450 Activities
In order to investigate the effects of the HFD as a representative model of type 2 diabetes on cyp450 activities, in vitro incubations were performed in presence of various microsomes with several probe drugs of CYP450s including bupropion, repaglinide, tolbutamide, bufuralol, chlorzoxazone, ebastine, midazolam and dodecanoic acid, which were used as markers of the functional orthologs of human CYP2B6, CYP2C8, CYP2C9, CYP2D6, CYP2J2, CYP2E1, CYP3A4/5 and CYP4A11, respectively. This study employed cocktails of probe substrates already accepted and validated to investigate the impact of diabetes on specific cyp450 activities. The production of each specific metabolite was quantified from substrate probes and variations of in vitro probes reactions can be inferred to affect all substrates metabolized by the same enzymes.
Assay conditions were previously optimized by standard incubations with probes (buffer, incubation period, protein contents and drug concentrations). All incubations were performed in triplicate. The incubation mixture containing microsomes [5 µL for liver (~20 mg proteins/mL) and 50 µL for kidney, lungs and heart microsomes (~7-14 mg proteins/mL)], NADPH-regenerating system solution (NADP 6.5 mM, G6P 16.5mM, MgCl 2 5 mM and 0.2 U G6PD) and 100 mM phosphate buffer PO 4 (pH 7.4) or 100 mM TRIS buffer (pH 7.4, for tolbutamide and dodecanoic acid) were pre-incubated in a shaking bath for 10 min at 37 • C. Reaction was initiated by the addition of substrates (bupropion, tolbutamide, bufuralol, chlorzoxazone, midazolam, ebastine, dodecanoic acid or repaglinide) to the incubation mixture (total final volume of 500 µL). Bupropion, chlorzoxazone, ebastine and midazolam were incubated together as a cocktail as previously described whereas the other probe substrates were tested separately. The substrate concentrations used span a range from 38-620 µM for bupropion, Pharmaceutics 2017, 9, 40 5 of 18 2.5-40 µM for bufuralol, 50-800 µM for chlorzoxazone, 0.125-2 µM for ebastine, 1.25-20 µM for dodecanoic acid, 0.25-4 µM for midazolam, 0.85-14 µM for repaglinide and 25-400 µM for tolbutamide (5 different concentrations were used with liver microsomes while one concentration (>2 km) to ensure saturation was selected to investigate cyp450 activities in extra-hepatic microsomes). After 30 min, the reaction was stopped using 1000 µL of ice-cold internal methanol containing isotope-labeled internal standard probe metabolite(s). Reaction mixtures were put on ice for 10 min, and following a centrifugation at 13,000 rpm for 10 min, and then the supernatant was transferred for analysis. For dodecanoic acid metabolite analysis, the solution was evaporated to dryness at 50 • C under a gentle stream of nitrogen, reconstituted with 200 µL of methanol and transferred to an injection vial for analysis.

Chromatographic Conditions for the Metabolites of Bupropion, Midazolam and Ebastine
This analysis was performed on a Thermo Scientific Acclaim RSLC Polar Advantage C16 column (75 mm × 3.0 mm, 3 µm) and Phenomenex Security Guard Cartridge (C12, 4 mm × 2 mm) operating at 50 • C. The mobile phase was a gradient elution consisting of (A) 0.1% formic acid in acetonitrile and (B) 10 mM ammonium formate in water adjusted to pH 3; ratio A:B varied from 25:75 to 60:40 (v/v), at a flow rate 500 µL/min (a total run time of 10 min). A 10 µL aliquot of the extract was injected into LC-MSMS system.
A Thermo Scientific TSQ Quantiva Triple Quadrupole mass spectrometer (San Jose, CA, USA) was interfaced with a Thermo Scientific Ultimate 3000 XRS UHPLC system (San Jose, CA, USA) using a pneumatic assisted heated electrospray ion source. MS detection was performed in positive ion mode using selected reaction monitoring (SRM). Selection of optimal transitions and collision energy and tube lens voltage conditions for the metabolites and their respective internal standard are listed in Table S1. 2.5.2. Chromatographic Conditions for the Metabolites of Chlorzoxazone, Tolbutamide, Dodecanoic Acid, Bufuralol and Repaglinide These analyses were carried out on a Phenomenex Luna PFP (2) column (150 mm × 3.0 mm I.D., 3 µm) with a Phenomenex PFP security guard cartridge operating at 40 • C. A mobile phase in isocratic mode was composed of acetonitrile and 0.01% formic acid having a fixed ratio of 40:60 (v/v) for chlorzoxazone, tolbutamide and dodecanoic acid, while a ratio of 50:50 was used for bufuralol and repaglinide. The flow rate was 0.30 mL/min (total run time 10 min). A 10 µL aliquot of the extract was injected into the LC-MSMS system.
The HPLC system consisted of a Shimadzu Prominence series UFLC pump and auto sampler (Kyoto, Japan). The tandem MS system used was a Thermo TSQ Quantum Ultra (San Jose, CA, USA). The mass spectrometer was interfaced with the HPLC system using a pneumatic assisted heated electrospray ion source. MS detection was performed in negative ion mode for chlorzoxazone, tolbutamide, and dodecanoic acid metabolites, and in positive ion mode for bufuralol and repaglinide metabolites using selected reaction monitoring (SRM). Selection of optimal transitions and collision energy and tube lens voltage conditions for the metabolites and their respective internal standard are shown in Table S2.

Statistical Analysis
Calibration curves were calculated from the equation y = ax + b, as determined by weighted 1/x and 1/x 2 linear regressions of the calibration lines constructed from the peak-area ratios of metabolites to the internal standard (XCalibur software, Thermo Fisher, San Jose, CA, USA). Relative expression levels of CYP450s mRNAs were analyzed by one-way analysis of variance followed by the Dunnett post-hoc test. A difference with p < 0.05 was considered statistically significant. Enzyme kinetic Pharmaceutics 2017, 9, 40 6 of 18 parameters were determined by non-linear regression analysis using Michaelis-Menten equation and Lineweaver-Burk double reciprocal plot and data points were expressed as the mean ± S.D., K m and V max values and the 95% confidence interval for the intrinsic clearance. Data were analyzed using GraphPad Prism 5 (GraphPad Software, La Jolla, CA, USA) and SAS statistical software (Version 9.4 of the SAS System for Windows, Copyright©, SAS Institute Inc., Cary, NC, USA).

Animal Model
Although no animal model exactly reflects human T2D, some have similar features. Human T2D is a heterogeneous disorder with a complex interplay between genetic, epigenetic and environmental factors. On one hand, diabetic animal models including chemically-induced or surgically-provoked develop hyperglycemia primarily by cytotoxic actions on beta cells rather than through insulin resistance. On the other hand, transgenic/knockout model are more useful to investigate the role of a specific candidate gene unlike heterogeneity as seen in humans. Moreover, the observed diabetes conditions are less stable; chemicals can produce toxic actions and development of digestive problems can be observed which could also affect CYP450 activities. Consequently, a validated nutritionally (high-fat diet; HFD) obese mouse model has been selected to characterize the effects of obesity-induced diabetes on CYP450 activities [51].

Hepatic Activities
As shown in Figure 2 ( Table 2), HFD induced variations in hepatic cyp450 activities in an isoform-dependent manner. A significant decrease in midazolam metabolism, a marker of cyp3a subfamily, was observed following HFD treatment; the intrinsic clearance of 1-hydroxymidazolam was reduced in LDR and HDR (23 and 40 µL/min/mg prot) groups compared to ND (107 µL/min/mg prot) (p < 0.001) (Table S4). This was mostly explained by a decrease in V max ; formation of 1-hydroxymidazolam decreased from 0.32 (ND) to 0.06 nmol/mg protein/min (LDR and HDR, p < 0.001) ( Table 2). HFD treatment was also associated with a diminished hepatic activity of cyp2c subfamilies compared to ND group. The intrinsic clearance of tolbutamide (used as a probe of CYP2C9 in human) was significantly reduced from 0.80 to 0.54 and 0.57 µL/min/mg prot in ND, LDR and HDR groups, respectively (p < 0.05) (Table S4). Similarly, HFD affected also the repaglinide hydroxylation (CL int to M1-hydroxyrepaglinide) from 10.6 µL/min/mg prot in ND compared to 1.0-1.4 µL/min/mg prot in HFD groups) (Table S4). Cyp2b activity as measured by bupropion hydroxylation tended to be slightly decreased in HDR group (CL int in HDR vs. ND reduced by 22%). In contrast, no significant effect was observed on the hepatic hydroxylation of bufuralol (cyp2d), dodecanoic acid (cyp4a), chlorzoxazone (cyp2e1) and ebastine (cyp2j) as demonstrated by comparable pharmacokinetics values ( Table 2).

Hepatic Activities
As shown in Figure 2 ( Table 2), HFD induced variations in hepatic cyp450 activities in an isoform-dependent manner. A significant decrease in midazolam metabolism, a marker of cyp3a subfamily, was observed following HFD treatment; the intrinsic clearance of 1-hydroxymidazolam was reduced in LDR and HDR (23 and 40 µL/min/mg prot) groups compared to ND (107 µL/min/mg prot) (p < 0.001) (Table S4). This was mostly explained by a decrease in Vmax; formation of 1hydroxymidazolam decreased from 0.32 (ND) to 0.06 nmol/mg protein/min (LDR and HDR, p < 0.001) ( Table 2). HFD treatment was also associated with a diminished hepatic activity of cyp2c subfamilies compared to ND group. The intrinsic clearance of tolbutamide (used as a probe of CYP2C9 in human) was significantly reduced from 0.80 to 0.54 and 0.57 µL/min/mg prot in ND, LDR and HDR groups, respectively (p < 0.05) (Table S4). Similarly, HFD affected also the repaglinide hydroxylation (CLint to M1-hydroxyrepaglinide) from 10.6 µL/min/mg prot in ND compared to 1.0-1.4 µL/min/mg prot in HFD groups) (Table S4). Cyp2b activity as measured by bupropion hydroxylation tended to be slightly decreased in HDR group (CLint in HDR vs. ND reduced by 22%). In contrast, no significant effect was observed on the hepatic hydroxylation of bufuralol (cyp2d), dodecanoic acid (cyp4a), chlorzoxazone (cyp2e1) and ebastine (cyp2j) as demonstrated by comparable pharmacokinetics values ( Table 2).    (Table 3) illustrates the effects of diabetes induced by HFD on the formation rate of cyp450 probe metabolites measured in extra-hepatic tissues. A tissue-dependent modulation of cyp450 activities by HFD was observed. In the kidneys, HFD treatment produced a significant decrease in cyp3a activity; formation rate of 1-hydroxymidazolam was three-and five-times lower in LDR and HDR groups, respectively, compared to ND group (p < 0.001) ( Table 3(A)). In contrast, the cyp3a activity measured in the lung was not affected by the HFD indicating a tissue-dependent modulation of cyp450 by HFD. No cyp3a activity was detectable in mouse hearts.

Extra-Hepatic Activities
Our data showed that the formation of hydroxybupropion was approximately 20-50-times greater in the lungs compared to the liver and kidneys (regardless of the diet groups, p < 0.0001) ( Table 3(B)). Renal and lung microsomes displayed a slight decrease in the hydroxylation of bupropion in HDR group vs. control diet group (p < 0.01) (Figure 3). Although the magnitude of activity was low, a similar observation was made for the hydroxylation of bupropion in the heart (0.150 vs. 0.037 and 0.025 pmol/mg protein/min in ND vs. LDR and HDR, respectively, p < 0.001) ( Table 3(C)). In addition, ebastine (cyp2j) and dodecanoic acid (cyp4a) metabolisms were reduced by HFD, particularly in HDR, in renal and lung tissues. Ebastine hydroxylation in heart microsomes was not affected by HFD, whereas cyp4a activity could not be detected. In contrast, our results showed Figure 2. Hepatic cyp450 activities of cyp2b (bupropion), cyp2c (tolbutamide and repaglinide), cyp2d (bufuralol), cyp2e (chlorzoxazone), cyp2j (ebastine), cyp3a (midazolam) and cyp4a (dodecanoic acid) in C57BL6 mice fed a normal diet (ND) or a HFD (LDR, low-diet responders; and HDR, high-diet responders after an eight-week period of treatment). Data were expressed as the mean ± S.D. LDR or HDR vs. ND; * p < 0.05. N/F; activity not found, N/A; Not Available (in vitro incubation could not be determined) and dotted line represents the limit of quantification. Table 2. Hepatic microsome activities for cyp2b (bupropion), cyp2c (tolbutamide and repaglinide), cyp2d (bufuralol), cyp2e (chlorzoxazone), cyp2j (ebastine), cyp3a (midazolam) and cyp4a (dodecanoic acid) in C57BL6 mice fed a normal diet (ND) or a HFD (LDR, low-diet responders; and HDR, high-diet responders after an eight-week period). Results are expressed as mean N-fold differences in cyp gene relative to the average expression of housekeeping genes and a calibrator (2 −ΔΔCt sample ). ND low-diet responders; and HDR, high-diet responders. NQ = Not Quantifiable (35 < Ct < 38), ND = Not Detectable (Ct > 38). Each experiment was perform in triplicates. One-way ANOVA was performed with Dunnett post-hoc test. LDR or HDR versus ND; * p < 0.05, ** p < 0.01, ᵵ p < 0.001. Repaglinide → Hydroxyrepaglinide 0.0020 ± 0.0001 0.0010 ± 0.0002 7, 9, 40 8 (Table 3) illustrates the effects of diabetes induced by HFD on the formation rate of cyp450 probe metabolites measured in extra-hepatic tissues. A tissue-dependent modulation of cyp450 activities by HFD was observed. In the kidneys, HFD treatment produced a significant decrease in cyp3a activity; formation rate of 1-hydroxymidazolam was three-and five-times lower in LDR and HDR groups, respectively, compared to ND group (p < 0.001) (Table 3(A)). In contrast, the cyp3a activity measured in the lung was not affected by the HFD indicating a tissue-dependent modulation of cyp450 by HFD. No cyp3a activity was detectable in mouse hearts.

Extra-Hepatic Activities
Our data showed that the formation of hydroxybupropion was approximately 20-50-times greater in the lungs compared to the liver and kidneys (regardless of the diet groups, p < 0.0001) (Table 3(B)). Renal and lung microsomes displayed a slight decrease in the hydroxylation of bupropion in HDR group vs. control diet group (p < 0.01) (Figure 3). Although the magnitude of activity was low, a similar observation was made for the hydroxylation of bupropion in the heart (0.150 vs. 0.037 and 0.025 pmol/mg protein/min in ND vs. LDR and HDR, respectively, p < 0.001) ( Table 3(C)). In addition, ebastine (cyp2j) and dodecanoic acid (cyp4a) metabolisms were reduced by HFD, particularly in HDR, in renal and lung tissues. Ebastine hydroxylation in heart microsomes was not affected by HFD, whereas cyp4a activity could not be detected. In contrast, our results showed that DIO mouse were associated with an increase in tolbutamide (~20-50%) and bufuralol (~90-110%) metabolic activities in kidneys, while in lung microsomes, the hydroxylation of tolbutamide and bufuralol tended to decrease (p < 0.01 and p < 0.001, respectively, in HDR group). Overall, cyp450 activities tended to be reduced in HFD group with greater effects being observed in HDR.

Discussion
This study demonstrated that expression and activities of major cyp450s involved in the metabolism of drugs were modulated in DIO C57BL6 mice used as a model of T2D (Table S5). First, we demonstrated that cyp3a expression and activities were decreased by HFD. Second, cyp2c activities were reduced in all organs tested (except for tolbutamide in the kidneys). Finally, cyp2b activity, largely expressed in the lungs, was also decreased by HFD.
C57BL/6 mice were divided into three groups: ND, LDR and HDR. LDR were less obese and developed intermediate severity of insulin resistance, while HDR were more obese and developed severe insulin resistance. The extent of changes in cyp450 activities by HFD tended to be gradually decreased in LDR compared to HDR and some effects being only observed in HDR group. This observation suggests that modulation of some cyp450 activities happens at early stage of pre-diabetes and similar pathways of regulation involved in pre-diabetes development can also intervene on cyp450 activities in an isoform-dependent manner.
Obesity and diabetes have been shown to alter the expression and activity of hepatic CYP450s [14,28,29,[35][36][37][38][40][41][42][55][56][57]. Clinical studies and animal experiments report mostly on hepatic CYP3A since it is the most important isoenzyme involved in the metabolism of prescribed drugs [58,59]. In humans, a significant decrease in CYP3A activity has been reported in diabetic human liver microsomes compared to healthy subjects [56], while a significant increase in CYP2E1 activity has been reported in obese T2D human liver microsomes [33]. In our study, we observed a significant decrease in hepatic cyp3a mRNA expression levels in HFD groups compared to normal diet group. Similarly, decreased levels of cyp3a11 mRNA expression have been reported in DIO mice [38,59,60]. Preliminary results from our clinical study using oral CYP450 probe cocktail demonstrate that oral clearance of midazolam was significantly reduced in subjects with T2D compared to subjects without T2D [61]. Our results on RNA transcripts are also consistent with phenotypic findings showing that cyp3a activity determined via hydroxylation of midazolam was significantly reduced in the two HFD groups compared to control group. In agreement with our finding, CYP3A activity has been reported to be also decreased in Zucker diabetic fatty (ZDF) rats using midazolam and testosterone as probes [40].
There are no or very limited data pertaining to the effects of T2D on cyp2c and cyp2b families. We observed that cyp2c and cyp2b subfamilies mRNA expression levels were not changed in the liver by HFD compared to normal group, except for cyp2b9 which was increased in HFD groups. Our results are in agreement with those reported by Yoshinari et al. and Guo et al. who demonstrated unchanged relative mRNA levels of cyp2c and cyp2b subfamilies in the liver [59,60]. In contrast, it has been reported a decrease of cyp2b10 mRNA levels in CD1 mice fed with HFD [38]. However, the relative hepatic levels of cyp2b10 mRNA measured in our study were very low or below the limit of quantification yielding a comparison analysis unreliable. In addition, our results showed that hepatic cyp2c metabolic activities, using repaglinide and tolbutamide as markers, were significantly decreased in HFD groups. Kim et al. reported that CYP2C catalytic activity determined with diclofenac was decreased in chemically induced diabetes in rats [62]. This finding, in agreement with our study, suggest that CYP2C activities are impaired, thereby lower metabolic clearance can be anticipated for drugs metabolized by CYP2Cs, particularly CYP2C8 or CYP2C9, under conditions associated with pre-diabetes or diabetes.
Our results indicate that cyp2e1 mRNA expression levels and activity remained unchanged and comparable among all groups (HFD and ND). In the same way, expression levels of cyp2e1 mRNA were reported to be unchanged in DIO mice [38,59]. In addition, activity of cyp2e1 was also unaffected in DIO and db/db mice [37,38], but increased activity was shown in ZDF [42]. No alteration in hepatic levels of cyp2e1 was found in streptozocin-induced diabetic mice [41]. These differences in cyp2e1 activity modulation could be function of the animal model of diabetes used and the stage of the disease.
Our data demonstrated a tissue-specific modulation of cyp2b activities by HFD. Indeed, no significant difference was observed for the hydroxylation of bupropion in liver microsomes. However, bupropion hydroxylation was significantly decreased in HFD groups, particularly in the lungs, the heart and kidneys (HDR group only).
To our knowledge, no study has been conducted to assess the impact of HFD induced obesity as a model of early diabetes stage in humans on CYP450s expression or activity in extrahepatic tissues. Although the magnitude of metabolic capacity and significance in total body clearance of extrahepatic CYP450s metabolism are much lower in comparison to hepatic CYP450s, extrahepatic CYP450s metabolism may affect the local exposure to xenobiotics and thus, influence their pharmacological and toxicological effects. For instance, renal cyp450 metabolites of arachidonic acid, 20-HETE and EET, play an important role in the control of blood pressure and the development of acute kidney injury [63]. In fact, arachidonic acid is metabolized by CYP4A and CYP4F families to 20-HETE (vasoconstrictor) and by CYP2C and CYP2J families to EETs (vasodilator). Our results showed a significant decrease in renal activity of cyp4a and cyp2j in the HDR group using dodecanoic acid and ebastine hydroxylation as marker, respectively. This finding indicates that diabetes could influence homeostasis by affecting local biotransformation of endogenous compounds. In high fat diet induced hypertension rats, CYP2C and CYP4A activities were found to be decreased [64]. The discrepancy observed for CYP2C activity can be explained by using different substrate markers (arachidonic acid being not specific for CYP2C but, also a CYP2J substrate).
Little is known about the effects of obesity and diabetes on CYP450 catalytic activities in the lungs, in both humans and animals. Our data showed extensive metabolic activities in lung microsomes for midazolam, bupropion, ebastine and chlorzoxazone corresponding to cyp3a, cyp2b, cyp2j and cyp2e1, respectively. In lung microsomes, HFD was associated with modulation of cyp450 catalytic activity for certain isoforms independently of patterns observed in the liver or in the kidneys. These findings support the concept that CYP450s expressed in the lungs may contribute to drug metabolism for drugs administered intravenously or locally as well as contribute to the first pass metabolism.
In conclusion, the major finding of this study was that HFD affects CYP450 expression and activities in an isoform-and tissue-dependent fashion. Our results clearly indicate that cyp3a and cyp2c metabolic activities were reduced in DIO-T2D mice (Table S5). In humans, these two CYP450 isoenzymes are involved in the metabolism of 80% of medications prescribed in clinical settings. We speculate that modulation of hepatic CYP450s by T2D diabetes may alter drug pharmacokinetics leading to intersubject variability in drug response. In addition, modulation in CYP450s expressed in extra-hepatic organs can cause variations in tissue concentrations of drugs or endogenous compounds leading to impaired pharmacological action of drugs as well as disruption of homeostasis. Therefore, variation in hepatic and extra-hepatic CYP450s makes patients with pre-diabetes and obesity more prone to adverse drug effect, toxicity or inefficacy (e.g., for prodrugs).

RT-qPCR
real time quantitative polymerase chain reaction T2D type 2 diabetes VHDR very high-diet responders