Influence of Quercetin on Diabetes-Induced Alteration in CYP3A Activity and Bioavailability of Pioglitazone in Rats

Problem statement: Quercetin-a common bioflavonoid, is present in herb al preparations consumed by diabetic patients along with routine an ti-diabetic agents. We recently showed that quercetin increases the bioavailability of pioglita zone in non-diabetic rats. Thus, present study investigated whether this pharmacokinetic interacti on is also evident in diabetic animals, especially because diabetic subjects have altered Gastrointest inal (GI) function and CYP3A activity. Approach: The study was carried out in alloxan-induced (40 m g kg, i.v.) diabetic rats. After 2 weeks of diabetes induction, rats were treated for 2 week s with quercetin (10 mg kg , p.o.) or vehicle (5% methyl cellulose, 10 mL kg ). At the end of 4 weeks, these rats were used to i nvestigate: (1) GI function in terms of gastric emptying and intestinal transit of semisolid barium sulphate meal; (2) CYP3A activ ity in liver and intestinal microsomes by erythromycin N-demethylase assay; (3) plasma levels of orally an d intravenously administered pioglitazone (10 mg kg , p.o.; 5 mg kg, i.v.). Results: The results revealed that diabetic rats exhibited: (1) delayed gastric e mptying and intestinal transit; (2) decreased CYP3A activity and (3) a significant increase in the oral and intravenous AUC 0-∞ of pioglitazone as compared to non-diabetic rats. Quercetin treatment prevented the diabetes-induced GI dysfunction, whereas diabetes-induced decrease in CYP3A activity and inc reased bioavailability of pioglitazone remained unaffected. Conclusion: The results suggested that quercetin attenuated GI ysfunction but did not affect the bioavailability of pioglitazone in diabe tic rats contrary to the increase reported in nondiabetic rats. However, the safety of co-joint use of quercetin containing herbs and pioglitazone in clinical practice requires further pharmacokinetic substantiation.


INTRODUCTION
Several herbal preparations are often used by the diabetic patients along with the routine antidiabetic agents with an intention to produce better glycemic control and to reduce diabetic complications [1,2] . Such herbal preparations often contain bioflavonoid that are antioxidant in nature and thus help to eliminate hyperglycemia-induced oxidative stress [1,3] . Some of the bioflavonoid are reported to inhibit the activity of CYP3A-a class of CYP450 isoenzymes [4][5][6] and incidentally this class of enzymes is responsible for metabolizing pioglitazone and nateglinide, the often used oral antidiabetic agents [7,8] .
Recently, we have shown that quercetin-a bioflavonoid, decreases the metabolism of pioglitazone by inhibiting the CYP3A activity in non-diabetic rats and increases its bioavailability [9] . However, the CYP3A content and its activity in liver and intestinal tissues are low in diabetic condition [10] . Moreover, type 2 diabetes mellitus is associated with polymorphism of CYP3A4 [11] . In addition, diabetic subjects often exhibit variety of Gastrointestinal (GI) dysfunctions such as delayed gastric emptying and intestinal transit [12][13][14] . Under these circumstances, the drug-drug interaction between quercetin and pioglitazone evident in nondiabetic condition cannot be straightway extrapolated to diabetic condition.
Hence, the present study investigated: (1) the functional status of GIT; (2) the CYP3A activity in intestinal and liver tissues and (3) the pharmacokinetics of pioglitazone in alloxan-induced diabetic rats treated with vehicle or quercetin.

MATERIALS AND METHODS
Animals: Subjects were young, healthy adult male Sprague-Dawley rats (250-270 g), purchased from National Institute of Nutrition, Hyderabad (AP), India. Rats were housed in groups of 3-4 per cage in opaque polypropylene cages and maintained at 25±2°C with 12 Experimental design and animal model: Rats were divided randomly into three body weight-matched groups (n = 12). One group was non-diabetic control group and was administered saline (1 mL kg −1 , i.v.); diabetes was induced in others by administering alloxan (40 mg kg −1 , i.v.) dissolved in cold saline (1 mL kg −1 ). Induction of diabetes was checked 48 h after injection of alloxan by glucometer. Two weeks after the induction of diabetes, diabetic rats were treated for further two weeks orally with vehicle (5% methylcellulose, 10 mL kg −1 ; diabetic control group) or with quercetin at a daily dose of (10 mg kg −1 ; quercetin treated group). At the end of 4 weeks, 3-4 animals each from a group were used for oral and intravenous pharmacokinetic studies and estimation of hepatic and intestinal CYP3A activity. The animals used for oral pharmacokinetic studies were later employed for GI function studies. The study design is represented schematically in Fig. 1. animals each from a group were used for oral and intravenous Pharmacokinetic (PK) studies and estimation of hepatic and intestinal CYP3A activity. The animals used for oral PK studies were employed for GI function studies.
The dose of alloxan was selected on the basis of previous report [15] . The dose of quercetin was selected from previous reports and our preliminary studies in such a way that treatment of diabetic rats for four weeks did not cause reversion of diabetes [16,17] . The duration of diabetes required for precipitation of gastroparesis was selected on the basis of previous report [18] and our previous observations (unpublished data).
Estimation of blood glucose levels: Blood glucose levels were measured by GOD-POD method using commercially available glucometer (Ascensia Entrust, Bayer Healthcare LLC, USA) and expressed in mg dL −1 . The rats showing fasting blood glucose levels more than 250 mg dL −1 after two days of alloxan administration were retained in the diabetic group.

Assessment of CYP3A activity in liver and intestinal microsomes:
The influence of various treatments on the CYP3A activity in liver and intestinal microsomes was studied using the method of erythromycin N demethylase assay. In brief, at the end of four weeks, overnight fasted rats were euthanized by pentobarbitone overdose and the livers were perfused with 0.1 M Phosphate Buffered Saline (PBS) and isolated. In addition, a piece of intestine ~25 cm) was isolated and cleaned with PBS. The intestinal and liver microsomes were prepared by methods reported by us earlier [9] .
The mixture of microsomal suspension (0.1 mL, 25%), erythromycin (0.1 mL, 10 mM) and potassium phosphate (0.6 ml, 100 mM, pH 7.4) was incubated at 37°C. The reaction was initiated by adding NADPH (0.1 mL, 10 mM) and terminated after 10 min, by adding ice-cold trichloroacetic acid solution (0.5 mL, 12.5% w/v). The mixture was centrifuged (2000×g) to remove proteins. 1.0 mL of NASH reagent (2 M ammonium acetate, 0.05 M glacial acetic acid, 0.02 M acetylacetone) was mixed with 1.0 ml of supernatant and heated in a water bath at 50°C for 30 min. After cooling the absorbance was measured at 412 nm. The activity was calculated from standards (1-100 µM formaldehyde) prepared by substituting sample with standard solution, which were run in parallel. All the samples were run in duplicate. The CYP3A activity was expressed as nM of formaldehyde obtained per milligram of protein per min.

Assessment of gastric emptying and intestinal transit:
Gastric emptying of a non-nutrient semi-solid meal was assessed by previously reported method [19] . In brief, at the end of four weeks and overnight fasting, a suspension of 7.5 g barium sulphate in 10 ml water was given orally to rats in a volume of 2 mL/100 g body weight. After 30 min, rats were euthanized by pentobarbitone overdose. The abdomen was cut open and esophageal and pyloric ends of stomach were clamped with a string to prevent any leakage of its residual content. The stomach was weighed along with its contents and then incised, the contents removed and stomach weighed again. The gastric barium sulphate content was calculated from the weight difference between the filled and empty stomach and the amount of BaSO 4 suspension administered was determined by weighing the dose (syringe) before administration. The percent gastric emptying rate was calculated by the following formula: Where: % GE = Percent Gastric Emptying W1 = Wt of BaSO 4 suspension administered W2 = Wt of stomach before washing-Wt of stomach after washing Intestinal transit was determined as the percentage of movement of barium sulphate in the intestine in relation to the whole length of the gut as determined by visual inspection of the test meal passage in the intestine.

Pharmacokinetic and bioavailability studies in rats:
At the end of four weeks, rats were fasted overnight and were fed 4 h after dosing pioglitazone orally. For intravenous studies animals were not fasted. The animals had free access to water during all these studies. Each rat was given either a single intravenous dose (5 mg kg −1 ) or oral dose (10 mg kg −1 ) of pioglitazone (n = 3-4 each). The intravenous doses were administered as a bolus via the saphenous vein and the oral doses by oral gavage. Pioglitazone was administered as a solution in a vehicle composed of 0.1 M citric acid in dose volume of 10 mL kg −1 for oral administration and in (30% propylene glycol+3% DMSO) mixture at a dose volume of 2 mL kg −1 for intravenous studies. Blood samples (0.3 mL) were withdrawn through tail vein from unrestrained animals [20] and collected in heparinized tubes at 0.5, 1, 2, 4, 8 and 24 h after oral administration and 0.083, 0.5, 1, 2, 8, 12 and 18 h after intravenous administration of pioglitazone. Plasma was prepared immediately by centrifugation and stored at -20°C until samples were analyzed.
Pioglitazone concentration was determined as reported by us earlier [9] . In brief, to 100.0 µL of plasma sample, 50.0 µL of rosiglitazone (12.5 mg in methanol) solution as internal standard and 100.0 µL of acetonitrile were added to precipitate the proteins. The mixture was vortex mixed for 5 min after which it was centrifuged at 10,000×g for 10 min and 20.0 µL of the supernatant was injected onto the HPLC system for analysis. The UV detector was set at 269 nm. C18 (2) column (4.6×250 mm, 100°A) Luna, PHENOMENEX, USA was set at 30°C. The flow rate was 1. Statistical analysis: Wherever required the data is analyzed by using one-way ANOVA followed by Newman-Keuls test for multiple comparisons. p<0.05 was considered statistically significant.

Induction of diabetes:
Single intravenous dose of alloxan (40 mg kg −1 ) was sufficient to induce diabetes in the rats and the blood glucose levels were above the threshold (250 mg dL −1 ) in all the animals during four weeks. The mortality due to diabetes was about 20%. The blood glucose levels were checked weekly to ensure maintenance of diabetic state. The initial and final blood glucose values along with body weights, water intake and urine output are shown in Table 1.    Pharmacokinetic studies: The mean plasma concentration-time profiles of pioglitazone after its oral and intravenous administration at a dose of 10 and 5 mg kg −1 respectively, to non-diabetic rats, vehicle treated diabetic rats and quercetin treated diabetic rats are shown in Fig. 6 and 7 and some relevant pharmacokinetic parameters are shown in Table 2 and 3. One-way ANOVA revealed a significant influence of all the treatments on the oral and intravenous pharmacokinetics of pioglitazone. The AUC 0-∞ (81% increase), T 1/2, Vd and MRT of pioglitazone was significantly increased (p<0.05) and K el and Cl was significantly decreased (p<0.05) in diabetic rats after its    intravenous administration. Similarly, the AUC 0-∞ of pioglitazone was significantly increased (p<0.05; 73.0% increase) in diabetic rats after its oral administration. However, other pharmacokinetic parameters were not significantly affected.
Quercetin treatment had no influence (p>0.05) on the diabetes-induced changes in the oral and intravenous pharmacokinetics of pioglitazone.

DISCUSSION
Pioglitazone is known to sensitize the tissues to insulin to produce its antidiabetic action [21] . If given along with a sulphonylurea or insulin, it often leads to severe hypoglycemia [22,23] . In view of this risk factor, its bioavailability is therapeutically crucial in view of possible pharmacokinetic drug-drug interactions. We had earlier shown that administration of pioglitazone along with quercetin to non-diabetic rats increases its bioavailability, as it is principally metabolized by CYP3A and quercetin inhibits the same. However, it is difficult to extrapolate such possibility in diabetic condition, particularly because diabetic rats have altered GI function and CYP3A activity. Therefore, the present investigation was carried out to study the fate of pharmacokinetic drug-drug interaction between quercetin and pioglitazone in diabetic condition.
These studies revealed that the Cmax and Area Under The Curve (AUC) of orally given pioglitazone were higher in alloxan-induced diabetic rats as compared to non-diabetic rats and the results in nondiabetic rats were in accordance with the previous reports [24] . However when given intravenously, AUC was higher, clearance was lower and Cmax remained unaffected in diabetic rats as compared with nondiabetic group. These observations indicate that the bioavailability of pioglitazone is as such higher in diabetic group as compared to non-diabetic group. The studies further revealed that pretreatment with quercetin did not influence the increased bioavailability of pioglitazone in diabetic rats and this is contrary to our previous observation that quercetin increases the same in non-diabetic rats [9] .
The GI function studies revealed that the diabetic group exhibited significant delay in gastric emptying and intestinal transit and thus mimicked the condition of diabetic gastroparesis seen in humans after long standing diabetes mellitus. In addition, it was observed that treatment with quercetin significantly prevented these diabetes-induced changes in GI function. Recently, we have reported that reduced NO levels in GI tissues is the pivotal reason for diabetes-induced GI dysfunction [19] . Quercetin is reported to enhance the bioavailability of NO in diabetic rat aortas [25] . Thus, quercetin might have produced beneficial effects in diabetic rats through modulation of NO levels in the GIT.
The delay of gastric emptying and intestinal transit is normally expected to postpone the Tmax, while improvement by quercetin is expected normalize it. However, the Tmax in diabetic group was same as that of non-diabetic group, indicating that prevailing GI function had no contribution to the bioavailability of pioglitazone and hence quercetin-induced improved GI function did not show any change in the Tmax. It is known that diabetic gastroparesis affects the emptying of solid meal while the emptying of liquid meal remains unaltered [26,27] . Incidentally, in the present investigation pioglitazone was administered in the solution form.
The diabetes-or quercetin-induced changes in CYP3A activity could be responsible for the changes in Cmax and reduction in clearance of pioglitazone. The erythromycin N-demethylase assay revealed decreased CYP3A activity in liver and intestinal microsomes obtained from diabetic group. These observations are in accordance to the earlier reported changes in the CYP3A activity in diabetic condition [10] . The CYP3A activity in these tissues was uninfluenced by quercetin treatment to diabetic group. In our earlier study we found that quercetin significantly inhibited CYP3A activity in non-diabetic rats, which is not evident in diabetic rats in the present investigation. Western blotting studies have reported that the content of CYP3A of liver and intestine is as such low in diabetic condition [10] . Thus, inadequate availability of CYP3A and decreased metabolism could be the possible reason for higher Cmax and AUC of pioglitazone in diabetic condition. For the same reason, quercetin-induced inhibition of CYP3A was not evident in diabetic group, which is otherwise seen in non-diabetic group in our earlier studies.
Thus, these studies reveal that the bioavailability of pioglitazone is higher in diabetic condition but remains unaffected by quercetin treatment. Part of this observation is contrary to a clinical report wherein type II diabetic patients exhibited normal clearance of pioglitazone [7] . However, the study did not assess CYP3A activity and/or the functional status of GIT in these patients.

CONCLUSION
The present investigations revealed that quercetin treatment improved diabetes-induced GI dysfunction with no concurrent influence on the hepatic and intestinal CYP3A activity and hence the bioavailability of pioglitazone. However, extensive clinical pharmacokinetic studies are necessary to extrapolate these observations for therapeutic enrichment.