Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Stereoselective pharmacokinetics and placental transfer of fluoxetine and norfluoxetine in pregnant sheep… Chien, Caly 2004

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2004-901637.pdf [ 10.25MB ]
Metadata
JSON: 831-1.0091754.json
JSON-LD: 831-1.0091754-ld.json
RDF/XML (Pretty): 831-1.0091754-rdf.xml
RDF/JSON: 831-1.0091754-rdf.json
Turtle: 831-1.0091754-turtle.txt
N-Triples: 831-1.0091754-rdf-ntriples.txt
Original Record: 831-1.0091754-source.json
Full Text
831-1.0091754-fulltext.txt
Citation
831-1.0091754.ris

Full Text

Stereoselective Pharmacokinetics and Placental Transfer of Fluoxetine and Norfluoxetine in Pregnant Sheep Under Steady-State Conditions by Caly Chien B.Sc.(Pharm.), The University of Alberta, Edmonton, Canada, 1997 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Faculty of Pharmaceutical Sciences) (Division of Pharmaceutics and Biopharmaceutics) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 2004 © Caly Chien Abstract Fluoxetine (FX) is one of the most popular SSRIs used for the treatment of depression. Its use in treating depressed pregnant woman has increased substantially in recent years. A considerable amount of attention in the literature regarding the use of FX in pregnancy has been devoted to the assessment of birth outcomes following in utero FX exposures while studies that focus on pharmacokinetics during pregnancy are relatively sparse. Previous FX experiments in pregnant sheep using single dose intravenous (i.v.) bolus administration have focused primarily on the acute exposure and effects of the drug on both the mother and fetus. In the present studies, experiments were conducted to investigate the stereoselective pharmacokinetics of FX and norfluoxetine (NFX) under steady-state conditions in nonpregnant sheep and pregnant sheep at late gestation via continuous 8-day i.v. infusion of FX to the ewe. Drug infusions were conducted at two dose levels (i.e. low dose and high dose) in nonpregnant sheep that allowed the examination of nonlinear pharmacokinetics of FX in sheep. The pregnant sheep infusion experiment was to determine the impacts of pregnancy on the stereoselective pharmacokinetics of FX and NFX, and to assess the exposure of the fetus to FX and NFX enantiomers at steady-state. Moreover, a continuous 4-day direct fetal i.v. FX infusion experiment was conducted to study the pharmacokinetics of FX and NFX in fetal lambs in attempt to explore intrinsic drug elimination capability by the fetus. In nonpregnant sheep, FX and NFX enantiomers exhibited significant pharmacokinetic differences under steady-state conditions, confirming the observations made in previous i.v. bolus experiments. For FX, the R-enantiomer had a lower steady^ state concentration (Css) ii level but a longer elimination half-life (t|/2p) than the S-enantiomer while for NFX, both the R- and S-enantiomers had similar C s s levels but the ti/2p was slightly longer for R-NFX compared to S-NFX. FX and NFX also displayed stereospecific protein binding, with free fractions of the R-enantiomers that were about 2 to 3-fold higher than those of the S-enantiomers. In comparison to humans, quantitative differences exist but qualitatively the overall stereoselective disposition of FX and NFX are similar between sheep and humans. Using the literature reported value for hepatic blood flow in sheep, the intrinsic clearances (CL int) for the FX enantiomers were determined. CL i n t of R-FX was estimated to be -50% higher than that of S-FX, indicating that, in addition to stereoselective protein binding, stereoselective metabolism also played a pivotal role in the overall stereoselective disposition of FX. Moreover, FX displayed a disproportional change in pharmacokinetics in sheep upon long-term administration, characterized by a decrease in total body clearance (CL TB) and increase in dose normalized area under the curve (AUC) with higher dose. This nonlinearity may be due to inhibition of a CYP2D6-like enzyme in the sheep liver, as this is the major drug metabolizing enzyme for FX in humans. In pregnant sheep, the stereoselective pharmacokinetics of FX and NFX under steady-state conditions were similar to those observed in nonpregnant animals. No remarkable changes in the S/R ratio of FX and NFX were observed. Similarly, no change in FX clearance was observed as opposed to the previous finding of a higher clearance in pregnant sheep following single i.v. bolus administration. Protein binding did not appear to differ between pregnant and nonpregnant sheep for either FX and NFX. FX and NFX exhibited moderate degrees of placental transfer upon chronic administration, with a fetal-to-maternal (F/M) AUC ratios of about 0.40 and 0.33, respectively. When concentrations were corrected for the in difference in free fraction between the maternal and fetal compartments, the F/M ratio still remained less than unity, suggesting the presence of other nonplacental routes of elimination for FX in the fetus. Analyses of amniotic and fetal tracheal fluids showed that excretion of FX and NFX into these fluids resulted in levels not exceeding those in fetal plasma and significant accumulations were therefore unlikely. Comparisons of F/M ratios between the R- and S-enantiomers revealed no difference between R-FX and S-FX, while S-NFX was slightly higher (-25%) compared to R-NFX, indicating that NFX distributed between the mother and the fetus in a stereoselective manner. Following fetal i.v. infusion of FX, alterations in fetal blood gas status, characterized by decreases in PO2 and O2 saturation, and increases in PC02 and pH, were noted. Due to variability between animals, the majority of these effects did not reach statistical significance. The stereoselective pharmacokinetics of FX and NFX in fetal lambs were similar to those observed in adult sheep, but the extent of stereoselectivity in terms of the S/R ratio was less. The differential steady-state concentrations of the FX enantiomers observed in the fetal circulation were likely due to stereoselective protein binding in fetal plasma and differential maternal-to-fetal placental transfer of the enantiomers preceded by maternal stereoselective metabolism. Based on the F/M ratios for NFX (-0.62) following fetal FX infusion, it appears that the fetus has limited metabolic capability, if any, to biotransform FX to NFX. It is not fully known, however, whether other metabolic pathways such as the formation of trifluoromethylphenol via O-dealkylation are functional in the fetus. Renal excretion of FX and NFX did not contribute substantially to their overall clearance in the fetus (<1% of CLJB). In comparison to maternal and nonpregnant adult sheep, weight normalized CLTB, C L r e n a i , and Vd s s were significantly higher in the fetus. Unbound fractions of FX and NFX iv were also higher in the fetus owning to the lower concentration of plasma protein in fetal plasma. The intrinsic capability of the fetus to eliminate FX was assessed by calculating the maternal and fetal placental and nonplacental clearances based on the 2-compartment model for the maternal-fetal unit. The results indicated that both placental and nonplacental routes were equally important in fetal elimination of FX. It is concluded that other elimination routes are likely present in the fetus but these remain to be elucidated. Table of Contents Abstract ii Table of Contents vi List of Tables x List of Figures xii List of Abbreviations xviii Acknowledgments xxiv C H A P T E R 1 I N T R O D U C T I O N 1 1.1 Background 3 1.1.1 Physicochemical Properties 3 1.1.2 Pharmacology and Therapeutic Uses 4 1.1.3 Pharmacokinetics 4 1.1.4 Metabolism 5 1.1.5 Stereoselective Properties of Fluoxetine and Norfluoxetine 8 1.1.6 Depression and FX in Pregnancy 11 1.1.7 Effects of Pregnancy on Pharmacokinetics 13 1.1.7.1 Absorption 15 1.1.7.2 Distribution 15 1.1.7.3 Protein Binding 16 1.1.7.4 Metabolism 17 1.1.7.5 Excretion 19 1.1.7.6 Placental Transfer 19 1.1.8 Placental Transfer and Pharmacokinetics of FX in Pregnancy: Previous Studies .21 1.2 Rationale 24 1.3 Hypotheses 26 1.4 Objectives .27 C H A P T E R 2 M A T E R I A L S , I N S T R U M E N T A T I O N A N D A N A L Y T I C A L M E T H O D O L O G I E S 28 VI 2.1 Materials 28 2.2 Instrumentation 30 2.2.1 Gas Chromatography-Mass Spectrometry 30 2.2.2 Physiological Monitoring 31 2.2.3 Other Equipment 31 2.3 Analysis of FX and NFX Enantiomers in Biological Fluids 32 CHAPTER 3 STEADY-STATE STEREOSELECTIVE PHARMACOKINETICS OF FX AND NFX IN NONPREGNANT SHEEP 34 3.1 Methods .' 35 3.1.1 Animal and Surgical Preparation 35 3.1.2 Experimental Protocol 36 3.1.3 Determination of FX and NFX Plasma Protein Binding : 38 3.1.4 Pharmacokinetic Analysis 39 3.1.5 Statistical Analysis 40 3.2 Results 41 3.2.1 Stereoselective Pharmacokinetics 41 3.2.1.1 Fluoxetine 41 3.2.1.2 Norfluoxetine 49 3.2.2 Nonlinear Pharmacokinetics 52 3.2.2.1 Fluoxetine 52 3.2.2.2 Norfluoxetine 58 3.3 Discussion 60 3.3.1 Stereoselective Pharmacokinetics 61 3.3.1.1 Fluoxetine 61 3.3.1.2 Norfluoxetine 66 3.3.2 Nonlinearity 72 3.3.3 Mechanism of Stereoselective Pharmacokinetics of FX 76 3.4 Summary 83 v i i ' CHAPTER 4 AN INVESTIGATION OF THE EFFFECTS OF PREGNANCY ON THE STEREOSELECTIVE DISPOSITION OF FX AND NFX IN PREGNANT SHEEP 84 4.1 Methods 85 4.1.1 Animal and Surgical Preparation 85 4.1.2 Experimental Protocol 85 4.1.3 Determination of FX and NFX Plasma Protein Binding 86 4.1.4 Pharmacokinetic Analysis 86 4.1.5 Statistical Analysis 87 4.2 Results 87 4.2.1 Fluoxetine 87 4.2.2 Norfluoxetine 94 4.3 Discussion 98 4.4 Summary 120 CHAPTER 5 STUDY OF THE PLACENTAL TRANSFER OF FX AND NFX FOLLOWING MATERNAL INTRAVENOUS INFUSION OF RACEMIC FX TO STEADY-STATE 121 5.1 Methods 122 5.1.1 Animal and Surgical Preparation 122 5.1.2 Experimental Protocol 123 5.1.3 Pharmacokinetic Analysis 124 5.1.4 Statistical Analysis 125 5.2 Results .- 125 5.2.1 FX and NFX Levels in Fetal Arterial Plasma 125 5.2.2 FX and NFX Levels in Amniotic and Fetal Tracheal Fluids 131 5.3 Discussion 136 5.4 Summary 153 viii CHAPTER 6 DISPOSITION OF FX AND NFX IN FETAL LAMBS UNDER STEADY-STATE CONDITIONS 155 6.1 Methods 156 6.1.1 Animal and Surgical Preparation 156 6.1.2 Experimental Protocol 156 6.1.3 Physiological Recording and Monitoring Procedures 158 6.1.4 Determination of FX and NFX Plasma Protein Binding 159 6.1.5 Pharmacokinetic Analysis 159 6.1.6 Statistical Analysis 160 6.2 Results 160 6.2.1 Stereoselective Disposition of FX and NFX in the Fetus 162 6.2.2 Comparative Pharmacokinetics of FX and NFX in Adult, Maternal, and Fetal Sheep 169 6.2.3 Fetal Blood Gas Results 172 6.3 Discussion 175 6.4 Summary 193 CHAPTER 7 O V E R A L L SUMMARY AND CONCLUSIONS 194 REFERENCES 202 IX List of Tables Table 1.1 Stereoselective properties of FX and NFX enantiomers. Table 1.2 Pregnancy-associated physiologic changes and their potential influence on drug pharmacokinetic parameters. Table 3.1 Pharmacokinetic parameters of the FX enantiomers in non-pregnant sheep after high dose (HD) and low dose (LD) 8-day racemic FX infusions and the corresponding /.-values. Table 3.2 Pharmacokinetic parameters for the renal elimination of FX and NFX in nonpregnant sheep after high dose (HD) and- low dose (LD) 8-day racemic FX infusions and the corresponding p-values. Table 3.3 Pharmacokinetic parameters of NFX enantiomers in non-pregnant sheep after high dose (HD) and low dose (LD) 8-day racemic FX infusions and the corresponding/^ -values. Table 3.4 Comparison of FX pharmacokinetic parameter values determined from different dosing experiments in nonpregnant sheep. Table 3.5 Comparison of renal FX and NFX elimination kinetic parameter values obtained from different dosing experiments in non-pregnant sheep. Table 3.6 Comparison of pharmacokinetic parameters of NFX between different dosing experiments in nonpregnant sheep. Table 3.7 Comparison of R versus S relationships of FX and NFX enantiomers in sheep and humans. Table 3.8 Comparative extraction ratio (E) and intrinsic clearance (CLjnt) values for the FX enantiomers in sheep and humans. Table 4.1 Comparison of pharmacokinetic parameters and S/R ratios for FX enantiomers in nonpregnant and pregnant sheep and the corresponding p-values. Table 4.2 Comparison of pharmacokinetic parameters and S/R ratios of the NFX enantiomers in nonpregnant and pregnant sheep and the corresponding /^ -values. Table 4.3a Pharmacokinetic parameters derived from simulation of the dextromethorphan and dextrorphan plasma concentration-time profiles in human CYP2D6 poor metabolizers (PM) during and following pregnancy. 9 14 45 48 50 53 57 59 65 81 91 96 106 X Table 4.3b Comparative dextromethorphan/dextrorphan (DM/DX) meta- 106 bolic ratios determined at various individual time points as well as from AUC data. Table 4.4 Comparative pharmacokinetic parameters of pindolol enantio- 119 mers in pregnant and nonpregnant human subjects. Table 5.1 Pharmacokinetic parameters of total, R- and S-FX in fetal lambs 128 following an 8-day maternal infusion of racemic FX. Table 5.2 Pharmacokinetic parameters of total, R- and S-NFX in fetal 129 lambs following an 8-day maternal infusion of racemic FX. Table 5.3 Pharmacokinetic parameters of FX and NFX in amniotic and 132 fetal tracheal fluids following 8-day maternal infusion of racemic FX. Table 5.4 Literature data comparing the effect of physicochemical 139 properties of various therapeutic agents on their degree of placental transfer. Table 6.1 Pharmacokinetic parameters for the FX and NFX enantiomers in 164 fetal lambs after a 4-day fetal infusion of racemic FX. Table 6.2 Comparison of FX enantiomers pharmacokinetic parameters 170 values in nonpregnant adult, pregnant and fetal sheep. Table 6.3 Comparison of NFX enantiomers pharmacokinetic parameter 171 values in nonpregnant adult, pregnant and fetal sheep. Table 6.4 Mean fetal arterial blood gas status at 0700 h on each day of a 4- 177 day fetal FX i.v. infusion and post-infusion period. Table 6.5 Average values of fetal and maternal placental (CL f m and CLmf, 186 respectively) and nonplacental clearances (CLf0 and C L m 0 , respectively) of FX and other drugs studied in pregnant sheep during late gestation. Table 6.6 Comparative kinetic parameters for FX N-demethylation and O- 190 dealkylation in human liver microsomes. Table 7.1 Mean (%CV) summary of plasma pharmacokinetic parameters 195 for fluoxetine and norfluoxetine in sheep. XI List of Figures Figure 1.1 Chemical structure of fluoxetine enantiomers. 3 Figure 1.2 A -^demethylation of fluoxetine to norfluoxetine. 5 Figure 1.3 Proposed schematic diagram of FX metabolic pathway. 7 Figure 2.1 Representative total ion chromatogram of FX, NFX, and internal 33 standard in SIM mode, showing chromatographic separation of the FX and NFX enantiomers in a run time of 22 minutes. Figure 3.1 Schematic diagram of the experimental protocol used in the 37 nonpregnant adult sheep experiments. Figure 3.2 Representative plasma concentration versus time profiles of FX 42 and NFX enantiomers in a nonpregnant sheep (E7139) for a high dose 8-day i.v. infusion experiment. The inset figure shows the concentration-time profile during the first 6 hours of infusion. Figure 3.3 Representative plasma concentration versus time profiles of FX 42 and NFX enantiomers in a nonpregnant sheep (E8107) for a low dose 8-day i.v. infusion experiment. Figure 3.4 Comparison of the average plasma concentration versus time 43 profiles for total FX from the high and low dose FX 8-day i.v. infusion experiments in nonpregnant adult sheep (n = 6). Figure 3.5 Comparison of the average plasma concentration versus time 43 profiles of total NFX from the high and low dose FX 8-day i.v. infusion experiments in nonpregnant adult sheep (n = 6). Figure 3.6 Representative plot of the cumulative amount of FX excreted in the 47 urine of a nonpregnant sheep (E7139) for a high dose FX 8-day i.v. infusion experiment. Figure 3.7 Representative plot of the cumulative amount of NFX excreted in 47 the urine of a nonpregnant sheep (E7139) for a high dose FX 8-day /.v. infusion experiment. Figure 3.8 Total body clearance of Total FX, R-FX, and S-FX following high 54 dose infusion, low dose infusion, and single i.v. bolus administration of racemic FX. ** A significant dose-related decrease in clearance is observed for S-FX. X l l Figure 3.9 Dose normalized AUC values of Total-FX,'R-FX, and S-FX (Chart 56 A), and Total NFX, R-NFX, and S-NFX (Chart B) following high dose infusion, low dose infusion, and single i.v. bolus administration of racemic FX. * and ** denote significant differences from the i.v. bolus data with p-values of <0.01 and <0.05, respectively. Figure 3.10 Simulations of NFX enantiomer concentration-time profiles 68 following an 8-day continuous i.v. infusion of FX. The solid (—) and dotted lines (---) denote the concentration-time curves of R-NFX and S-NFX, respectively. The S/R ratio for Vd was fixed at 0.5 in all cases. In Panel A, the S/R ratios offm and C L m were 0.5 and 1, respectively. In Panel B, the S/R ratios offm and C L m were 1 and 0.5, respectively. In Panel C, the S/R ratios offm and C L m were 0.5 and 0.5, respectively. In Panel D, the S/R ratios offm and C L m were 1 and 1, respectively. Panel C closely resembles the concentration-time profiles observed for the NFX enantiomers in the present 8-day FX infusion experiments. Figure 3.11 Interrelationships between the S/R ratios offm and AUC (Panel A); 70 S/R ratios of C L m and AUC (Panel B); S/R ratios of C L m and t,/2 (Panel C). The S/R ratio of C L m is fixed at 0.5 in (A); S/R ratio of fm is fixed at 0.5 in (B) and (C). The arrow sign (—») denotes the point when a change in the direction of stereoselectivity occurs. Figure 3.12 Plot of extraction ratio versus predicted CL; n t value calculated from 79 equation 3.11 and the true CLjn, value calculated using equation 3.13. QII and fu were assumed to be 1.5 l/min and 0.01, respectively, in both cases. A significant deviation (>30%) between the predicted and true value occurred when the extraction ratio was greater than 0.3. Figure 4.1 Representative plasma concentration versus time profiles of FX 88 and NFX enantiomers in a pregnant ewe (E4237) obtained for an 8-day maternal i.v. infusion of racemic FX. Figure 4.2 Comparison of average plasma concentration versus time profiles 88 of total FX and NFX in pregnant sheep after 8-day maternal i.v. infusion of racemic FX (n=5). Figure 4.3 Comparison of average plasma concentration versus time profiles 90 of total FX of pregnant and nonpregnant sheep after 8-day i.v. infusion of racemic FX. Figure 4.4 Comparison of average plasma concentration versus time profiles 90 of FX enantiomers in pregnant and nonpregnant sheep after an 8-day i.v. infusion of racemic FX. xiii Figure 4.5 Schematic diagram of average S/R FX concentration ratio versus 93 time profiles from the nonpregnant and pregnant studies illustrating the dependency of the S/R ratio to time to steady-state in all cases (error bars have been omitted for clarity). Figure 4.6 Plot of the S/R FX concentration ratio versus the corresponding 93 total FX concentration data from the nonpregnant and pregnant studies, showing the plateau of S/R ratio at steady-state concentrations in each case. Figure 4.7 Comparison of the average plasma concentration versus time 95 profiles of total NFX of pregnant and nonpregnant sheep during and following an 8-day i.v. infusion of racemic FX, showing longer elimination half-lives of NFX in pregnant sheep compared to nonpregnant sheep. Figure 4.8 Comparison of the average plasma concentration versus time 95 profiles of the NFX enantiomers in pregnant and nonpregnant sheep during and following an 8-day i.v. infusion of racemic FX. Figure 4.9 Simulations of the plasma concentration-time profiles for 105 dextromethorphan and dextrorphan. Panel A illustrates the simulated profiles in poor metabolizer of CYP2D6 based on the literature data (Schadel et al.,1995). Metabolic ratios determined at 2 h and 5 h (=Tmax) are 4.06 and 4.08, respectively. Panel B illustrates the comparative profiles between pregnant and nonpregnant subjects. C m a x is delayed to 8 h when ka is reduced by 50% to mimic the gastric delaying effect in pregnancy. Predicted values for the metabolic ratios at 2 h and 8 h during pregnancy is 4.05 and 4.45, respectively Figure 4.10 Metabolism of dextromethorphan by the CYP2D6 and CYP3A4 108 pathways to dextrorphan, 3-methoxymorphinan, and 3-hydroxy-morphinan (adapted from Capon et al., 1996; Wadelius et al., 1997). Figure 4.11 Plot of S/R FX AUC ratio versus the corresponding total body 116 clearance value (CLTB) of FX enantiomers (data from the non-pregnant (closed symbol) and pregnant sheep (open symbol) experiments). Figure 5.1 Average plasma concentration versus time profiles of total FX and 126 NFX in fetal sheep (n=5) following an 8-day maternal i.v. infusion of racemic FX, showing concentration differences between FX and NFX from day-1 to 8 of the infusion period. XIV Figure 5.2 Representative plasma concentration versus time profiles of FX 126 and NFX enantiomers in a fetus (E4237) following an 8-day maternal i.v. infusion of racemic FX, showing stereoselective disposition of FX enantiomers in the fetus. Figure 5.3 Average plasma concentration versus time profiles of total FX in 130 the mother and fetus (n = 5) following an 8-day maternal i.v. infusion of racemic FX, with an average F/M AUC ratio of 0.40 ± 0.08. Figure 5.4 Average plasma concentration versus time profiles of total NFX in 130 the mother and fetus (n = 5) following an 8-day maternal i.v. infusion of racemic FX, showing accumulation of NFX in the fetus with an average F/M AUC ratio of 0.33 + 0.11. Figure 5.5 Average amniotic fluid (AMN), fetal tracheal fluid (TR) and fetal 133 arterial plasma (FA) concentration versus time profiles of total FX in the fetus following an 8-day maternal i.v. infusion of racemic FX, showing comparable concentrations of FX in TR and FA but lower in AMN. Figure 5.6 Average amniotic fluid (AMN), fetal tracheal fluid (TR) and fetal 133 arterial plasma (FA) concentration versus time profiles of total NFX in the fetus following an 8-day maternal i.v. infusion of racemic FX, showing the degree of accumulation of NFX in the following order: TR ~ FA > AMN. Figure 5.7 Representative amniotic fluid concentration versus time profiles of 135 FX and NFX enantiomers in E4237 following an 8-day maternal /.v. infusion of racemic FX, showing stereoselective disposition of both FX and NFX enantiomers in the amniotic fluid. Figure 5.8 Representative fetal tracheal fluid concentration versus time 135 profiles of FX and NFX enantiomers in E4237 following an 8-day maternal i.v. infusion of racemic FX, showing stereoselective disposition of both FX and NFX enantiomers in the fetal tracheal fluid. Figure 5.9 Placental transfer of therapeutic agents in pregnant sheep in 140 relation to their physicochemical properties.. Fetal-maternal exposure ratios (F/M ratio) were determined following maternal drug administration (see Table 5.4 for individual values). Plot A: log P vs. F/M ratio; Plot B: unionized fraction vs. F/M ratio; Plot C: molecular weight vs. F/M ratio. Only compounds with molecular weight of 100-600 are included in plots A and B. Figure 5.10 Compartmental model of the maternal-fetal unit with continuous 143 drug input into the maternal compartment (Szeto et al., 1982b). XV Figure 5.11 Relationship of F/M ratios between sheep and human. In Plot A, 149 no correlation is observed when all 19 compounds are included for analysis (Table 5.4). In Plot B, significant linear correlation was observed between sheep and human values with 14 compounds when outliers are excluded (p<0.01, Spearman rank correlation). Compounds that fit the correlation include (1) fluoxetine, (2) indomethacin, (3) labetalol, (4) meperidine, (5) midazolam, (6) morphine, (7) nifedipine, (8) norfluoxetine, (9) propranolol, (10) quinidine, (11) quinine, (12) ritodrine, (13) salicylic acid, and (14) valproic acid. Figure 6.1 Schematic diagram of the experimental protocol used for (A) the 157 preliminary 2-step fetal infusion study, where infusion rate at step 2 is double the rate of step 1; (B) the 4-day fetal infusion study (L.D. = loading dose) followed by a 4-day post-infusion period. Figure 6.2 Plasma concentration versus time profiles of total FX and NFX in 161 the fetus (E8105) after step-wise fetal i.v. infusion of racemic FX at two different rates, showing steady-state concentrations of FX after each infusion rate and continuous accumulation of NFX during the entire infusion period. Figure 6.3 Comparative average plasma concentration versus time profiles of 163 total (i.e. R plus S enantiomers) FX and NFX in the fetuses (n=4) after a 4-day fetal i.v. infusion of racemic FX, illustrating the concentration differences between parent drug and metabolite. Figure 6.4 Representative plasma concentration versus time profiles of FX 163 and NFX enantiomers in a fetus (E9140) after a 4-day fetal i.v. infusion of racemic FX, showing stereoselective steady-state concentrations of the FX enantiomers and a low level of NFX accumulation in the fetal plasma. Figure 6.5 Average plasma concentration versus time profiles of total FX (i.e. 166 R plus S enantiomers) in the mother and fetus (n=4) after a 4-day fetal i.v. infusion of racemic FX, with F/M AUC ratio averaging 2.87 ±0.93. Figure 6.6 Plot of the time course of the cumulative amount of FX excreted in 166 urine of a fetus (E8105) from the 2-step racemic FX fetal i.v. infusion experiment, demonstrating a linear relationship between renal elimination rate and dose. Figure 6.7 Representative plot of the amount of FX enantiomers excreted in 167 fetal urine versus AUC during the urine collection time intervals tn-tn+i for a 4-day FX fetal i.v. infusion experiment (E9168). XVI Figure 6.8 Average plasma concentration versus time profiles of total NFX 167 (i.e. R plus S enantiomers) in the mother and fetus after a 4-day fetal i.v. infusion of racemic FX, showing higher accumulation of NFX in maternal plasma compared to fetal plasma (F/M AUC ratio = 0.62 ± 0.13; n=4). Figure 6.9 Mean fetal arterial P0 2 (A) and PC0 2 (B) during the first 24 h of 173 fetal FX infusion (closed symbols, n = 5) or maternal FX infusion (open symbols, n = 13). # and * denote significant differences (p<0.05) from pre-infusion day in fetal and maternal infusion experiments, respectively. Figure 6.10 Mean fetal arterial pH (A) and oxygen saturation (B) during the 174 first 24 h of fetal FX infusion (closed symbols, n = 5) or maternal FX infusion (open symbols, n = 13). # and * denote significant differences (p<0.05) from pre-infusion day in fetal and maternal infusion experiments, respectively. Figure 6.11 Mean fetal arterial lactate (A) and glucose (B) concentrations 176 during the first 24 h of fetal FX infusion (closed symbols, n = 5) or maternal FX infusion (open symbols, n = 13). # and * denote significant differences (/?<0.05) from pre-infusion day in fetal and maternal infusion experiments, respectively. Figure 6.12 Plot of the maternal S/R FX concentration ratio versus the 179 corresponding total FX concentration, a compilation of all data from the nonpregnant and pregnant (following either maternal or fetal i.v. FX infusion) sheep experiments. Figure 6.13 A representation of various placental and non-placental drug 183 clearances in the 2-compartment pharmacokinetic model of the maternal-fetal unit (CL m o - maternal nonplacental clearance; CLf0 -fetal nonplacental clearance; CLmf - placental clearance from the mother to the fetus; CLfm - placental clearance from the fetus to the mother; k0 - drug infusion rate to the mother; k0' - drug infusion rate to the fetus). (Adapted from Szeto et al., 1982b). Figure 6.14 Mean maternal plasma concentration versus time profiles of (A) 192 FX and (B) NFX during the initial 24 hours following maternal or fetal FX i.v. infusion. XVII List of Abbreviations 5-HT °C rig pi um pM AAG ADME AIC AMN ANOVA AUC ,0-1 AUC' AUC°'O T A U C m A U M C P BW C b c f Cm c t C L H CLff 0-ra 5-Hydroxytryptamine or serotonin Micron Degree Celsius Microgram Microliter Micrometer Micromolar Approximately oci-acid glycoprotein Absorption, distribution, metabolism, elimination Akaike's Information Criterion Amniotic Analysis of variance Area under the plasma concentration vs. time curve Area under the curve from time zero to the last time point sampled Area under the curve from time zero to infinity Area under the curve of metabolite Area under the first moment curve from time zero to infinity Terminal elimination rate constant Body weight Drug concentration in buffer Fetal drug concentration Concentration of metabolite or maternal drug concentration Last time point sampled Hepatic clearance Fetal total clearance CLfm Placental clearance from the fetus to the mother C L f 0 Fetal nonplacental clearance CLjnt Intrinsic clearance C L m Elimination clearance of metabolite CL n lm Maternal total clearance CL n lo Maternal nonplacental clearance CL m f Placental clearance from the mother to the fetus C L n o n . renal Non-renal clearance C L r e n a l Renal clearance of the total drug CLTB Total body clearance based on total drug concentrations Cmax Maximal plasma concentration c P Plasma concentration of the total drug Css Steady-state concentration Css(m) Steady-state plasma concentration of metabolite C u Plasma concentration of the unbound drug C Y P Cytochrome P-450 enzyme d Day Da Dalton DM Dextromethorphan DPHM Diphenhydramine DPMA Diphenylmethoxyacetic acid DX Dextrorphan E Extraction ratio EDTA Ethylenediaminetetraacetic acid EI Electron impact EM Extensive metabolizer eV Electron volt F Systemic bioavailability F/M Fetal-to-maternal FA Fetal femoral arterial plasma fa Total fraction of the administered dose absorbed fm Fraction of drug converted to metabolite fu Unbound fraction FX Fluoxetine g Gram GC Gas chromatography GC-MS Gas Chromatography Mass Spectrometry GFR Glomerular filtration rate GI Gastrointestinal h Hour HD High dose HPLC High performance liquid chromatography i.d. Internal diameter i.e. id est; that is i.v. Intravenous k Overall first order elimination rate constant ka First order absorption rate constant kf- Formation rate constant of metabolite from parent drug kg Kilogram Kj Inhibitor affinity for an enzyme, the inhibitory constant km Elimination rate constant of metabolite Km Substrate concentration at which the reaction velocity is at half -maximal, a Michaelis-Menten parameter k0 Infusion rate 1 Liter X X LC Liquid chromatograph LD Low dose logP Logarithm of partition coefficient (octanol-to-water) LOQ Limit of quantitation of the assay M Molar (moles/litre) MA Maternal arterial plasma mEq Milliequivalent mg Milligram min Minute ml Milliliter mm Millimeter mM Millimolar mm Hg Millimeter of mercury MRT Mean residence time MS Mass spectrometry MW Molecular weight n Number of subjects or animals NFX Norfluoxetine ng Nanogram nmol Nanomole o.d. Outer diameter 0 2 Sat Oxygen Saturation PBS Phosphate-buffered saline P C 0 2 Partial gas pressure of carbon dioxide PD Pharmacodynamic P-gP P-glycoprotein pH Negative logarithm of hydrogen ion concentration PK Pharmacokinetic pKa Negative logarithm of acid association constant PM Poor metabolizer pmol picomole Po 2 Partial gas pressure of oxygen psi Pound per square inch QI Bottom quartile Q3 Top quartile QC Quality control Q H Hepatic blood flow r2 Coefficient of determination rs Spearman rank correlation R-FX R-fluoxetine R-NFX R-norfluoxetine RSD Relative standard deviation SC Schwarz Criterion S/R S-to-R ratio S.D. Standard deviation S.E. Standard error of mean S-FX S-fluoxetine SIM Single ion monitoring S-NFX S-norfluoxetine SSRI Selective serotonin reuptake inhibitor x Dosing interval t|/2p Apparent terminal elimination half-life TCAs Tricyclic antidepressants TEA Triethylamine TFAP (S)-Trifluoroacetylprolyl chloride TFMP Trifluoromethylphenol tm a x Time of occurrence of maximal plasma concentration TR Fetal tracheal UGT UDP-glucuronosyltransferase Vd Volume of distribution Vd m Volume of distribution of metabolite Vd s s Apparent steady-state volume of distribution Vmax Maximal velocity of an enzymatic reaction, a Michaelis-Menten parameter X u Cumulative amount of drug or metabolite collected in urine XXl l l Acknowledgments I would like to express my sincere acknowledgment and thanks to my research supervisors, Drs. K. Wayne Riggs and Dan W. Rurak, for their guidance, encouragement, friendship, and patience throughout my graduate training. The free thinking and independent learning atmosphere that they provided in the lab have been the most valuable and enjoyable experience of my Ph.D. program. Many thanks to the members of my graduate research committee, Drs. Frank S. Abbott, Thomas K.H. Chang, Timothy F. Oberlander, and Ronald E. Reid for their valuable time and suggestions. The help from Dr. Janna L. Morrison for getting me through most of the sheep experiments is greatly appreciated. Recognitions must also be given to Mr. Eddie Kwan and Ms. Nancy Gruber for their technical assistance with sheep studies. A very special thanks my friend, Dr. Sam Au-Yeung, for sharing my joyous and depressed moments with me over these years. Thanks to other members of our lab, Ms. Nancy Chung, Dr. John Kim, Dr. Sanjeev Kumar, Ms. Karen Lo, Ms. Agatha Ng, Ms. Manisha Ramaswamy, Mr. Swamy Subramanian, Mr. Vincent Tong, and Dr. Harvey Wong for their help and friendship. I would also like to thank Mr. Albert Wong for allowing me to work at his pharmacy part-time with great flexibility while accomplishing my intensive Ph.D. work. Moreover, I would like to acknowledge the financial support received from Rx&D Health Research Foundation/MRC Graduate Research Scholarship in Pharmacy and UBC Graduate Fellowship during the course of my graduate studies. These studies were funded by Canadian Institutes of Health Research. This thesis is dedicated to my family, my mom and dad, my brothers Collins and Johny, and my grandmother. A special dedication is made to my beautiful wife, Agatha Ching, for her love, patience, support, encouragement, prayers, and trust. A final tribute is given to my Lord Jesus Christ for His countless blessings throughout every path of my life. The completion of this Ph.D. work is not possible without Him. "Unless the Lord builds the house, its builders labor in vain. Unless the Lord watches over the city, the watchmen stand guard in vain.'" Psalm' 127:1 X X I V CHAPTER 1 INTRODUCTION Preface Although it has been long known that stereoisomers of a chiral drug often exhibit pronounced differences in their pharmacokinetic and pharmacodynamic properties both in quantitative and qualitative terms, more than 500 drugs are currently marketed as racemic mixtures without relevant pharmacokinetic and pharmacodynamic information for each individual stereoisomer (Aboul-Enein and Wainer, 1997). This neglect of stereochemistry in pharmaceutical research was wide spread until Ariens' famous critical review of sophisticated scientific nonsense was published (Ariens, 1984). It was Ariens' review that finally incited drug researchers to consider the importance of stereoselective differences. The advent of efficient methods for large-scale stereoselective organic synthesis and the introduction of chromatographic systems for stereoselective analysis now make it much easier to study enantiomer drugs and their metabolites in biological fluids. As a consequence our knowledge of differences in individual enantiomer pharmacodynamic and/or pharmacokinetic properties as a result of stereoselective interaction with optically active biological macromolecules is mounting (Campbell, 1990; Wright and Jamali, 1993; Srinivas et al., 2001). The study of stereoselective pharmacokinetics in the context of pregnancy, however, remains very limited. During pregnancy, a number of physiological functions undergo substantial changes which will significantly alter the pharmacokinetics/pharmacodynamics of drugs as well as the subsequent degree of exposure to the developing fetus (Loebstein et al., 1997). The naive 1 assumption that for therapeutic purposes a pregnant woman is no more than a nonpregnant woman with a fetus attached is seriously flawed. Although it is not totally clear whether pregnancy will affect the stereoselective properties of drug disposition, physicians should be aware of these changes since this knowledge will define the principles that guide the therapeutic use of racemic drugs in pregnant women. Unfortunately, information on stereoselective pharmacokinetics and fetal drug exposure in human pregnancy is fragmentary and generally unsatisfactory. There is not a single drug with which a reasonably complete picture can be presented. In an experimental setting, drugs can only be given ethically for a clinical need so that standardized conditions can seldom be achieved. Well-designed experiments using appropriate animal models are therefore necessary to provide detailed pharmacokinetic and fetal drug exposure data that cannot be obtained ethically and practically in human subjects. This thesis is an investigation of the stereoselective pharmacokinetics and placental transfer of the selective serotonin reuptake inhibitor fluoxetine (FX) and its metabolite, norfluoxetine (NFX) in chronically catheterized pregnant sheep. Fluoxetine was chosen as the study compound based on its 1) extensive use in the pregnant population, 2) unique stereoselective pharmacokinetic and pharmacodynamic properties, and 3) nonlinear characteristics upon multiple dosing. Our previous studies involving a single i.v. bolus dose of racemic fluoxetine revealed significant stereoselective disposition and considerable fetal exposure of fluoxetine in sheep (Kim et al., 2004). The present study was undertaken to further examine the stereoselective pharmacokinetics and placental transfer of fluoxetine in the mother and fetus under steady-state conditions, a more clinically relevant situation in human therapeutics. It is hoped that the information obtained from our studies will add to the 2 knowledge currently available in the literature and allow for a more rational approach to drug therapy in pregnancy. 1.1 B A C K G R O U N D 1.1.1 Physicochemical Properties Fluoxetine (FX), (±)-A/-methyl-3-phenyl-3-[(a,a,a-trifluoro-p-tolyl)-oxy] propylamine contains one asymmetric center which gives rise to the R- and S-enantiomers (Figure 1.1). The commercially available drug is a racemic mixture of 2 optical isomers. It has a low molecular weight of 309.33 Daltons (Da). It is a weakly basic amine of high lipophilicity, with pKa and octanol/water partition coefficient (Log P) of 9.97 and 4.05, respectively (Florey et al., 1990; Hansch et al., 1990). S-fluoxetine R-fluoxetine F I G U R E 1.1 Chemical structure of fluoxetine enantiomers. 3 1.1.2 Pharmacology and Therapeutic Uses FX belongs to a class of therapeutic agents called the Selective Serotonin Reuptake /nhibitors (SSRI). It acts by inhibiting the serotonin reuptake pump thereby increasing the amount of serotonin (5-HT) available at the cerebral receptor sites, which is related to antidepressant action. The 1988 introduction of fluoxetine revolutionized the treatment of depression. Due to its low affinity for norepinephrine and dopamine reuptake sites and the lack of effects on the muscarinic, cholinergic, histaminic, and alphai-adrenergic receptors at usual therapeutic dosages, FX is devoid of life-threatening side effects of traditional tricyclic antidepressants (TCAs), such as cardiotoxicity and CNS toxicity, as well other intolerable side effects that include dry mouth, abnormal vision, memory impairment, constipation, and urinary retention (Gram, 1994). The favorable safety profile of the SSRIs makes FX widely used in the treatment of depression as well as other psychological disorders, including anxiety, obsessive-compulsive disorders, bulimia nervosa and premenstrual dysphoric disorders. FX is used at dosages of 20 to 80 mg/day (PDR, 2002). The recommended starting dose of 20 mg has been shown to be the most frequently used dose in the majority of patients and also the best tolerated (Beasley et al., 2000). 1.1.3 Pharmacokinetics In humans, FX is well absorbed with a bioavailability of -72-90% after oral administration. Following a single oral 40 mg dose, peak plasma concentrations of FX ranging from 15 to 55 ng/ml were observed after 6 to 8 hours (PDR, 2002). It is highly protein bound with 95% of the drug bound to plasma proteins (Hiemke and Hartter, 2000). Alpha 1-acid- glycoprotein (AAG) is believed to be the major plasma protein responsible for 4 FX binding (Holladay et al., 1998). FX is highly distributed into tissues and has a volume of distribution (Vd) of 14-100 1/kg. The lungs show the highest accumulation of FX, and lysosomal trapping is considered to play a role in the high Vd of FX (Daniel and Wojcikowski, 1997). The elimination half-life in humans is about 1-4 days, which is the longest among other agents in this class of drugs. Urinary excretion of the intact drug is low, with less than 11% of administered dose excreted unchanged; therefore metabolism accounts for the majority of the clearance of fluoxetine (Catterson and Preskorn, 1996). Fluoxetine exhibits nonlinear pharmacokinetics based on evidence of a prolonged half-life (to approximately 4-5 days) and a disproportionate increase in plasma concentration, which occurred after administration of multiple versus single doses of the drug (DeVane, 1994; Hiemke and Hartter, 2000). 1.1.4 Metabolism Fluoxetine undergoes extensive biotransformation. One of its major metabolites is the N-demethylated product, norfluoxetine (NFX) (Figure 1.2). C F 3 C F 3 CYP2D6, CYP2C9, CYP3A4 -CH, C H 2 C H 2 N - C H 3 H C H 2 C H 2 N - H I H FIGURE 1.2 7V-demethylation of fluoxetine to norfluoxetine. 5 This A -^demethylation pathway is mediated by cytochrome P450 (CYP) enzymes. Data from clinical pharmacokinetic studies suggest a pivotal role for CYP2D6 in the clearance of FX because the area under the FX concentration versus time curve was higher in CYP2D6 poor metabolizers than extensive metabolizers (Hamelin et al., 1996; Fjordside et al., 1999), while in vitro studies show that multiple CYP enzymes are involved in the formation of NFX (von Moltke et al., 1997). Scaled intrinsic clearance values, used to estimate the contribution of individual CYP enzymes in human liver microsomes, suggest that CYP2D6, CYP2C9 and CYP3A4 contribute the greatest amount of FX A -^demethylation (Margolis et al., 2000; Ring et al., 2001). Previous investigations from our laboratory using purified human cDNA-expressed CYP isozymes demonstrated that in addition to CYP2D6 and CYP2C9, CYP2C18 and CYP2C19 are also involved in the formation of NFX, however no activity was observed with CYP3A4 (Kim, 2000). Like the parent drug, NFX is also chiral in nature with both R-and S-NFX being formed in the body. The elimination half-life of NFX is about 7-15 days in humans, which is considerably longer than that of the parent drug. Another important feature of NFX is its preserved pharmacological activity in 5-HT reuptake inhibition (Brosen and Skjelbo, 1991; Wong et al., 1993). While much is known about the conversion of FX to NFX, limited study has been conducted to investigate the other metabolic pathways of FX. A schematic diagram illustrating FX's proposed metabolic pathway is shown in Figure 1.3 (Altamura et al., 1994). Following conversion to NFX it has been reported that FX undergoes further metabolism to form FX carboxylic acid. FX, NFX and FX carboxylic acid all undergo O-dealkylation to give /?-trifluoromethylphenol (TFMP). Following the removal of the TFMP moiety, O-dealkylated FX is further metabolized to give benzoic acid, which is then conjugated with 6 o I CHCH 2CH 2NCH 3 Glu Fluoxetine O CHCH 2CH 2NHCH 3 CF, O CHCH2CH2NH l Glu CF, O I CHCH2COO Glu Glucuronide Conjugate Norfluoxetine O CHCH 2CH 2NH 2 CF-, Fluoxetine Carboxylic Acid O CHCH2COOH OH ,1 CHCH2COOH Benzoic Acid -OH P-Trifluoromethylphenol (TFMP) | Beta-oxidation COOH ^ CONHCH2COOH Hippuric Acid FIGURE 1.3 Proposed schematic diagram of FX metabolic pathway. 7 glycine to form hippuric acid and excreted into urine. This benzoic acid-glycine conjugate comprises -20% of the administered FX dose recovered in human urine. About 11% of the administered dose of FX is excreted as free FX, 7% as free NFX, and 7% and 8% as FX and NFX TV-glucuronides, respectively. TFMP has been quantified in human and rat plasma and urine, with free urine TFMP levels exceeding those in plasma (Urichuk et al., 1997). This metabolite is also observed in adult sheep urine but its contribution to the overall elimination of FX remains to be determined (Chien et al., unpublished data). A fairly recent in vitro study reports that CYP2C19 and CYP3A4 are the major cytochrome P450 isoforms responsible for fluoxetine O-dealkylation (Liu et al., 2002). 1.1.5 Stereoselective Properties of Fluoxetine and Norfluoxetine It is important to be aware that though optical isomers share the same physical and chemical properties, they do not necessary have similar pharmacokinetic and pharmacodynamic properties. Table 1.1 summaries the stereoselective properties of fluoxetine and norfluoxetine enantiomers. In terms of pharmacological activity, both S- and R-enantiomers of FX are equally active as serotonin reuptake pump inhibitors (Gram, 1994). However, the pharmacokinetics of the S- and R-enantiomers of FX are significantly different from one another. The clearance of S-FX is significantly lower than that of R-FX resulting in a higher area under the plasma FX concentration-time curve (AUC) and steady-state concentration for S-FX compared to R-FX. Interestingly, while having a smaller clearance value, S-FX has a shorter elimination half-life instead of a longer one compared to R-FX in poor metabolizers (PM) of CYP2D6, and approximately equal in extensive metabolizers (EM) (Fjordside et al., 1999; Eap et al., 2001). A significant difference in the Vd of S-FX and R-8 FX should be present in order to explain the difference in elimination half-life based on the theoretical relationship between elimination half-life, clearance and volume of distribution. Unfortunately, information regarding the Vd of FX enantiomers is not available in the literature. FX is highly bound to plasma protein but the stereoselective aspect of its binding characteristics has never been explored. TABLE 1.1 Stereoselective properties of FX and NFX enantiomers. PK-PD property FX NFX Pharmacological activity S = R S > R by 20 fold Clearance (CL) a S <R N/A Half-life (t, /2)a S < R In PM c S = R In EM c S > R in PM S = R in EM AUC a S>R S>R Steady-state cone. (Css) b S>R S = R in PM S > R in EM Volume of distribution (Vd) N/A N/A Protein binding (fu) N/A N/A CYP2D6 inhibition (K;) S <R S <R a Data based on a single dose study with single oral administration of 60 mg F X to 12 subjects (6 E M s and 6 P M s ) (Fjordside et al., 1999). b Data based on a multiple dose study with daily administration o f 20 mg F X to 11 subjects (8 E M s and 3 PMs) for up to 23 days (Eap et al., 2001). c P M , C Y P 2 D 6 poor metabolizer; E M , C Y P 2 D 6 extensive metabolizer. FX is also known to be a potent inhibitor of CYP2D6 at therapeutic doses, which is of clinical significance in terms of potential drug interactions (Otton et al., 1993). The degree of CYP2D6 inhibition is also different between the FX enantiomers with S-FX a more potent 9 inhibitor (~6-fold) (inhibition constant (Ki), 0.22 ± 0.11 uM) compared to R-FX (K i ; 1.38 ± 0.48 pM) (Stevens and Wrighton, 1993). The principal metabolite of FX, NFX is also pharmacologically active in serotonin reuptake inhibition. Unlike the parent drug, pharmacological activity of NFX is significantly different between its enantiomers with S-NFX being equally active compared to the parent drug while R-NFX has only 1/20 of the activity of FX (Fuller et al., 1992; Wong et al., 1993). In PMs, S-NFX has a longer elimination half-life but a similar steady-state plasma concentration (Css) compared to R-NFX. In EMs, the elimination half-life is the same between NFX enantiomers but C s s is higher for S-NFX compared to R-NFX (Fjordside et al., 1999; Eap et al, 2001). No information is available regarding the clearance, volume of distribution or protein binding of NFX. Nonetheless, NFX has also been shown to be capable of inhibiting CYP2D6 with S-NFX being a stronger enzyme inhibitor compared to R-NFX (K,: S-NFX = 0.31 ± 0.04 uM versus R-NFX = 1.48 ± 0.27 uM) (Brosen and Skjelbo, 1991; Stevens and Wrighton, 1993). Pharmacokinetics determines the availability of a drug to its site of action. The site of action can either be an enzyme or a receptor. Fluoxetine and norfluoxetine act by binding to the 5-HT reuptake transporter to elicit their pharmacological effects. Either therapeutic and/or toxic effects will result from this interaction. It is important to point out that this pharmacokinetic-pharmacodynamic (PK-PD) relationship of fluoxetine and norfluoxetine is stereoselective in nature. Under certain circumstances, changes to their stereoselective pharmacokinetic properties may negatively affect their therapeutic and toxicity profiles. Additionally, the stereoselective inhibitory effects of fluoxetine and norfluoxetine enantiomers on CYP2D6 not only complicate their own pharmacokinetic-pharmacodynamic 10 profiles (i.e. nonlinearity), but also potentially lead to interactions with other drugs, which are substrates of the same enzyme. Inhibitory effects on enzyme activity are frequently concentration dependent, thus pharmacokinetics plays a pivotal role in contributing to this phenomenon. This concentration dependent interaction is often of clinical significance for drugs with a narrow therapeutic index. In summary, complex stereoselective components exist between the pharmacokinetic-pharmacodynamic relationship of fluoxetine and norfluoxetine. Because of this, changes in their stereoselective pharmacokinetics may have significant clinical implications in their therapeutic effects, toxicity and drug interactions. 1 . 1 . 6 Depression and F X in Pregnancy Depression occurs at a similar rate in pregnant and nonpregnant women, with approximately 20% of women experiencing depressive symptoms during pregnancy (Kessler et al., 1993; Burt and Stein, 2002). Since depressed patients tend to neglect themselves, a depressive episode during pregnancy increases the risk for suicide and poor self-care. Women with depression often present with decreased appetite and consequently lower-than-expected weight gain in pregnancy, factors that have been associated with negative pregnancy outcomes (Bhatia and Bhatia, 1999; Nonacs and Cohen, 2002). In addition, pregnant women with depression are also more likely to smoke and to use either alcohol or illicit drugs, behaviors that further compromise the health of the developing fetus (Zuckerman et al., 1989; Zuckerman et al., 1990; Hoffman and Hatch, 1996). Higher rates of admission to neonatal ICU, lower birth weight and premature delivery have been more commonly associated with women from suffering depression during pregnancy than those 11 who were not (Steer et al., 1992; Chambers et al., 1996; Grush and Cohen, 1998; Orr et al, 2002). Uncontrolled depression during pregnancy also triples the risk for depression in the postpartum period (O'Hara, 1995; Cooper and Murray, 1998; Newport et al., 2002). It is generally agreed that drugs of any type should only be taken during pregnancy when the risk to the mother and the fetus of no treatment outweigh the risks of taking the drug (Hendrick and Altshuler, 2002; Nonacs and Cohen, 2002). Thus, when psychotherapy has not been effective, or when the illness can affect fetal prognosis and the future mother-child relationship, pharmacological treatment of depression during pregnancy may be necessary (Robert, 1996; Wisner et al., 1999). The SSRIs have become the most commonly prescribed class of antidepressants worldwide in the general adult population as well as in pregnant women. The increasing use of SSRIs in pregnant women has raised concerns regarding their safety during pregnancy and resulted in a number of studies to investigate their reproductive risks. The largest amount of information on prenatal SSRI exposures is available for FX (Altshuler, 2002). Neither retrospective nor prospective studies have revealed an increased risk of birth defects, poor perinatal condition, neurodevelopment delay, or poor cognition and language development associated with the use of FX in pregnancy (Chambers et al., 1996; Goldstein et al., 1997; Loebstein and Koren, 1997; Addis and Koren, 2000; Cohen et al., 2000; Nulman et al., 2002). However, several reports of neonatal toxicity, cardiac arrhythmia, and neonatal withdrawal syndrome, characterized by irritability, constant crying, shivering, increased tonus, eating and sleeping difficulties and convulsions, have been reported in newborns exposed prenatally to FX (Spencer, 1993; Mhanna et al., 1997; Mohan and Moore, 2000; Nordeng et al., 2001; Abebe-Campino et al., 2002). These adverse events have been largely attributed to in utero 12 FX exposure in late gestation, which is the period when maternal adaptive changes are most likely to influence drug pharmacokinetics. Furthermore, it was observed that some of these neonates had plasma FX and NFX concentrations that were above the usual adult therapeutic ranges at a given dosage, suggesting that alterations in the pharmacokinetics of FX might be involved during the perinatal period. While the incidence of these adverse events appears to be low, further research should focus on the safety of FX and its pharmacokinetics during the last trimester of pregnancy in order to establish a causal link between drug exposure and toxicity. 1.1.7 Effects of Pregnancy on Pharmacokinetics During pregnancy, the body undergoes a number of physiological changes, which may modify the pharmacokinetic processes of drug absorption, distribution, metabolism and excretion. Important among these are maternal adaptations that serve to increase the delivery of blood and nutrients to the fetus and to remove metabolic waste products from both mother and fetus (Nation, 1980). These adaptations generally result in a marked increase in both the intravascular and extravascular volumes; the enhancement of cardiac output, renal blood flow, and the glomerular filtration rate; and the augmentation of hepatic and other metabolic activities in the mother, the fetus, and the placenta (Krauer et al., 1980). Consequently, the volume of distribution, biotransformation and renal clearance of xenobiotics can vary significantly in comparison with corresponding values obtained in the nonpregnant state. These pregnancy-associated physiologic changes and their therapeutic implications are summarized in Table 1.2. 13 c o C5 u "a, B s cu a s-01 13 o <H-H o s o CH cn O t>0 CO 3 ai TJ CD 13 cd CD -4—< -4—» O co '-3 'a & TJ co CD co * p op e 53 CD ^ <+-! — C O CD -4. f i CD 43 o .2 GO "cd 2 ' TJ S3 CD CD CD > 3 2 O O O , 0 co HH CD TJ CD " 3 CD CD r>H <H-H <+H o 0 o DO t-H TJ a 3 43 00 O SH 53 ±z 2 c§ 5- u OH CD SH O S3 3 JD JD 3 JS 'Lo 'co CO CO O O OH PH 03 >H r +-* C+_( fi. o I-H C O c° cd TJ +e CD O CD S3 OH a cd TJ O 53 O U to CD 43 cd O i-i '.3 cd £ 8 43 PH cd O CO S3 _o "cd S--4—» S3 CD CD S3 8 ^ § > T J I H TJ O CD CO O ft i i «J .O CD 43 T J CD 43 CD 3 S3 0-3 TJ S3 cd CD CO O TJ TJ — CD cd CO > 03 CH CD CD o 3 _S3 SH .O 00 G co CD ~ CD T J S3 CD a s CO CD CO CD O TJ CD 43 a 3 ."35 CD . — l -H ' 3 'o ^ "3 CD cd C+H "5J o w CD TJ 2 § TJ CD CD 3 cd 3 CD CD s CD CD s 2 o f*H <H-H 3 & o co " § C-t—i O CD ' CD — 3 1 3 CD r " r - H O (—1 E 0 C+H S3 O \£3 & O CO 43 cd TJ CD CO cd H—» CD In O cd CD 3 in -4—» TJ CD co cd CD o CD TJ > TJ CD CO cd <D HH CD 3 o 3 O CD 3 £ 2 o 2 TJ C3 OH H_ CD a S3 O OH S T ) CD CD TJ CO CD 00 S 3 fi 43 O S T J « a T J 00 C O CD cd co O H CD CD CD 3 3 T J ~ CD T J 3 o Jrt W 3 cd CO r r CD cd 43 h u 'C H SH CD O TJ CO ' . a -a CD CD S-H cd CD o 00 5 TJ TJ CD CO cd CD HH o 3 o B 'Lo ,2 ,3 co Id « fi Z u Id cd ,CD H— CH—1 CO ^ 3 Ol S cd • TJ CD CO cd o 3 3 o , 17:3 cd CD i3 00 CD cd CD 43 ° a CD 3 CD OX) e « 43 CD OX o 00 0 ^ g | C 'Lo co O _ rt 2 TJ 00 +3 CD co TJ T3 cd CD CD CD >, >, CD -2 CD 13 "CD Q Q Q o a l - H o > s 43 OH 3^, CD a > 1 fcH 0 M a O H-» CO > 2? cd '53 n b CD . a •£ TJ CD 43 W 43 TJ $3H S3 3 03 ^—' CD 0 2 o C M ^ 43 3 TJ O ^ o id a 43 43 Id ^ Vi 0 >- 2 w C2 TJ >H CD CD 43 23 a v ° CH ^ O o M m 3 TJ CD td > t3 cd 1 CD 3 O > H CD o I i d a Id "3 co >H O 3 O u "2 - 3 ao =3 ii H O CD O cd OH a 43 -3 T3 CD T 3 CD 'i3 CD HH cd co Cd o cd CH f CH CD 43 o 3 3 fc cd -4—» 3 CD CD cd ^ 3 TJ .2 00 "cd . a "to c <u " 3 00 S G TJ ' o CD C 2 cd CH ^ o cd 3 4 1.1.7.1 Absorption During pregnancy, both the tone and motility of the maternal gastrointestinal tract may be reduced as a consequence of physical displacement by the enlarging uterus; these changes may result in delayed gastric emptying and prolonged intestinal transit. The latter effects may in turn influence either the bioavailability and effectiveness of drugs with acid-labile properties or the site-specific absorption of drugs from the intestinal tract (Reynolds and Knott, 1989). 1.1.7.2 Distribution Several changes in body compartments occur during pregnancy and may have important effects on the distribution of drugs. The normal weight gain in pregnancy of up to 12 kg is a result of the combined increases in total body water, fat stores, and of course, the products of conception (Reynolds and Knott, 1989). The 6-8 liter increase in total body water is accounted for by a rise in plasma volume, by the fetus, uterus, and amniotic fluid, and by a large expansion of the extracellular fluid space. There is a large inter-individual variation in maternal plasma volume as well as in the extracellular accumulation of water. The rise in plasma volume may increase by up to 50 per cent, and this change is dependent on the number and size of the developing fetuses. Extracellular volume may rise by 5 liter or more at the 40lh week of gestation in cases of generalized edema (Krauer et al., 1980). The increase in total body water is more profound on the distribution of polar drugs, which are confined to the extracellular space, than on already widely distributed lipophilic drugs, whose distribution is probably more influenced by changes in protein and lipid binding. Body fat increases by 3-4 kg (largely in subcutaneous deposits) during the first two trimesters and this 15 store is maintained throughout the remaining period of gestation. During the third trimester, plasma concentrations of lipoproteins, glycerol, and non-esterified free fatty acid gradually increase and peak at around delivery (Reynolds and Knott, 1989). Thus, not only will fat-soluble drugs have an increased volume of distribution, but may also compete with other endogenous compounds (e.g. fatty acids) for protein-binding sites. This may result then, in alterations in the unbound fraction of the drug. 1.1.7.3 Protein Binding Albumin is the major binding protein for acidic drugs such as anticonvulsants, salicylates, and nonsteroidal anti-inflammatory compounds as well as for some neutral drugs including warfarin and diazepam. On the other hand, basic drugs such as local anesthetics, opioids, (3-blockers, tricyclic antidepressants, and SSRI antidepressants are principally bound with high affinity to ai-acid glycoprotein (AAG), which has a lower binding capacity than albumin (Jeffries and Bochner, 1988). During the first half of pregnancy plasma albumin concentrations fall from around 35 to 25 g/1, partly due to plasma dilution (Krauer et al., 1984). Animal and human studies have shown that, in addition to hypoalbuminemia, an increase in plasma concentration of endogenous ligands, including free fatty acids, may compete for albumin binding sites and thus further reduce its binding capacity in late pregnancy (Reynolds and Knott, 1989). Consequently, unpredictable increases in the free fraction have been reported for many drugs and this is often associated with a lower total plasma drug concentration. A serious underestimation of the free, therapeutically active concentration of a drug will result if drug binding is not be accurately predicted. Reduced albumin binding can contribute to the increased distribution volume and clearance for drugs 16 that undergo restrictive elimination. In contrast to albumin, the concentration of AAG does not vary consistently during pregnancy. Further, the effect of pregnancy on AAG concentrations is poorly documented with no readily apparent trends in drug binding (Wood and Wood, 1981; Krauer et al., 1984). 1.1.7.4 Metabolism Hepatic metabolism in pregnancy Smoking, ethanol consumption, and progesterone are potent enzyme inducers and may significantly enhance drug metabolism, while estrogens may competitively inhibit enzyme induction. Theoretically, it is possible that the balance between these increasing hormone levels during pregnancy may dictate the net effect on enzyme activity. A high estrogen/progesterone ratio may lead to enzyme inhibition while a low estrogen/ progesterone ratio may lead to enzyme induction (Krauer, 1984). Progressive changes in this ratio, combined with the influence of environmental agents, may explain the apparently inconsistent findings for individual drugs whose elimination rates may be raised, lowered, or unchanged during pregnancy. Finally, the cholestatic effect of estrogen may interfere with the clearance of drugs that undergo biliary excretion (Harrison and Gibaldi, 1976). While the overall changes of pregnancy on hepatic physiology have the potential to alter drug metabolism, the relevance of this is not presently apparent. Placental and fetal hepatic metabolism There is some in vitro evidence that both the human placenta and fetal liver are capable of significant drug biotransformation (Pelkonen, 1980). Further, the metabolite profiles appear qualitatively similar to those generated by adult human liver and this may 17 also be the case in fetal liver. It is thought that differences in the rate of principal metabolite formation between the adult liver and the placenta may relate to differences in microsomal enzyme concentrations as well as to and their specific regulation (Juchau, 1980). Similar to hepatic microsomal enzymes, those in the placenta that are responsible for the metabolic conversion of endogenous substances have also been found to be induced or inhibited by certain drugs (Juchau, 1980; Pasanen 1999). In contrast to oxidative metabolism, the conjugative capabilities of the placenta have not been clearly demonstrated (Juchau, 1980; Pasanen 1999). It is also likely that certain substances may only cross the placenta after biotransformation. Studies with human fetal liver demonstrate the ability to oxidize endogenous and exogenous substrates, using the cytochrome P450 system, by the end of the first trimester (Loebstein et al., 1997; Hines and McCarver, 2002). In general, the human fetal liver has the adult complement of enzymes, but the activities at term are only about one-half those of the adult and they are not as readily induced (Shimada et al., 1996; Loebstein et al., 1997). In late gestation, the human fetal liver is deficient in UDP-glucuronyltransferase activity (Burchell et al., 1989; de Wildt et al., 1999b), however sulfate conjugation appears to be well developed with the activity of these enzymes even exceeding those in adults (Juchau et al., 1980; McCarver and Hines, 2002). Overall, fetal metabolism contributes very little to metabolite formation and the fetus relies largely on the maternal system for the clearance of parent drug and metabolite(s) (Rane and Tomson, 1980; Ring et al., 1999). However, fetal metabolism may be important when the metabolites produced do not readily cross the placenta as their accumulation in the fetus may have adverse consequences. 18 In summary, the clinical significance of potential fetal hepatic and placental metabolism remains to be clearly defined. Certainly a more complete understanding of the roles played by enzymatic reactions in both the placenta and fetal liver in fetal parent drug and/or metabolite exposure is required. 1.1.7.5 Excretion Of all the physiological systems which are modified by pregnancy, renal function perhaps undergoes the greatest change. Within the first few weeks, of gestation, the glomerular filtration rate increases by up to 50% and remains elevated until after delivery (Davison and Noble, 1981). Plasma creatinine and urea clearances increase during pregnancy. Creatinine clearance, which closely reflects the glomerular filtration rate, increases to a maximum at 34 weeks, then declines with consequent raised plasma concentrations just prior to delivery. Renal plasma flow starts to rise by about 15 weeks and may have doubled by 26 weeks (Reynolds and Knott 1989). In parallel with these changes, polar, water-soluble drugs, which are predominantly eliminated by renal excretion, are cleared more rapidly during pregnancy and have shorter half-lives (Chow and Jewesson, 1985). 1.1.7.6 Placental Transfer Movement of drug substances across the placenta is generally bi-directional, although net transfer occurs from mother to fetus in most instances. The transplacental passage of a .drug is regulated by many factors, principally the physiological characteristics of the maternal-placental-fetal unit and the physiochemical properties of the drug (Mirkin, 1975). Physiologic factors include 1) hemodynamic changes in either maternal or the fetal 19 contribution to total placental blood flow, 2) the thickness and maturation of the placental membranes, and 3) the metabolic activity of placental tissues. Although active and facilitated transport of some substances across the placenta has been demonstrated, the transplacental passage of most drugs occurs primarily by simple diffusion (Reynolds and Knott, 1989). The rate of diffusion across the placenta is directly proportional to the maternal-fetal concentration gradient and the surface area of the placenta and is inversely proportional to the thickness of the placental membranes (Mihaly and Morgan, 1984). Early in pregnancy the placental membranes are relatively thick, and this characteristic tends to reduce permeability. The thickness of the trophoblastic epithelium decreases in the last trimester, and any passage of drugs is thereby greatly facilitated (Garland, 1998). This progressive thinning of the placenta as gestation advances results in increased transplacental diffusion and alternative routes of drug elimination from the mother. A drug with a low molecular weight of <600 Da, a high degree of lipid solubility, a low degree of ionization, and a low affinity for protein binding is able to transverse the placenta more readily whereas a drug with a molecular weight of > 1,000 Da may be restricted by its size (Reynolds and Knott, 1989). Only the unbound (free) fraction of a drug is subject to placental transfer; therefore, the more a drug is bound to maternal plasma protein, the less it is available for crossing to the fetus. In general, drugs with a low degree of protein binding reach higher total concentrations in the fetus whereas drugs that are highly protein bound achieve'higher maternal than fetal total concentrations (Szeto, 1993). Collectively the physiological changes described above have the potential to alter drug disposition during pregnancy and may therefore have an impact on the pharmacokinetic disposition of FX and NFX in pregnant women. The effect of pregnancy on the clearance of 2 0 FX is of particular interest as this is a major factor that needs to be considered for its long-term dosing in the treatment of depression. The high degree of plasma protein binding exhibited by FX is also important since progressive changes in protein binding over the course of pregnancy may also affect its fetal-placental transfer and consequently the overall exposure of the fetus to this drug. Therefore, it is prudent to systemically study these physiological effects of pregnancy on the disposition of FX and NFX to improve our understanding of the appropriate use of FX in the treatment of pregnant women with depression. 1.1.8 Placental Transfer and Pharmacokinetics of FX in Pregnancy: Previous Studies Earlier studies have been conducted in our laboratory to examine the stereoselective pharmacokinetics of FX placental transfer in chronically-catheterized nonpregnant and late gestational age pregnant sheep following a single i.v. dose of the drug (Kim, 2000; Kim et al., 2004). The stereoselective disposition of FX previously described in humans (Fjordside et al., 1999) was also observed in sheep. Differences in plasma concentration, systemic clearance, volume of distribution and free fraction were noted between the FX enantiomers. Based on the estimation of free drug clearance values, it was suggested that differential protein binding of the FX enantiomers appeared to explain the stereoselective disposition of the drug in sheep. In addition, the metabolic profile of FX, in general, was found to be similar between humans and sheep. YV-demethylation of FX to form NFX, which is a major metabolic pathway of FX in humans, was also observed in sheep. The glucuronide conjugate of FX was also identified in sheep urine although the amount eliminated appeared to be less compared to that reported in human subjects (Altamura et al., 1994). 21 FX pharmacokinetics in pregnant sheep exhibited a number of statistically significant differences compared to nonpregnant sheep, which included a higher total body clearance, shorter half-life and lower steady-state volume of distribution. It was postulated that these differences could be due to an increased activity of a CYP2D6-like enzyme and/or an increase in the plasma concentration of ai-acid glycoprotein during pregnancy in sheep. A pregnancy-related induction of CYP2D6 occurs in the human (Wadelius et al., 1997), which suggests that changes in FX pharmacokinetics similar to those in the sheep may occur during pregnancy in human. FX was found to be extensively transferred across the sheep placenta. The fetal-to-maternal (F/M) AUC ratio in pregnant sheep after maternal i.v. bolus administration was 0.57 + 0.17 for FX and 0.51 ±0.13 for NFX. FX did not appear to accumulate in amniotic and fetal tracheal fluids in contrast to a number of amine drugs studied earlier in sheep (Riggs et al., 1987; Wright etal., 1991; Kumar et al., 1997). The metabolic capacity of FX in ovine fetuses was also explored. Following fetal i.v. bolus administration of FX, no NFX was detected in fetal plasma. In vitro microsomal studies using fetal sheep liver also showed no evidence of fetal NFX formation. Thus, metabolic pathway(s) for FX appear to be inactive in the fetal lamb although further studies are required to fully characterize the expression and capacity of the fetal enzymes involved in FX metabolism. The contribution of other routes of elimination of FX in fetal lambs such as renal elimination and placental metabolism, were not determined in these previous studies. An investigation of the placental transfer of FX in pregnant women during the perinatal period was also conducted as part of this previous work. The average maternal exposure period for FX was 9 months (range 1 to 48 months) at the time of delivery with a 22 daily dosage of 10-30 mg. The mean serum cord-to-maternal (F/M) ratios of racemic FX and NFX at birth were 1.04 ± 0.29 and 1.04 ±0.16, respectively (Kim, 2000). Similar umbilical cord to maternal blood concentration ratios of FX and NFX (0.83 ± 0.27 for FX and 0.81 ± 0.24 for NFX) have also been reported by others (Stowe, 2000). The ratios reported by Kim (2000) were significantly higher in the human studies than those observed in the pregnant ewes, where fetal exposure was estimated as the ratio of fetal and maternal plasma AUC following a single FX dose to the ewe. It was suggested that this difference is related to the longer-term of drug administration regimen in the human study, which allows for the attainment of steady-state, compared to the single i.v. bolus injection protocol employed in the sheep project. In the same study, a negative correlation was observed between the apparent oral clearance and exposure time for FX in humans. This finding confirms the results by other investigators that the metabolism of FX is mediated by CYP2D6, and also that CYP2D6 is inhibited at drug concentrations in the therapeutic range (Brosen and Skjelbo, 1991; Steven and Wrighton, 1993). In addition, the S/R ratio of steady-state FX concentrations was found to be negatively correlated to the oral clearance of FX but positively correlated to the exposure time to FX in humans during chronic therapy. The relationship of the S/R ratio of steady-state FX concentrations with respect to clearance and time suggests that the stereoselective disposition of FX involves other CYP450 isozyme(s) when the CYP2D6-mediated metabolism is inhibited over time. Interestingly, a similar correlation between the S/R AUC ratio and the systemic clearances of the FX enantiomers was observed in sheep after single i.v. bolus FX administration. However, further experiments will be necessary to fully characterize this phenomenon in sheep. 23 1.2 RATIONALE Pharmacokinetic evaluations during pregnancy are particularly complex because of the markedly altered and rapidly changing physiological state of the pregnant woman as well as the technical and ethical difficulties associated with assessing fetal drug pharmacokinetics in utero. Other issues such as the contributions of placental and fetal tissues as determinants of drug elimination, and the changing relative contributions of fetal-placental and maternal-placental drug clearance during the various stages of pregnancy must also be considered in order to assess the maternal-placental-fetal pharmacokinetic process in a quantitative manner. Certain conditions such as depression have a high incidence in pregnant women and frequently require pharmacological intervention. A thorough understanding of drug pharmacokinetics in pregnancy is of great importance to allow drugs to be safely used in this special population. Ideally, studies performed in humans will provide us with the most relevant clinical information. However, such data is very limited due to ethical limitations on the design of human placental transfer studies. For the most part such information has usually been obtained from single maternal and umbilical blood samples at delivery following acute single dose administration of drug or chronic dosing to the mother during labor (Bourget et al., 1995; Kim et al., 2004). For these reasons animal species have been commonly employed to examine placental drug transfer and subsequent fetal drug exposure. However, serial determinations of maternal and fetal steady state drug concentrations over time are not practical in small laboratory animals due to limited blood volumes and the inability to chronically implant catheters or monitoring devices in the mother and fetus. As a large farm animal species the chronically instrumented pregnant sheep has been extremely useful in addressing both the effects and kinetics of drugs in the fetus in late gestation (Rurak 24 et al., 1991). The frequent fetal sampling achieved in the sheep model is not possible in either human or small animal models. Compared to other species (e.g. monkeys), sheep do not often abort or deliver prematurely after intrauterine surgery. The fetus in the last trimester is conveniently large (0.5-5 kg) and the investigator can perform intricate surgeries and obtain serial fetal blood samples large enough for drug analysis. Further, the ewe is a very docile animal and will not interfere with exteriorized indwelling catheters (Simone et al., 1994). While there are clear differences between sheep and humans in placental structure and intrinsic activity of drug metabolizing enzymes, the capability of chronically instrumented sheep studies to provide detailed information on maternal-fetal drug disposition appears to be far more important than any drawbacks. Previous experiments involving single i.v. bolus FX administration to pregnant sheep have contributed much to our understanding of the kinetics of FX transfer to the fetus, as well as the general impact of pregnancy on the disposition of FX. These studies have also shown that the stereoselective disposition of FX is qualitatively similar between sheep and humans. This is very uncommon compared to other species. However, the design of the single dose i.v. bolus experiments did not permit a complete assessment of several important pharmacokinetic features of FX such as nonlinearity, time dependent changes of clearance, and inter-relationships in stereoselectivity (S/R ratio). Such information can only be obtained by administering the drug for a longer period of time to steady state. Clinically FX is usually given repeatedly in humans to achieve steady state plasma drug concentrations, hence pharmacokinetic studies of FX at steady state in sheep should provide information of more clinical significance compared to the previous single dose i.v. bolus data. Additionally, it was speculated that the discrepancy observed between humans and sheep in the F/M ratio 25 . of FX was attributed to the difference in the duration of drug exposure between the two groups. More experiments involving long-term FX administration in pregnant sheep are necessary to test this assumption. To our knowledge, such experiments have not yet been performed for FX in any species. Lastly, after a decade of clinical use the pharmacokinetics of FX in pregnancy is still poorly understood. Many articles have been published on the clinical pharmacokinetics and use of FX but only a very few of these address the safety of FX use in pregnancy. With the expiration of the patent on the brand product of fluoxetine, Prozac®, in the United States in September 2001 (Elliott, 2001; Pabst, 2001), the emergence of competing generic substitutes has made this drug more affordable than ever for many patients, hence its use is expected to increase. With the increase of SSRIs in pregnancy it is our hope that through a series of well-designed long-term animal experiments the maternal-fetal disposition and placental transfer of FX can be examined in a more systematic and mechanistic way. Such experiments will tremendously further our understanding of the behavior of FX, and hopefully provide a more rational approach to it use for the treatment of depression and other psychiatric disorders in pregnancy. 1.3 HYPOTHESES 1. Fluoxetine (FX) and norfluoxetine (NFX) enantiomers undergo significant stereoselective pharmacokinetics at steady-state in nonpregnant sheep. 2. Pregnancy-induced physiological changes will alter the stereoselective disposition of FX and NFX at steady-state in sheep. 3. The exposure of the fetus to FX will be greater with steady-state administration, compared to acute single i.v. bolus dosing. 4. The clearance of FX and NFX enantiomers in fetal lambs involve both placental and nonplacental routes of elimination. 1.4 OBJECTIVES 1. To examine the stereoselective pharmacokinetics of FX and NFX at steady-state in nonpregnant sheep. 2. To investigate the effect of pregnancy on the stereoselective disposition of FX and NFX at steady-state in pregnant sheep. 3. To study the fetal exposure of FX and NFX enantiomers by examining their placental transfer, and distribution in amniotic and fetal tracheal fluid at steady-state. 4. To study the stereoselective disposition of FX and NFX in fetal lambs, with specific focus on protein binding, hepatic metabolism, and renal excretion. 27 CHAPTER 2 MATERIALS, INSTRUMENTATION AND ANALYTICAL METHODOLOGIES 2.1 Materials The reference standards, chemicals, reagents and other material used during the study and sample analysis, along with information on their purity (where applicable) and source, are listed below. Racemic fluoxetine hydrochloride [(R,S)-7V-methyl-3-(jL7-trifluoromethylphenoxy)-3-phenylpropylamine (>99% purity)] was generously provided by Novopharm (Toronto, ON, Canada). Racemic norfluoxetine hydrochloride [(R,S)-3-(p-trifluoromethylphenoxy) -3-phenylpropylamine (>99% purity)] and (R)-norfluoxetine hydrochloride [(R)-3-(/> trifluoromethylphenoxy)-3-phenylpropylamine (>99% purity)] were generously provided by Eli Lilly and Co. (Indianapolis, IN, U.S.A.). The internal standard [2-(diphenylmethoxy)-7V-methylethylamine hydrochloride] for the gas chromatographic mass spectrometric (GC-MS) assay method was synthesized in our laboratory. Optically pure fluoxetine isomers, R-fluoxetine [(R)-A -^methyl-3-(p-trifluoromethylphenoxy)-3-phenylpropylamine] and S-fluoxetine [(S)-A -^methyl-3-(/7-trifluoromethyl-phenoxy)-3-phenylpropylamine] were separated from racemic fluoxetine hydrochloride in our laboratory as described previously (Kim, 2000). (S)-(-)-trifluoroacetylprolyl chloride (97% purity) was obtained from Aldrich Chemical Co. (Milwaukee, WI, U.S.A.). Triethylamine (TEA), toluene, isopropanol, acetone, methanol and n-hexane were purchased from Caledon Laboratories (Georgetown, ON, Canada) and were of distilled in glass GC or HPLC grade. 28 Deionized high purity water was produced on-site by reverse osmosis and subsequent filtration using a Milli-Q® water system (Millipore, Bedford, MA). Ultra high purity (UHP) helium for GC-MS analysis and filtered nitrogen for sample concentration were obtained from Praxair (Mississauga, ON, Canada). Glassware for extraction and sample preparation (15 mL Pyrex® disposable culture tubes), polytetrafluoroethylene (PFTE) lined crew caps, and 1.0 N sodium hydroxide solution were obtained from VWR Canlab (Vancouver, BC, Canada). Materials used during the animal experiments were as follows: Veramix® sheep sponges (Tuco Products Co., Orangerville, ON); pregnant mares' serum gonadotropin (Ayerst Laboratories, Montreal, QC); thiopental sodium injectable (1 g/vial) USP (Abbott Laboratories, Montreal, QC, Canada); halothane, USP (Ayerst Laboratories, Montreal, QC, Canada); sterile water for injection, USP; 0.9% sodium chloride injection, USP; 0.9% sodium chloride irrigation, USP (Baxter, Mississauga, ON, Canada); injectable ampicillin (250 mg/vial) (Novopharm, Toronto, ON, Canada); injectable atropine sulfate (0.6 mg/ml) (Glaxo Laboratories, Montreal, QC); heparin (1000 units/mL) (Organon Canada Ltd., West Hill, ON, Canada); lidocaine 2% (Astra Pharma Inc., Mississauga, ON, Canada). All injectable formulations were obtained from BC Children's and Women's Hospital Pharmacy, Vancouver, BC. Other materials used during the animal experiments included: disposable syringe needles and plastic disposable Luer-Lok® syringes for drug administration and biological fluid sample collection (Becton-Dickinson Canada, Mississagua, ON, Canada); 500-mL sterilized evacuated glass containers for drug solution preparation (Baxter, Mississauga, ON, Canada); silicone or polyvinyl rubber tubing for catheter preparation (Dow Corning, 2 9 Midland, MI, U.S.A.); Foley bladder catheters (American Latex Corp., Sullivan, IN, U.S.A.) and urine drainage bags (Bard Medical, Covington, GA, U.S.A.) for urine collection in adult sheep; empty Viaflex® containers for fetal urine collection (Baxter, Mississauga, ON, Canada); nylon syringe filters (0.22 pm) (MSI, Westboro, MA, U.S.A.), heparinized blood gas syringes (Marquest Medical Products Inc., Englewood, CO, U.S.A.); and lavender-top K 3 EDTA BD VacutainerIM (Becton Dickinson, Oakville, ON, Canada) for plasma separation. Materials used for protein binding studies included Plexiglass® dialysis cells and cellophane dialysis membrane sacks with a molecular cutoff of 12,000 daltons (Sigma Chemical Co., St. Louis, MO, U.S.A.). The following reagent-graded chemicals were also used: potassium phosphate monobasic anhydrous, sodium phosphate dibasic heptahydrate (BDH Chemical Co., Toronto, ON, Canada), and sodium chloride (Sigma Chemical Co., St. Louis, MO, U.S.A.). 2.2 Instrumentation 2.2.1 Gas Chromatography-Mass Spectrometry A Hewlett-Packard (HP) model 5890 Series II gas chromatograph equipped with a split-splitless capillary inlet system, a HP model 7673 autoinjector, a HP model 5971A quadrupole mass selective detector, and a Vectra 486 25T computer with HP Chemstation model G1030A workstation software was utilized for the GC-MS analysis of FX and NFX in electron impact (EI) mode (Hewlett-Packard, Avondale, PA, U.S.A.). The gas chromatograph was equipped with a DB5MS (30m x 0.18mm i.d.; film thickness 0.18 pm; 5% phenylmethylsilicone; J&W Scientific, Folsom, CA, U.S.A.) cross-linked fused silica 30 capillary column for FX and NFX analysis. Samples were injected onto the GC-MS via a deactivated glass single taper inlet liner (4mm i.d., 6.5mm o.d., 900 pi;-Agilent Technologies Canada Inc., Mississauga, ON, Canada) and a Thermogreen LC-2 silicone rubber septum (J&W Scientific, Folsom, CA, U.S.A.) in the splitless mode. 2.2.2 Physiological Monitoring Maternal and fetal blood samples were analyzed for pH, partial pressure of oxygen (p02), partial pressure of carbon dioxide (pC02), and whole blood base excess (BE) with an IL 1306 pH/blood gas analyser (Allied Instrumentation Laboratory, Milan, Italy) and temperature corrected to 39.5°C for fetal samples and 39°C for maternal samples. Blood Hemoglobin (Hb) and oxygen saturation (02 sat) were measured with an OSM-2 Hemoximeter (Radiometer, Copenhagen, Denmark). Glucose and lactate were determined with a glucose/lactate 2306 STAT plus analyser (Y.S.I. Inc., Yellow Springs, OH, U.S.A.). 2.2.3 Other Equipment The following equipment was used for sample preparation: a vortex-type mixer and incubation oven (Isotemp model 350) (Fisher Scientific, Springfield, MA, U.S.A.); IEC model 2K centrifuge (Damon/IEC, Needham, MA, U.S.A.); rotating mixer (Labquake model 415-110, Lab Industries, Berkeley, CA, U.S.A); Zymak Turbo Vap® LV Evaporator (Zymak Co., Hopkinton, MA, U.S.A.). Other equipment included a syringe pump (Model 944, Harvard Appartus, Millis, MA, U.S.A.) and water bath (Isotemp model 220, Fisher Scientific, Springfield, MA, U.S.A.) for drug infusion and protein binding experiments, respectively. 31 2.3 Analysis of FX and NFX Enantiomers in Biological Fluids Enantiomers of FX and NFX in all biological fluids were measured by a chiral assay using GC-MS developed previously in this lab (Kim, 2000). Briefly, appropriate volumes of biological fluids (e.g. plasma, urine) were basified with 0.5 ml of 1.0 N NaOH solution, and extracted for 20 min with 7 ml of 0.05 M TEA in hexane:isopropanol 98:2 v/v mixture on a rotary mixer. The organic extract was separated and dried under a gentle stream of nitrogen at 35°C for 30 min. The residue was reconstituted in 80 pi of a 5% (S)-(-)-7V-trifluoroacetylprolyl chloride - toluene mixture and derivatized by heating at 65°C for 1 h, which converted the FX and NFX enantiomers into their diastereomers and allowed baseline separation of all peaks using gas chromatography. Following sample cooling a 2 pi aliquot of the final sample was injected into the GC-MS in splitless mode. Chromatographic separations were performed on a DB5MS (30m x 0.18mm i.d.; film thickness 0.18 pm; 5% phenylmethylsilicone) cross-linked fused silica capillary column with helium as the carrier gas at a 15 psi column head pressure. The GC operating conditions were as follows: Injection port temperature 260°C; oven temperature 145°C initially and increased by 40°C/min to 215°C, 1.5°C/min to 239°C and finally by 14.4°C/min to 300°C where it was held for 2.5 min. Total chromatographic run time was 22 minutes. The mass spectrometer was operated in electron impact ionization mode (EI energy = 70 eV) with selected ion monitoring (SIM) at a transfer line temperature of 300°C. Ions monitored for FX, NFX and the internal standard [2-(diphenylmethoxy)-A -^methylethylamine] were m/z 341, 327 and 252 respectively. The method was linear over a concentration range of 1.0-100 ng/ml (r2>0.99) with intra- and inter-day coefficient of variation (CV) <15%. The limit of quantitation (LOQ) of this assay was 1.0 ng/ml for each FX and NFX enantiomers using a 200 pL 32 biological fluid sample volume. A representative total ion chromatogram is shown in Figure 2.1. 40000 35000 30000 g 25000 C ro "O 20000 C 3 ^ 15000 10000 5000 0 4 15.00 S - N F X R - N F X S - F X I R - F X . Internal S t a n d a r d L u 16.00 17.00 18.00 19.00 Time 20.00 21.00 22.00 FIGURE 2.1 Representative total ion chromatogram of FX, NFX, and internal standard in SIM mode, showing chromatographic separation of the FX and NFX enantiomers in a run time of 22 minutes. 33 CHAPTER 3 STEADY-STATE STEREOSELECTIVE PHARMACOKINETICS OF FX AND NFX IN NONPREGNANT SHEEP Preface One of the biggest challenges of stereoselective pharmacokinetic studies is the selection of an appropriate animal model. High interspecies variations in the stereoselective disposition of several drugs have been reported (Ruelius, 1987; Jamali et al., 1989). For example, cyclophosphamide did not exhibit stereoselective pharmacokinetic in humans, however, metabolism of the (+)-enantiomer is higher in mice and the opposite is observed in rats and rabbits (Cox et al., 1978). Similar phenomena have been observed with other racemic compounds such as disopyramide, nicotine, oxazepam, fenfluramine, propranolol, and nonsteroidal anti-inflammatory drugs (Cook et al., 1982; Crooks and Godin, 1988; Sisenwine et al, 1982; Jori et al, 1978; Walle, 1985; Jamali, 1988). Human data on the stereoselective pharmacokinetics of FX are very limited in the literature (Fjordside et al., 1999; Eap et al., 2001). Previous studies conducted in this lab revealed significant stereoselective disposition of FX enantiomers in sheep (Kim, 2000), which was similar to that observed in humans. The primary objective of this chapter is to extend our investigation of the stereoselective pharmacokinetics of FX and NFX in sheep under steady-state conditions. Such data will allow us to assess the stereoselective pharmacokinetics of FX and NFX in sheep in greater detail, which is of great importance before further comparisons can be drawn between sheep and humans. Further, it will enable us to decide if extrapolations to the human situation are appropriate or not based on the observations in sheep. The results 34 from this study will also serve as the foundation to further our studies on the effect of pregnancy on the pharmacokinetics of FX and NFX, as well as the mechanisms behind the stereoselective disposition. Moreover, it has been shown that FX exhibits nonlinear pharmacokinetics upon multiple dosing in humans. The mechanism of this nonlinearity is thought to be related to the inhibition of CYP2D6, which is one of the principal metabolic enzymes responsible for FX metabolism. However, this phenomenon has never been described or characterized in any animal model before. Therefore, the secondary objective of this chapter was to investigate whether or not FX exhibits nonlinear pharmacokinetics in sheep. 3.1 Methods 3.1.1 Animal and Surgical Preparation Six nonpregnant ewes (Dorset and Suffolk breeds) were used in this study. The study was approved by University of British Columbia Animal Care Committee and the procedures performed on sheep conformed to the guidelines of the Canadian Council on Animal Care. Surgery was performed aseptically under halothane (1-2%) and nitrous oxide (60%) in oxygen anesthesia after induction with intravenous (i.v.) sodium pentothal (1 g) and intubation of the ewe. Silicone rubber catheters were implanted in femoral artery and vein for blood sampling and drug administration, respectively. Intramuscular injections of ampicillin (500 mg) were given to the ewe on the day of surgery and for 3 days postoperatively. All catheters were flushed daily with approximately 2 ml of heparinized (12 unit/ml) sterile 0.9% sodium chloride to maintain their patency. After surgery, the ewes were kept in holding pens with other sheep and allowed free access to food and water. A 35 minimum of 3 days were allowed for the ewes to recover before experimentation. One day before the experiment, a Foley® bladder catheter was inserted via the urethra and attached to a urine drainage bag for cumulative urine collection. 3.1.2 Experimental Protocol Figure 3.1 is a schematic illustration of the experimental protocol. The infusion study was conducted at two FX dose levels, namely the high dose and the low dose experiments. Racemic FX infusion solution was prepared in sterile water for injection to give a final concentration of 1.523 g/ 550ml (2.77 mg/ml) and 0.762 g/ 550 ml (1.39 mg/ml) for the high dose and low dose experiments, respectively. The resulting solution was sterilized by filtering through a 0.22-pm nylon filter into a capped sterile evacuated glass container before administration. In the high dose experiment (n=6), each animal received a 70 mg i.v. bolus loading dose over one minute followed by a continuous i.v. infusion of racemic FX via the maternal femoral vein at a rate of 6.92 mg/h. In the low dose experiment (n=5), a 35 mg i.v. bolus loading dose over one minute followed by continuous infusion of racemic FX at 3.46 mg/h (half of the rate used in the high dose experiments) was given. In both cases, the infusion continued for 8 days in order to achieve a desired total FX steady-state concentrations (Css) of 100 ng/ml and 50 ng/ml in high and low dose experiments, respectively. In three cases the same animals (Ewe # 106, 6117, 7107) were used for the high and low dose experiments following an appropriate washout period. Serial blood samples from the maternal femoral arterial catheter (5 ml) were collected at 0, 5, 15, 30 min, 1, 2, 4, 6, 12, 24, 36, 48, 60, 72, 84, 96, 108, 120, 132, 144, 156, 168, 180 h during the 36 3-4 days 8 days 4 days (Recovery Period) (Infusion Period) (Post-infusion Period) DDI D D I ^ > DDI t t t Surgery Loading Dose + Infusion begins Infusion terminates F I G U R E 3.1 Schematic diagram of the experimental protocol used in the nonpregnant adult sheep experiments. 37 infusion, and at 0, 5, 10, 20, 30, 45 min, 1, 2, 3, 4, 6, 9, 12, 24, 36, 48, 60, 72, and 96 h post-infusion. The initial zero time blood sample was collected 10 minutes prior to i.v. bolus dose administration. Urine samples, were collected immediately before dosing, then at twelve-hour intervals during the drug infusion period, and up to 72 hours after the discontinuation of the infusion. All blood samples collected for drug and metabolite analysis were placed into K 3 E D T A Vacutainer® tubes, gently mixed, and centrifuged at 2000g for 10 min to separate the plasma. The plasma supernatant was transferred into clean borosilicate test tubes with polytetrafluoroethylene (PTFE)-lined caps. The volume of urine collected at each time interval was recorded and a 10 ml aliquot kept and transferred into a clean borosilicate test tube for subsequent FX and NFX measurement. All biological fluid samples were stored at -20°C until the time of GC/MS analysis. 3.1.3 Determination of FX and NFX Plasma Protein Binding Plasma protein binding of FX and NFX was determined ex vivo using the equilibrium dialysis procedure described by Kim (2000). Briefly, dialysis was carried out at 39°C using 1 ml Plexiglass® dialysis cells and cellophane dialysis membranes (MW cutoff of 12,000 Da). The dialysis membrane was boiled in distilled water for 30 minutes and soaked in isotonic phosphate buffer for a minimum of 1 hour prior to use. Aliquots (800 pi) of plasma sample were dialyzed against an equal volume of the isotonic phosphate buffer (pH 7.4), which was prepared by dissolving 1.79 g of K H 2 P O 4 , 14.35 g of Na 2 HPCv7H 2 0, and 3.9 g of NaCl in 1 liter of distilled water. The equilibration time was 4 hours after which the 38 plasma and buffer were analyzed for FX and NFX. Plasma unbound fraction (fu) was calculated as: %/ u =~xl00 ' (3.1) Cp where Cb and C p are drug or metabolite concentrations in buffer and plasma at the end of the dialysis experiment, respectively (Wright et al., 1996). 3.1.4 Pharmacokinetic Analysis All pharmacokinetic parameters were calculated by standard pharmacokinetic procedures (Gibaldi and Perrier, 1982). Briefly, terminal elimination half-lives (ti/2p) of the parent drug and metabolite were obtained from model fitting of the post-infusion data using the nonlinear least-squares regression software WinNonlin Version 1.1 (Scientific Consulting, Inc., Apex, NC). Model fittings were carried out using a weighting factor of 1/predicted y. Goodness of fit was evaluated based on Akaike's Information Criterion (AIC) and Schwarz Criterion (SC) values. The model associated with the smallest values of AIC and SC was regarded as giving the best fit. The area under the plasma concentration-time curves (AUC0"00) was calculated as the sum of area under the curve (AUC0"1) to the last time point sampled (Ct) as determined by the linear trapezoidal rule, and the estimated area to infinity (AUCt"cc) was determined by the relationship Q/p, where (3 is the terminal elimination rate constant. The area under the first moment curve (AUMC°'~) was calculated in a similar fashion with the portion from t*Q to infinity determined by (t|ast*Ct/p + Ct/p ). Total body clearance (CL T B) and renal clearance (CL i e n ai) were determined from C L T B - total dose/AUC°"OT and CLrenai - XU/AUC°"OT, respectively, where X u is the cumulative amount of 39 drug (FX) or metabolite (NFX) collected in urine from 0 h of infusion to 96 h post-infusion (Note: total dose = i.v. bolus dose + total amount of FX infused over 8 days). Mean residence times (MRT) of the parent drug (Cheng and Jusko, 1991) and metabolite (Veng-Pedersen and Gillespie, 1987) were calculated as follows: MRT p a r e m d r u g AAUMC°"M^ k„-x 2 AUC 0 / parent drug 2(k0-x + D b o l u s ) (3.2) MRT n K . t a b o m c r A U M C ° ' ^ J metabolite MRT p a r e n t d r u g (3.3) where k0, x, and Dboius a r e the infusion rate, infusion duration, and initial bolus loading dose of the parent drug, respectively. Steady-state volume of distribution of the parent drug (Vdss) was calculated by CLTB*MRTp arent drug-3.1.5 Statistical Analysis A paired t-test was used for statistical analysis of the pharmacokinetic parameters when comparing 1) R- versus S-enantiomers, and 2) parent drug versus metabolite. Comparisons of the mean values of the parameters from the high and low dose infusion and i.v. bolus experiments were performed using one-way ANOVA followed by a post hoc Tukey's test. The significance level chosen was p<0.05. The achievement of steady-state for plasma FX concentrations was evaluated by testing whether the slope of plasma concentration versus time curve was significantly different from the value of zero. Unless otherwise specified, all data are reported as the mean ± standard deviation (S.D.). 40 3.2 Results The mean body weight of nonpregnant ewes in the high dose and low dose experiments were 57.9 ± 10.3 kg and 59.3 ± 10.0 kg, respectively. 3.2.1 Stereoselective Pharmacokinetics 3.2.1.1 Fluoxetine Figures 3.2 and 3.3 show the representative plasma concentration versus time profiles of FX and NFX enantiomers in a nonpregnant ewe for a high dose and low dose experiment, respectively. Significant stereoselective pharmacokinetics of the FX enantiomers was observed in both experiments. Differences in the concentrations of FX enantiomers were observed as early as 15 minutes after the start of the infusion. The concentration of S-FX remained higher than that of R -FX throughout the entire duration of the infusion and post-infusion periods. The time course of average total FX plasma concentrations of the high and low dose experiments are shown graphically in Figure 3.4. As illustrated in Figures 3.2, 3.3, and 3.4, FX and its enantiomers progressively declined from peak plasma levels obtained with the initial i.v. bolus loading dose and reached a minimum at around 12 hours. Following this they increased gradually and reached steady-state during day-5 to day-8 of the infusion period in both sets of experiments [slopes calculated from the plasma concentration versus time curve between day-5 to -8 were not significantly different from zero (/>>0.05)]. The average C s s of total FX was 151 ± 39.2 ng/ml for the high dose experiment and 62.0 ± 30.8 ng/ml for the low dose experiments. The drug infusion was terminated after day-8 and the post-infusion pharmacokinetic profile was followed for 4 days. Upon discontinuation of the infusion, FX exhibited an apparent biphasic log plasma concentration 41 1000 Time (day) F I G U R E 3 . 2 Representative plasma concentration versus time profiles of FX and NFX enantiomers in a nonpregnant sheep (E7139) for a high dose 8-day i.v. infusion experiment. The inset figure shows the concentration-time profile during the first 6 hours of infusion. 100 R-FX m— S-FX o - R-NFX e-S-NFX O) c c o c CD U c o o Time (day) F I G U R E 3 . 3 Representative plasma concentration versus time profiles of FX and NFX enantiomers in a nonpregnant sheep (E8107) for a low dose 8-day i.v. infusion experiment. 42 1000 0 2 4 6 8 10 12 Time (day) FIGURE 3.4 Compar i son o f the average p lasma concentration versus t ime profiles for total F X from the h igh and l o w dose F X 8-day i.v. infusion experiments i n nonpregnant adult sheep (n = 6). 1000 c o ro -4—1 c fl> u c o o 100 « — High Dose e— Low LDose Time (day) FIGURE 3.5 C o m p a r i s o n o f the average p lasma concentration versus t ime profiles o f total N F X from the h igh and l o w dose F X 8-day-i.v. infus ion experiments i n nonpregnant adult sheep (n = 6). 43 versus time elimination profile, which was best described by a two-compartment pharmacokinetic model based on AIC and SC values. The pharmacokinetic parameters calculated for the FX enantiomers (mean ± S.D.) and their corresponding S/R ratios (mean ± S.D.) from the high and low dose experiments are summarized in Table 3.1. With the exception of terminal elimination half-life (ti/2p), significant differences were found in all pharmacokinetic parameters between the FX enantiomers. The elimination half-life of R-FX was similar to that of S-FX in both experiments (High dose: 18.4 ± 7.7 h versus 14.8 ± 3.3 h, p - 0.20; Low dose: 20.1 ± 8.5 h versus 19.7 ± 8.8 h, p - 0.18). Mean residence time (MRT) of R-FX was significantly shorter than that of S-FX (High dose: 24.6 ± 7.0 h versus 34.9 ± 8.6 h,/?<0.0001; Low dose: 33.4 ± 19.9 h versus 44.1 ± 22.1 h, /?<0.01). C s s was significantly lower for R-FX compared to S-FX (High dose: 35.8 ± 7.6 ng/ml versus 115.2 ± 31.6 ng/ml, /?<0.0005; Low dose: 12.6 ± 4.6 ng/ml versus 49.4 ± 26.7 ng/ml, /?<0.05). Similarly, area under the plasma concentration versus time curve (AUC) of R-FX was also significantly lower than that of S-FX (High dose: 6086 ±1311 pg*h/l versus 18056 ± 5448 ug*h/l,/?<0.001; Low dose: 2289 ± 880 pg*h/l versus 8080 ±4318 pg*h/l,p<0.05). R-FX had a higher volume of distribution at steady-state (Vdss) per kilogram compared to S-FX (High dose: 44.9 ±11.3 1/kg versus 22.7 ± 8.3 1/kg, /?<0.001; Low dose: 76.7 ± 32.6 1/kg versus 31.0 ± 13.6 1/kg, p<0.0\). The weight-normalized total body clearance (CL-re) of R-FX was significantly higher than the clearance value of S-FX (High dose: 1.91 ± 0.61 1/h/kg versus 0.68 ± 0.28 1/h/kg, /?<0.001; Low dose: 2.55 ± 0.68 1/kg versus 0.79 ± 0.28 1/kg, p<0.005). The free fraction of the FX enantiomers was determined from plasma samples obtained at 144, 168 and 192 h during the 44 •D > o < E c O si \-cc x LL co x LL CC X LL CO X LL 0. X LL CO X LL UL X LL CO X LL X LL CO X LL CC X LL CO X LL oi X LL CO X LL CC JD E E c < CO CM "Ct o oo CJ) CN LO 00 "Ct CD o oo Is-o d d d d d +1 CD Is- Is- Is- CD CO CN CN CO LO CM LO CO LO "— CN CN cb CO CM +1 o Is- Is- CO CD CO 00 CN "Ct CD CJ) CO O CD O d d d d T "~ d +1 CO CO o CO CD CO CO o CD LO CD CJ) CM CN CN +1 O o O CO CD Is- CO cb- d CD LO Is-' CN 00 CM CN CN 00 CM +1 CO CO LO CO ^ CD CO CO CO CO csi "Ct 00 CN "Ct "Ct -<t "Ct CD ^ t +1 CD LO o CD CD CD CO CO to CD CO LO "Ct •* CM Is- CJ) CJ) CD O -ct o CD Is- -ct 00 CD OO LO CM CM +1 CD 00 CM O LO CD CD 01 O O o T— 00 CO CO CD ) CD CD o O CD Is- LO co CD "Ct CD +1 CD 00 Is- LO ^ t OM CD oS 1^  CO CD o CN CJ) CM f2 CO +! LO CN en Is- LO 00 00 CD "Ct CD ^ t CD 00 1^  LO CO CO CM CO CM CO +1 LO Is- ro Is- CD CD CD CD 00 LO cb LO 00 CO •>cf CN CN CO CO CO +1 "3- CO •* o CJ) CD q CD cb 00 CO -ct Is-CM 00 CM CM CM CN +1 CO LO Is-; oo CD CO CO CD cb d cb cb d -ct cb +1 •<t CD -ct CD CO Is- -ct Is-CO LO d 1^  Is-' d 00 CN CM +1 Is- Is- CD o CD o 00 LO 00 T— T— CN CN T— CD Is- Is- I s- I s-Q Q Q Q Q X X X X X CD O 5 X « CO ^ +1 o o o d V CC J CD i ^ £ cu , CL LO I s- I s-I s- CD •<cf O I s-oo d C\i CO CN 00 00 o CD 00 00 CD CD CO CO CO LO "Ct "Ct "Ct CD CO ^ t O CO o O OO O I s- CD Is-; d T ~ d d d CM CO o I s- O o o o LO LO CO CM CO CN CN CO o CN cb CO LO CM CO d LO CO CO CD "Ct ^ "Ct I s-•* CO •St CO CD CM CN I s- cb cb I s-^ o O) t^ Is-Q CD o oo oi CD LO LO O OO 00 Is- CM CO O O Is- CD O LO "Ct CD CD 00 LO 00 "Ct CD I s- CD O CD LO O I s- 00 LO LO 00 co CM CO T— CO CM O O I s- 00 LO ^ t CM CN CN CD CO CJ) CJ) CO CN "Ct LO "Ct CN CO :D CD CM t^ 1 +1 CD CD CD CD -ct O "Ct LO 00 "- •"t Is-q CM "Ct t^ CO cb CM LO co o CO oo o Is- ca Is- CN CO ^ t T— CD cb CM " oo +i Is- LO "Ct CN Is- LO cb CO ^ t T— d cb CN 00 CN Is- ^ t Is-T— CO o c 00 CD LO CO a) Q Q a _ i _ i _ j "Ct d o +i "Ct CD d +1 oo Is-o +1 CD O +1 Is-CD d CC DC CO 45 FX constant rate infusion (i.e. at steady-state). R-FX had a higher free fraction compared to S-FX (High dose: 2.42 ± 0.82 % versus 0.71 ± 0.23%, /?<0.005; Low dose: 4.33 ± 1.50 % versus 2.37 + 1.30 %,/?<0.005). Due to premature detachment of the Foley® catheter during the early experimental period in some animals, complete urine collection was obtained from only three sheep in the high dose experiments (Ewe # 106, 7139, and 7250). All urine was successfully collected from the five sheep in the low dose experiments. Figure 3.6 illustrates a representative plot of the time course of the cumulative amount of FX excreted in the urine of a nonpregnant sheep in a high dose experiment. The cumulative amount of the unchanged FX excreted in urine reached a maximum 2 days after the termination of the drug infusion. The residual amount of drug collected over the last 48 hours period was less than 1% of the total amount excreted. Pharmacokinetic parameters for FX renal elimination obtained in the nonpregnant sheep in the high (n=3) and low dose (n=5) experiments are summarized in Table 3.2. In both sets of experiments, mean weight normalized renal clearance (CLr e nai) of R-FX appeared to be higher than that of S-FX (High dose: 0.97 ± 0.40 ml/h/kg versus 0.66 ± 0.38 ml/h/kg, p=0.\6; Low dose: 2.64 ± 3.58 ml/h/kg versus 0.57 + 0.38 ml/h/kg, /?=0.24) but the difference was not statistically significant. The percentage of unchanged FX recovered in urine was less than 1% of the total administered dose and did not differ significantly between R-FX and S-FX (High dose: 0.04 ± 0.02 % versus 0.08 ± 0.03 %,p=0A5; Low dose: 0.10 ± 0.12 % versus 0.07 ± 0.04 %,/?=0.49). 46 0 2 4 6 8 10 12 14 Time (day) FIGURE 3.6 Representative plot of the cumulative amount of FX excreted in the urine of a nonpregnant sheep (E7139) for a high dose FX 8-day i.v. infusion experiment. 1400 1200 °> 1000 800 \ c 3 O E < CD > IS 600 3 E E 3 o 400 200 -e—R-NFX -B—S-NFX -A—Total NFX Time (day) FIGURE 3.7 Representative plot of the cumulative amount of NFX excreted in the urine of a nonpregnant sheep (E7139) for a high dose FX 8-day i.v. infusion experiment. 4 7 X LL o E < X u. CO X LL z 01 x Ll_ CO X LL d. X LL. LT X LL CO X LL ct CO E CM O O X CO d O) d Q X o d d Q X CO S CD o +i ° +i ro a . o CO ^ +1 LO d E t/i ° £ c ni, ro •o § • d co d D o d o d CD d cn d co o c\i d LO d cn d cn d 1 ^ T— CO O CD LO CO O a Q _ l _J _1 CN o ° ° d ° ° +1 X o go O CO s +1 cn d CM cn d o +i o d Q_ eg CO co o d cn CNJ d cu E to ° S c 4. ro "g <P" .!= CO ro . D_ co 48 3.2.1.2 Norfluoxetine Representative plasma concentration versus time profiles obtained for the NFX enantiomers in a nonpregnant ewe for the high and low dose experiments are presented in Figures 3.2 and 3.3, respectively. Unlike the parent drug, no apparent differences were observed in the plasma concentration-time profiles of the NFX enantiomers. The mean total NFX plasma concentration versus time profiles obtained for the high and low dose experiments are illustrated in Figure 3.5. NFX appeared rapidly in plasma with quantifiable levels observed at 5 minutes after the beginning of FX infusion in all animals. The concentration of NFX continued to accumulate gradually and reached steady-state between day-6 to day-8 of the infusion period. Mean C s s of total NFX was 143.2 ± 38.4 ng/ml for the high dose experiments and 57.8 ± 23.7 ng/ml for the low dose studies. The concentration of NFX was comparable to that of FX at steady-state. Upon discontinuation of the 8-day FX infusion, NFX levels remained relatively constant for 2 to 3 hours before starting to decline. The elimination of NFX was slower than that of FX with considerable levels of the metabolite measured in all samples collected 4 days post infusion (range 2.4 - 35.7 ng/ml for NFX versus 0 - 10.3 ng/ml for FX). The pharmacokinetic parameters of the R and S NFX enantiomers (mean ± S.D.) and their corresponding S/R ratios (mean ± S.D.) for the high and low dose experiments are summarized in Table 3.3. The terminal elimination half-life was longer for R-NFX compared to S-NFX (High dose: 26.0 + 5.6 h versus 21.4 + 4.2 h,p<0.05; Low dose: 31.0 ± 3.5 h versus 26.1 + 4.8 h,/?<0.05). Similarly, the MRT of R-NFX was also significantly higher than that of S-NFX (High dose: 124.8 ± 5.4 h versus 108.5 ± 3.6 h, p<0.05; Low dose: 126.1 + 16.7 h versus 105.7 ± 10.0 h, /?<0.05). In contrast to the parent drug, the average C s s of R-NFX and S-NFX were not significantly different from one another 4 9 o 3 < ~ X 5= Z a + ^ X o X ra O < cc x LL z c/> x LL z CC CC to X LL z a: x LL z CO X X LL z CO X LL z 1^  CM CN CN cn CO o co CD cn CO CM "- "- d d *— 1 "-CM CM CO LO CN CD cn o 1^  CO o CM CN CM CN CO in CO CO CO CM CO CN O) T— LD LO LO CO CM cn o cn o o CM CM CO CM CN CO CO r- CN CN a> CO CM CO co LO O CO CN 00 LO CO o 00 cn CM CN CO CD LO CO CD CD o LO o 1— d d d d o CD cn CN LO CM CO CD CD CD CO CD CM CN CN CN CM CM CD cn CO LO CO CD cn o CD CO LO CD LO CN CO CD LO o o CN o CN l-» oo 00 CM CD CO LO co o LO CD Tf CD CM LO CD O CN CM o CO CM CD cn f~ CN CO iri CO CD 00 CO CD cn LO CD CD r~ LO LO o 00 cri CO CO LO LO CD CO CD CD CN LO oi CO 00 00 o o o o o CD CN _ LO CO CO cri cri d CO CM CM CM CO CN CN CD 00 LO CD o CD CO T4 CM CM CM CN _ o CO LO o o o CO d CO CD CO CM CN CN CN CO CN CN 1^  CD o cn o CO LO CO c CN CM ro CD 1^  1^  1^  r- TT CD Q D Q O O Q X X X X X X +1 LO o o o CO d . co ^ ro to > a Q. ~ X •o LL ' ro CO 0. CM •3- CO CD o •<* CN O) cn o T— CM CM CO CO CN CN LO LO CD CN O O oq T— CD 1^  CD CN LO cn CO CM O CO CD O CN in CO '— T— T— cn CO CD CD in o LO cn co CN 00 CD cn CD LO 00 CD in CD d d d d d d o o> o oo CM CD o CD 00 CO CM CM CN 00 CO o O CO CO O) CD CD CO CD •* o CD T— CO LO LO LO cn o CD 00 CD CD CD cn CD CO o LO a> CO CM r~ cn CD CD 00 LO CO cn CD d CN CM CD CN CM CN iri CM 00 +1 CO LO CM CM O) + 15.8 CM LO d CN CM LO CO oi CO CO + 15.8 00 CO CN o CN o d o CO cn iri o d +i 00 CO in CN cn CM cri LO CD CM CD +l cn 00 CN CD CO CM CN CN CM d CN CO CM CM CO CO CN +l CD CO CO 1^  O in CM cri CM d CO CN CO iri CO CO CO +l CO o c ro CD LO 00 CD Q Q Q _J _i _] +1 +1 00 o +1 o +1 CO cn d oo o d +l o +1 CC CC CO 50 (High dose: 74.4 ± 20.7 ng/ml versus 68.8 ± 19.9 ng/ml, p = 0.35; Low dose: 31.9 ± 15.8 ng/ml versus 25.9 ± 8.6 ng/ml, p = 0.23). Similarly, no difference was observed between the AUC values of R-NFX and S-NFX (High dose: 13658 ± 4122 pg*h/l versus 12265 ± 3169 u.g*h/l, JP = 0.14; Low dose: 6566 ± 3226 u.g*h/l versus 5130 ± 1788 pg*h/l,/? = 0.17). The AUC ratio of NFX to FX was significantly higher for the R-enantiomer compared to the S-enantiomer (High dose: 2.26 ± 0.44 versus 0.73 ± 0.26, /xO.0001; Low dose: 2.81 ± 0.91 versus 0.69 ± 0.14, p<0.01). The combined AUC values of the parent and metabolite showed a significant difference between the R- and S-enantiomers with the R-enantiomers having a lower combined AUC value compared to the S-enantiomers (High dose: 19743 ± 5 1 5 2 pg*h/i versus 30321 ± 7206 pg*h/l,/?<0.0005; Low dose: 8855 ± 4015 pg*h/l versus 13210 ± 6076 pg*h/l, jO<0.05). Similar to the parent compound, the free fraction was significantly higher for R-NFX compared to S-NFX (High dose: 2.02 ± 0.47 % versus 1.14 ± 0.35 %, /7<0.005; Low dose: 4.15 ± 1.62 % versus 3.00 ± 1.14 %,/?<0.05). Figure 3.7 illustrates a representative plot of the time course of the cumulative amount of NFX collected from the urine of a nonpregnant sheep in a high dose experiment. Overall, the excretion of NFX was close to complete 4 days after the termination of the 8-day FX infusion, with the amount of metabolite collected over the last 48 h accounting for less than 5% of the total amount excreted. The renal elimination kinetic parameters for R and S NFX in the nonpregnant sheep are summarized in Table 3.2. No significant difference was observed between the mean weight normalized renal clearance (CLr e nai) of R-NFX and S-NFX (High dose: 1.46 ± 1.03 ml/h/kg versus 1.07 ± 0.57 ml/h/kg,/? = 0.28; Low dose: 0.99 ± 1.07 ml/h/kg versus 0.81 ± 0.75 ml/h/kg, p = 0.29). Similar to FX, the percentage of total NFX recovered in urine was less than 1% of the total administered FX dose. The percentage 51 of R-NFX excreted appeared to be higher than that of S-NFX (High dose: 0.15 ± 0.10 % versus 0.11 ± 0.06 %, p = 0.17; Low dose: 0.09 ± 0.05 % versus 0.07 ± 0.05 %,p = 0.08) but the difference was not statistically significant. 3.2.2 Nonlinear Pharmacokinetics 3.2.2.1 Fluoxetine The effect of dosing on the pharmacokinetics of FX was examined by statistically comparing the pharmacokinetic parameters of total FX, and the R and S enantiomers determined from the current high and low dose 8-day infusion experiments, as well as from the previous i.v. bolus studies (Kim, 2000). The results of these comparisons are summarized in Table 3.4. No significant difference was found in the elimination half-lives values of total FX, R-FX and S-FX obtained from these three different dosing experiments. MRTs of total FX and S-FX were significantly prolonged in the high and low dose infusion experiments when compared to the i.v. bolus experiments {p <0.01). Similarly, MRTs of R-FX obtained from the infusion experiments were also longer compared to the i.v. bolus experiments although only the low dose infusion and i.v. bolus comparison was statistically significant (c<0.05). An increase in Vd s s of FX was observed in the infusion experiments compared to the previous z'.v. bolus studies, however, this was only significant for the low dose infusion experiments (p<0.01). Moreover, the Vd s s determined from the low dose infusion experiments appeared to be higher than those from the high dose infusion studies but this was only statistically significant for R-FX (p <0.01). An interesting dose-related change in clearance was observed among different dosing experiments as illustrated in Figure 3.8. The CLTB of total FX, R-FX, and S-FX did not differ significantly between 52 T A B L E 3.4 Comparison of FX pharmacokinetic parameter values determined from different dosing experiments in nonpregnant sheep. FX Pharmacokinetic Parameter Total/ Dosing Mean ± S.D. p-Value Dosing Regimen Comparison1 Enantiomer Regimen* High Low i.v. bolus High 14.2 ±3.2 - NS -Total FX Low i.v. bolus 19.9 + 9.4 n/as 0.19 NS - -High 18.4 ±7.7 - NS NS ti/2n (h) R Low 20.1 ±8.5 0.82 NS - NS i.v. bolus 20.2 ±6.0 NS NS -High 14.8 + 3.3 - NS NS S Low 19.7 ±8.8 " 0.33 NS • - NS i.v. bolus 14.8 ±5.0 NS NS -High 32.2 ±8.3 - NS ** Total FX Low 41.7 ± 21.6 <0.01 NS - ** i.v. bolus 9.3 + 3.4 ** ** -High 24.6 + 7.0 - NS NS MRT (h) R Low 33.4 + 19.9 <0.05 NS - * i.v. bolus 10.0 ±3.7 NS * -High 34.9 ±8.6 - NS ** S Low 44.1 +22.1 <0.01 NS - ** i.v. bolus 8.8 + 3.2 ** ** -High 30.8 + 10.0 - NS NS Total FX Low 44.6 ± 19.1 <0.01 NS - ** i.v. bolus 14.4 ±3.7 NS ** -High 44.9 ± 11.3 - ** NS Vds s (l/kg) R Low 76.7 ± 32.6 <0.01 ** - ** i.v. bolus 18.5±4.7 NS ** -High 22.7 ±8.3 - NS NS S Low 31.0 ±13.6 <0.01 NS - ** i.v. bolus 11.6 + 2.9 NS ** -High 1.00 ±0.39 - NS NS Total FX Low 1.19 + 0.40 0.06 NS - NS i.v. bolus 1.66 ±0.51 NS NS -High 1.91 ±0.61 - NS NS CLTB (l/h/kg) R Low 2.55 ±0.68 0.19 NS - NS i.v. bolus 1.97 ±0.48 NS NS -. High 0.68 ± 0.28 - NS ** S Low 0.79 ±0.28 <0.01 NS - ** i.v. bolus 1.44 + 0.51 ** ** -Total FX High 0.039 ±0.011 - NS NS Low 0.033 ±0.016 0.05 NS - NS /.v. bolus 0.018 + 0.008 NS NS -Dose normalized High 0.010 ±0.002 - NS NS AUC (h/l per dose R Low 0.007 ±0.003 0.27 NS - NS unit) /.I/, bolus 0.007 ±0.002 NS NS -High 0.029 ±0.009 - NS * S Low i.v. bolus 0.026 ±0.014 0.011 ±0.005 <0.05 NS * * * High 0.95 ±0.32 - ** -Total FX Low i.v. bolus 2.76 + 1.34 n/a§ <0.01 ** -High 2.42 ± 0.82 - * * /u (%) R Low 4.33 ±1.50 <0.05 * - NS i.v. bolus 4.9 + 0.6 * NS -High 0.71 ±0.23 * * S Low i.v. bolus 2.37 ±1.30 2.3 ±0.4 <0.05 * * NS NS ' High, low and i.v. bolus refer to the high and low dose FX i.v. infusions and the previous i.v. bolus experiments (Kim, 2000), respectively. • From one-way ANOVA with post hoc Tukey comparison (*p<0.05, **p<0.01, NS = not significant). Could not be determined due to insufficient data. 53 3.50 • High (n=6) Total FX R-FX S-FX FIGURE 3.8 Total body clearance of Total FX, R-FX, and S-FX following high dose infusion, low dose infusion, and single i.v. bolus administration of racemic FX. ** A significant dose related decrease in clearance is observed for S-FX. 54 the high and low infusion experiments (Figure 3.8, Table 3.4). Decreases of 40% and 20% in CLTB for total FX were observed in the high and low dose infusion experiments, respectively, compared to i.v. bolus experiments, however, these differences were not statistically significant (p = 0.06). Interestingly, C L T B of R-FX remained unchanged in all three experiments (p = 0.19) whereas the C L T B of S-FX was significantly decreased by 53% and 45%) in the high and low dose infusion experiments, respectively, compared to the i.v. bolus experiments (p <0.01) (Figure 3.8, Table 3.4). Corresponding to the change in C L T B , the dose normalized AUC values of S-FX obtained from the infusion experiments were significantly higher than that from the i.v. bolus experiment (p<0.05) (Figure 3.9, Table 3.4). The free fraction of total FX determined from the high dose infusion experiment was significantly lower compared to the low dose infusion studies (p<0.0\). The free fraction values of R- and S-FX, determined from the high dose infusion experiments, were also significantly lower than those obtained from the low dose infusion and i.v. bolus experiments (p<0.05). Free fractions of R-FX and S-FX did not differ significantly between the low dose infusion and i.v. bolus experiments. The comparison of the renal elimination kinetic parameters obtained for FX in the three different studies is summarized in Table 3.5. Total FX CLrenai was significantly higher in the i.v. bolus experiments compared to both infusion values (/?<0.05). In agreement with the lower CLrenai values, a relatively lower percentage of the total drug was excreted in urine in the infusion studies compared to the i.v. bolus experiments (p = 0.05). CL r e n a i and the % amount excreted for S-FX were also significantly higher in the i.v. bolus experiments compared to both infusion studies (p<0.01). On the other hand, no differences were observed for the renal elimination of R-FX among these three different studies (p>0.05). 55 o < T3 CU N "ro E u. O z CD 10 O Q 0.060 0.050 A T ^ A c D CD CO o T3 u. CD Q. 0.040 -0.030 - 0.020 A 0.010 A 0.000 • High (n=6) H Low (n=5) D i.v. bolus (n=6) ** p< 0.01 vs. i.v. bolus Total FX R-FX S-FX 0.060 B ?? £ 8! ro o O CD z a O CO O Q 0.050 0.040 A 0.030 0.020 A 0.010 0.000 • High (n=6) 0 Low (n=5) • i.v. bolus (n=6) ** p<0.01 vs. i.v. bolus Total NFX R-NFX S-NFX FIGURE 3.9 Dose normalized AUC values of Total FX, R-FX, and S-FX (Chart A), and Total NFX, R-NFX, and S-NFX (Chart B) following high dose infusion, low dose infusion, and single i.v. bolus administration of racemic FX. * and ** denote significant differences from the i.v. bolus data with/j-values of <0.01 and <0.05, respectively. 56 TABLE 3.5 Comparison of renal FX and NFX elimination kinetic parameter values obtained from different dosing experiments in nonpregnant sheep. Pharmacokinetic Total/ Dosing Mean ± S.D. p-value Dosing Group Comparison5 Parameter Enantiomer Group* High Low i.v. bolus High 0.74 + 0.37 - NS * Total FX Low 1.01 +0.97 <0.05 NS - * i.v. bolus 3.24 + 1.47 * * -High 0.97 ± 0.40 - NS NS R-FX Low 2.64 + 3.58 0.68 NS - NS i.v. bolus 2.16± 1.01 NS NS -High 0.66±0.38 -• NS ** S-FX Low 0.57±0.38 <0.01 NS - ** CL r e n a l (ml/h/kg) i.v. bolus 4.08 + 1.83 ** ** -High 1.27 10.79 - NS ** Total NFX Low 0.8910.89 <0.01 NS - ** i.v. bolus 17.7018.22 ** ** -High 1.4611.03 - NS ** R-NFX Low 0.9911.07 <0.01 NS - ** i.v. bolus 13.14 16.28 ** ** -High 1.07 + 0.57 - NS ** S-NFX Low 0.81 +0.75 <0.01 • NS - ** i.v. bolus 23.60 + 13.01 ** ** -High 0.06 ± 0.02 - NS NS Total FX Low 0.08 + 0.07 0.05 NS - NS i.v. bolus 0.1910.09 NS NS -High 0.04 1 0.02 - NS NS R-FX Low 0.1010.12 0.51 NS - NS i.v. bolus 0.11 10.05 NS NS -High 0.0810.03 - NS ** S-FX Low 0.0710.04 <0.01 NS - ** Amount excreted as i.v. bolus 0.2710.13 ** ** -% of FX dose High 0.1210.10 - NS ** Total NFX Low 0.0810.05 <0.01 NS - ** i.v. bolus 0.96 + 0.58 ** ** -High 0.1510.10 - NS * R-NFX Low 0.09 + 0.05 <0.05 NS - * i.v. bolus 0.7610.48 * * -High 0.11 ±0.06 - NS ** S-NFX Low i.v. bolus 0.07 ± 0.05 1.1510.67 <0.01 NS ** ** ** t H igh , low and i.v. bolus refers to the high and low dose F X i.v. infusion and the previous i.v. bolus experiments ( K i m , 2000), respectively. * From one-way A N O V A with post hoc Tukey comparison (* /?<0.05, ** £><0.01, N S = not significant). 57 3.2.2.2 Norfluoxetine Statistical analysis was also performed to examine the effect of dosing on the pharmacokinetics of NFX by comparing the pharmacokinetic parameters of total NFX, R-NFX, and S-NFX obtained from the high and low dose infusions, and previous i.v. bolus experiments; these results are summarized in Table 3.6. The elimination half-life (ti/2p) of total NFX in the high dose infusion experiments was significantly shorter than that in the low dose infusion studies (p<0.05). The elimination half-life (ti/2p) of total NFX could not be determined in the i.v. bolus experiments therefore it is not included in the analysis. Although a significant difference was observed for the t i /2p of total NFX, no statistically significant differences were found in the t|/2p values of the NFX enantiomers among the high dose infusion, low dose infusion, and i.v. bolus experiments. However, the MRTs of total NFX, R-NFX and S-NFX were significantly prolonged by more than 10-fold in the high and low dose infusion experiments compared to the i.v. bolus experiments (p<0.0\). Metabolite to parent area ratios (NFX/FX AUC ratio) for the total drug increased by 27-30% in the infusion experiments compared to i.v. bolus experiment but this difference was not statistically significant. The NFX/FX AUC ratios of R-enantiomer were significantly higher (>2-fold) in the high and low dose infusion experiments compared to the i.v. bolus studies (p<0.01). On the contrary, no statistically difference was found in NFX/FX AUC ratios of S-enantiomer among these experiments. The dose normalized AUC values of total NFX, R-NFX and S-NFX showed significant increases in the high and low dose infusion studies compared to the i.v. bolus study (p<0.01 in all cases) and these effects are illustrated graphically in Figure 3.9. The free fraction of NFX in the high dose experiments appeared to be the lowest among all free fractions determined in the other experiments. The free fraction 58 T A B L E 3.6 Comparison of pharmacokinetic parameters of NFX between different dosing experiments in nonpregnant sheep. NFX Pharmacokinetic Parameter Total/ Dosing Group* Mean + S.D. p-value Dosing Group Comparison5 Enantiomer High Low i.v. bolus High 23.7 ±4.9 - * -Total NFX Low i.v. bolus 32.9 ±5.9 n/as <0.05 * - ; High 26.0 ±5.6 - NS NS ti/2|i (h) R Low 31.0 ± 3.5 0.11 NS - NS i.v. bolus 48.8 + 30.2 NS NS -High 21.4 ±4.2 - NS NS S Low 26.1 ±4.8 0.54 NS . - NS i.v. bolus 26.2 ±14.1 NS NS -High 114.3 + 4.2 - NS ** Total NFX Low 113.7 + 12.5 <0.01 NS - ** i.v. bolus 8.5 ±4.0 ** ** -High 124.8 + 5.4 - NS ** MRT (h) R Low 126.1 ± 16.7 <0.01 NS - ** i.v. bolus 8.5 ±4.2 ** ** -High 108.5 ±3.6 - NS ** S Low 105.7 ± 10.0 <0.01 NS - ** i.v. bolus 8.2 ±3.8 ** ** -High 0.041 ± 0.011 - NS ** Total NFX Low 0.037 ±0.016 <0.01 NS - ** i.v. bolus 0.016 ±0.005 ** ** -Dose normalized High 0.022 ±0.007 - NS ** AUC (h/l per dose R Low 0.021 ±0.010 <0.01 NS - ** unit)' i.v. bolus 0.008 ±0.002 ** ** -High 0.020 ±0.005 - NS ** S Low 0.016 ±0.006 <0.01 NS - ** i.v. bolus 0.008 ±0.003 ** ** -High 1.12 ± 0.31 - NS NS Total NFX Low 1.15 + 0.18 0.12 .NS - NS i.v. bolus 0.88 ±0.15 NS NS -High 2.26 ±0.44 - NS ** N FX/FX AUC ratio R Low 2.81 ±0.91 <0.01 NS - ** i.v. bolus 1.08 ± 0.11 ** ** -High 0.73 + 0.26 - NS NS S Low i.v. bolus 0.69 ±0.14 0.74 ±0.19 0.91 NS NS NS NS High 1.64 ±0.39 - ** -Total NFX Low i.v. bolus 3.76 ± 1.45 n/as <0.01 ** ; ; High 2.02 ± 0.47 - * NS /u (%) R Low 4.15 ± 1.62 <0.05 * - NS i.v. bolus 3.8 ±0.4 NS NS -High 1.14 + 0.35 - ** ** S Low i.v. bolus 3.00± 1.14 3.0 ±0.5 <0.01 ** ** NS NS High , low and i.v. bolus refers to the high and low dose F X i.v. infusion and the previous i.v. bolus experiments ( K i m , 2000), respectively. * F rom one-way A N O V A with post hoc Tukey comparison (* /?<0.05, ** p<0.0\, N S = not significant). * Could not be determined due to insufficient data. 59 of total NFX in the high dose infusion experiments was significantly lower than that in the low dose infusions (p<0.01). Similarly, the free fraction for R-NFX in the high dose infusions was significantly lower than that in the low dose infusion experiments (p<0.05) while no statistically significant difference was observed compared to the i.v. bolus value. For S-NFX, the free fraction in the high dose infusion experiment was significantly lower compared to both the low dose infusion and i.v. bolus experiments (p<0.0\). There were no differences in the free fractions of R-NFX and S-NFX between low dose infusion and i.v. bolus experiments. Renal elimination kinetics parameters from the three different dosing experiments were also compared for total NFX, and R- and S-NFX and the results are summarized in Table 3.5. In general, CL r e nai and the percentage of dose excreted in urine for total NFX, R-NFX and S-NFX were significantly higher in the i.v. bolus experiments compared to the infusions (/?<0.05), suggesting diminished renal elimination of NFX upon long-term FX administration. 3.3 Discussion Pharmacokinetic studies in human subjects have shown that FX and NFX exhibit stereoselective disposition, with the S-enantiomers of both FX and NFX having higher plasma concentrations than the R-antipodes (Hiemke and Hartter, 2000). Stereoselective disposition of FX has also recently been demonstrated in sheep following a single i.v. bolus injection of FX (Kim, 2000). In the present study, the disposition of the FX and NFX enantiomers was further investigated under steady-state conditions in nonpregnant adult sheep. 60 3.3.1 Stereoselective Pharmacokinetics 3.3.1.1 Fluoxetine In agreement with the previous findings, we observed a significant difference in the pharmacokinetic disposition of FX enantiomers in sheep following 8-day continuous racemic FX infusion. In both, high dose and low dose infusion experiments, C s s and AUC of S-FX were 3.19 to 3.78- and 2.92 to 3.44-fold higher, respectively, than those for R-FX, and the mean C L T B was 2.81 to 3.23-fold lower than that for R-FX (Table 3.1). The S/R ratios for C s s , AUC, or CLTB were either significantly larger or smaller then the value of one. Interestingly, the degree of pharmacokinetic parameter differences between FX enantiomers observed in the present infusion studies were more dramatic in comparison with the results reported from the previous single i.v. bolus studies. Approximately 3-fold differences in AUC and CLTB for R-FX and S-FX were observed in this study whereas only 1.4-fold differences in AUC and CLTB were observed in the i.v. bolus experiments. These differences appear to be related to the dose as well as the duration of administration of FX. The effect of dose on the stereoselective pharmacokinetics of FX will be discussed in further detail in a later section of this chapter. In agreement with the differences in AUC, there was also a significant difference in MRT for FX enantiomers with MRT of S-FX being approximately 1.4-fold longer than that of R-FX. The terminal elimination half-lives in the present infusion study did not differ significantly between the R- and S- FX enantiomers whereas the previous i.v. bolus experiment reported a longer ti/2 for R-FX compared to S-FX (20.2 h versus 14.8 h). Apparently, the R versus S relationship for elimination half-life does not agree with that for clearance, since the clearance value for R-FX was higher than that of S-FX, and it is 61 generally assumed that a larger clearance value would be associated with a shorter elimination half-life. However, it is important to realize that elimination half-life is a dependent variable and relates directly to the volume of distribution and inversely to clearance. Volume of distribution and clearance are determined by different physiological factors, and the two parameters are independent of each other (Wilkinson, 1987). The apparent relationship among t\/2, Vd, and C L T B can be illustrated by the following equation: t l /2 = 0.693 x - ^L (3.4) C L T B Drugs with a higher volume of distribution (i.e. exceeding blood/plasma volume) are extensively distributed outside of blood/plasma into other tissues and, therefore, less is available in the systemic circulation for clearance by eliminating organs such as the kidney and liver. Apparently, the lack of a significant difference in 11/2 between R-FX and S-FX is due to a larger Vd term for R-FX, which offsets the effect of a higher CLTB value, resulting in a ti/2 value that is similar to that for S-FX (Table 3.1). The stereoselective difference in the distribution of FX enantiomers has not been reported in other species to our knowledge. Protein binding and tissue binding are thought to be the key components in determining the stereoselective difference between R-FX and S-FX in terms of their volume of distribution. In theory (physiological approach), the relationship between Vd, protein binding and tissue binding can be described by the following equation (Wilkinson, 1983): Vd t o t a, =V b +V T (3.5) V/T j where Vb is the blood volume, VT is the volume of the other tissues of the body, and/r and/t, are fractions of the drug present in the unbound form in tissue and blood, respectively. We 62 have shown that the volume of distribution differed significantly by about 2-fold between the F X enantiomers, which was paralleled by a 2-3 fold differences in free fraction. Intuitively, the difference in the volume of distribution could be regarded as a result of the difference in protein binding of the R- and S- enantiomers. Due to the unavailability of tissue binding data, the possibility of unequal tissue binding in relation to the stereoselective difference in volume of distribution of the F X enantiomers cannot be ruled out definitively, however, the extent of its contribution will likely be of minor significance. F X is extensively metabolized in sheep and the urine data showed that less than 1% of the administered dose was excreted as unchanged drug in urine. Renal clearances of the F X enantiomers were approximately one thousandth of the estimates for CLTB (Tables 3.1 and 3.2). Our data confirm the previous findings, from the single i.v. bolus studies that renal excretion is a minor pathway for the elimination of F X in sheep. The fact that there is no stereoselective difference observed for this pathway indicates that renal elimination mechanism of F X involves only glomerular filtration, a process primarily based on the physiochemical properties of the substance to be eliminated. Nevertheless, the low C L r e r , a i may also suggest the presence of tubular reabsorption of F X in the kidney but further study is required. In humans, the primary route of F X elimination is hepatic metabolism to inactive metabolites, which are subsequently excreted by the kidney (Hiemke and Hartter, 2000). A L 4 C - F X disposition study showed that less than 5% of the total radioactivity was recovered as unchanged drug in urine (Altamura et al., 1994). Our results are in good agreement with the data reported for humans and suggest that F X metabolism in humans and sheep is similar. Nevertheless, it has been shown that sheep are metabolically more efficient than human in terms of drug metabolism. Previous studies on other compounds such as diphenhydramine 63 (Kumar et al., 1997), valproic acid (Gordon et al., 1995), metoclopramide (Riggs et al., 1990), and labetalol (Yeleswaram et al., 1993a) have shown that the rate of drug elimination in sheep is generally higher than that reported in humans. The t\a of total FX in sheep, which ranged from 14 to 20 hours, was much shorter than the values of 1-3 days reported for human volunteers taking the drug orally (Hiemke and Hartter, 2000). Apparent oral clearance (CL/F) of total FX in humans averages 36 1/h and 11 1/h after single and multiple dose administration, respectively (Bergstrom et al., 1988). Assuming a bioavailability of 80% and body weight of 70 kg for a normal human subject, the weight normalized total body clearance is about 0.64 1/h/kg and 0.20 1/h/kg after single dose and multiple dose administration, respectively. This compares to clearance values ranging from 1.00 to 1.66 1/h/kg in sheep (Table 3.4), which suggests that the metabolic rate of FX in this species is about 3 to 5-fold higher than that in humans. From a qualitative perspective, the stereoselective pharmacokinetics of FX in sheep resembles that in humans as summarized in Table 3.7. Similar R versus S-FX enantiomers relationships between sheep and humans are observed for most major pharmacokinetic parameters such as ti/2, AUC, C S S , and CLTB-Other kinetic parameters such as MRT, Vd, protein binding, and renal excretion have not been determined for FX enantiomers in humans, therefore these comparisons cannot be performed. Knowing the fact that there are high interspecies variations in the stereoselective disposition of drugs (Ruelius, 1987; Jamali et al., 1989), the striking similarity observed for the FX enantiomers between sheep and human is exceptional and deserves further study to allow a more thorough comparison between these two species. 64 I CO co CD E X CD "O c o co M— CL.E £ CO co-g o 0 8 CD X 0 "55 E CD CD CL CD g o o CD E CD J= D_ C o Q_ E o O to cu >-LU CL c c CO CO II A or or CO 11 0T c o ro c • or CD O < O CD 2*L CD CD C C JD c ' 0 o CL O X LL 4) JD "CD 0 ~o LU 0 . C C CO CO V II or ct: co A QT c o CD C LU or CD a o ZD < to 0 >-CO CO V II or: or co 1  or CD C O X LL co 0 LU CL C C CO CO » A or or CO A or o TO o ZD < X LL X CD o Z> < X CD C T3 C C o CL CD o ON ON '—•• ON O O T CN 13 13 CJ tj a. CO W — ' s CO 3 CJ CO CO 0 CU -a CO O ca 0 CJ 0 "a <u "ab c £ CO c a 0 0 T3 -a CJ CJ CO CO ca CCS -0 - D CO CS 4^ CO P * * * 65 3.3.1.2 Norfluoxetine There was a relatively small difference between R- and S-NFX in terms of their plasma concentration versus time profiles (Figures 3.2 and 3.3). The S/R ratios for both C s s and AUC were close to unity in both the high and low dose infusions. Nevertheless, statistically significant differences in other parameters were found between NFX enantiomers (Table 3.3). The longer t i / 2 and MRT of R-NFX indicates that the pharmacokinetics of R-NFX is different from those of S-NFX, and hence a stereoselective process is involved in NFX disposition. In addition, the metabolite to parent drug AUC ratio (NFX/FX AUC ratio) was also found to be significantly higher for the R-enantiomer. It has been shown that this ratio approximates the ratio of formation clearance (CLf) to elimination clearance of the metabolite (CLm) under steady-state conditions (Lane and Levy, 1981). The difference in this ratio for the NFX enantiomers therefore suggests that stereoselectivity was involved in the formation and elimination processes of NFX. When a drug is given by continuous i.v. infusion, the steady-state concentration of the metabolite can be described by the following equation (Houston, 1981): f x k • -r = Jj? ° . • . . • . . (2 6) m ) CL According to Equation 3.6, CSS(m), the steady-state plasma concentration for a metabolite is a function of both fm, the fraction of drug converted to metabolite, and C L m , the elimination clearance of metabolite; k0 is the rate of infusion for the parent drug. The determination of/™ and C L m for the NFX enantiomers requires administration of preformed metabolites therefore these parameters are not quantifiable in this study. However, the R versus S relationship for these parameters can be estimated indirectly. Based on the understanding that 1) N-demethylation of FX to NFX is the major metabolic pathway for FX elimination and 2) there 66 is negligible renal elimination of the unchanged drug, the higher C L T B value determined for R-FX suggests a higher fm value for R-NFX relative to S-NFX. If this is true, C L m of R-NFX must be higher than that of S-NFX in order to allow both enantiomers of NFX to achieve a similar concentration at steady-state, as was observed in our study. A panel of simulated NFX concentration-time profiles illustrating the relationships among C S S ( M ) , fm and C L M are presented in Figure 3.10. Panels A to C suggest that enantiomers of metabolite with different fm and C L M values may result in either different or similar CSS(m) levels, depending on the magnitude of their differences. In addition, it is worth mentioning that the Vd of R-NFX is assumed to be higher than that of S-NFX based on the differences observed between R- and S-NFX in their free fractions (Table 3.3). Although Equation 3.6 suggests that Vd does not affect the ultimate level of CSS(m), it does play an important role in determining the elimination half-life which eventually affects the time required to attain CSS(m) as illustrated in Panel D. Among the four panels, Panels C and D mimic the concentration time-profile of NFX in our infusion studies where similar steady-state levels of NFX enantiomers were observed (Figures 3.2 and 3.3). Since no apparent difference was detected for the time to achieve C s s for R-NFX and S-NFX, Panel D may not be the best scenario to describe the pharmacokinetic process of the NFX enantiomers. Therefore we speculate that Panel C most closely reflects our results and suggests that R-NFX has/m and C L m values higher than those of S-NFX. The sum of FX and NFX AUCs provides a measurement of the total body exposure to the pharmacologically active species that contribute to the overall pharmacodynamics of FX. In the case of NFX, the S-enantiomer retains pharmacological activity equal to that of the parent drug. Our data show that the S-enantiomer had a higher summed FX and NFX AUC 67 Time (day) Time (day) FIGURE 3.10 Simulations of NFX enantiomer concentration-time profiles following an 8-day continuous i.v. infusion of FX. The solid (—) and dotted lines (---) denote the concentration-time curves of R-NFX and S-NFX, respectively. The S/R ratio for Vd was fixed at 0.5 in all cases. In Panel A, the S/R ratios of fm and C L m were 0.5 and 1, respectively. In Panel B, the S/R ratios of/m and C L m were 1 and 0.5, respectively. In Panel C, the S/R ratios of/m and C L m were 0.5 and 0.5, respectively. In Panel D, the S/R ratios of fm and C L m were 1 and 1, respectively. Panel C closely resembles the concentration-time profiles observed for the NFX enantiomers in the present 8-day FX infusion experiments. 68 ratio and therefore contributes significantly to the overall pharmacological profile of FX relative to the R-antipode. It is important to realize that factors such as protein binding displacement and stereoselective inhibition of FX metabolism may affect the magnitude of this combined S/R AUC ratio, thereby altering the stereoselective pharmacodynamic and toxicological profile of FX. In comparison to humans, the elimination half-life of NFX was much shorter in sheep (23.7-32.9 h versus 7-15 days), indicating a mark interspecies difference in terms of the rate of metabolism. Similar to the parent drug, the R versus S relationships of the NFX enantiomers between sheep and humans are generally similar but a few exceptions are noted (Table 3.7). The elimination half-life of R-NFX was longer than that of S-NFX in sheep whereas the opposite was observed in humans. Similar R- and S-NFX AUC values were found in sheep while the AUC of R-NFX was smaller than that of S-NFX in humans. The discrepancies in the R versus S relationships for t\n and AUC may or may not be attributable to interspecies differences in the underlying pharmacokinetic processes for these enantiomers. The AUC of a metabolite (AUCm) can be defined by the following equation (Houston, 1981): A U C (3.7 ) As discussed previously, we have rationally predicted that the fm and C L m values were higher for R-NFX compared to S-NFX. Making this assumption, the relationships between the S/R ratio of A U C m versus the S/R ratios of fm and C L m based on Equation 3.7 are depicted in Figure 3.11. As shown in Panel A when the S/R ratio offm is equal to 0.5, the S/R ratio for AUCm is approximately equal to 1, which is the case being observed in sheep. When the fm S/R ratio is changed from 0.5 to 1, and from 1 up to infinity, the S/R A U C m ratio will be 69 FIGURE 3.11 Interrelationships between the S/R ratios offm and AUC (Panel A); S/R ratios of C L m and AUC (Panel B); S/R ratios of C L m and t,/2 (Panel C). The S/R ratio of C L m is fixed at 0.5 in (A); S/R ratio offm is fixed at 0.5 in (B) and (C). The arrow sign (—») denotes the point when a change in the direction of stereoselectivity occurs. 70 larger than 1 (i.e. A U C m of R-NFX is smaller than the A U C m of S-NFX). From the literature data, the S/R ratio of NFX AUC in humans ranges from 1 to 2 only (Fjordside et al., 1999; Eap et al., 2001), suggesting that the fm S/R ratio will be less than 1 in humans. Apparently, the direction for the R versus S relationship for fm remains the same (R > S) between sheep and humans. It is the difference in the magnitude of the S/R ratio that leads to a difference in the A U C m S/R ratio. The same logic applies to the R versus S relationship of CL m , which is illustrated in Panel B of Figure 3.11. When C L m of R-NFX is greater than S-NFX, depending on the degree of their difference, the S/R ratio of A U C m will be equal to 1 (sheep) or larger than 1 (human). Again, the R versus S relationship of C L m remains the same in sheep and human (where R > S in both species) and it is the deviation in the magnitude of their S/R ratios that leads to the difference in AUC (Table 3.7). The relationship between C L m and metabolite Ua is depicted in Panel C based on Equation 3.4, and the same assumptions and logic are applied to explain the difference in ti/2 S/R ratio between sheep and humans. Although little is known about the stereoselective relationship of the Vd of the NFX enantiomers, previous protein binding data suggest that the R versus S relationship for Vd is similar between sheep and humans (Kim, 2000). Another interesting observation that is found between sheep and human is that the R versus S relationship of NFX in sheep closely resembles that in human subjects with PM CYP2D6 status (see NFX t i / 2 and C s s comparison in Table 3.7). This finding brings up a series of questions regarding the metabolic status of sheep in relation to the impairment of CYP2D6. Two possibilities can be generated from the above observations: 1) sheep do not have CYP2D6-like enzymatic activity and therefore, behave like PM of CYP2D6 in humans; or 2) sheep have an orthologous enzyme that processes CYP2D6-like activity which is 71 similar to that in humans, but this enzyme is inhibited upon long-term FX administration thereby converting the animal to a metabolizing status similar to PMs of CYP2D6. The key to the above possibilities can be found by examining the nonlinear pharmacokinetics of FX in sheep. 3.3.2 Nonlinearity Data in the literature indicates that FX exhibits nonlinear pharmacokinetics in humans based on a comparison of single-dose versus steady-state pharmacokinetics in volunteers (Altamura et al., 1994). These studies showed that the elimination t\n was prolonged from 1.9 to 5.7 days and C L T B decreased from 35.5 to 10.8 1/h after multiple-dose administration. This change in pharmacokinetics leads to a larger AUC at steady-state than is observed after a single dose. It has been shown that such nonlinearity is largely attributed to FX inhibitory effects on CYP2D6, which is the major metabolizing enzyme responsible for its biotransformation (Ring et al., 2001) i.e. a phenomenon of self-inhibition. It is assumed that FX is not likely to exhibit nonlinear pharmacokinetics in a species that lacks the CYP2D6 enzyme or CYP2D6-like activity. Based on this assumption, it is tempting to examine whether FX exhibits nonlinear pharmacokinetics in sheep in order to assess whether this species has a CYP2D6-like enzyme that possesses activity similar to that found in humans. The first sign of nonlinearity in FX pharmacokinetics in sheep, relates to the deviations observed between the FX C s s values (High dose: 151 ng/ml, Low dose: 62 ng/ml) and the targeted C s s (High dose: 100 ng/ml, Low dose: 50 ng/ml) that we originally planned to achieve in the 8-day infusion experiments. This target C s s was determined based on the total body clearance value obtained from our previous i.v. bolus studies (Kim, 2000): 72 ^ ss (target) = 7^ : (3-8) C L T B /.V. where k0 is the constant i.v. infusion rate and CLTB i-v. is the total body clearance value obtained with i.v. bolus dosing. From our current studies, it is apparent that there was a change in the pharmacokinetics of FX upon long-term administration compared to acute i.v. bolus dosing. ANOVA analyses comparing the pharmacokinetic parameters of FX following high or low dose 8-day continuous infusions or single bolus administration of FX were performed to further examine this phenomenon. As illustrated in Figure 3.8, a progressive decrease in CLTB of total FX was noted in the high dose infusion studies compared to the low dose infusion and single i.v. bolus experiments. This suggests inhibition of FX metabolism upon long-term dosing, which is believed to be the cause of the discrepancy between the observed C S S and target C S S in our current infusion studies. Interestingly, upon closer examination of the clearance of individual FX enantiomers, a significant reduction in clearance was found for S-FX but not R-FX, indicating a stereoselective mechanism behind this nonlinear kinetic process. Moreover, the significantly prolonged MRT observed in the infusion studies, suggests that persistence of the drug upon long-term administration parallels its decrease in clearance. Renal clearance of FX also appeared to be decreased in the infusion studies compared to the single bolus data, suggesting that renal elimination of FX may involve a carrier mediated process (such as tubular reabsorption) that is saturable at higher drug concentrations. However, the reduction of renal clearance will not have any significant impact on the pharmacokinetics of FX due to its low contribution to the overall total body clearance (i.e. CLTB). The distribution characteristics of FX also appeared to be altered during the infusion studies. Volume of distribution increased by 2-3 fold in the infusion 73 experiments compared to the single bolus studies. Concentration-dependent protein binding, characterized by an increase unbound free fraction due to the saturation of binding sites, is the most common phenomenon that causes an increase in Vd at high drug concentration. However, we observed an unusual decrease in unbound fraction in the high dose infusion experiment compared to the low dose infusion and single bolus experiments. Protein binding of FX appears to follow an atypical dose dependent phenomenon with more drug binding to plasma protein at higher concentrations. This is an unusual finding because it is against conventional saturable protein binding theory, which states that binding decreases as concentration of the substrate increases. It is difficult to judge whether this is an experimental artifact because a downward tendency for Vd, which is consistent with the increase in protein binding, was found in the high dose infusion experiments compared to the low dose infusions. Previous in vitro protein binding experiments using spiked sheep plasma at drug concentrations ranging from 100 to 10,000 ng/ml did not reveal this concentration-dependent change in binding (Kim, 2000). Concurrently, a substantial increase in Vd was observed upon long-term FX dosing, which may be attributable to an increase in lysosomal trapping of FX in lungs. Drug accumulation in tissue involving lysosomal trapping has been reported for numerous basic drugs including promazine and amitriptyline (Daniel and Wojcikowski, 1997). This trapping mechanism serves as a depot for extravascular distribution of FX in a time dependent manner where drug accumulation occurred progressively over time. Given the finite availability of lyzosomal vesicles, saturation of this uptake mechanism may happen at high drug concentration. Upon saturation, unbound drugs that are excluded from the lyzosomal vesicles diffuse back into the vascular system and become available for binding with plasma protein, which might explain why there was a 74 decrease in the free fraction observed at high FX concentration in vivo. And because of the involvement of peripheral tissues, this type of redistribution event is not expected to happen in an in vitro situation. The comparison of NFX data between different dosing experiments in sheep yielded some interesting results. Elimination half-life, due to its hybrid characteristics with Vd and CLTB , did not differ significantly between different dosing experiments. However, significant accumulation of NFX was observed with a disproportionate increase in AUC and a significant prolongation of MRT in the infusion studies relative to the i.v. bolus experiments. The accumulation of NFX led to an increase in overall total metabolite to parent (NFX/FX) AUC ratio by about 25-30% upon long-term FX dosing. Interestingly, a significant increase in this ratio was observed for the R-enantiomers but not the S-antipodes. This difference is due to a similar percentage increase in dose normalized AUC of R- and S-NFX but a non-parallel rise in dose nonnalized AUC for FX with a larger increase for the S-enantiomer compared to the R-enantiomer (Figure 3.9). The above observations suggest nonlinearity in NFX disposition, which has never been described or characterized before in the literature. It is proposed that the major metabolic pathways of NFX involve deamination and (9-dealkylation to FX carboxylate and p-TFMP, respectively, yet the metabolic enzymes responsible for the biotransformation of NFX remain poorly understood (Altamura et al., 1994). It is speculated that the nonlinear pharmacokinetic characteristics of NFX may be attributed to the inhibition of CYP2D6 by both FX and NFX. Fjordside et al., 1999 have reported smaller AUC values of NFX in PlVls of sparteine compared to their EM counterparts after a single oral dose of FX in human volunteers. These results suggest a link between CYP2D6 and NFX metabolism. Nevertheless, more experiments such as in vitro metabolism 75 studies of NFX and in vivo inhibition studies (i.e. co-administration of NFX with known cytochrome P-450 inhibitors) will be necessary to systemically delineate the exact mechanism of NFX metabolism. While inhibition of metabolism is considered to be the most likely reason to explain nonlinearity, nonlinear protein binding may also play a role in the nonlinear pharmacokinetics of NFX. Similar to FX, an unusual decrease in unbound fraction was observed at high drug concentrations. Drugs that are known to exhibit nonlinear pharmacokinetics due to nonlinear protein binding include disopyramide and valproic acid (Upton and Williams, 1986; Gomez Bellver et al., 1993; Wong et al., 2001). Obviously, the protein binding of NFX should be investigated in greater detail in a mechanistic manner in order to help us better understand its role in the overall pharmacokinetics and pharmacodynamics of this active and long half-life FX metabolite. 3.3.3 Mechanism of Stereoselective Pharmacokinetics of FX The mechanism responsible for the stereoselective disposition of FX in humans has been studied (Stevens and Wrighton, 1993; von Moltke et al., 1997; Margolis et al., 2000; Ring et al., 2001). Yet, most of these studies have focused on the differential biotransformation of FX enantiomers to their respective N-desmethyl metabolites. Previous in vitro studies performed with human liver microsomes and cDNA-expressed CYPs suggest that CYP2D6 and CYP2C9 play a pivotal role in the formation of NFX (Margolis et al., 2000; Ring et al., 2001). It has been shown that both CYP2D6 and CYP2C9 contribute to the formation of R-NFX, whereas only CYP2D6 is responsible for the conversion to S-NFX. Upon chronic dosing, the CYP2D6-mediated metabolism of the FX enantiomers is likely to 76 be inhibited by their counterparts (i.e. one enantiomer inhibiting metabolism of the other enantiomer) as well as by NFX. As a consequence, the formation of R-NFX is mediated by CYP2C9 while the formation of S-NFX involves multiple high K m enzymes (e.g. CYP3A4 and CYP2C19), resulting in a lower clearance value of S-FX (Ring et al., 2001). O-dealkylation to p-TFMP has been identified as an alternate route of elimination for FX (Altamura et al., 1994; Urichuk et al., 1997). Studies performed in vitro using human liver microsomes have shown that the formation of />TFMP mainly involves CYP2C19 and CYP3A4 (Liu et al., 2001), but the stereoselective aspect of these metabolic pathways is not well understood. In sheep, the observed stereoselective disposition of FX after a single i.v. bolus injection of racemic FX seems to be related to differential protein binding of the FX enantiomers (Kim, 2000). R-FX was found to be less protein bound compared to S-FX, therefore having a higher free fraction and being more readily available for clearance by the liver. Kim (2000) also suggested that the difference in clearance of the FX enantiomers was largely due to differences in their degree of plasma protein binding since free intrinsic clearances (CLint) between the FX enantiomers did not differ significantly. In fact, the reported value for CLjn, of R-FX is actually lower than that of S-FX (40.2 vs. 62.6 1/h/kg), which is contrary to what was described previously in humans (Stevens and Wrighton, 1993), but the author did not attempt to address this discrepancy at all. Upon closer examination of his report, the method that Kim (2000) employed to estimate CL i n t seems to be problematic. CLjnt was determined based on the assumption that FX is a low clearance drug with a low extraction ratio (E < 0.3). According to the well-stirred model (Wilkinson, 1987) of hepatic drug elimination, hepatic clearance (CLH) is defined as: 77 C L H = Q H x E (3.9) where Q H is the hepatic blood flow, and E is the extraction ratio which can be further defined as (Wilkinson, 1987): f CT E = ^ i2! (3.10) ( Q H + / „ C L I M ) For a drug that has a low extraction ratio and is predominately eliminated by the hepatic metabolism with negligible renal excretion, CLTB approximates the product of the unbound fraction in plasma (fu) and C L ; N T as shown in Equation 3.11: C L T B = C L H = / U C L I M (3.11) The assumption of FX as a low extraction ratio drug without knowing its actual value is subject to error because Equation 3.11 will underestimate the true C L I N T value for drugs that have either a medium (E = 0.3-0.7) or high (E > 0.7) extraction ratio. Such a deviation from the true value will amount to 30% even when E = 0.3 (Figure 3.12). Because of this, Equation 3.11 should not be used to calculate CL, n t when extraction ratio is greater than 0.3. Although the extraction ratio of FX in human has never been published, the literature reported values for apparent clearance (CL T B/F = 36 1/h) and systemic bioavailability (F = 0.7) obtained following a single oral dose administration of FX in human (van Harten, 1993) allow us to estimate extraction ratio using the following equation (assume CLTB ~ CLM, QH = 1.5 l/min; Wilkinson and Shand, 1975): CT CT QH QH We found that the extraction ratio of total FX was approximately 0.28, therefore fulfilling Kim's assumption that FX is a low extraction ratio drug. However, it is questionable if this assumption truly applies to sheep. First, our data have shown that sheep are more efficient at 78 700 Extraction Ratio FIGURE 3.12 Plot of extraction ratio versus predicted CL j n t value calculated from equation 3.11 and the true CL j n t value calculated using equation 3.13. Q H and fu were assumed to be 1.5 l/min and 0.01, respectively, in both cases. A significant deviation (>30%) between the predicted and true value occurred when the extraction ratio was greater than 0.3. 79 eliminating drugs than humans (see earlier discussion in Section 3.3.1.1); hence it is likely that the extraction ratio of FX in sheep will be different from that in humans. Second, hepatic blood flow in sheep is reported to be approximately 60% higher than that in the human (2.4 l/min vs. 1.5 l/min) (Apatu and Barnes, 1991; Katz and Bergman, 1969). As illustrated above in Equations 3.9 and 3.10, hepatic blood flow is one of the important components in the estimation of a drug's extraction ratio, a difference in hepatic blood flow between sheep and human would be expected to influence the estimation of this ratio in sheep. Third, in vitro metabolism data indicated that CLj n l of R-FX (ranged 0.41 - 5.54 ul/min/mg) was higher than that of S-FX (ranged 0.25 - 0.84 ul/min/mg) (Stevens and Wrighton, 1993). Although we have just calculated that racemic FX has a low extraction ratio of 0.28 based on the in vivo literature data, this value may not be applicable to the individual enantiomers due to large stereoselective differences in their pharmacokinetics. Indiscriminant use of racemic data to interpret stereoselective data will lead to a false estimation of the true CLj n t if one of the enantiomers in fact has an extraction ratio larger than 0.3. It is therefore, imperative to appropriately evaluate FX's pharmacokinetic data in sheep before a rational conclusion can be reached regarding its mechanism(s) of stereoselectivity. Since hepatic blood flow was not measured in these current studies, the literature value of hepatic blood flow of 2.9 1/h/kg in sheep (Apatu and Barnes, 1991) was used to calculate the extraction ratio using Equation 3.12. CLj n t was determined from Equation 3.13 by rearranging Equation 3.10: C L ^ Z ^ i ) • < 3 ' 1 3 ) The results are summarized in Table 3.8. Similarly, extraction ratio and CLj n t values for the FX enantiomers in humans calculated based on literature values from various reports 80 C/J B c CO E X 0 N O JD CD 0 > CO c 0 X LU X CD CD LL CM O CO O CL 0 0 JZ CO 0 N O JD CD O O 0-co 13 o CQ 0 CD c CO X O o LL CT) CO o CO c g CO 13 CD T3 00 1> t-0 o E CD CD E CD CD 0-JZ 0. X LO LL CNJ C\i CO d O X CD oo LL 1 CO CC o X LL CO X LL I CC X LL I CC CC CO o o' "* O 00 CD O) CD CD S-CD O CO CN CD CD CN 03 CD -cf CD LU ^ , co oa -a o 0 0 O II CQ o -a c 3 o jo a 3 CQ C © © ft -3 > C3 T3 CD o -a o _o CJ o II LL o o o ft CJ 4 ^ 0 0 M j2 _a ON CN II I a O CO > CJ s CJ 3 CJ 2 CJ a CJ E CJ ft X CJ c o o o © CO -a X ft CJ 43 C E o -a £3 O c 3 3 =3 , CJ ON ON ON T3 o CJ e CO a . C4-I o o jo X co e o CJ ft X CJ 0 0 a lo o CJ w 3 -a c c cd s~ o o ft a >N -a 3 3 (T1 0 0 c o -o 0 0 e o -a Q O Q a C3 3 cr W 0 0 a LO ° nj 1 3 CN CJ i 1 CO -S O T3 0 0 O 0 0 lo c o T3 cj co CO X) w u 81 (van Harten, 1993; Fjordside et al., 1999) are also included for purposes of comparison. It was observed that extraction ratios for both FX enantiomers are > 0.3 in the single i.v. bolus experiments in sheep. This indicates that FX is a medium extraction ratio drug in sheep as opposed to the previous assumption that the drug has the same extraction ratio in both sheep and humans. In addition, C L j m of R-FX was found to be more than 50% higher than that of S-FX, suggesting that, in addition to differential protein binding, stereoselective metabolism is involved in the stereoselective disposition of FX. In contrast to R-FX the extraction ratio and CLi m values of S-FX were observed to be lower in the infusion study compared to the single i.v. bolus study, resulting in a larger difference between their metabolic rate and more prominent stereoselective differences in their disposition at steady-state. This observation is important in supporting the hypothesis that inhibition of the metabolism of S-FX occurred upon long-term FX administration in sheep, suggesting the presence of a CYP2D6-like enzyme. In humans, a stereoselective difference in the extraction ratio between FX enantiomers is observed and such a difference is more apparent for PMs compared to EMs. Due to a lack of CYP2D6 activity, PMs will have a lower extraction ratio compared to the EM counterparts (Table 3.8). Interestingly, the extraction ratio of R-FX (0.36) in EM is within the middle part of the spectrum (E = 0.3 - 0.7) while S-FX (0.29) falls into the lower range of the spectrum (E < 0.3) and the racemic drug also has an extraction ratio value of less than 0.3. It can be concluded then, that one has to be cautious in the interpretation of the disposition of a racemic drug that exhibits stereoselective pharmacokinetics. Further, the pharmacokinetics of enantiomers should be evaluated individually rather than making generalizations based on racemic data. 82 3.4 Summary In summary, we have shown that the enantiomers of FX and NFX exhibit significant pharmacokinetic differences in sheep under steady-state conditions, confirming the observations made in previous i.v. bolus experiments (Kim, 2000). In comparison to humans, quantitative differences exist between the two species, however qualitatively the overall stereoselective disposition of FX and NFX is similar. Our data revealed a dose-dependent decrease in clearance upon long term dosing in sheep, suggesting the presence of an orthologous enzyme(s) comparable to human's CYP2D6. From our combined studies the stereoselective disposition of FX involves the following mechanisms: 1) differential protein binding, 2) stereoselective metabolism and 3) stereoselective enzyme inhibition. As mentioned earlier, selection of an appropriate animal model for stereoselective pharmacokinetic drug studies is a challenge given that substantial interspecies variability is usually present in stereoselective drug disposition. Given the similarity between our sheep FX data and that available for humans, sheep may be an animal model of choice for other stereoselective drug disposition studies. There is no doubt, however, that this requires further testing, particularly with respect to metabolic enzyme identification, enzyme kinetics and protein binding mechanisms. Nevertheless, it opens up new opportunities for allometric pharmacokinetic studies and drug metabolism research by including sheep as a suitable model in addition to those that are commonly used for drug research at present (i.e. mice, rats, dogs, monkeys, and rabbits). Finally, our results provide an important framework for further investigations into the effect of pregnancy on the stereoselective pharmacokinetics of FX. 83 CHAPTER 4 AN INVESTIGATION OF THE EFFECTS OF PREGNANCY ON THE STEREOSELECTIVE DISPOSITION OF FX AND NFX IN PREGNANT SHEEP Preface As discussed in the general introduction of this thesis, changes in maternal physiology occur normally during pregnancy and have the potential to alter the absorption, distribution, and elimination of drugs used therapeutically in pregnant women. These physiological changes include: plasma volume expansion and increases in extracellular fluid space and total body water; decreased plasma albumin concentration; a compensated respiratory alkalosis; increased cardiac output with regional blood flow changes; increased renal blood flow associated with increased glomerular filtration; changes in hepatic drug metabolizing enzymes; and changes in gastrointestinal function. These changes begin in early gestation but are most pronounced in the third trimester. Further maternal physiological changes occur intrapartum with some normalizing themselves within 24 hours of delivery, while others are sustained and return to normal some 12 weeks postpartum. The purpose of this chapter is to examine the effects of pregnancy on the stereoselective disposition of FX and NFX during the last trimester of pregnancy in chronically instrumented sheep. Pharmacokinetic comparisons are made between this data and that obtained for nonpregnant sheep (Chapter 3). Mechanisms of some pregnancy-induced changes on stereoselective drug disposition are also explored based on the current literature data available for other racemic drugs. 84 4.1 Methods 4.1.1 Animal and Surgical Preparation Five pregnant ewes (Dorset and Suffolk breeds) at gestational ages ranging from 123 to 125 days and term -145 days were used in these experiments. The studies were approved by the University of British Columbia Animal Care Committee and the procedures performed conformed to the guidelines of the Canadian Council on Animal Care. To obtain time-dated ewes, the estrus cycle was synchronized by inhibition of spontaneous ovulation. This was accomplished by the administration of a medroxyprogesterone acetate pessary (Vermix Sheep Sponge) intra-vaginally for two weeks. Ovulation was then, induced by removal of the pessary and intramuscular injection of 500 I.U. of pregnant mare's serum gonadotropin. The ewes were subsequently placed with the ram for 1-2 days to result in time-dated pregnancies. Pregnancy was confirmed by ultrasound at -70-120 days gestation. The general surgical procedure was similar to that described previously in Chapter 3, Section 3.11 for nonpregnant ewes. Blood collection and drug administration were performed via silicone rubber catheters implanted in maternal femoral artery and vein, respectively. A minimum of 3 days was allowed for the ewes to recover before experimentation. Due to the increased risk of infection in the pregnant ewe and/or fetus secondary to long-term Foley catheter placement, urine collection was not performed in these experiments. 4.1.2 Experimental Protocol The dosing and sampling protocols used were identical to for the nonpregnant sheep experiments (Figure 3.1) except that the FX infusion was conducted only at the high dose level. Each ewe received a 70 mg i.v. bolus loading dose over one minute followed by a 85 continuous i.v. infusion of racemic FX via the maternal femoral vein at a rate of 6.92 mg/h for 8 days to achieve steady-state concentrations. As in the nonpregnant ewes, serial blood samples (5 ml) were collected from the maternal femoral arterial catheter at 0, 5, 15, 30 min, 1, 2, 4, 6, 12, 24, 36, 48, 60, 72, 84, 96, 108, 120, 132, 144, 156, 168, 180 h, 192 h during the infusion, and at 5, 10, 20, 30, 45 min, 1, 2, 3, 4, 6, 9, 12, 24, 36, 48, 60, and 72 h post-infusion. The zero time blood samples were collected within 10 minutes prior to loading dose administration. All blood samples collected for drug and metabolite analysis were placed into K 3 EDTA Vacutainer® tubes, gently mixed, and centrifuged at 2000g for 10 min to separate the plasma. The plasma supernatant was transferred into a clean borosilicate test tube with a polytetrafluoroethylene (PTFE)-lined cap. All samples were stored at -20°C until the time of analysis by GC/MS. 4.1.3 Determination of FX and NFX Plasma Protein Binding Plasma protein binding of FX and NFX was measured ex vivo using the equilibrium dialysis procedure as.described in Chapter 3, Section 3.1.3. Plasma unbound fraction (fu) was calculated as described in Chapter 3, Section 3.1.3, Equation 3.1. 4.1.4 Pharmacokinetic Analysis All pharmacokinetic parameters were calculated by standard pharmacokinetic procedures as described in Chapter 3, Section 3.1.4. Since no urine was collected in these experiments, renal clearance (CLr e nai) and the corresponding pharmacokinetic parameters were not determined. 86 4.1.5 Statistical Analysis Paired t-tests were used for statistical analysis of the pharmacokinetic parameters when comparing R- versus S-enantiomers in the pregnant ewes. Comparisons of the mean values of the parameters between the pregnant and nonpregnant infusion experiments were performed using two-sample t-tests. The significance level used was p<0.05. The achievement of steady-state for plasma FX and NFX concentrations was evaluated by testing whether the slope of plasma concentration versus time curve was significantly different from the value of zero. Unless otherwise specified, all data are reported as the mean ± standard deviation (S.D.). 4.2 Results The mean body weight of pregnant and nonpregnant sheep in the high dose FX infusion experiments were 71.7 ± 9.9 kg and 57.9 ± 10.3 kg, respectively. The mean body weight of pregnant ewes was significantly higher than that of nonpregnant sheep (/?<0.05). 4.2.1 Fluoxetine Representative plasma concentration versus time profiles of the R and S enantiomers of FX enantiomers obtained in pregnant sheep are illustrated in Figure 4.1. A comparison of the average plasma concentration versus time profiles of total FX and NFX is shown in Figure 4.2. As in the nonpregnant studies a two-compartment model best described the post-infusion concentration time data based on AIC and SC values using a weighting factor of 1/predicted y. Similar to the findings in the nonpregnant animals, FX exhibited significant stereoselective pharmacokinetics in pregnant sheep. FX reached steady-state plasma 87 1000 100 10 Time (day) 12 FIGURE 4.1 Representative plasma concentration versus time profiles of FX and NFX enantiomers in a pregnant ewe (E4237) obtained for an 8-day maternal i.v. infusion of racemic FX. 1000 q 1 -I , , , , , , , , , , 0 2 4 6 8 10 12 Time (day) FIGURE 4.2 Comparison of average plasma concentration versus time profiles of total FX and NFX in pregnant sheep after 8-day maternal i.v. infusion of racemic FX (n=5). 88 concentrations during day-6 to day-8 of the infusion period compared to day-5 to -8 in nonpregnant ewes. A comparison of the average plasma total FX and FX enantiomers concentration-time profiles in pregnant and nonpregnant sheep is presented in Figures 4.3 and 4.4, respectively. Steady-state total FX concentrations (C S S ) averaged 208.2 ± 85.0 ng/ml and 151.0 ± 39.2 ng/mL in the pregnant and nonpregnant ewes, respectively, and were not significantly different. Similarly there were no significant differences between the two groups in the steady-state concentrations of the R and S enantiomers (Figure 4.4). The pharmacokinetic parameters calculated for the FX enantiomers (mean + S.D.) from the pregnant ewe experiments are summarized in Table 4.1. Also included in this table are the S/R ratios for the various pharmacokinetic parameters as well as the mean data used for statistical comparisons between pregnant and nonpregnant sheep. The relationship between the S and R -FX enantiomers in the pregnant ewes was very similar to that observed in nonpregnant sheep (Figure 4.4). No difference was found in the terminal elimination half-life (ti/2p) while significant differences were observed in all other pharmacokinetic parameters (i.e. MRT, C s s , AUC, Vd s s, C L T B , and fu) between R-FX and S-FX (Table 4.1). A comparison of the individual enantiomers between the two study groups (i.e. R-FX(pregnant) versus R-FX (nonpregnant) ; S-FX(prcgnant) versus S-FX(n0npregnani)) revealed tendencies for increases in ti/2p, mean residence time (MRT), C s s , and AUC in pregnant sheep compared to the nonpregnant animals, but the difference was only significant for the ti/2p of S-FX (35.9 + 18.3 h versus 14.8 ± 3.3 h,/?<0.05). Likewise, the MRTs of R-FX and S-FX also tended to be longer in pregnant than nonpregnant sheep but the differences were not statistically significant (R-FX: 40.2 ± 21.0 h versus 24.6 ± 7.0 h,p = 0.12; S-FX: 57.8 ± 24.3 h versus34.9 ± 8.6 h, p = 0.06). As a result of the t|/2p prolongation, the time to steady-state 89 1 -I • , . , , , , , , , , , 0 2 4 6 8 1 0 12 Time (day) FIGURE 4.3 Comparison of average plasma concentration versus time profiles of total FX of pregnant and nonpregnant sheep after 8-day i.v. infusion of racemic FX. 1000 • - 100 +— R-FX (Pregnant) e— R-FX (Nonpregnant) S-FX (Pregnant) S-FX (Nonpregnant) 10 12 Time (day) FIGURE 4.4 Comparison of average plasma concentration versus time profiles of FX enantiomers in pregnant and nonpregnant sheep after an 8-day i.v. infusion of racemic FX. 90 X LL. OT X Li_ ti X LL OT — CO _ l X o cc X LL kg) OT X > LL CC X LL OT CD o X < LL ri X .—- LL E OT c X O LL ci X LL .£= OT 1-cc X LL ti FX , OT X LL ti E 70 2 "co E c < C O a> .O ro H O CD tr o +1 CD C O 2 LO 3 co m +i CO CO o CO C N C O L O T— CO T— C O C N +1 d +1 C O C O CD L O CO C O +1 CD CO CO 00 +1 o d +1 C O CO o Tf C N +1 CO C O ^ C O C O i * -+1 u ot ^ +1 CN l O C D CD C O C D C N C N m 00 C O 00 L O O d d d d d o +1 d O +1 d V o CD o L O 00 CO C N C O Tf d C N C O CM C O C N C N d +1 L O CO C N o •>t 00 C N C N d CM C N Tt C N CO o d d d d d d o +1 d o +1 d C N C O d +1 o V L O L O o L O L O o CO o CD d CM d o ai CT) C N C N m CO Ti-ed o CO C N d C N CD d CD o -H d o +1 d CO L O V L O a> L O C N Tf iri C O CO 00 iri CM +1 d LO 00 C N C N m Tf C O C O L O CO 00 o IT) oo C B in C N C N CT) Tf O in CD co C N Tf C N 00 CO a> .— CO C O T— T— C N +1 o -o • 0> 00 CO o CO L O CO CD CT) CT) CO C O C N L O C D 00 1^  C D o m +1 CM 00 C D h - C D C N C O d CO C O C O co +! CT) L O o o C N CD 1^  C N CD o o +1 o d V C O CD C D CT> CM T-C N C D iri iri CO CM T— C O C O Tf CO CO m C O +1 00 C O CO L O a> CO CD ai ai C N CD C O CO *t L O +1 o d o o +1 d C N o o V CO L O o CD L O in C N o o CO d C N T— L O C N C N CO +1 CD CO in CO in "Cr C O 00 C O d C D iri C D CD *~ C N C O CO +1 o o C N +1 d C D C D C O CM CD C O CT) Tf o C N co C N d C O L O CM CO C O +1 X , . LL In CO C N O r ~ o OT g o CO CM L O CO c Q ro C N C N ro OT CC i § in r ~ •<t CD LU LU LU LU LU +1 CC OT i 4 CL a CL Cu CL o d o d o v X L L *a> CO ro c/i > ro > i Q. CC OO o d o d CD o d CM d o d CD lit ra ra LU ra ° - C L a E . ¥ o « £ o > oo C N Tf d C N d c o o s 1> ro't-cc 5 ¥ E OT O O 91 was delayed by an additional -24 hours in the pregnant animals. The C s s of R - and S-FX also tended to be higher in the pregnant animals compared to nonpregnant ewes but again the differences were not statistically significant (R-FX: 41.1 ± 8.7 ng/mL versus 35.8 ± 7.6 ng/mL, p = 0.31; S-FX: 167.2 ± 76.7 ng/mL versus 115.2 ± 31.6 ng/mL, p = 0.16). Similarly, the AUC of R - and S-FX also tended to be higher in the pregnant animals, but again there were no significant differences (R-FX: 7630 ± 1753 pg*h/l versus 6086 ± 1311 pg*h/l,/> = 0.11; S-FX: 29343 ± 14207 pg*h/l versus 18056 ± 5448 pg*h/l,/? = 0.10). These changes appeared to be consistent with a lowering of the weight-normalized total body clearance (CLT B) of FX observed in pregnant sheep compared to nonpregnant sheep (R-FX: 1.20 ± 0.32 1/h/kg versus 1.91 ± 0.61 1/h/kg, /?<0.05; S-FX: 0.38 ± 0.22 1/kg versus 0.68 ± 0.28 1/kg, p = 0.08). Despite a generally known increase in plasma volume and altered plasma protein concentrations during pregnancy, no significant difference was found in Vd s s (R-FX: 48.2 ± 25.3 1/kg versus 44.9 ± 11.3 1/kg, p = 0.78; S-FX: 20.2 ± 10.9 1/kg versus 22.7 ± 8.3 1/kg, p = 0.66) or the steady-state unbound fraction (fu) of the FX enantiomers (R-FX: 2.71 ± 0.48 % versus 2.42 ± 0.82 %,p = 0.50; S-FX: 0.83 ± 0.26 % versus 0.71 ± 0.23 %,p = 0.47) between the pregnant and nonpregnant animals. The S/R ratios of FX pharmacokinetic parameters were comparable between pregnant and nonpregnant animals with no significant differences noted. Outwardly, at least, pregnancy did not cause any remarkable changes in the S/R ratios of the FX enantiomers. Nevertheless, subtle changes in the S/R ratios were observed, particularly with C s s , which appeared to be positively correlated with time and total FX concentrations as illustrated in Figures 4.5 and 4.6. The average C s s of FX in the pregnant animals was higher than in the nonpregnant animals, resulting in a generally higher S/R C s s ratio in pregnant sheep. A higher degree of 92 4.5 4 o 3.5 to a: sz 3 o ro L. 2.5 C CD O C 2 O O X 1.5 LL pr (75 1 0.5 0 • Nonpregnant • Pregnant • • • • • • • • Steady-state concentration between day-5 to 8 0 1 2 3 4 5 6 7 8 9 Time (day) F I G U R E 4.5 Schematic diagram of average S/R FX concentration ratio versus time profiles from the nonpregnant and pregnant studies illustrating the dependency of the S/R ratio to time to steady-state in all cases (error bars have been omitted for clarity). 4.5 c o re 3 e o § 2.5 o X 2 to 1.5 Css = 151.0 ±39.2 Css = 208.2 + 85.0 * • Q o > • Nonpregnant • Pregnant 50 100 150 Total FX Concentration (ng/ml) 200 250 F I G U R E 4.6 Plot of the S/R FX concentration ratio versus the corresponding total FX concentration data from the nonpregnant and pregnant studies, showing the plateau of S/R ratio at steady-state concentrations in each case. 93 variability in the pregnant animal data appeared to be responsible for the lack of statistical differences between the two groups. 4.2.2 Norfluoxetine Representative plasma concentration versus time profiles of the R- and S-NFX enantiomers in a pregnant ewe are shown in Figure 4.1. A comparison of the average plasma concentration versus time profiles of total FX and NFX is depicted in Figure 4.2. Similar to the findings in nonpregnant experiment, no apparent differences were observed in the plasma concentration-time profiles between R- and S-NFX. As in the nonpregnant studies, the concentration-time profiles for the metabolite data also followed two-compartment model pharmacokinetics. Figures 4.7 and 4.8 compare the average plasma concentration-time profiles of total NFX and the two NFX enantiomers, respectively in pregnant and nonpregnant sheep. Unlike the situation in nonpregnant sheep, NFX continued to increase and did not reach steady-state by the end of the infusion period. The average concentration of total NFX before the termination of the drug infusion was 194.7 ± 79.3 ng/ml, which tended to be higher than the mean C s s value of 143.2 ± 38.4 ng/mL in the nonpregnant sheep high dose experiment but the difference was not statistically significant (p = 0.19). Upon discontinuation of the FX infusion, NFX levels remained constant for up to 4 hours before starting to decline. Outwardly, the elimination of NFX appeared to be much slower than that of FX, with considerable amounts of the metabolite being measured in all samples collected 4 days post infusion. The pharmacokinetic parameters determined for the NFX enantiomers from the pregnant experiments are presented in Table 4.2. Also summarized in this table are the 94 1000 q •e— N F X - N o n p r e g n a n t 0 2 4 6 8 10 12 Time (day) FIGURE 4.7 Comparison of the average plasma concentration versus time profiles of total NFX of pregnant and nonpregnant sheep during and following an 8-day i.v. infusion of racemic FX, showing longer elimination half-lives of NFX in pregnant sheep compared to nonpregnant sheep. - • - R - N F X (Pregnant ) - e - R - N F X (Nonpregnant ) - • - S - N F X (Pregnant ) 1 -I , , . , , , , , 0 2 4 6 8 10 12 Time (day) FIGURE 4.8 Comparison of the average plasma concentration versus time profiles of the NFX enantiomers in pregnant and nonpregnant sheep during and following an 8-day i.v. infusion of racemic FX. 95 o Z> < ~ X ;F CT Z i + X o < X o < a) x i E E c < X L L z CO X LL. z x LL z X LL or x LL X LL z ct X LL z X LL z a: DC = 1 " +i T - CO CM £ co CM o CO +1 CO CN CD LO *- +1 18 8 C N " +1 0 0 £ LO CM CD H LO c£) CO CO 2 +i CO •* , 4 +1 T ~ -Ct CN + | o co CD cri CN + | O CO +1 CD LO O o +1 o d LO V d LO d o o o d LO V o T — O o +1 o o CM d CO V d - r -d +i d CD d CM O d o +1 o o d co V d o d LO +i o d CO V CO d co „ > S . CD g ct ^ "co V> > 2 Q. CC 2 X co •a LL £ Z ' ro OT CL -ct LO CO CD o 00 CN O CD CM -ct CO CN d 1 ^ CO CD -ct CO ^t CD LO • - LO CD CM •- CM v--00 CT) CM LO CN LO CD a> CO -ct O O CD CM CM 00 CD CD LO LO LO -ct CN CM -ct CO o LU 0. -ct -ct -ct CD a> CO -ct CM ^ t CD CO •ct O CD a> CM 00 LO CM LO CM CM CO CM CN CD CM CO CD LO ^ t LO CO CD d d d d d d L O ' CD 00 ^t CM LO LO N - ^t CO CO CN CN csi CN CN 00 CM LO 1 ^ CN CM CO -ct O o -ct . — LO LO 1 ^ CD LO o CD CO CN 00 T— CD CM +1 CN o CD CM CO CD O CD CD CN CD . — LO T— LO CD LO CM CN +1 00 ^t CO CD 00 O CD •*t LO CD LO oS LO T -CO T— CN CO O CM • -+1 LO CO a> O CO CD LO CM CO d CO -Ct CD CD -ct CD +1 CN CM CO 00 CN od d oo CM CO CM -ct CO LO CM CO CO CD CD O -ct CO CO CO CD ^t CD CD CD +1 CM O t~ CO CN LO CO c CN CN ro LO I— -ct CP LU LU LU LU 0- Q_ D. Q_ -ct • -d +l o V d o ^t d CO +1 o d ^t V o • -d o +i o d CM V d CD o CM +1 •*t d CD d CO o d o +1 o CO d V d T— d LO +1 O d -ct V d X q LL Z*a) CO a: CO a: "> ? CO ct CM d CN CM d CO d o d LO o o o d o v S S X 2 S LU CO °"CL o. E . o > 0- LU -ct d C O d o d V CD s l O > 96 statistical analyses for the comparison of the pharmacokinetic parameters and their corresponding S/R ratios between the pregnant and nonpregnant sheep. As in the nonpregnant ewes similar relationships between the S- and R-NFX enantiomers were observed in pregnant sheep. Significant differences were observed in t|/2p, MRT, NFX/FX AUC metabolic ratio, combined FX+NFX AUC, and/, between R-NFX and S-NFX. AUC did not differ significantly between R-NFX and S-NFX in both pregnant and nonpregnant sheep. A comparison of the individual enantiomers between the two study groups (i.e. R-N F X ( p r e g n a n t ) versus R-NFX ( nonpregnant) ; S-NFX (pregnant) versus S-NFX ( n 0 npregnant) ) revealed significant increases in t|/2p and MRT, and tendencies for increases in NFX AUC and combined FX+NFX AUC in pregnant sheep compared to the nonpregnant animals. The ti/2p of both R- and S-NFX were significantly prolonged in pregnant sheep compared to nonpregnant ewes (R-NFX: 68.0 ± 18.0 h versus 26.0 ± 5.6 h,p<0.0005; S-NFX: 52.1 ± 23.3 h versus 21.4 ± 4.2 h, /?<0.001). Similarly, the MRTs of R- and S-NFX were significantly longer in pregnant compared to nonpregnant sheep (R-NFX: 160.0 ± 13.3 h versus 124.8 ± 5.4 h, /K0.0005; S-NFX: 125.0 ± 11.9 h versus 108.5 ± 3.6 h, p<0.01). No significant difference was found in the AUC of R- and S-NFX between pregnant and nonpregnant sheep, although the values tended to be higher in the pregnant animals (R-NFX: 17999 ± 5192 pg*h/l versus 13658 ± 4122 pg*h/l, p = 0.16; S-NFX: 165,02 ± 7047 pg*h/l versus 12265 ± 3 1 6 9 pg*h/l, p = 0.22). The NFX/FX AUC metabolic ratio did not differ significantly between the two groups (R-enantiomer: 2.35 ± 0.39 versus 2.26 ± 0.44, p = 0.74; S-enantiomer: 0.61 ± 0.18 versus 0.73 ± 0.26, p = 0.41). However, the combined AUC value of FX and NFX, which is a measurement of exposure to all pharmacologically active species, tended to be higher in pregnant compared to nonpregnant animals (R-NFX: 25629 ± 97 6775 pg*h/l versus 19743 ±5152 pg*h/l, p = 0.14; S-NFX: 45845 + 20645 pg*h/l versus 30321 ± 7206 pg*h/l, p - 0.12). Finally, no significant changes were observed in fu of the NFX enantiomers between pregnant and nonpregnant animals (R-NFX: 1.94 ± 0.48 % versus 2.02 ± 0.47 %,p = 0.80; S-NFX: 1.38 ± 0.41 % versus 1.14 ± 0.35 %,p = 0.32). Comparison of the S/R ratios of NFX pharmacokinetic parameters between pregnant and nonpregnant animals only revealed a small (10%) but significant difference in stereoselectivity of the MRT values. 4.3 Discussion As mentioned previously, data on the pharmacokinetics of FX and NFX in human pregnant subjects are extremely limited. Numerous clinical studies have been conducted to evaluate the safety of FX and NFX in human pregnancy but plasma drug concentrations were not monitored. The question of pregnancy-related physiological changes and their impact on drug disposition and pharmacotherapy has been neglected in the medical arena until recently, when Hostetter et al., 2000, reported that two thirds of 34 pregnant women taking fluoxetine, sertraline, or paroxetine, for the treatment of depression, required an increase in their daily dose of medication to maintain euthymia. A recent study by Heikkinen and co-workers, who longitudinally examined FX and NFX pharmacokinetics in pregnant & postpartum women, found an increase in FX concentration and a corresponding decrease in NFX concentration in the postnatal period (Heikkinen et al., 2003). Similar observations have been made in several pregnant subjects studied by our laboratory (Kim, 2000). These results strongly suggest that significant changes in the pharmacokinetics of these antidepressants must have occurred during pregnancy and highlight the need of more pharmacokinetic data in this population. 98 However, ethical and technical issues associated with in utero fetal blood sampling for detailed pharmacokinetic analysis presently precludes this. Animal experiment thus becomes an extremely valuable alternative to help answer some of these questions. Recently our lab has conducted a pharmacokinetic study of FX and NFX in pregnant sheep following a single i.v. bolus injection of racemic FX (Kim, 2000; Kim et al., 2004). In that study, five pregnant sheep at a gestational age of between 135-140 days and six nonpregnant ewes were given a 100 mg i.v. bolus dose of racemic FX and serial blood samples were collected to characterize the concentration-time profiles of FX and NFX and their respective enantiomers. It was reported that the apparent terminal half-lives of the FX enantiomers were significantly shorter in pregnant sheep (6.6 ± 2.5 h for R-FX and 6.7 ± 0.7 h for S-FX) than those in nonpregnant ewes (21.1 + 6.0 h for R-FX and 14.8 + 5.0 h for S-FX). Total body clearance values also appeared to be higher in pregnant sheep (3.1 ± 0.8 L/h/kg for R-FX and 1.9 ± 0.2 L/h/kg for S-FX) compared to those of the nonpregnant animals (1.97 ± 0.48 L/h/kg for R-FX and 1.44 ± 0.51 L/h/kg for S-FX). Similar findings of increased drug clearance have been reported for several compounds during human pregnancy. Citalopram, which is also a SSRI anti-depressant, has been shown to have lower trough steady-state plasma concentrations during pregnancy in humans. The mean desmethylcitalopram/citalopram and didesmethyl-citalopram/desmethylcitalopram metabolic ratios were significantly increased by 23% and 54%, respectively, during pregnancy compared to ratios determined 2 months after the delivery, suggesting increased metabolism during gestation (Heikkinen et al., 2002). During pregnancy, the oral clearance of metoprolol, a cardioselective Pi-adrenoceptor blocker, has also been reported to exceed that following pregnancy by a factor of 2 to 13, resulting in a lower oral bioavailability of the drug in pregnant women (Hogstedt et al., 99 1985). In theory, the increased drug clearance reported by these studies could be related to one or more of the following changes during pregnancy: 1) protein binding, 2) renal elimination and 3) hepatic metabolism. As in our current studies, the free fractions of metoprolol were not significantly different in pre- and post-partum patients, suggesting that protein binding is unlikely to be the causal factor. FX, citalopram, and metoprolol are extensively metabolized by the liver with negligible amounts of unchanged drug being excreted renally. It is unlikely, then, that the increased clearance of these drugs during pregnancy is the result of increased glomerular filtration. Rather it would appear that the increase in drug clearance observed in pregnancy is the result of increased hepatic metabolism. A point of interest among these compounds is that they are metabolized predominantly by the same enzyme, cytochrome P450 2D6. Expression of the CYP2D6 enzyme is mainly under genetic control, and enzyme-inducing drugs have only a minor influence on its enzyme activity (Eichelbaum et al., 1986). However, the metabolic activity of CYP2D6 has been shown to increase in human pregnancy (Wadelius et al., 1997). The induction of CYP2D6 in pregnancy could explain the increase in clearance observed in the fore mentioned studies. In contrast to the observations from the previous single-dose FX studies (Kim, 2000), the present long-term FX infusion experiments did not show any change in systemic clearance between pregnant and nonpregnant animals. A possible explanation for this is that increased CYP2D6 activity in pregnancy may be offset by the inhibitory effect of FX on this enzyme upon long-term drug administration. In addition, there is no change in the NFX/FX AUC metabolic ratio between pregnant and nonpregnant animals, which lends further support to this argument. In the previous chapter it was demonstrated that FX was capable of inhibiting its own metabolism upon long-term administration based on the 100 differences in clearance between the single i.v. bolus versus steady-state data. In fact, our results for the infusion studies (Table 4.1) indicated a slight, but not statistically significant decrease, in clearance compared with the nonpregnant values. Wadelius et al., 1997, made a somewhat similar observation previously in human subjects. In their study, a single oral dose of dextromethorphan was given to seventeen female volunteers during and after pregnancy for CYP2D6 phenotyping. The metabolic ratio of dextromethorphan versus its metabolite dextrorphan, which served as a measurement of functional CYP2D6 activity, was determined during and after pregnancy. The study found that extensive metabolizers (EM) of CYP2D6 exhibited decreased dextromethorphan/dextrorphan metabolic ratios during pregnancy, suggesting the induction of CYP2D6 in pregnant women. Interestingly, an unexpected increase in dextromethorphan/dextrorphan metabolic ratio in poor metabolizers (PM) was observed during pregnancy, implying that metabolism of dextromethorphan was somehow compromised in this group of subjects. Due to the small number of subjects in the PM group (n=4), the authors did not perform any statistics or provide any explanations for this latter observation. It is possible that the experimental design of their study was a factor. Upon closer examination of their study, it was found that the determination of the dextromethorphan/dextrorphan metabolic ratio was based on a single blood sample collected 2 hours post oral dosing regardless of the subject's pregnancy and CYP2D6 metabolic status. The 2-hour time point was an approximation of the time to reach peak plasma-concentration (Cmax) of dextromethorphan and dextrorphan in CYP2D6 EMs. However, this experimental design contains three potential problems. First, the time to reach C m a x (i.e. Tm ax) varies between EMs and PMs. Studies have shown that the T m a x of dextromethorphan and its metabolites in PMs occurs at approximately 5 hours (range 4-6 h) following an oral dose, 101 which represents a delay of more than 200% compared to EMs (Schadel et al., 1995). The determination of the dextromethorphan/dextrorphan metabolic ratio at a time other than the actual Cmax then, may lead to either an under- or over-estimation of the true metabolic ratio. Second, pregnancy is known to cause delays in gastric emptying. Studies with acetaminophen have shown that rate of absorption could be delayed by up to 50% in pregnant compared to nonpregnant women (Simpson et al., 1988; Levy et al., 1994). The change in .gastric empting in pregnancy could also potentially be a confounding factor that affects the determination of dextromethorphan/dextrorphan metabolic ratio. Third, the use of Cmax for the determination of metabolic ratio is prone to errors because C m a x is a poor indicator of systemic exposure. Based on one compartment model pharmacokinetics with first order input, the Cm ax of a drug can be described as follows (Gibaldi and Perrier, 1982): k F T ) = K a ^ ( - k , m a x _ - k a , „ m } ( 4 A } ™x Vd(k a -k) V ) K J Equation 4.1 shows that C m a x is dependent on D, the orally administered dose; k, the first order elimination rate constant; ka, the first order absorption rate constant; F, the fraction of the dose that is available to the systemic circulation and Vd, the volume of distribution. Among these parameters, ka is subject to a high degree of intra-subject variability, especially during pregnancy. Area under the curve (AUC), which is the standard index for exposure, should be used instead for a more accurate determination of metabolic ratio. Unlike Cm ax, AUC is mainly dependent on systemic clearance and relatively unaffected by the rate of absorption (equation 4.2): FD AUC = (4.2) CT TI3 102 Metabolic ratios calculated based on AUC are relatively resistant to changes due to gastric emptying delays in drug absorption. However, estimation of the complete AUC requires an intensive blood sampling schedule that may not be suitable in human pregnant subjects. Under such circumstances, the single sample method is preferable but the time of sampling becomes very critical. According to statistical moment theory the concentration at the mean residence time (MRT) should be the best predictor of AUC (Zimmerman and Slattery, 1983). Alternatively, cumulative urine collection is a relatively non-invasive approach of sampling that can be used for the determination of the dextromethorphan/dextrorphan metabolic ratio. There are studies to support the use of urinary metabolic ratio as an index of functional CYP2D6 activity in human (Nagai et al., 1996; Capon et al., 1996; Streetman et al., 1999). To determine whether the increased dextromethorphan/dextrorphan metabolic ratios in PM subjects reported by Wadelius et al., 1997, were truly due to the effect of pregnancy on drug metabolism rather than a flaw in experimental design, a simulation was performed. Concentration-time profiles for dextromethorphan and dextrorphan in nonpregnant and pregnant PMs were mimicked and the metabolic ratios calculated at different time points compared. According to literature data (Schadel et al., 1995; Capon et al., 1996), the pharmacokinetics of dextromethorphan and dextrorphan in PMs follow one compartment model pharmacokinetics therefore, their concentration-time relationships can be described by equations 4.3 and 4.4, respectively (Kochak and Rakhit, 1987): C = k a / „ ( 1 - E ) D ( e - k , , e - k a l ) ( 4 3 ) (k a -k)Vd 103 CL = ( k a - k m ) . 1 L ED VdL e-k,„. _e-k.t + k . k f / , ( l - E ) D Vd,„ ( k - k . ) ( k m - k . ) e J + 1 (k a -k)(k m -k) e-k,+ 1 ( k . - k j f k - k j (4.4) - k , „ t Where C m is the concentration of the metabolite in plasma; km is the total elimination rate constant of the metabolite;^ is the total fraction of the administered dose absorbed; ka is the first order absorption rate constant; k is the overall first order elimination rate constant; E is the plasma hepatic extraction ratio; D is the dose of the parent drug; and Vd m is the apparent volume of distribution of the metabolite. The simulated concentration-time profiles obtained using the pharmacokinetic parameter data obtained from the study by Schadel et al., 1995, are illustrated in Figure 4.9. Pharmacokinetic parameters derived from the simulated curves are summarized in Tables 4.3a and 4.3b. Three observations can be made from these simulations: 1) the metabolic ratio of dextromethorphan/dextrorphan is under-estimated by approximately 25% when it is calculated based on Cmax as opposed to AUC; 2) the metabolic ratios determined at 2 h and 5 h (i.e. T m a x ) differ by 5% only; 3) the metabolic ratio determined at 2 h post dose remains unchanged when ka is reduced by 50% to mimic the gastric emptying delay that occurs in pregnancy. Despite intra-subject variability in the dextromethorphan/dextrorphan metabolic ratio averaging 37% (Kashuba et al., 1998), it appears unlikely that the 63% increase in this ratio during pregnancy reported by Wadelius et al., 1997, was an artifact. As mentioned earlier, we observed a slight reduction in FX clearance in pregnant sheep from the long-term FX infusion experiments. In both scenarios, an apparent decrease in drug metabolism was observed and it appears that a common mechanism might be involved. Because of this, we are tempted to propose several hypotheses to explain these observations. First, we suspect that under conditions when 104 FIGURE 4.9 Simulations of the plasma concentration-time profiles for dextromethorphan and dextrorphan. Panel A illustrates the simulated profiles in poor metabolizer of CYP2D6 based on the literature data (Schadel et al.,1995). Metabolic ratios determined at 2 h and 5 h (-Tmax) are 4.06 and 4.08, respectively. Panel B illustrates the comparative profiles between pregnant and nonpregnant subjects. C m a x is delayed to 8 h when ka is reduced by 50% to mimic the gastric delaying effect in pregnancy. Predicted values for the metabolic ratios at 2 h and 8 h during pregnancy is 4.05 and 4.45, respectively. 105 3 CO CD ct £ E CO CL CD Q CN Q. > O c ca cz a) I CD X Q 1 3 co CD Ct "D JD ZD CO CL CD a CM CL >-o "c ro c D) CD Q. C o LD CJ) _r °-ro CD *S a5 P-"a >-ro o o CO X a a x Q O o ro co E fe ro 0-CD oo LO o 00 LO oo c\i 00 CM LO c\i 00 E O CD LO E I— LO O •r- ° 00 CD LO CM 00 o O LO O £ o LO CM 00 O LO eg co o o CM o T -o o CM CC CL CC LO LO 00 c o < T3 CD c o a, ji & o 43 X CD , T 3 ii ; X Q * C ° co u [ED o c 12 •« CD >-) ro ro Q "O a) E CO C o "D CD C O ro ro o .a ro 0 X Q c CO c cn CD c ro c ai CD i _ Q_ C o Z J O o CL CD O O LO T-O T-00 CM O 00 LO 0) CM TJ- S- O "fr CD CD CO o oo LO CO CM LO o O CM CM CO LO CM <cf CD cb ->- CM CO LO CD 00 £ CM 106 CYP2D6 activity is absent/reduced (e.g. in PMs) or inhibited (e.g. inhibition by FX), pregnancy paradoxically suppresses CYP2D6 activity (instead of inducing its activity as observed in EMs) or enhances the CYP2D6 inhibition imposed by the enzyme inhibitor. This hypothesis may explain the lowering of FX clearance observed in our pregnant sheep study after long term FX administration, but falls short of explaining the increased dextromethorphan/ dextrorphan metabolic ratio reported in human PMs. The pregnant PM subjects recruited in the study by Wadelius et al., 1997, possessed CYP2D6*3 or CYP2D6*4 mutant alleles, which are caused by frameshift mutations or altered splicing sites of the gene, respectively, resulting in completely non-functional enzymes (Rau et al., 2002). Other alleles such as *9, *10 and *41, with alterations in amino acid sequence, are associated with a reduction of in vivo enzymatic activity (Raimundo et al., 2000) but none of these were detected in these particular pregnant individuals. Obviously, the increased dextromethorphan/dextrorphan metabolic ratio cannot be attributed to the suppression of CYP2D6 activity because there were simply no functional enzymes available for suppression in these pregnant subjects. The second part of this hypothesis postulates suppression of other metabolizing enzymes during pregnancy. In addition to the formation dextrorphan, the metabolism of dextromethorphan also involves a parallel pathway to form 3-methoxymorphinan (Figure 4.10). The formation of 3-methoxymorphinan is mediated by CYP3A4 enzyme (Jacqz-Aigrain et al., 1993; Yu and Haining, 2001). Similarly, the biotransformation of FX to NFX involves CYP2D6, CYP2C9 and CYP3A4 (Figure 1.2 in Chapter 1) although CYP2D6 is considered to be the principal metabolic enzyme for this demethylation pathway. It seems that pregnancy might have a negative effect on the activity of CYP3A4, resulting in the increased dextromethorphan/dextrorphan metabolic ratio in PMs 107 CH 3 0. Dextromethorphan \ L J Q \ CYP3A4 CYP2D6 N H H NH 3 - methoxynx)rphinan CYP2D6 V H NH 3 - hydroxymorphinan N H H Dextrorphan CYP3A4 NCH, FIGURE 4.10 Metabolism of dextromethorphan by the CYP2D6 and CYP3A4 pathways to dextrorphan, 3-methoxymorphinan, and 3-hydroxymorphinan (adapted from Capon et al., 1996; Wadelius etal., 1997). 108 reported by Wadelius et al., 1997, as well as the decreased FX clearance observed in our long-term infusion experiments through an orthologous enzyme in sheep. The co-administration of oral contraceptives has resulted in reduced clearance of CYP3A4 substrates such as diazepam, alprazolam, and imipramine (Harris et al., 1995); hence the alterations in hormone levels that occur during pregnancy could have similar effects on CYP3A4. Wadelius et al., 1997, attempted to assess CYP3A4 activity by. examining the dextro-methorphan/3-methoxymorphinan metabolic ratio but were unsuccessful because plasma 3-methoxymorphinan levels were below their assay limit of quantification. Instead, they calculated the dextrorphan/3-hydroxymorphinan metabolic ratio as an indirect estimate of CYP3A4 activity but showed no significant change in this ratio during pregnancy in all genotypes. The validity of using dextrorphan/3-hydroxymorphinan metabolic ratio to monitor CYP3A4 activity is, however, questionable and this may also be the case for the dextromethorphan/3-methoxymorphinan metabolic ratio, which, in theory, is supposed to be a better indicator of CYP3A4 activity. Several studies have shown that this latter ratio exhibits large intra-individual variability and the inability to discriminate moderate CYP3A4 inhibition suggesting that dextromethorphan is a suboptimal probe for accurately assessing CYP3A4 activity (Min et al., 1999; Kashuba et al., 1999). Midazolam, which is a specific CYP3A4 marker, is an effective probe for in vivo phenotyping for CYP3A4 activity using a single plasma concentration (Lin et al., 2001). It exhibits minimal transfer across placenta to the. fetus and has been used clinically in pregnant women for nocturnal sedation prior to elective cesarean section (Kanto et al., 1983). On this basis, experiments that involve the administration of midazolam to women during and after pregnancy could potentially provide considerable information on the relationship between pregnancy and CYP3A4 activity. 109 Apart from modifications in drug metabolism, important distributional changes that occur in pregnancy may also have a significant impact on pharmacokinetics. The plasma free fraction of many drugs is altered during pregnancy (Reynolds and Knott, 1989; Loebstein et al., 1997; Little, 1999). Because FX is highly protein bound (>95%), it is anticipated that changes in plasma protein binding may lead to alteration in the unbound fraction and subsequently the pharmacological response. Albumin and ct|-acid glycoprotein are the two major plasma proteins for drug binding. In transgenic mice with elevated serum ot|-acid glycoprotein, there is a significant reduction in volume of distribution and elimination half-life of FX, suggesting that oti-acid glycoprotein plays a larger role in its disposition (Holladay el al., 1998). In humans, maternal total plasma protein concentration falls particularly during the first trimester. The change is primarily due to a fall in the concentration of albumin while oti-acid glycoprotein remains unchanged (Wood and Wood, 1981; Krauer et al., 1984). The unbound fractions of FX were previously determined in pregnant and nonpregnant sheep following a single i.v. bolus dose of FX (Kim et al., 2004). The results indicated that no change in the unbound fraction occurred during pregnancy in sheep. In the same study, the plasma unbound fraction of FX determined from women during pregnancy and postpartum also showed no changes in protein binding. In agreement with these previous reports, our long-term FX infusion study also revealed no significant difference in the free fraction and volume of distribution of FX between pregnant and nonpregnant animals. Although it is not known whether the concentration of CM-acid glycoprotein is altered during pregnancy in sheep, the lack of a change in free faction of FX suggests that the concentration of cci-acid glycoprotein remained unchanged. In general, unbound fraction normally correlates with the concentration of the binding protein so a 110 change in protein concentration usually affects the unbound fraction. However, this relationship is not always predictive. For example the free fraction of lidocaine, which is a basic drug that binds predominantly to oti-acid glycoprotein (~ 65% bound), is greater in -pregnant women at delivery compared to nonpregnant controls (48% versus 37%, respectively; Wood and Wood, 1981). The difference in unbound fractions could not be explained on the basis of the concentrations of plasma oti-acid glycoprotein as they did not differ between the pregnant and nonpregnant groups. In the same study, the use of oral contraceptives was associated with an increased unbound lidocaine fraction of 37%. This suggests that endogenous sex hormones might act as binding inhibitors and cause the displacement of lidocaine from its protein binding sites leading to an increased free fraction during pregnancy. In addition, other factors such as changes in free fatty acids levels are also known to affect drug binding. The levels of free fatty acids increase by 2-fold during pregnancy (Nau et al., 1984), and this could significantly influence the amount of protein-bound drug. Propranolol is another example that shows a discordant correlation between drug binding and protein concentration in plasma. It is a basic drug that is highly bound to proteins in plasma (80-95%), and the major binding protein at therapeutic propranolol concentrations is oti-acid glycoprotein. The plasma protein binding of propranolol is moderately decreased in pregnant patients (Wood and Wood, 1981), and at the time of delivery the unbound propranolol fraction was 20%. While this value is significantly different from the observed value of 15% in control patients, the plasma concentration of ox-acid glycoprotein did not differ between the pregnant or control groups. However, it is important to be aware of the fact that a drug typically binds predominantly but not exclusively to a particular protein. Albumin exhibits low affinity but high capacity binding 111 for basic drugs like propranolol. The level of albumin falls during pregnancy and may then contribute, at least in part, to the change in free propranolol fraction. The negative correlation of the propranolol free fraction with albumin in fetal serum also appears to support this assumption (Krauer et al., 1986). It is important to point out that the data presented for FX as well as those cited from the literature were derived from analyses of unrelated groups of pregnant and nonpregnant animals or human subjects. Although this may give an approximate picture, it must be realized that large interindividual differences make the comparison of disparate pregnant-nonpregnant groups subject to considerable error and that meaningful information can only be gained from a crossover study using the same animal or subject. Moreover, most studies comparing protein binding between pregnant and nonpregnant state (either in humans or animals) have been conducted during the late period of gestation. Haram et al., 1983, reported a statistically significant drop in cti-acid glycoprotein concentration from 0.72 g/L at 8-11 weeks of pregnancy to 0.55 g/L at 16-27 weeks followed by a return to concentrations of 0.65 g/L, which are slightly lower than the 8-11 week values, by the end of the 37th gestational week. The fluctuation of ct|-acid glycoprotein concentration throughout pregnancy strongly suggests that gestational age differences should also be taken into account when comparing pregnant and nonpregnant values. While no apparent differences in the pharmacokinetics of FX were found between pregnant and nonpregnant animals, the disposition of NFX was subject to a few significant changes. Both Unp, and MRT of NFX were significantly prolonged in the pregnant animals. Theoretically, alterations in clearance and/or volume of distribution of NFX could have contributed to these changes. A change in clearance, however, is unlikely based on the lack 112 of significant difference in NFX concentrations between pregnant and nonpregnant animals. Alternatively, an increase in volume of distribution is more likely to happen in pregnancy inasmuch as the fetus could act as a deeper tissue compartment for drug distribution. Upon discontinuation of drug infusion, anything that transferred across the placenta to the fetus will slowly diffuse back to the maternal circulation for elimination, assuming that the fetus lacks the capacity for drug elimination. If this redistribution process is slower than the maternal intrinsic rate of elimination, it will become a rate-limiting step for the overall drug elimination from the mother and manifests as a prolongation of Uap and MRT. Nevertheless, these changes in t)/2p and MRT did not result in a significant increase in AUC of NFX. The combined FX plus NFX AUC value, which is the sum of exposure of all pharmacological active species, appeared to be higher in pregnant compared to nonpregnant animals, but the difference was not large enough to reach statistical significance. Overall, our sheep data suggest that pregnancy does not pose a significant risk of increasing the exposure of the pharmacological active species to both the mother and the fetus. However, the human data reveal lower combined FX and NFX concentrations during pregnancy (Heikkinen et al., 2003). The same has been reported for citalopram, which exhibits low maternal trough plasma concentrations during pregnancy (Heikkinen et al., 2002). It is not clear if these low blood levels during pregnancy could lead to. therapeutic failure, but clinicians should be aware of this possibility so that depression in pregnancy is not undertreated. The effect of pregnancy with respect to the stereoselectivity of drug disposition is poorly understood. While gender differences may alter the stereoselective pharmacokinetics of drugs (Gilmore et al., 1992; Sasaki et al., 1993; Gupta et al., 1995; Hooper and Qing, 1990), it is anticipated that the unique physiological changes that occur in pregnancy could 113 influence the stereoselective disposition of enantiomers in their own right. Generally speaking, the mechanisms of stereoselective drug disposition mainly involve, but are not limited to, stereoselective protein binding, stereoselective metabolism, and stereoselective renal elimination. Enantiomers usually bind to the same class of plasma protein but the binding affinity may differ between enantiomers. hi pregnancy, increased total plasma lipids and free fatty acids, altered concentrations of sex hormones and other endogenous substances may potentially influence drug binding. Displacement of drug from its plasma protein binding site(s) may occur when the concentration and affinity of the endogenous substance exceed those of the drug. Displacement may also have differential effects on enantiomers that exhibit different binding affinities to plasma protein, resulting in greater displacement for the enantiomer that has lower binding affinity thereby changing the stereoselectivity (or stereoselective index) of protein binding. For other racemic drugs that exhibit stereoselective pharmacokinetics because of stereoselective metabolism, the impact of pregnancy on certain metabolic enzymes has the potential to alter the metabolism of each enantiomer in a discriminating manner. An example of such interaction is the induction of CYP2D6 in pregnancy (Wadelius et al., 1997). In this example, the racemic drug is eliminated primarily by metabolism with one enantiomer being predominantly metabolized by CYP2D6 whereas a different enzyme exclusively metabolizes the other enantiomer. As a result the clearance of one enantiomer will be higher during pregnancy than that of its counterpart thus changing the AUC ratio of the enantiomers. Compared to hepatic metabolism, fewer drugs have been reported to exhibit stereoselective renal elimination (i.e. active transport and/or renal metabolism) (Jamali et al., 1989; Tracy, 1995). Although glomerular filtration rate (GFR) increases by approximately 114 60-80% during pregnancy, it is believed this has no impact on stereoselective drug elimination as glomerular filtration is a passive process, which depends on the physiochemical property of the drug (e.g. molecular weight, pKa, partition coefficient, etc.), and for enantiomers these are identical. Other than GFR, renal secretion and reabsorption increase modestly by approximately 20% during pregnancy (Little, 1999). If we consider the situation where a racemic drug is mainly excreted by the kidney and exhibits stereoselective elimination and further that one enantiomer has a lower C L r e n a | due to preferential tubular reabsorption then, the clearance of that particular enantiomer will be expected to be reduced during pregnancy. This would occur as a result of more of it is being reabsorbed relative to its antipode. This phenomenon does not, however, appear to have been investigated. The present study is the first to examine the effect of pregnancy on the stereoselective pharmacokinetics of FX and NFX. Our results showed a progressive increase of S/R ratio with respect to time (Figure 4.5). A positive correlation was also observed between S/R C s s ratio and total FX concentration (Figure 4.6). The mean FX steady-state level in pregnant animals was slightly higher than that of nonpregnant animals resulting in a higher S/R ratio in the former group. The difference in S/R ratio between pregnant and nonpregnant animal is unlikely to be the result of CYP2D6 induction during pregnancy as we have proposed earlier in the discussion that any increase CYP2D6 activity was likely to be abolished by the CYP2D6 inhibitory effect of FX. When the S/R C s s ratio was plotted against the C L T B a negative correlation resulted (/?<0.01, Spearman rank test; Figure 4.11). This finding suggests that reduced metabolism (probably through the inhibition of CYP2D6) plays a more important role in shifting the S/R ratios. For protein binding, no difference in the S/Rfu ratios of FX and NFX enantiomers was found between pregnant and nonpregnant sheep. 115 FIGURE 4.11 Plot of S/R F X C S S ratio versus the corresponding total body clearance value (CLTB) of F X enantiomers (data from the nonpregnant (closed symbol) and pregnant sheep (open symbol) experiments). 116 Similar findings were reported for other oti-acid glycoprotein bound drugs such as propranolol and verapamil. Walle and colleagues reported an S/R unbound fraction ratio of 0.86 for propranolol in serum of healthy human subjects, which was similar to the value of 0.91 determined by Belpaire and colleagues in maternal serum (Walle et al., 1983; Belpaire et al., 1995). Similarly, S/R unbound fraction ratio of verapamil calculated based on the data reported by Gross and colleagues was 1.42 in the serum of human volunteers, which was similar to the value of 1.36 determined by Belpaire and colleagues in maternal serum (Gross et al., 1988; Belpaire et al., 1995). Based on the above findings, it appears that pregnancy has no significant impact on the S/R unbound fraction ratio for racemic drugs that bind predominantly to ai-acid glycoprotein. The effect of pregnancy with respect to stereoselective renal elimination could not be explored in this study because maternal urine was not collected. Nevertheless, both FX and NFX were extensively metabolized and only minute amounts of the unchanged moieties (<1%) were excreted renally, so any changes occurring in this pathway during pregnancy should only affect their disposition to a very limited extent. However, for drugs that are predominantly eliminated by the kidney, pregnancy-related changes in renal function may be of great significance. The stereoselective pharmacokinetics of pindolol have been studied in pregnant women with hypertension in a recent report (Goncalves et al., 2002). Pindolol is marketed as a racemic mixture of the S-(-) and R-(+) enantiomers with the S-enantiomer being a more potent P-adrenoceptor blocking agent than the R-enantiomer. It is cleared approximately equally by both metabolism and renal excretion. Pindolol undergoes active secretion in the renal tubules with its CL ,enai exceeding the glomerular filtration rate. Goncalves et al., 2002, reported that the renal elimination of pindolol is stereoselective in hypertensive pregnant women and 117 similar observations have also been reported in nonpregnant healthy (normotensive) subjects (Hsyu and Giacomini, 1985; Somogyi et al., 1992). The authors compared the pharmacokinetics of racemic pindolol between pregnant and nonpregnant subjects, but deliberately failed to examine any change in stereoselectivity during pregnancy, as apparently this was not within the scope of their study objectives. To facilitate the investigation of the effect of pregnancy on stereoselective renal elimination, we have combined the data from Goncalves et al., 2002, and Somogyi et al., 1992, for further analysis. S/R ratios of the reported pharmacokinetic parameters were calculated and the results are summarized in Table 4.4. The results indicate that the CL r c nai of pindolol enantiomers and their corresponding S/R ratios were slightly lower in pregnant woman compared to nonpregnant subjects, suggesting that tubular reabsorption might have increased and that this process was more efficient for the S-enantiomer during pregnancy. Although these values have not been weight normalized, the reported ranges of body weight in the nonpregnant and pregnant groups were 52-78 kg and 68-97 kg, respectively (Somogyi et al., 1992; Goncalves et al., 2002), indicating that the difference in C L r e n a i is not due to a higher body weight in the nonpregnant subjects. Apart from renal clearance, we observed an interesting change in the direction of the S/R ratios for other parameters such as AUC and apparent total body clearance (CL/F). The S/R AUC ratio increased from 0.86 to 1.13 in nonpregnant and pregnant subjects, respectively, while the S/R CL/F ratio decreased from 1.17 in nonpregnant subjects to 0.89 in pregnant women. The authors commented that the increased CL/F in pregnant subjects was likely due to the induction of enzymes involved in the metabolism of pindolol. Further, our preliminary analysis shows a preferential increase in CL/F with respect to R-(+) pindolol compared to S-(-) pindolol, indicating a pregnancy-118 o CD It? =5 in C co "o CD S" CO ro CD X c CO c CD CD L _ Q. C o c CD c ro c CD CD > CO c _CD k_ CD Q. >. I ro or CO + ct CO o ro or CO + I CC co o "5 . .E CD o E ro ro E s ro o-.c Q_ CD i - •<- |s-00 CO CM 1 -CM 05 o CM 1^-00 oo co CD iri cn o CM CM 00 CO CD LO 00 E ^ D) o < cn CM LO 00 CM LO CO 1^-d oo CO 00 LO ro oo CM CD i r i LO CD < CJ LO d CM LO CM CM OO 00 T— co T t 00 CO 00 1-^ d d CM CO 00 LO oo CO O o C N C o O v i cd , — , T3 CN A — J-I r £ 13 6 0 o a o t/3 C N > CO T3 CU O p E _o c Cd p >1 T3 cu •a '> •3 <u 3 s -a cu CU > o cu cu •co-ed O T3 <4-l O O -a eo P oo c p C J p <; - o cu C .a •a cu CU > o o cu u p < -D •a cu •g '> -3 cu c 3 o P E -f 6 0 c cu 4 3 v i cd -O cu c5 •S 3 6 0 ^ c g E p ca c c n -a J0 cu T 3 CU 1—1 U *o p CO PH u c cu p 3 O E ca w i ca T3 a> ta o 13 o p p 1 „ a as J CU J o < u cu s 4= -a cu 43 p C J 119 induced change in stereoselective pharmacokinetics. The finding of a higher non-renal clearance value (CL n 0n-renai/F) for R-(+) pindolol in pregnant subjects suggests the involvement of stereoselective metabolism, which was overlooked by other reports on pindolol pharmacokinetics. 4.4 Summary In summary, this study has demonstrated that FX and NFX exhibit stereoselective pharmacokinetics in pregnant sheep. Increased clearance of FX was observed in the previous single i.v. bolus experiments in pregnant sheep but this effect was absent upon long-term FX infusion. The potent enzyme inhibitory effects of FX and NFX might have abolished any increased enzyme activity during pregnancy under steady-state conditions. The findings of decreased CLTB of FX in pregnant sheep at steady-state together with the increased dextromethorphan/dextrorphan metabolic ratio in CYP2D6 PMs during pregnancy suggests a reduction of CYP3A4 activity during pregnancy. Protein binding of FX and NFX did not differ between pregnant and nonpregnant sheep. This finding is consistent with the lack of significant change in oti-acid glycoprotein concentration during pregnancy. However, the relationship between unbound fraction and concentration of the binding protein does not always correlate, as exemplified by the alteration of binding of other a,-acid glycoprotein bound drugs like lidocaine and propranolol during pregnancy. The theoretical basis of interactions between pregnancy and stereoselective drug disposition has been discussed. Although no remarkable change in the S/R ratio of FX was found in our experiments, a review of the literature for other racemic drugs shows that pregnancy does have the potential to alter stereoselective drug disposition. The clinical significance of such phenomena, however, remains to be elucidated. 120 CHAPTER 5 STUDY OF PLACENTAL TRANSFER OF FX AND NFX FOLLOWING MATERNAL INTRAVENOUS INFUSION OF RACEMIC FX TO STEADY-STATE Preface The placental transfer of FX and FX in humans has been reported, but detailed pharmacokinetic studies are lacking. Studies of placental drug transfer and fetal pharmaco-kinetics cannot be readily undertaken in humans, in whom observations have been largely confined to the comparison of single point drug concentrations in maternal and umbilical cord plasma at birth. Interpretation of such data is difficult because the single point fetal-to-maternal concentration ratio is highly dependent on the time of sampling relative to the time of maternal dosing. Moreover, such observations contribute little to the understanding of placental transfer, or of fetal drug distribution and elimination. The pregnant sheep preparation is not limited in this way, and the increased flexibility of experimental design allows much greater insight into the fetal handling of drugs. The present study is an extension of the maternal infusion study in which fetal arterial blood, amniotic and fetal tracheal fluids were collected and analyzed for FX and NFX concentrations. The fetal exposure index (i.e. fetal-to-maternal AUC ratio) was calculated and components of fetal nonplacental drug elimination were explored. 121 5.1 Methods 5.1.1 Animal and Surgical Preparation The experimental setup is the same as the pregnant sheep experiments in the previous chapter. Five pregnant ewes (Dorset and Suffolk breeds) at gestational ages ranging from 123 to 125 days and term -145 days were used in this experiment. The study was approved by the University of British Columbia Animal Care Committee, and the procedures performed on the sheep conformed to the guidelines of the Canadian Council on Animal Care. The surgical procedure was similar to that described previously in Chapter 3, Section 3.11. Addition details for the surgical procedures on the fetus have been described previously (Morrison et al., 2001). Briefly, a midline abdominal incision was made in the ewe after shaving, cleansing and disinfecting the abdomen, flank and groin areas with 10% povidone-iodine solution. The uterus was exposed and access to the fetus gained by an incision in an area free of placental cotyledons and major blood vessels. Polyvinyl catheters were implanted in the fetal femoral arteries and lateral tarsal veins for drug sampling and administration, respectively. Another two catheters were placed in the amniotic cavity and anchored to the skin on the neck and the abdomen of the fetus. A non-occlusive catheter was inserted into the trachea via a separate uterine incision. Electrodes of Teflon-coated stainless steel wire were implanted on the dura of the parietal cortex for the measurement of the electroencephalograph and on the zygomatic bone of each eye for electroocular recordings. A Transonic blood flow transducer was placed around the main uterine artery for measurement of uterine arterial blood flow. The hysterotomy and laparotomy incisions were then closed. All catheters and electrodes were tunneled subcutaneously and exited through a small incision on the side of the maternal flank. The catheters were stored in a plastic pouch 122 attached to the back of the ewe when not in use. After surgery, 500 mg of ampicillin was administered intramuscularly to the ewe for 3 days, and via the tarsal vein to the fetus on the day of surgery, and to the amniotic cavity for the duration of the preparation. The ewe was kept in a holding pen with other sheep and allowed free access to food and water. A minimum of 3 days were allowed for the ewes to recover before experimentation. 5.1.2 Experimental Protocol The experimental protocol was the same as that used for the pregnant sheep experiments described in Chapter 4. Figure 3.1 in Chapter 3 is a schematic illustration of the general study protocol that was also used for the present experiment. Each animal received a 70 mg i.v. bolus loading dose over one minute followed by a continuous i.v. infusion of racemic FX via the maternal femoral vein at a rate of 6.92 mg/h. The drug infusion was continued for 8 days to achieve steady-state concentrations. Serial blood samples from the fetal femoral arterial (5 ml) were collected at 0, 5, 15, 30 min, 1, 2, 4, 6, 12, 24, 36, 48, 60, 72, 84, 96, 108, 120, 132, 144, 156, 168, 180 h, 192 h during the infusion, and at 5, 10, 20, 30, 45 min, 1, 2, 3, 4, 6, 9, 12, 24, 36, 48, 60, and 72 h post-infusion. Amniotic and fetal tracheal fluid samples (1 ml each) were also collected at 0, 12, 24, 36, 48, 60, 72, 84, 96, 108, 120, 132, 144, 156, 168, 180 h, 192 h during the infusion, and at 1, 2, 3, 4, 6, 9, 12, 24, 36, 48, 60, and 72 h post-infusion. All zero time samples were collected ~10 minutes prior to loading dose administration. During the experiment, the volume of fetal blood collected was replaced at intervals during the study by an equal volume of drug-free maternal blood. All blood samples collected for drug and metabolite analysis were placed into K 3 E D T A Vacutainer® tubes, gently mixed, and centrifuged at 2000g for 10 min to separate 123 the plasma. The plasma supernatant was transferred into a clean borosilicate test tube with polytetrafluoroethylene (PTFE)-lined cap. Amniotic and fetal tracheal fluid samples were transferred to borosilicate test tubes immediately upon collection. All samples were frozen at -20°C until the time of analysis. At each blood collection, an aliquot of whole blood sample (~1 ml) was collected for blood gas, acid-base, and glucose/lactate analyses using the instrumentation outlined in Chapter 2, Section 2.2.2. In addition, the maternal and fetal arterial pressure and heart rate, uterine blood flow, fetal electrocorticographic activity (ECoG), rapid eye movements (REM), amniotic pressure, and fetal tracheal pressure [to monitor fetal breathing movements (FBM)] were continuously monitored with a Grass K2G polygraph (Astro-Med, West Warwick, RI) and digitally stored using a computerized data acquisition system (Chart® v4.2, AD Instruments, Grand Junction, CO). The values from the physiologic and behavioral state measurements have been reported separately (Morrison et al., 2001; Morrison et al., 2002; Morrison et al., 2004). 5.1.3 Pharmacokinetic Analysis All pharmacokinetic parameters were calculated by standard pharmacokinetic procedures as described in Chapter 3, Section 3.1.4. In addition, the MRT of FX and NFX in the fetal circulation following maternal FX administration was calculated as the ratio of AUMCfetai to AUCfetai minus the MRT m a i e m ai value (Kim 2000). The fetal-to-maternal exposure ratio (F/M ratio), which is also known as the index of fetal exposure, was calculated as AUCfetai divided by AUCmatemai- The combined fetal-to-maternal exposure ratio (Combined F/M ratio), which is a measurement of the exposure to the pharmacologically 124 active species, was calculated as the sum of AUCfetai for FX and NFX divided by the sum of the AUCmatemai for FX and NFX. 5.1.4 Statistical Analysis A paired t-test was used for statistical analysis of the pharmacokinetic parameters when comparing R- versus S-enantiomers. The Spearman rank correlation was used to test the correlation between two variables. The achievement of steady-state for plasma FX and NFX concentrations was evaluated by testing whether the slope of plasma concentration versus time curve was significantly different from the value of zero. The significance level used was p<0.05 in all cases. Unless otherwise specified, all data are reported as the mean ± standard deviation (S.D.). 5.2 Results 5.2.1 FX and NFX Levels in Fetal Arterial Plasma The average plasma concentration versus time profiles of total FX and NFX are shown in Figure 5.1. There was rapid transfer of FX from the mother to the fetus across the placenta with FX detected in fetal plasma as early as 5 minutes following the initiation of maternal dosing. During the initial infusion period, there was a decline in the FX level in the fetus in the first 12 hours. After that, FX concentrations started to increase gradually over the next 5 days and reached steady state during day-6 to 8 of the infusion period. In contrast to FX, there was a delay in the appearance of NFX in fetal plasma. Plasma levels of NFX were below the limit of quantitation in the first 2 hours following the initiation of maternal dosing. After that, NFX began to increase slowly and continued to accumulate in the fetal circulation 125 1000 q 1 -I , , , , . r , , , , , 0 2 4 6 8 10 12 Time (day) FIGURE 5.1 Average plasma concentration versus time profiles of total FX and NFX in fetal sheep (n=5) following an 8-day maternal i.v. infusion of. racemic FX, showing concentration differences between FX and NFX from day-1 to 8 of infusion period. 0.1 -I . , . 1 . , . , 1 • 1 0 2 4 6 8 10 12 Time (day) FIGURE 5.2 Representative plasma concentration versus time profiles of FX and NFX enantiomers in a fetus (E4237) following an 8-day maternal i.v. infusion of racemic FX, showing stereoselective disposition of FX enantiomers in the fetus. 126 over the entire infusion period without reaching steady-state. Upon termination of the maternal drug infusion, the concentration of FX persisted for approximately one hour and then began to decline whereas the concentration of NFX remained steady for up to 9 hours post-infusion before any signs of elimination occurred. Figure 5.2 depicts representative plasma concentration versus time profiles of FX and NFX enantiomers in a fetal lamb, showing stereoselective pharmacokinetics of FX enantiomers in the fetus. The stereoselective relationship between enantiomers of FX and NFX with respect to their steady-state concentrations in the fetus was similar to that in adult sheep (i.e. R-FX < S-FX and R-NFX = S-NFX). The pharmacokinetic parameters for total FX, R-FX and S-FX are summarized in Table 5.1 while those for total NFX, R-NFX and S-NFX are summarized in Table 5.2. The elimination of FX appeared to be slower in the fetus compared to the mother (maternal t, /2p = 31.9 h for R-FX and 35.9 h for S-FX, maternal MRT = 40.2 h for R-FX and 57.8 for S-FX; Table 4.1 in Chapter 4). The elimination t| / 2p for total FX, R-FX and S-FX was 35.8 ± 22.9 h, 35.0 ± 23.5 h, and 36.7 ± 23.3 h, respectively. MRTs for total FX, R-FX and S-FX were 100.2 ± 11.6 h, 103.6 ± 13.5 h, and 99.2 ± 12.1 h, respectively. No differences in ti/2p and MRT were observed between R-FX and S-FX. C s s of R-FX was significantly lower than that of S-FX (14.7 ± 6.5 ng/ml vs. 61.4 ± 33.4 ng/ml, respectively; /?<0.05) with a mean S/R ratio of 4.09 ± 0.85). Similarly, R-FX had a lower AUC compared to S-FX (3059 ± 1248 pg*h/l vs. 12497 + 7107 pg*h/l, respectively; p<0.05), and the mean S/R AUC ratio was 3.83 ± 0.90. Figure 5.3 shows the comparative average plasma concentration versus time profiles of total FX in the mother and the fetus. The fetal concentration of total FX was consistently lower than the maternal concentration throughout the entire infusion period with an average F/M ratio of 0.40 ± 0.08. The combined F/M ratio 127 o ZD < LL g "§ s c !Q E o O O < o < LT + X E CT) c n C O CT) C O C O C O C N d d d d CT> d LU CL IG C N CO 00 CO L D ™ L O P • C D O C N O C O CM L O C O CM CM 1^  L O L U UJ L U LU CL CL CL 0. +1 C D d CO "3-CO 00 +1 CD L O L O L O +1 C N d L O C N 00 C N C O C O d L O +1 L O CD CO 00 OL X LL OT X LL tr x LL OT X LL tr x LL OT X LL tr x LL X LL tr CD E CD CM CD T - CT) T r Tf CO CO CO CO d d d d d d C O T - o C O CM CM C O C O C M C O d d d d d d r ~ CD CT) CO L U Q_ CN o CM C O C D Tf d C N d CJ) Tt" T r X CO L O co LL CM OT o CT> o X ^ Tf LL CO d tr o O) ^ 00 o CM C N T— L O CO -a- Tf CO Tf d d d d d d CO CT) CM o CT) LO C O CO -a- C O C O d d d d d d CN ^ +1 CO C N O T— CD CT) L O N - L O T r L O CO O CO O Tf C N C N Tf CO TT CO CT) r _ r - r- L O d C O d Tf' CO CM +1 CT) CT) CD *n CO CO O +1 L O CM CD 1^  L O L O C O co L O CD CO L O C O CO T r CT) 00 L O T - o CD C N 00 L O L O L O C D T - C N CO CM o r -C O CM L O C O c C N CM CO L O r ~ T T CD LU L U L U L U CL CL CL D_ o d CN d CO > CD tr (0 t o > B -~ X T3 LL CD OT (0 CL 128 o z> < X LL X LL z o Z> < CJ < 3 o o i -JO E CD co CM CO co CD CD CM CO d d d d d o CM 00 00 1^ CM ° > ™ CD CO CO CO CM CM -ct CO CO CM CO d d d d d d CO LO -ct CM 00 o. + | oo LO o CD o -ct CM CO •"ct ^t CO 00 CO LO •<*• LO CO CM 1^ o o CO CM LO CO CM CM LO h~ LLI LU LU LU LU _CD CO ^ +1 X LL z ct CO X LL z ct X LL z CO X LL z co x LL z ct X LL z CO X LL z ci E c < -ct CD CO oo o CD CM LO CD T t CO LO d d d d d d CD ^ ^ CO CM , _ CM CO CM o 1^ o CO CM LO CO CM CM r~- LO r~- -ct LU LU L U LU LU 00 CM CD CD LO T— 00 -ct CD CM 00 d CM CM CD CM 1^ - -ct o CD CM LO co CO CO CO d d d d d d 00 CO o CD -ct CO CO CM CM d d d d d d -ct CD CM CO CD o 1^ -ct CO CM CD O CD oo -ct CD ^ t 1^ ^ t CD CO LO 00 +1 CD CM CD CD LO 00 CD O CO -ct 00 ^ct CO CD CO CD CO CO LO £2 co +i CD CO CO LO CO LO cb co CO CM •*t CO CM CD LO q CO -ct CM CD -ct CD ^t d LO ^ t in O CO ^ +1 o +] C O d o +i o +1 o +1 CD C N d +1 o TO i _ CC CO 129 1000 co c c o c Ci u c o o 100 10 F X - M o t h e r -e- F X - F e t u s 10 12 Time (day) FIGURE 5.3 Average plasma concentration versus time profiles of total FX in the mother and fetus (n = 5) following an 8-day maternal i.v. infusion of racemic FX, with an average F/M AUC ratio of 0.40 ± 0.08. 10 12 Time (day) FIGURE 5.4 Average plasma concentration versus time profiles of total NFX in the mother and fetus (n = 5) following an 8-day maternal i.v. infusion of racemic FX, showing accumulation of NFX in the fetus with an average F/M AUC ratio of 0.33 ± 0.11. 130 had a similar value of 0.36 + 0.05. Both the F/M ratio and combined F/M. ratio did not differ significantly between R-FX and S-FX. For NFX, elimination seemed to be much slower in the fetus compared to the mother (see Table 4.2 in Chapter 4 for maternal data). Elimination t,/2p's for total NFX, R- NFX and S-NFX were 53.4 ± 21.0 h, 54.4 ± 20.9 h, and 52.5 ± 22.6 h, respectively. MRTs for total NFX, R-NFX and S-NFX were 94.8 ± 27.2 h, 94.5 ± 19.8 h, and 93.0 ± 30.1 h, respectively. MRT and t|/2p values for R- and S-NFX were not significantly different (p = 0.84). R- and S-NFX were equally distributed in the fetal plasma, with respective AUC values of 5148 + 1975 ug*h/l and 5927 ± 2548 |ag*h/l and an S/R ratio of 1.15 ± 0.18 (p = 0.13). Figure 5.4 shows the comparative average plasma concentration versus time profiles of total NFX in the mother and the fetus. Total NFX concentration was lower in the fetus compared to the- mother with a mean F/M ratio of 0.32 ± 0.09. Interestingly, the F/M ratio for R-NFX was significantly lower than that of S-NFX (0.29 ±0.09 versus 0.36 ± 0.09; p<0.0\) with an S/R ratio of 1.28 ± 0.18, suggesting a stereoselective transfer process in the maternal-fetal unit. Fetal exposure to total NFX was slightly lower compared to total FX, with a mean total NFX/FX AUC ratio of 0.83 ± 0.32. However, this ratio was significantly greater for R-NFX compared to S-NFX with respective values of 1.81 ± 0.60, and 0.56 ± 0.21 (p<0.0\), resulting in a mean S/R ratio of 0.31 ±0.07. 5.2.2 FX and NFX Levels in Amniotic and Fetal Tracheal Fluids FX and NFX were also detected in amniotic (AMN) and fetal tracheal (TR) fluids. The pharmacokinetic parameters for both compounds in these fluids are summarized in Table 5.3. Figures 5.5 and 5.6 illustrate the average concentration-time profiles of total FX and NFX in amniotic fluid, fetal tracheal fluid and fetal arterial (FA) plasma, respectively. 131 6 o ZD < < LL EE ci o ZD < CC o ZD < < o ZD < X CO , - O CO o , -LL O CM LO CO CD CD CO "- * - d d d d X CD -ct CO O CM 1^ LL CD CM CM LO CO cp tk "- CM "- "- * -X LL X X LL tk X CD CM CD CD ^ t LL CD CM CM LO "Ct CO d d d d d d X LL CO X LL or CD JO E Z ro E c < CD o CO CD CM 1^ CM CO CD q o d d CO 1^ CO CD CM CD LO O -ct CM -ct -ct o CO LO CD CD CO CO CD CM LO co CO CO eo CM o o CO CM LO CO CM CM 1^ LO t^- ^ t LU UJ LU LU LU CD •ct 1^ 00 O LO CO CO CD CD O LO CO CM CM -ct CN CD CM CD £ O LO 00 +1 X CO CM 00 "Ct ^ t LL CO T— LO 00 CD tk "- " - d d d d o CD LO I - CD T t CO LO CO CO CD LO d d d d d d o CD co O CD o CD CM LO CO LO LO ^ t LO CM ^ t +1 CO CD 1^ LO CM T— 00 O CD LO T— CM CO CD CO "Ct CD LO CN CM "— CM CD CO CO CN 00 O LO •St o CD O CO CD CO "Ct ro° 0 CO s +1 o d o V o d CO > i CO CC 3 ^ ro co > & Q. - X •O LL CO ' .!= CO ro X LL z CO x LL z CC X LL z tk 3 £ P X LL X LL z tk x LL X LL Z tk 3 £ P -7 CO JO E ID Z ro E c < CD LO LO •ct r— T t T t "- CD q CD q "^ "^ d d CD CO CO CD CD "Ct CO CM o o CO CM LO CO CN CN LO "Ct LU LU LU LU UJ CD CO CD CD CO LO CD LO 00 CM "- CO "-^  d "- "— o LO CO o CD CO CO T— LO CO "Ct CO 1^ 00 LO LO CM CD CN -ct LO CO CO CD CO LO 1^ o CD LO CO CN CO LO CO LO CD 00 00 CO CO "Ct CD CO CN CO CM CO 1^ o CO LO ^t CM CO CD d d d d d d CO CM 00 o o q CD CO LO CD d d d d d CO CD CO CD CD LO CM "Ct LO d d d d d d o CD CD CD CN CD T— CM O CD CO o o O T t CO CO CM CM CM "— "-^t CD 00 CM CD CO CO T— CD "Ct CD CM •ct CD T— 1^ O CD CO CM "— CM CO CM +1 LO CD T- CO ^ t 1^ CM CD LO CD O LO CD CD LO LO •ct CM LO -ct <D CO 2 +1 o +1 CM O +1 CD d o +1 d or CO o d CC m t o I CO Q. 7. X CO Z •s OT CL 1000 E "5) c o c cu u c o o 100 6 Time (day) 10 12 FIGURE 5.5 Average amniotic fluid (AMN), fetal tracheal fluid (TR) and fetal arterial plasma (FA) concentration versus time profiles of total FX in the fetus following an 8-day maternal i.v. infusion of racemic FX, showing comparable concentrations of FX in TR and FA but lower in AMN. 1000 n 1 -I , , , , , , , , , , , 0 2 4 6 8 10 12 Time (day) FIGURE 5.6 Average amniotic fluid (AMN), fetal tracheal fluid (TR) and fetal arterial plasma (FA) concentration versus time profiles of total NFX in the fetus following an 8-day maternal i.v. infusion of racemic FX, showing the degree of accumulation of NFX in the following order: TR ~ FA > AMN. 133 Throughout the infusion period, FX levels in amniotic fluid were lower than those in fetal plasma. The mean AUC in amniotic fluid for total FX was 7351 ± 4830 ug*h/l, which was about 50% of the AUC value in fetal plasma. The AUC of R -FX in amniotic fluid tended to be less than that of S-FX (2611 ± 1518 ug*h/l versus 4739 ± 3347 u.g*h/l, S/R ratio = 1.72 ± 0.51) but the difference was not statistically significant (p = 0.07). The average amniotic fluid-to-fetal arterial plasma (AMN/FA) ratio for total FX was. 0.54 ± 0.20. This ratio was significantly higher for R-FX compared to S-FX (0.94 ± 0.31 versus 0.44 ± 0.19; p<0.0\), suggesting disproportional distribution of FX enantiomers in amniotic fluid. The extent of distribution of NFX in amniotic fluid was comparable to FX. The mean AUC in amniotic fluid for total NFX was 4504 ± 1334 p.g*h/l, which was about half of that in fetal plasma. The AUC of R-NFX in amniotic fluid was significantly higher than that of S-NFX (2623 ± 880 ug*h/l versus 1882 ± 516 |_ig*h/l; /><0.05), with an average S/R ratio of 0.75 ± 0.14. This is an interesting finding because S/R ratios for the NFX enantiomers approximated unity in both maternal and fetal plasma. The average AMN/FA ratio for total NFX was 0.56 ± 0.21. Similar to FX, this ratio was significantly higher for R - compared to S-NFX (0.70 ± 0.29 versus 0.43 ±0.14; /?<0.05) with the S/R ratio averaging 0.64 ±0.10. A representative amniotic fluid concentration versus time profile for the FX and NFX enantiomers is depicted in Figure 5.7. Concentrations of FX in tracheal fluid were higher than those in amniotic fluid and comparable to fetal plasma. The mean AUC in fetal tracheal fluid for total FX was 15122 ± 8770 p.g*h/l, which did not differ from that in fetal plasma. The AUC of R -FX in fetal tracheal fluid was significantly lower than that of S-FX (4705 ± 2279 |jg*h/l versus 10417 ± 6544 |ag*h/l; /><0.05), with an average S/R ratio of 2.05 ± 0.55. The average fetal tracheal 134 ••— R-FX 0.1 -i , , , , . , . 1 , , , 0 2 4 6 8 10 12 Time (day) FIGURE 5.7 Representative amniotic fluid concentration versus time profiles of FX and NFX enantiomers in E4237 following an 8-day maternal i.v. infusion of racemic FX, showing stereoselective disposition of both FX and NFX enantiomers in the amniotic fluid. FIGURE 5.8 Representative fetal tracheal fluid concentration versus time profiles of FX and NFX enantiomers in E4237 following an 8-day maternal i.v. infusion of racemic FX, showing stereoselective disposition of both FX and NFX enantiomers in the fetal tracheal fluid. 135 fluid-to-fetal arterial plasma (TR/FA) ratio for total FX approximated unity (1.07 ± 0.28). Uneven distribution of the FX enantiomers in fetal tracheal fluid was observed, with the TR/FA ratio for R-FX being higher compared to S-FX (1.67 ± 0.43 versus 0.91 ± 0.27; p<0.01). The distribution of NFX in fetal tracheal fluid was comparable to FX, with mean AUC value for total NFX of 11368 ±4112 pg*h/l. The AUC of R-NFX in fetal tracheal fluid was higher than that of S-NFX (6615 ± 2348 pg*h/l versus 4753 ± 1809 pg*h/l; /?<0.01), with an average S/R ratio of 0.72 ± 0.08. This value was lower than the usual S/R ratios for the NFX enantiomers in either adult or fetal plasma. The average TR/FA ratio for total NFX was 1.35 ± 0.42. Similar to FX, this ratio was higher for R- compared to S-NFX (1.71 ± 0.56 versus 1.04 ± 0.30; p<0.0\) with the S/R ratio averaging 0.61 ± 0.06. Representative fetal tracheal fluid concentration versus time profiles illustrating the stereoselective distribution of FX and NFX are shown in Figure 5.8. 5.3 Discussion Kim, 2000, measured FX and NFX in cord (obtained from the umbilical cord at birth) and maternal venous plasma samples from 8 pregnant women who received a 10-30 mg oral dose of the drug daily during the course of pregnancy. The fetal FX concentration ranged from 7.9 to 94.3 ng/ml, with a mean of 45.4 ng/ml. These values are somewhat lower than those measured in fetal lambs in our current studies, which ranged from 40.4 to 110.9 ng/ml (mean = 76.1 ng/ml). On the other hand, fetal concentrations of NFX in human cord plasma ranged from 27.8 to 193.8 ng/ml (mean of 84.4 ng/ml), and were much higher than the concentrations measured in fetal sheep (range 29.4 to 92.4 ng/ml, mean = 49.5 ng/ml). When the concentrations of FX and NFX were added together, the combined concentration was 136 comparable between the human and sheep studies with mean values of 129.8 ng/ml and 125.6 ng/ml, respectively, indicating that they were exposed to the same levels of pharmacologically active moieties. This finding has significant implications in the interpretation of the physiological measurements that were collected in parallel with the current study (Morrison et al., 2001). In this report we demonstrated that maternal FX administration in pregnant sheep resulted in decreases in low voltage ECoG, REM and FBM in the fetal lambs. Postnatally, diminished weight gain following birth was also observed in these FX exposed lambs (Morrison et al., 2002). In two clinical studies, poor neonatal adaptation has been reported in human infants exposed to FX in the third trimester (Chambers et al., 1996; Cohen et al., 2000), and there is also a report of decreased facial and heart rate responses to painful stimuli in infants exposed prenatally to FX and other SSRIs (Oberlander et al., 2002). These postnatal effects could be due to alterations in fetal behavioral states, which are thought to be associated with normal brain development during the fetal and postnatal periods. Although the mechanisms underlying these effects elicited by FX and their relationships with drug concentrations remain to be elucidated, the comparable concentrations of the active moieties observed in the two species infers that similar fetal behavioral changes might occur in human fetuses exposed to FX. The present study demonstrates rapid but limited maternal-fetal transfer of FX and NFX in pregnant sheep. Our F/M ratios for FX (0.40 ± 0.08) and NFX (0.32 ± 0.09) are slightly lower than the values determined previously from the maternal single i.v. bolus experiment (0.59 for FX and 0.63 for NFX) (Kim et al., 2004). In general, following a single dose of drug to the mother fetal exposure is substantially affected by placental permeability (Mihaly and Morgan, 1984). However, with multiple dosing or after steady-state infusion, 137 placental permeability will not determine fetal exposure, as the rate of placental drug transfer will not influence the final equilibrium achieved, but merely the time at which it is reached. The higher F/M AUC ratio obtained from the single i.v. bolus dose studies as opposed to steady-state infusion suggests that placental permeability does not seem to affect the transfer of FX and NFX. The transport of substances across the placenta may occur by simple diffusion, facilitated diffusion, active transport, and pinocytosis. For the transport of drugs, simple diffusion is usually the most important mechanism. This mode of transfer is largely dependent on the physicochemical properties of the drug. FX has a relatively low molecular weight and high lipid solubility, both of which favor placental transfer by this process. Elliott et al., 1994, reported a significant relationship by multiple regression between physicochemical properties (i.e. Log P, pKa, and MW) and the mean drug/antipyrine transport ratios obtained experimentally from an in vitro cotyledon human placenta model. However, in the in vivo situation the extent of placental transfer of drugs does not always appear to parallel their physicochemical properties as demonstrated bythe data presented in Table 5.4. This table lists the physiochemical properties and F/M ratios of all therapeutic agents that have been studied in the pregnant sheep model as well as human data where available. The lipid solubility of the drug, as indicated by the octanol/water partition coefficient (Log P), is generally considered a good index of its ability to cross the placenta, i.e. the more lipid soluble the drug the more readily it should cross the placenta. However, as illustrated in Figure 5.9A, a plot of Log P versus F/M ratio based on the data from Table 5.4 does not reveal a significant correlation between lipophilicity and extent of transfer. For example, diphenhydramine and indomethacin have comparable log P values of 3.26 and 3.49, respectively, but there is a 20-fold difference in their F/M ratios (0.85 for diphenhydramine 138 co ct 2 ^ cz .2 CO CO g .9 c o CO CO co C L CO U CO ° E | I co S n n ^ CJ> ^ CM CJ) ° ) CO c ^ CD CO O) g g -o CO CJ) CD ; CD CO to CO <: CO o CO X co" 00 CO CD § _ c CO CO E to CO >. CD CD CO SZ > is S g ? o — CM CD CD • - CD ts CO "5 O) c 'E E o CD CD CD CO _ C L - CD — CD ro •<-8 w oo ^ : CD O CM to ro ro T? .-e E E k: CO CD CO to CD CD CO OJ O) 0 CO • or N CO c CM" CD ro to m CD CO 00 "D 1^ CD CD c ro • -sz 0) 0 CO E ro CO 0> CO CD les CD CO >-D ) 15 CD m N S S J CD ro o CD CO CD O o - CM CM g 3 « CO 00 ir> CD CO CD — CD ro ^ ro ^ •5; c ™ to ~ cu -0 ro CD £ < CD m _ . 0. CO o -+ co I » s | ? 3 ) CO CO I: ™ co ro -til +T* *L7i m *> to 2 o s a w CO ro co ti) Q) co CD S co DC > _ : _• CM O co ro to -2 ^ ? 0 5 N ~ 15 ~ ! m O) 0. CO % . . CM 1 1 1 CD CD fe CD 72 • - - CO _ — c ^ ro sz ro — o ~ CD CO CD > , ro _ " O CD o g-S E CD CD £-3 « CD . CD ~ ?> nf ~ > ro co "5 F. ro tu CO 3 CD ro T~ Z _ r co" " 8 « O -o CM co CO o CO 9 £ O CO "5 ro E co o m o o co 0 0 0 0 o d o d o CO o » 0 0 0 0 0 0 co -ct- co o "<cr 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 * o s * CD CD CO CD ^ t •ct LO CO LO CO CO 0 CD CO 0 CO CO 0 ^ t CO 0 CN CM CM 0 CO LO CM 00 v- 0 •— •— CO CO CO CD CO CO O CO T t LO 1^ ^ t ^ t CO CO CM 00 1^ CO CO CO "Cf CO •ct T t CO CO •— CM CM CD d "Ct CD CM LO CO LO d cri 00 d LO 00 •ct h - ' d d d LO LO d 10 LO d •* •ct ^ t cb ^ 00 CO •ct •ct LO 00 CO CD 00 LO CO O LO 0 CM LO CD CN CO "Ct 0 ^ t CT> CM CO CD LO CM CM CO 00 CO LO LO -^ CM CD CM CM •ct CO CO CM CO CO CO CO CO CM CM CO CM CO CM CO CM co CM CO CO CO CO CM CM •- •ct o d • - Tf d d o o n o s Q d d -^t o o co d 1^ 1 - P r- CO CD LO ^ d ^ CM 00 P LO CD CD T -CO CO CO CO 1^ d O CO CO •ct O T - 1 d CM ^ CD LO CD CD ° 0) S O) CO 00 06 in in s CD CO d CD d CD cb L O C M C D ' s f C O r ^ C O C M O CD CD CD CD LO LO CO r -S O ) S C O c O O ) 0 ) c D C ( ! ^ ^ ^ 0 1 C O C O C O i n c O O CO cb O d CD 00 CD •ct O CD CD cb in -ct ^ co 00 s CN O • - T t CO N CO in CM CM s T - CN CN CD T - t^ LO n N T - g c < ) 0 * N g ( D f c * i n i D O ) l N N 0 l 9 l 0 O C I ) ^ 9 N 9 C 0 0 ) ^ T l ; d ^ ^ i N ^ d r i c M ' i N n o ^ n o ' n c v i c i r i c M T t c M N < Q _1 < o CD o ro . T t C D C D C O T t c O " c f C O C D C O O C N C N i - O C D O O O C N C D d cb •— CM d CM "ct o sz CL • CO < CD E ro o Jo S id £ .E .H < < < CD E < o o E o ro E O CD O Q SZ — c: o CO <z Q. SZ 3 b HI u. CD 2 m & k F c g b r o CD TO CD X £ XL CO O c JS 'ro to g ro 73. CD CO cz <z 1 5 CD CO V, E 2 E Q. ro 2 o O NJ p ro co : o N O ™ cr _ C L 2 *o co 5 • i ° ^ .9- 3 C L Cp » » 5 o l o £ g o N ro x -a o -a o £ o „, ro <D ro = .— a} c o E o E I o E JF f I ro o o . o o o r : o r Q r : c o c o > > CD CD 139 1 . 2 - , 1 . 0 - 0 o 0 . 8 - o re 0 . 6 -i . o LU 0 . 4 o 0 . 2 -0 0 . 0 -o o o o o °o 2 3 4 LogP 1 . 2 - i 1 . 0 • < » 0 . 8 o 2 0 . 6 -I 0 . 4 oo o 0 2 1 . ° o.o § - £ — , -0 2 0 1 . 2 -1 . 0 - o o o o 0 . 8 - a o * ° o 0 ra 0 . 6 - 0 s L L 0 . 4 -o O 0 o 0 ° 0 0 . 2 -0 . 0 -O O Q 0 B o o 4 0 6 0 8 0 1 0 0 % Unionized Fraction 0 2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0 Molecular Weight FIGURE 5.9 Placental transfer of therapeutic agents in pregnant sheep in relation to their physicochemical properties. Fetal-maternal exposure ratios (F/M ratio) were determined following maternal drug administration (see Table 5.4 for individual values). Plot A: log P vs. F/M ratio; Plot B: unionized fraction vs. F/M ratio; Plot C: molecular weight vs. F/M ratio. Only compounds with molecular weight of 100-600 are included in plots A and B. 1 4 0 and 0.04 for indomethacin) in sheep. The degree to which the drug is ionized, as determined by the pKa of the ionizable group, may limit the extent of drug transfer across placenta. Thus neutral species are expected to cross the placenta more efficiently while ionized species are less efficient for transfer. Biehl et al., 1978 have shown that with fetal acidemia, the F/M ratio of lidocaine, which has a pKa of 7.9 increased significantly from 0.76 to 1.21 because more drug was ionized and trapped in the fetal plasma; a mechanism known as "ion-trapping." Since fetal blood pH (-7.35) is slightly lower than maternal blood pH (-7.45), ion trapping can occur even under normal physiological conditions and contribute to the F/M gradient particularly for drugs that possess pKa values close to the physiological pH. However, many other drugs which have pKa values outside the range of normal physiological pH are either completely ionized or unionized in the plasma therefore ion trapping becomes less relevant in those situations. Figure 5.9B depicts the relationship between the percentage ionization of drug at physiological pH versus F/M ratio, showing that no apparent correlation exists between the two parameters. For example, FX and albendazole sulphoxide have unionized fraction of 0.1% and 98.9 % at physiological pH, respectively, but their F/M ratios do not differ significantly (0.40 versus 0.46, respectively). The sheep placenta has a distinctive molecular cut off at which it becomes a significant barrier to transfer for substances having a molecular weight (MW) > 600 Da (Faber and Thornburg, 1983; Rurak et al., 1991). Molecules with a MW less than 600 Da appear to cross the placenta freely but no correlation is observed between molecular size and the extent of transfer, i.e. a low MW * a higher F/M ratio (Figure 5.9C). For example, FX and methadone have almost the same MW (-309 Da) but the F/M ratio of FX is 3-fold higher than that of methadone (0.40 vs. 0.14). Overall, these data have shown that the in vivo F/M 141 ratio is not merely a reflection of drug transfer across the lipid bilayer membrane of the placenta. Instead, it is a complicated physiological process within the maternal-fetal unit, which cannot simply be explained or predetermined by the physicochemical characteristics of the drug alone. For most drugs following maternal administration, fetal plasma concentrations are found to be lower than the maternal plasma concentrations at steady-state. Hill & Abramson, 1988, reported that in humans the degree of maternal and fetal protein binding is the most important factor in determining the F/M drug ratio at steady-state. When differential plasma protein binding of FX and NFX in the mother (see Chapter 4 for maternal fu values) and the fetus (see Chapter 6 for fetal/, values) is taken into consideration, the F/M ratio of unbound drug increases to 0.71 for FX and 1.0 for NFX. It appears that differential protein binding is sufficient to explain the maternal-fetal concentration differences for NFX but not for FX. Szeto et al., 1982b, use a two compartment pharmacokinetic model (Figure 5.10) to propose that the difference in fetal and maternal unbound steady-state concentrations could be due partly to fetal drug elimination, which is commonly denoted as fetal nonplacental clearance (CLf0). Common routes of fetal drug elimination include: 1) fetal drug metabolism, 2) fetal renal excretion, 3) excretion in amniotic fluids, and 4) excretion in fetal tracheal fluid (Rurak et al., 1991). Preliminary experiments involving incubation of FX in ovine fetal hepatic subcellular fractions did not show any formation NFX from FX following a 30-min incubation (Kim et al., 2004). Moreover, direct i.v. bolus injection of FX to the ovine fetuses resulted in no detectable level of NFX in the fetal plasma (Kim et al., 2004). These findings suggest the absence of fetal metabolic capability to convert FX to NFX. In our infusion experiments, the presence of NFX in the fetal circulation and other fluids is probably due to the placental transfer of the preformed metabolite from the maternal 142 Maternal Compartment C L , mf / u ( m o t h e r ) / u ( f e t u s ) C L fin Fetal Compartment C L n Placenta Case 1. C s s ( m o , i i e r ) = C S S ( f c t u S ) in the absence of fetal nonplacental clearance. Protein binding is assumed to be the same between the mother and the fetus. Maternal Compartment C L , mf ^(mother) ^./u(fetus) C L fm Fetal Compartment Placenta C L n C L fo Case 2. C ss(moiiicr) > C S S ( f L . t U S ) in the presence of fetal nonplacental clearance (CLf 0 ) and difference in the protein binding between the mother and the fetus. FIGURE 5.10 Compartmental model of the maternal-fetal unit with continuous drug input into the maternal compartment (Szeto et al., 1982b). 143 circulation. In adult sheep, hepatic metabolism is the major elimination route for FX whereas renal excretion accounts for less than 1 % of the total body clearance. While the fetal liver is generally less developed for drug metabolism, studies have shown that the fetal kidney has the capacity for efficient renal tubular secretion of meperidine and cimetidine (Szeto et al., 1979; Mihaly et al., 1983). Thus, a compensatory mechanism might take place in the fetal kidney whereby renal excretion becomes the predominant factor for FX elimination rather than the liver, however further experiments are necessary to assess this possibility. Significant excretion of drug into the amniotic fluid could not account for the large gradient because FX and NFX concentrations were low in amniotic fluid, with the AMN/FA ratio averaging 0.54 and 0.56 for total FX and total NFX, respectively (Table 5.3). This is in agreement with findings in previous maternal i.v. bolus experiments where insignificant accumulation of FX and NFX in amniotic fluid was observed (Kim et al., 2004). A lack of accumulation in amniotic fluid has also been reported for several acidic drugs in sheep, such as valproic acid (Gordon et al., 1995), diphenylmethoxyacetic acid (DPMA), a diphenhydramine metabolite (Tonn et al., 1995), and indomethacin (Krishna et al., 2002). In contrast, most amine drugs including metoclopramide (Riggs et al., 1987), ritodrine (Wright et al., 1991), labetalol (Yeleswaram et al., 1992), and diphenhydramine (Kumar et al., 1997) have been found to accumulate substantially in amniotic fluid with concentrations approximately 4-15 times higher than those in fetal plasma. In spite of being amines, FX and NFX do not appear to follow these same general distribution characteristics in amniotic fluid. As fetal urine is a major source of amniotic fluid, low drug concentrations in amniotic fluid imply that fetal renal excretion of FX and NFX might also be low. Intra-membranous pathways are also important in fluid and solute exchange between amniotic fluid and fetal 144 blood (Gilbert & Brace 1989). However, the low amniotic concentrations of FX and NFX suggest that this pathway is of minor importance. Protein binding may also be a factor in determining the distribution of drugs in amniotic fluid. The protein binding of FX and NFX in pooled fetal plasma following single i.v. bolus administration to the fetus ranged from 90-94% (Kim et al., 2004). Indomethacin and DPMA are also both highly protein bound in fetal plasma (-98%) and exhibit low concentrations in amniotic fluid (Anderson et al., 1980a; Kumar et al., 1999a). FX and NFX were detected in fetal tracheal fluid at levels similar to those in fetal plasma. This is in contrast to the results for most amine drugs such as diphenhydramine, metoclopramide, labetalol and ritodrine, which are present in high concentrations in tracheal fluid (ranging 2- to 15-fold higher than fetal plasma) following either maternal or fetal drug administration (Riggs et al., 1987; Wright et al., 1991). The reason for the difference between FX and the other basic lipophilic drugs studied is unclear. There is extensive accumulation of FX and other basic lipophilic psychotropic drugs in the rat lung via lysosomal trapping and phospholipid binding (Daniel and Wdjcikowski, 1997). The former process involves diffusion of the non-ionized fonn of the drug into the acidic interior of lysosome where protonation of the drug occurs, preventing efflux from the lysosome. As FX is more lipophilic than the other basic drugs that have been studied, it may be that lysosomal trapping is more important for this compound resulting in limited efflux and accumulation in lung fluid. However, to determine if this might be a factor, measurement of FX in subcellular components of the lung in comparison to other drugs is necessary. The possibility of placental drug metabolism in addition to fetal drug elimination should be considered. Several drug biotransformation processes have been identified in 145 animal and human placentas, but the overall role of the placenta in xenobiotic metabolism still remains unclear (Juchau, 1980; Pasanen, 1999). As such, the relative importance of placental compared to fetal elimination of FX requires further investigation. Active transport of FX from fetus to mother could theoretically be responsible for the unbound drug gradient from mother to fetus. Specific transporters are found in placenta that facilitates the transfer of nutrients and other endogenous substances at the maternal-fetal interface (Ganapathy et al., 2000). Specificity of these transporters in the placenta is not strictly restricted to endogenous substances. Xenobiotics that bear significant structural similarity to these substrates have the potential to be recognized by the transporters expressed in the placenta. Thus placental transporters could play a crucial role in the distribution of drugs across the maternal-fetal interface. Numerous transporters have been identified in placental tissues with P-glycoprotein (P-gp) the most extensively studied in terms of drug disposition (von Moltke and Greenblatt, 2000; Lin and Yamazaki, 2003). P-glycoprotein-mediated transport occurs unidirectionally facilitating the efflux, and not the influx of various substances. This drug efflux transporter appears to have evolved as a protective mechanism in animals against the potential toxic effects of environmental toxins and other xenobiotics. Animal studies with P-gp-knockout mice have provided strong supporting evidence for an important role of placental P-gp in protecting the fetus from potentially harmful and therapeutic agents (Lankas et al., 1998; Smit et al., 1999). Intravenous administration of the P-gp substrates digoxin, saquinavir, and taxol to pregnant animals led to a 7-fold greater transfer of these drugs across the placenta into the fetus in P-gp knockout mice than in wild-type type mice. An in vitro placental transfer study of glyburide, which is a P-gp substrate, showed that the drug only minimally crossed the human placenta (Elliott et al., 1991). P-gp 146 was proposed to be the underlying mechanism as neither its pKa, solubility nor molecular weight precludes it from being able to cross the placenta. A similar observation was made for midazolam, which is also a P-gp substrate and exhibits a low degree of placental transfer in humans (Tolle-Sander et al., 2003; Kanto et al., 1983). Recent studies regarding the interactions between antidepressants and P-gp showed that FX and NFX are moderate P-gp inhibitors, suggesting the presence of substrate-recognition between P-gp and FX (Weiss et al., 2003). However, conflicting results were observed in a blood-brain-barrier uptake study where the brain-to-tissue distribution ratios of FX and its metabolites were indistinguishable between P-gp knockout mice and wild-type mice (Uhr et al., 2000). The expression of P-gp in the sheep placenta has not yet been confirmed, therefore it is difficult to attribute any of our findings to P-gp at present. It is clear that further studies are indicated to examine the role of P-gp in the placental transfer of FX. Our F/M ratios in sheep for FX and NFX were much lower than the cord blood/maternal venous plasma concentration ratio reported in humans with the latter approximating unity (Kim, 2000). This could be due to one or more of the following reasons. First, the epitheliochorial sheep placenta differs from the hemochorial human placenta in having a lower permeability to hydrophilic polar substances (Faber and Thornberg, 1983). Second, the human data are based on a single point detennination of the fetal-to-maternal drug concentration ratio, which changes constantly with respect to the time of dosing in non-steady-state conditions (Anderson et al., 1980b). Third, the extensive maternal and fetal hemodynamic changes that occur during and immediately after delivery may affect the disposition of the drug in the mother and/or the fetus (Hamshaw-Thomas et al., 1984) and hence extrapolation of the concentration ratio at delivery to the antepartum 147 intrauterine conditions may be erroneous. The last two points are exemplified in a recent study by Heikkinen et al., 2003, who reported a different newborn-to-maternal plasma concentration ratio for FX and NFX. They reported a mean F/M ratio of 0.65 for FX and 0.72 for NFX, which are about 30-40% lower compared to those reported previously by Kim (2000). Discrepancies in these results indicate that the single point determination of fetal-to-maternal drug concentration ratios in humans or animal species is subject to high variability. In addition, from a pharmacokinetic point of view the extent of fetal exposure to a drug is more accurately reflected by the F/M AUC ratio than by a comparison of fetal and maternal plasma drug concentrations at any time, t, after drug administration. Although care should be exercised in extrapolating results from animals to the clinical situation, our data indicate that the fetus could be exposed to upwards of 70-100%) of the concentrations of the pharmaco-logically active form (i.e. unbound drug) of FX and NFX present in the maternal plasma. The placental transfer of several compounds has been studied in both sheep and humans (Table 5.4), but no one has attempted to explore the correlation between the two species in this regard. On this basis, we decided to investigate the possibility of extrapolating the F/M ratio determined in sheep to that in humans. A total of nineteen compounds for which placental transfer in both sheep and human has been studied were used to test for a correlation. In our initial trial no correlation was found as illustrated in Figure 5.11, plot A. Attempts were made to identify outliers from the original data set based on the following two criteria: 1) human-to-sheep F/M ratio <1 (based on the presumption that the sheep placenta cannot be equally or more permeable to drugs compared to the human placenta); 2) human-to-sheep F/M ratio > Q3 + (Q3-Q1) x 1,5, where QI indicates the bottom quartile and Q3 the top quartile of the human-to-sheep F/M ratio (Zar, 1996; Ogawa et al., 2003). QI, Q3, and 148 1.C 1.6 -1.4 -.2 1 2 * 1 g 0.8 E x 0.6 0.4 0.2 0 o o o o 0.2 0.4 0.6 Sheep F/M Ratio 0.8 03 0£ E 3 X 0.2 0.4 0.6 Sheep F/M Ratio 0.8 FIGURE 5.11 Relationship of F/M ratios between sheep and human. In Plot A, no correlation is observed when all 19 compounds are included for analysis (Table 5.4). In Plot B, significant linear correlation was observed between sheep and human values with 14 compounds when outliers are excluded (/?<0.01, Spearman rank correlation). Compounds that fit the correlation include (1) fluoxetine, (2) indomethacin, (3) labetalol, (4) meperidine, (5) midazolam, (6) morphine, (7) nifedipine, (8) norfluoxetine, (9) propranolol, (10) quinidine, (11) quinine, (12) ritodrine, (13) salicylic acid, and (14) valproic acid. 149 the upper outlier limit are determined as 1.6, 6.5, and 13.8, respectively. Using these criteria, methadone, lidocaine, metoclopramide, cimetidine, and ranitidine were identified as outliers. Upon exclusion of these outliers, a second analysis was performed and the result revealed a significant linear relationship for the F/M ratio between sheep and humans as illustrated in Figure 5.1 IB (rs = 0.69; p< 0.01, Spearman rank correlation). When subject to linear regression analysis, the data set (outliers excluded) yielded the following regression equation: y = 1.59x + 0.32 (r2 = 0.54; p< 0.01 when H 0 : p=0, but not statistically significant when H 0 : P=l *). This regression equation tends to predict a higher F/M ratio in human with respect to the value in sheep, which is not illogical considering the placental structure in the two species. However, factors that influenced the correlation remain to be explored. Variability in human studies could be a causal factor accounting for some of the deviations. Empirically, drugs that follow the correlation might share similar physicochemical and ADME (absorption, distribution, metabolism and elimination) properties but this hypothesis requires further testing. Extensive pharmacokinetic analysis of the fetal FX concentrations was not attempted in our current study because the bidirectional placental flux of FX as well as fetal nonplacental clearance could not be calculated from the available data. The significantly longer apparent elimination ti/2p of FX and NFX in the fetus when compared to the ewe could be due to a small fetal-to-maternal transplacental clearance of the drug. It may be also be related to the extensive binding of FX to and consequent slow release from proteins and/or tissues in the fetus. Reduced drug elimination capacity in the fetus or recirculation of FX from the amniotic fluid compartment back to the fetus could also be an additional factor. In a * H 0 = null hypothesis, (3 = slope of the regression line 150 clinical case study, Heikkinen et al., 2003, reported that FX and NFX levels in newborns at 4 days post-delivery were about 60-80% of the levels at birth. The mothers of these infants had received oral dose of FX for at least 2 weeks before delivery. In the present study we observed that FX and NFX were excreted in amniotic and tracheal fluids and declined at a rate similar to that in fetal plasma (Figures 5.5 and 5.6). This indicates the rate-limiting step of drug elimination in fetal plasma could reside in these fluid compartments. Similar observations have been made with labetalol and metoclopramide in pregnant sheep following maternal administration (Riggs et al., 1988; Yeleswaram et al., 1992). The elimination ti/2 of labetalol was comparable among fetal plasma, amniotic fluid and fetal tracheal fluid, but longer compared to maternal plasma (Yeleswaram et al., 1992). For metoclopramide the rate of elimination was substantially slower in the fetus compared to the mother with the drug accumulating and persisting longer in the amniotic and tracheal fluids (Riggs et al., 1988). Stereoselectivity in placental drug transfer is poorly characterized, partly due to the limited number of chiral compounds that have been studied in. this regard. In the present study, stereoselectivity in the F/M ratio was observed. The F/M ratio for R-NFX was significantly lower than that for S-NFX (0.29 vs. 0.36) whereas no difference was observed between R-FX and S-FX (0.39 vs. 0.41). The 25% higher F/M AUC ratio for S-NFX compared to R-NFX indicating that NFX distributed between the mother and the fetus in a stereoselective manner. A clinically important consideration for this difference in F/M AUC ratio lies in the fact that S-NFX is 20-times more active than R-NFX in terms of serotonin reuptake inhibition, and hence the fetus is actually exposed to more pharmacological active species than one would normally predict based on racemic data. A similar finding was observed for albendazole sulphoxide (ABZSO), which is a racemic mixture used as an 151 anthelmintic drug in veterinary medicine. Studies with pregnant sheep showed that the AUC ratio between mother and fetus for (+)-ABZSO and (-)-ABZSO was 2.5 and 1.4, respectively (Capece et al, 2002). In the maternal plasma, the AUC ratio between (+)-ABZSO and (-)-ABZSO was 2.8 whereas in the fetal plasma the ratio decreased to 1.6. In our experiments we also observed a difference in the S/R AUC ratio of NFX enantiomers between the mother (0.91, Table 4.2) and the fetus (1.15, Table 5.2). As discussed earlier, differential protein binding between enantiomers could affect the concentration gradient between the mother and the fetus. When S/R AUC ratio is corrected for the free drug, the ratio becomes 0.65 for the mother and 0.91 for the fetus, which results in a corresponding F/M ratio of 0.94 for R-NFX and 1.32 for S-NFX. These results suggest that protein binding alone is not sufficient to explain the stereoselective difference in F/M ratios. Other mechanisms must, therefore, be in effect to cause the difference in disposition of NFX enantiomers between the mother and the fetus. Drug levels in the fetus are affected by both placental and fetal nonplacental clearances. These clearance processes could occur in a stereoselective manner. We observed that the excretion of NFX in amniotic and fetal tracheal fluid showed significant preference for R-NFX compared with S-NFX (/?<0.05). This is in contrast to what we observed in adult plasma where the S/R NFX AUC ratio approximated unity. The higher excretion of R-NFX into these fluids corresponds with a lower fetal plasma level of R-NFX, resulting in a smaller F/M AUC ratio for R-NFX compared to S-NFX. The mechanism of stereoselectivity behind these excretion processes is unclear. Drug entry into amniotic fluid is mainly through renal excretion, therefore stereoselective renal excretion could be involved, but further studies are necessary to examine this possibility. The difference in F/M ratio between enantiomers might also suggest an enantioselective transport through the placental membranes. As we 152 have discussed earlier, P-gp is expressed in placental tissues and plays an important role in controlling drug transfer across placenta. The 3-dimensional interaction between protein and substrate allows for the possibility that one enantiomer will be a better substrate than the other. As such, there may be a difference in efflux between enantiomers. Little regard has been paid to stereoselective placental transfer as most drugs cross the placenta by simple diffusion. Since enantiomers have the same physicochemical properties the degree of placental transfer should be the same. However, more research is now emerging to indicate that the role of drug transporters in this process is more important than originally perceived (Ganapathy et al., 2000; Young et al., 2003). It is clear then, that further studies examining the stereoselective aspects of drug transporters and their relationships with placental drug transfer and fetal drug disposition should be emphasized. 5.4 Summary In summary, we have found moderate placental transfer of FX and NFX in pregnant sheep. Physicochemical properties of the drug alone cannot provide a satisfactory explanation for the concentration gradient between the maternal and fetal circulations. The protein binding corrected F/M AUC ratio for FX has a value of less than unity, suggesting the presence of fetal nonplacental FX elimination. FX and NFX were excreted into amniotic and fetal tracheal fluids with levels not exceeding those in fetal plasma. There are possibilities for fetal renal excretion, active placental transport, and placental and fetal metabolism, although previous in vivo and in vitro experiments suggest that the later two elimination mechanisms are probably of minor significance. The extent of placental transfer in sheep is less than that in humans probably due to the permeability characteristics of the sheep placenta. An empirical correlation for F/M ratios between sheep and human was 153 derived from several compounds that were studied for their placental transfer in both species, however this relationship requires further testing. Stereoselectivity in the F/M ratio was observed for NFX. Factors causing the difference in S/R ratio between the mother and the fetus could involve stereoselective fetal drug elimination (e.g. renal excretion) and stereoselective carrier mediated transfer (e.g. P-glycoprotein). 154 CHAPTER 6 DISPOSITION OF FX AND NFX IN FETAL LAMBS UNDER STEADY-STATE CONDITIONS Preface In a previous study conducted in our lab with direct i.v. bolus injection of FX to fetal lambs, no NFX was found in fetal plasma (Kim et al., 2004). This finding suggests that the fetus may be incapable of eliminating FX by metabolism. In human newborns from mothers who received FX during pregnancy, the rate of elimination of FX was found to be much slower compared to the adult, suggesting that fetal elimination of FX is impaired during the perinatal period (Heikkinen et al., 2003). In the present study, racemic FX was administered to the fetus by infusion to allow further investigation of the fetal disposition of FX and NFX under steady-state conditions. Moreover, the stereoselective pharmacokinetics of FX and NFX and the underlying mechanisms were examined in the fetus. Fetal urine was collected to determine the role of renal elimination in the fetal nonplacental clearance of FX and NFX. Comparative analyses of pharmacokinetic parameters in fetal, maternal and nonpregnant adult sheep were performed to assess the ontogeny of FX and NFX pharmacokinetics. Finally, computations of maternal and fetal placental and nonplacental clearances using the method described by Szeto et al., 1982b, which utilizes data from both maternal and fetal infusion experiments, were attempted to estimate the contribution of fetal nonplacental clearance to the overall clearance of the drug in the maternal-fetal unit. 155 6.1 Methods 6.1.1 Animal and Surgical Preparation Five pregnant ewes (Dorset and Suffolk breeds) at gestational age ranging from 119 to 126 days and term -145 days were used in these experiments. The study was approved by University of British Columbia Animal Care Committee and the procedures performed on sheep conformed to the guidelines of the Canadian Council on Animal Care. The surgical procedure was similar to that described previously in Chapter 5, Section 5.11 except that brain electrodes and blood flow transducers were not implanted. In addition, the fetal bladder was cannulated through a suprapubic incision to allow fetal urine collection. After surgery, ampicillin of 500 mg was administered intramuscularly for 3 days to the ewe, and via the tarsal vein to the fetus at the time of surgery and daily into the amniotic fluid for the duration of the preparation. A minimum of 3 days were allowed for the ewes to recover before experimentation. 6.1.2 Experimental Protocol A preliminary experiment was performed in one pregnant sheep to determine the appropriate infusion rate to achieve the desired steady-state concentration. A step-wise fetal i.v. infusion experiment at two different rates (kQ) was performed (k0 = 2.1 mg/h for step 1 and 4.2 mg/h for step 2) for the purpose of achieving a target FX concentration of 50-100 ng/ml. Each infusion step lasted for 4 days. Figure 6.1 A is a schematic illustration of the experimental protocol used for this preliminary 2-step fetal infusion study. Serial blood samples were collected from both the fetus and the mother over the entire course of the infusion. Cumulative fetal urine samples were collected every 12 hours until the end of the 156 A 3-4 days 4 days 4 days DDI 0 • > 0 D t Surgery t t Start of Infusion Step 1 Start of Infusion Step 2 B 3-4 days 4 days 4 days t Surgery t t Start of Infusion (without L.D.) End of Infusion F I G U R E 6.1 Schematic diagram o f the experimental protocol used for ( A ) the p re l iminary 2-step fetal infusion study, where infusion rate at step 2 is double the rate o f step 1; (B) the 4-day fetal infusion study ( L . D . = loading dose) fo l lowed by a 4-day post- infusion per iod. 157 infusion. Figure 6. IB is a schematic illustration of the final experimental protocol selected for the fetal infusion studies based on the targeted FX plasma concentration. Each animal received a continuous i.v. infusion of racemic FX HCI via the fetal tarsal vein for 4 days to achieve steady-state concentrations. Serial blood samples from the maternal and fetal femoral arteries (5 ml) were collected at 0, 15, 30, 45 min, 1, 2, 4, 6, 12, 24, 36, 48, 60, 72, 75, 78, 81, 84, 96 h during the infusion, and at 1, 2, 3, 4, 6, 9, 12, 18, 24, 27, 30, 36, 48, 51, 60, 72, 75, 84, and 96 h post-infusion. The zero time samples were collected within 10 minutes prior to starting the drug infusion. The volume of fetal blood collected was replaced by an equal volume of drug-free maternal blood at intervals during the study. Fetal urine samples were collected immediately before dosing, then at 4 h intervals from 73 to 77 hours during the infusion, and from 1 to 5, 7 to 11, 25 to 29, 49 to 53, 73 to 77 hours after the discontinuation of the infusion. All blood samples collected for drug and metabolite analysis were placed into K 3 EDTA Vacutainer® tubes, gently mixed, and centrifuged at 2000g for 10 min to separate the plasma. The plasma supernatant was transferred into clean borosilicate test tubes with polytetrafluoroethylene (PTFE)-lined caps. The volume of fetal urine collected at each time interval was recorded and a 10 ml aliquot was taken and transferred into a clean borosilicate test tube for subsequent drug and metabolite analysis. The remaining urine was infused into the amniotic cavity. All samples were stored at -20°C until the time of analysis. 6.1.3 Physiological Recording and Monitoring Procedures An aliquot of whole blood sample (~1 ml) was collected from the fetus pre-dose and at 15, 30, 45 min, 1, 2, 4, 6, 12, 24, 48, 72, and 96 h during the infusion (the same sampling 158 schedule repeated at the initiation of the second step of the preliminary 2-step infusion experiment), and at 24-hour intervals during the post-infusion period for blood gas, acid-base, electrolytes.and glucose/lactate analyses. The equipment used for these measurements has been previously described in Chapter 2, Section 2.2.2. 6.1.4 Determination of FX and NFX Plasma Protein Binding Plasma protein binding of FX and NFX was measured ex vivo using the equilibrium dialysis procedure as described in Chapter 3, Section 3.1.3, except that the dialysis of fetal plasma was performed at a temperature of 39.5°C, the physiological temperature of fetal lamb. Plasma unbound fraction (fu) was calculated as described in Chapter 3, Section 3.1.3, Equation 3.1. 6.1.5 Pharmacokinetic Analysis All pharmacokinetic parameters were calculated using standard pharmacokinetic procedures as described in Chapter 3, Section 3.1.4. The fetal-to-maternal exposure ratio (F/M AUC ratio) was calculated the same way as described in Chapter 5, Section 5.1.4. Fetal renal clearance (CL r e n a|) was determined from the slope of the plot of cumulative drug excreted in the urine from tn to t„+i versus AUC between the same collection intervals. Fetal weight at the time of experiment was estimated using the following equation for fetal ovine growth described by Gresham et al., 1972: log intrauterine weight (g) = log birth weight (g) - 0.0153*(At) where At is the time difference, in days, between the day of delivery and the day of experiment. 159 6.1.6 Statistical Analysis A paired t-test was used for statistical analysis of the pharmacokinetic parameters when comparing R - versus S-enantiomers. Comparative analyses of the pharmacokinetic parameters obtained from the fetal, maternal, and nonpregnant adult sheep were performed using one-way ANOVA and post hoc Tukey's tests. The achievement of steady-state for plasma FX and NFX concentrations was evaluated by testing whether the slope of plasma concentration versus time curve was significantly different from the value of zero. The significance level used was p<0.05 in all cases. All pharmacokinetic data are reported as the mean ± standard deviation (S.D.). Fetal blood gas data were analyzed using repeated measures ANOVA followed by post hoc Tukey's tests to determine the effect of time on changes in measurements from pre-infusion values. All physiological data are presented as mean ± standard error (S.E.). 6.2 Results Mean fetal birth weight averaged 2945 ± 340 g and the mean time between the day of delivery and the day of the experiment 15.5 ± 1 days, therefore the estimated average intrauterine weight at the time of experiments was 1707 ± 219 g. Figure 6.2 shows the fetal plasma concentration versus time profiles of total FX and NFX from the preliminary 2-step infusion experiment. FX reached steady-state at approximately 1-2 days after the initiation of each infusion step. The mean C s s of total FX during the step 1 and step 2 infusions were 40.2 ± 5.4 ng/ml and 101.7 ± 2.3 ng/ml, respectively. The lack of proportionality between infusion rate and C s s (i.e. 2 times k0 resulted in 2.5 times increase in C s s) could be related to the nonlinear profile of FX pharmacokinetics discussed earlier in Chapter 3. In contrast to 160 2X initial infusion rate Total F X « — Total N F X 8 Time (day) F I G U R E 6 . 2 Plasma concentration versus time profiles of total (racemic) F X and N F X in the fetus (E8105) after step-wise fetal i.v. infusion of racemic F X at two different rates, showing steady-state concentrations of F X after each infusion rate and continuous accumulation of N F X during the entire infusion period. 161 the parent drug, NFX did not reach steady-state and continued to accumulate over the entire duration of the infusion. Concentrations of total NFX at the end of each infusion step were 4.6 ng/ml and 30.1 ng/ml, respectively. From this preliminary experiment, an infusion rate of 2.1 mg/h targeting a C S S of total FX of 50-100 ng/ml was selected for all subsequent infusion experiments following the protocol outlined in Figure 6.1b. Figure 6.3 depicts the mean plasma concentration versus time profiles of total FX and NFX in fetal lambs from these 4-day fetal infusion experiments. Steady-state was achieved between day 2-4 and the average C S S of total FX was 81.2 ± 8.6 ng/ml (n = 4). NFX was below the assay limit of detection in the fetal plasma during the first 6 hours of the infusion. After that, NFX slowly accumulated in the fetal circulation over the entire infusion period. The concentration of total NFX at the end of infusion averaged 18.1 ± 3.7 ng/ml. 6.2.1 Stereoselective Disposition of FX and NFX in the Fetus Figure 6.4 depicts representative plasma concentration versus time profiles for the FX enantiomers in a fetus, showing a consistent concentration difference between the FX enantiomers. Fetal pharmacokinetic parameter data for the FX and NFX enantiomers and their corresponding S/R ratios are summarized in Table 6.1. The stereoselective pharmacokinetics of FX enantiomers observed in the fetus is consistent with the findings in the adult sheep. With the exception of elimination ti/2p and MRT, significant differences were found in most parameters (e.g. A U C , C S S , Vd s s, and C L T B ) between R-FX and S-FX. The ti/2p of R-FX was not significantly different from that of S-FX (48.9 ± 12.2 h versus 43.4 ± 6.4 h, S/R ratio = 0.93 ± 0.24; p = 0.46). Likewise, the MRT of R-FX was comparable to that of S-FX (25.4 ± 10.5 h versus 20.7 ±5 .4 h, S/R ratio = 0.82 ± 0.15,/? = 0.11). The A U C 162 Time (day) FIGURE 6.3 Comparative average plasma concentration versus time profiles of total (i.e. R plus S enantiomers) FX and NFX in the fetuses (n=4) after a 4-day fetal i.v. infusion of racemic FX, illustrating the concentration differences between parent drug and metabolite. FIGURE 6.4 Representative plasma concentration versus time profiles of FX and NFX enantiomers in a fetus (E9140) after a 4-day fetal i.v. infusion of racemic FX, showing stereoselective steady-state concentrations of the FX enantiomers and a low level of NFX accumulation in the fetal plasma. 163 CD X o E c < 1/h/kg) S-FX 86.3 CD 15.9 26.7 33.8 E, ro X CD f^ T-_ J o L L tk •sf CN ci CO CM "Cf -cf CM m •FX CN 00 CO CO CD q CD oo CM ox CO ^ O <=> •FX CO o CO CO f^ O CO CD a: "Cf LO f^ LO Tf "o FX O cn o O LO CM r~ ro CO CO CM CM "^  CM on' < •FX •>cf CO "cf CD q CD L L cc Tf CO CO "~ cb l/h/kg) S-FX 9.74 14.91 13.08 11.29 12.25 —• _7 FX CD CM CM o CO t~-00 o cc "Cf "* CM d CN CD cb X cn cn CO h ~ CO 1/kg) S-F. LO LO CM CO CM CO CM CD t^ -CN ib rf CM X 00 00 q "Cf > R-F Tf ci CO -cf CD -cf LO CD T* cb CD "Cf X CD "cf CD CN (ng/ml) L L CO cri •* CD -cf -cf Tf d LO (ng/ml) C J R-FX o LO <J> •* f^ R-FX CM -ct cri CM CM "Cf CO cb CO !c *o> S-FX 5761 4519 4944 4529 4938 AUC R-FX 3960 2773 3199 3043 3244 X CO 00 CO LO xT L L c o CD CM LO CD ^f CM d CM MRT MRT R-FX , - r~ CM in Tf R-FX cri CN f~i l~-i CN r~-i CM cri CN X r~ ^ CN m Tf _r_~ L L CO csi LO CC CO CN d •* cb •cf R-FX i _ q CO CO CD R-FX CD "Cf CO CO CD CN i n cb o "Cf +1 +1 m L U CD LU coP O CO +1 CD o d 00 +l o T- d CD d CM d T— +1 o d CO V CN d CO o d LO +l o d LO V CO d CO o d T— +i o d CD V CD d O d T— +1 o d -cf V LO d cn -— ci "-+l o LO V Tf CO o d T— +i o d CO V i n LO -— d +1 CM d CO d "Cf CM d CD +1 "Cf CO d CD d X o LL > CD ro CO =S or 5? CO , a . CC o o ro O ZD < X o ro O ZD < o o 2 X LL z CO x LL z ct X LL z CO X LL z cc X LL z CO x LL z ct x LL Z CO X LL z tk x LL z CO z tk X LL z CO x LL z tk x LL z CO X LL z tk E c < "- i n o CD LO m d d i n i n CN Tf o CD CO CO d d CO h-i LO d CO f-"Cf CO CD CD o CD cn o IO "Cf LO -cf i n CO Tf LO O CD CO CD f^ i r i d d d o CM CD 00 CO CO CO CM d d d d d CD CO CM -cf CM CO i n i n -cf d d d d d o CM CO CD CO CO CM CD d d d d d f^ CO -cf CD LO o CM CD d d d d d 1138 m 00 CO CD f^ CD f^ o o Tf o CD CO CO i n CD i n L U FQ E9 FQ o CO CO i n CD Tf x— o CD CO CN +[ CN CD CD i n cq Tf d CO i r i cb o o o +1 CD Tf o o "Cf d cb d CM d CO CN CO T— -cf CN • - 1 f^ i n CO "Cf CO CD cb cb d d d i n CO i n "Cf in CD f~ "Cf d CM Ni CD CD m CD CD +1 coQ 0 CO o +1 00 o +1 i n d oo d o +l CD CD CC CC CO CO "Cf O d o d V CD O d Z CD < ^ CO > > 4 ct co 164 of R-FX was significantly lower than that of S-FX (3244 ± 509 jug*h/l versus 4938 ± 583 |ig*h/l, S/R ratio = 1.53 + 0.08; p<0.01). Similarly, C s s was significantly lower for R-FX compared to S-FX (33.4 ± 6.3 ng/ml versus 47.7 + 2.7 ng/ml, S/R ratio = 1.45 ± 0.19; p<0.0\). R-FX had a higher Vd s s compared to S-FX (463.4 ± 61.2 1/kg versus 245.3 ± 21.1 1/kg, S/R ratio = 0.54 ± 0.10; p<0.0\). CLTB of R-FX was significantly higher than that of S-FX (18.87 ± 4.38 1/h/kg versus 12.25 ± 2.23 1/h/kg, S/R ratio = 0.66 ± 0.03; p<0.01). The comparative average plasma concentration versus time profiles of total FX between the mother and fetus are depicted in Figure 6.5. The F/M AUC ratio for total (racemic) FX averaged 2.87 + 0.93. This represents a fetal-to-maternal gradient of similar magnitude, but with the order of the enantiomers opposite that found in the maternal infusion experiments. The F/M AUC ratio of R-FX was significantly higher than that of S-FX (3.16 ± 0.97 versus 2.72 ± 0.91, S/R ratio = 0.85 ± 0.08; /?<0.05). Stereoselective protein binding was observed in the fetal plasma, with the fu of R-FX being significantly higher than that of S-FX (4.79 ± 0.66 versus 1.27 ± 0.40, S/R ratio = 0.28 ± 0.12; /?<0.01). Excretion of FX was observed in fetal urine as illustrated in Figure 6.6, which depicts the time course of cumulative amount excreted in fetal urine from the 2-step infusion experiment. The renal elimination of FX in the fetus followed a linear process with the elimination rate increasing proportionally with infusion rate. Figure 6.7 illustrates the graphical method of determining CL r e nai from the slope of the plot of the amount of the FX enantiomers excreted in the urine versus AUC during the urine collection time interval tn-tn+i in a 4-day fetal infusion experiment. The straight line obtained from the slope of graph confirms the linearity of the fetal renal elimination process. Fetal CL r e nai averaged 52.1 ± 50.3 ml/h/kg and 33.8 ± 36.1 ml/h/kg for R-FX and S-FX, respectively (Table 6.1). The 1 6 5 1 -I , , • , , , , , , 0 1 2 3 4 5 6 7 8 9 Time (day) FIGURE 6.5 Average p lasma concentration versus t ime profiles o f total F X (i.e. R plus S enantiomers) i n the mother and fetus (n=4) after a 4-day fetal i.v. infusion o f racemic F X , w i t h F / M A U C ratio averaging 2.87 ± 0.93. 0.35 i Time (h) FIGURE 6.6 P lo t o f the t ime course o f the cumulat ive amount o f F X excreted i n urine o f a fetus (E8105) f rom the 2-step racemic F X fetal i.v. infusion experiment, demonstrating a linear relationship between renal e l imina t ion rate and dose. 166 AUC(t1-t2) (ug*h/L) FIGURE 6.7 Representative plot of the amount of FX enantiomers excreted in fetal urine versus AUC during the urine collection time intervals tn-tn+i for a 4-day FX fetal i.v. infusion experiment (E9168). 100 i 1 -I , , , r-n , 1 ' i : , , 0 1 2 3 4 5 6 7 8 9 Time (day) FIGURE 6.8 Average plasma concentration versus time profiles of total NFX (i.e. R plus S enantiomers) in the mother and fetus after a 4-day fetal i.v. infusion of racemic FX, showing higher accumulation of NFX in maternal plasma compared to fetal plasma (F/M AUC ratio = 0.62 ± 0.13; n=4). 1 6 7 CLrenai of R-FX tended to be higher than that of (S/R ratio = 0.61 ± 0.06) but the difference was not statistically significant (p=0.08) due to high variability among the four animals. Overall, CL r e nai accounts for less than 1% of C L T B , indicating that renal excretion of FX is a minor route of elimination in the fetus. Unlike the parent drug, no apparent difference was observed in the plasma concentration-time profiles of the NFX enantiomers as illustrated in Figure 6.4. Both the t|/2p and MRT of R-NFX tended to be longer compared to S-NFX but their differences were not statistically significant (t,/2p: 62.7 ± 7.4 h versus 46.8 ± 5.6 h, p=0.06; MRT: 129.0 ± 13.4 h versus 108.5 ± 7.6 h,p=0.09) (Table 6.1). The AUC of R-NFX was significantly higher than that of S-NFX (1313 ± 291 |ig*h/l versus 1100 + 289 ug*h/l, S/R ratio = 0.84 ± 0.08; p<0.05). The comparative plasma concentration versus time profiles of total NFX between the mother and the fetus are shown in Figure 6.8. Both R-NFX and S-NFX have the same average F/M AUC ratios of 0.62 ± 0.13, resulting in an S/R ratio approximating unity. The NFX/FX AUC ratio was significantly higher for the R-enantiomer compared to the S-enantiomer (0.41 ± 0.10 versus 0.23 ± 0.07, S/R ratio = 0.55 ± 0.07; p<0.01). Similar to FX, stereoselective protein binding of NFX was found in the fetal plasma, with the/, being significantly higher for R-NFX compared to S-NFX (6.41 ± 0.71 % versus 5.09 ± 0.52 %, S/R ratio = 0.81 ± 0.17; p<0.05). NFX was excreted in fetal urine but to a lesser extent compared to the parent drug. CL r e n a i values for the NFX enantiomers were calculated in a manner analogous to that of FX. As for the parent drug, there was no significant difference in C L r e n a | between R-NFX and S-NFX (30.3 + 16.6 ml/h/kg versus 25.1 ± 21.1 ml/h/kg, S/R ratio = 0.78 ± 0.40; p = 0.43). 168 6.2.2 Comparative Pharmacokinetics of FX and NFX in Adult, Maternal, and Fetal Sheep Comparisons of the FX and NFX enantiomer pharmacokinetic parameters values were made between sheep at different developmental periods to examine the effects of any associated changes in physiological and biochemical processes on drug disposition. The results of the ANOVA analyses on the pharmacokinetic data from fetal sheep (from the current chapter), maternal sheep (from chapter 4) and nonpregnant adult sheep (from chapter 3) are summarized in Table 6.2 for FX and Table 6.3 for NFX. The ti/2p of R-FX and S-FX in fetal lambs were similar to those in the maternal sheep, but significantly longer compared to nonpregnant adult sheep (/?<0.01). On the other hand, the MRT of R-and S-FX were comparable between the fetal lambs and nonpregnant adult sheep. The MRT for FX in maternal sheep was significantly larger than in the fetus, but not different from the that in the nonpregnant ewes. Weight normalized Vd s s of the FX enantiomers was significantly higher in fetal lambs, with values that were approximately 10-fold larger compared to maternal and nonpregnant adult sheep (/?<0.01). Similarly, the weight normalized CL-TB'S of R- and S-FX in the fetal lambs were much higher than the values in maternal and nonpregnant adult sheep (cxO.Ol). Protein binding of R-FX was lower in the fetal plasma with fu being significantly higher in the fetal lambs compared to both maternal and nonpregnant adult sheep (/?<0.01). For S-FX, however, this difference was only observed between the fetus and nonpregnant ewes (p<0.05). The higher fu is consistent with the observation of low a,]-acid glycoprotein concentration in fetal plasma compared to adult plasma (Wood and Wood, 1981; Hill and Abramson, 1988). In spite of its low importance relative to the overall systemic clearance, C L r e n a i in the fetus appeared to be much higher compared to adult sheep, but due to the high 169 E ' o O g CD •o < o d v D co CD CN -ct CN +1 +1 CD CD • * 00 co "cf o d cb ~ d +1 + 1 CD «3 3 CO UO d o m d CN T -+1 +1 CN •>* d co t CN o d o d +1 +i CD CN Tf CO •ct o d CD . ° cN +1 +1 CN n . ° $ C N o d o o +1 +1 o d o o +1 +1 00 CO CD CO d d CO o o o +1 +1 +1 CN T - CD K CN CN "Ct CO O O O +1 +1 +1 i - CO h-S CO CM d d co CO oo d CD o Q CD CU i_ - i o o o o o o CD CU i_ - i o o o o CD CD i -M o o I I § o o oS CD CD i -N o p o o o o o o CC CC CC cc cc cc CO CD p O S ~ £ -2 ^ < '3 c ca I u- to CO CC > o co — CU a -a O I c3 12 cu a. c 170 E ' o o 3-1 o <5 o Q Q CO +1 c ro CD Q . 3 O o o a cu ,_ .E co I X -g| ro ro E ro ro Q-co CD CO i -k co CO CD O Tr CM CO CD LO CO Tr' CO CN LO +1 +1 +1 +1 +1 +1 o o Tr CO CD CN CO CD CN CD CN CN LO CD TI-1 1 rr CO T r C0 CO +1 +1 o o ci cri CD CM h co i i : CD ° - CD ! I co +1 + l +1 o  0 0 LO 0 0 11° CN O o o tr I -rr Tr cn TT co d d +1 +1 CD LO CN CO CNJ CN CO w o o (5 CD CO S CM T - O o d d +1 +1 +1 CO T - CO N CD CM o d d o o tr o < X co T -• * s o d d +1 +1 +1 IM • * i -O O) •* OJ r- CO CD CD CD Q . C= ~ -3 o o a, Z 5 LL T- CN T f LO d d +1 +1 CO CJ) CO o i r i CD CD i -I I 3 i l l tr I I 3 i l l P LO T - CN CJ -o e c/) P >i C3 CD CJ PH *5 o c 3 &o s * x o CD VI p o V ) CD CD P C3 P 60 CD • S o -^ d ^ VI •~ o. X * C B c^ 5 -P I oo CD i O CD _ O ca j -c/) ^ '§ E p <~ o -a CJ CD >. o CD CD P ° o IS V) C^ CD d < as £>> « - 5 " z -a < 3 ^ C3 RA § ? 2 C CD P M P «3 & S "° 17 variability among animals, statistically significant differences were not achieved. The ti/2p of the NFX enantiomers in the fetuses were similar to those in maternal sheep, but higher compared to the nonpregnant adults (/?<0.01). In contrast, the MRTs of the NFX enantiomers in the fetal lambs were similar to the values obtained in the nonpregnant adult sheep, but higher compared to maternal sheep (p<0.05). NFX/FX AUC ratios were significantly lower in the fetal lambs compared to both maternal and nonpregnant adult sheep O<0.01). These ratios were less than 20% and 40% of the adult values for R-NFX and S-NFX, respectively, indicating decreased metabolic efficiency in the fetuses. Similar to FX, the protein binding of NFX was significantly lower in the fetal plasma. Overall, the/, of both R- and S-NFX in the fetal lambs were about 3 to 4-fold higher than the respective values in maternal and nonpregnant adult sheep. Urinary excretion of NFX was higher in the fetuses, which had higher CLrenai compared to the adult sheep although only the comparison for R-NFX was statistically significant (p<0.05). 6.2.3 Fetal Blood Gas Results Mean fetal arterial blood gas and pH values during the first 24 h of drug infusion (pooled data from Day 1 of Step #1 of the preliminary 2-step infusion experiment and Day 1 of the 4-day infusion studies) are illustrated in Figures 6.9 and 6.10. Data from the maternal FX /.v. infusion experiments are also included for comparison (from Morrison et al., 2002). Compared with the pre-infusion values, fetal pH, PO2, and oxygen saturation tended to decrease, whereas PCO2 tended to increase after fetal drug infusion, but none of these changes were statistically significant. Moreover, the changes were less than those occurring during the maternal FX infusion experiments (Morrison et al., 2002). Mean fetal PO2 172 56 54 O) X 52 E E g 50 D_ 48 46 Fetal Infusion Maternal Infusion B 4 5 Time (h) 12 24 FIGURE 6 . 9 Mean fetal arterial P 0 2 (A) and P C 0 2 (B) during the first 24 h of fetal FX infusion (closed symbols, n = 5) or maternal FX infusion (open symbols, n - 13). # and * denote significant differences (/?<0.05) from pre-infusion day in fetal and maternal infusion experiments, respectively. 173 7.37 i 7.36 -7.35 7.34 H i . 7.33 7.32 7.31 7.30 H 7.29 Fetal Infusion Maternal Infusion A 12 24 - • — Fetal Infusion Maternal Infusion FIGURE 6.10 Mean fetal arterial pH (A) and oxygen saturation (B) during the first 24 h of fetal FX infusion (closed symbols, n = 5) or maternal FX infusion (open symbols, n = 13). # and * denote significant differences (/?<0.05) from pre-infusion day in fetal and maternal infusion experiments, respectively. 174 dropped from 23.8 ± 0.6 mm Hg at pre-infusion to a minimum of 21.0 ±1.8 mm Hg at 0.75 h after infusion, and remained steady over the next 24 h period. A similar decrease in 02 saturation was observed, with an initial decrease from 60.7 ± 2.9 to 47.8 ± 3.5 % at 4 h post-infusion. Mean fetal PCO2 increased gradually and reached a maximum of 51.5 ± 0.6 mm Hg at 4 h post-infusion. Mean fetal blood pH decreased slightly from 7.352 ± 0.010 to 7.319 ± 0.009 at 30 min after infusion, but returned to pre-infusion levels after 24 h. The pH decrease appeared to be associated with a similar decrease in base excess, which was 1.2 ± 0.3 mEq/L before FX administration and -0.3 ± 0.4 mEq/L at 30 min. Figure 6.11 illustrates the changes in mean fetal lactate and glucose concentration during the first 24 hours. Both lactate and glucose gradually increased and reached the highest concentrations of 1.47 ± 0.38 mM and 1.07 ± 0.14 mM, respectively, at 6 h post-infusion (p<0.05), then returned to pre-infusion levels after 24 h. A similar pattern of changes in fetal blood gas status was observed during the first 24 hours of Step #2 of the preliminary 2-step infusion experiment. The daily blood gas status monitored at 0700 h throughout the infusion and post-infusion period are summarized in Table 6.4. No significant changes in any fetal blood gas parameters were observed. 6.3 Discussion In general the present experiments demonstrated that the FX enantiomers exhibit significant stereoselective disposition in the fetus following fetal drug infusion to steady-state. Overall, the stereoselective relationships between R-FX and S-FX are similar to those previously observed in adult sheep, but the magnitude of the S/R ratios was less. For example, S/R ratios for AUC and C s s averaged 1.5 in the fetal infusion experiments 175 A 0.2 A 0.0 4 5 6 12 24 Time (h) FIGURE 6.11 Mean fetal arterial lactate (A) and glucose (B) concentrations during the first 24 h of fetal FX infusion (closed symbols, n = 5) or maternal FX infusion (open symbols, n -13). # and * denote significant differences (/?<0.05) from pre-infusion day in fetal and maternal infusion experiments, respectively. 176 I ro oq ~ P O CN TT _; o • O S CN Tt E X o ^ r--L O . O O N r> ro O oo oo - 5 d n P d d oo _^ r— ( N * ! I S -CN T f TT N O N O TT C> _ ' ^ N O 1 . oo d ro CN CN £ d ^ P oo ro _ O ro P ro O r~ d ' oo P o ^ ^ C N \ C N ^ O N C N cn ^ T t CN TT °° oq oo CN • 00 T t cri O N TT N O N O O — ' oo r-T t i-i d N O RT T f ro d o ro O T f oo T t ro O ° --1 ,-<• d ' O N ^ «=> d d 1 oo o ^ _ _ N O ^ P d d r~- o cn cn d C N o C N co « C N ro ^ n q - 6 r- o C O oo ro 00 O N o N O O N o O N d d d d d - 3 O N C N o © r o O N O o 1 - 1 O N d d d d ro T f N O o O N T T C O N D C O o O N d d d d d C O ro NO NO ON oo o ro O ON d d d d d o CN CN ro N O oo ro ro N O oo o O N d d d d d C N N O O N o O L O O N '—' N O -—1 O N d d d d d o C O O N C O C N O N N O o N O o O N d d d d d CN oq ro C O o CN ro CN Tt" oo o O N d d d d oo CN N O oo ro O N oo o o • O N d d d d d N O ro co CN C N O O C O oo < 1 o d d d d d CD p j C U t L j C U t L j C U p J CD p j CD |_L| CDpLj CULL] ca LTD O X CL, W PQ T 3 X ca c/T T J -C H CD c C J ex o CD c/} o ca X) a o W u c PQ c CH o o C+H ts L O CO tur I I c alu ca alu in > gen L O gen o o d ox V r i la >^  CD 00 ca LH Q . -O are o sion ca cS ca d 177 compared to 3 to 4 in the maternal infusion experiments. In chapter 3, it was shown that stereoselective protein binding and metabolism were key factors in determining the stereoselective disposition of FX in adult sheep. Although the extent of protein binding was lower in fetal plasma, stereoselectivity was preserved with the S/R ratio averaging 0.28, which is comparable to the adult value (0.30). Stereoselective metabolic formation of NFX is primarily due to the difference in intrinsic clearance between FX enantiomers. Previous in vitro metabolism studies using fetal ovine hepatic microsomes did not show any formation of NFX from FX (Kim et al., 2004). Thus, it is believed that the fetal liver does not contribute significantly to the observed difference in fetal enantiomer concentrations. Instead, the apparent difference in concentration between R- and S-FX in the fetus at equilibrium is more likely to be a byproduct of stereoselective metabolism from the maternal side, with subsequently placental transfer. This is supported by the finding of comparable S/R AUC ratios in the mother (1.80) and the fetus (1.53) in the present study. Moreover, the lower S/R ratios for C s s and AUC correspond with the lower FX concentrations in maternal plasma (mean total FX C s s of 33.4 ng/ml; Figure 6.5). These results also parallel the positive correlation observed between FX concentration and S/R ratio (Figure 4.6) in the adult sheep experiments. Figure 6.12 is a modification of Figure 4.6, from which the maternal FX concentrations and S/R ratios from the present fetal infusion experiment were included to further illustrate the presence of a positive correlation between FX concentration and S/R ratio. A previous study with direct i.v. bolus injection of FX to fetal lambs resulted in no measurable level of NFX in fetal plasma (Kim et al., 2004). In contrast, NFX was detected in fetal plasma and urine during the current fetal steady-state FX infusions, with the mean 178 4.5 •• • • DS • m a ^ • • Nonpregnant • Pregnant (Maternal Infusion) • Pregnant (Fetal Infusion) 50 100 150 Total FX Concentration (ng/ml) 200 250 FIGURE 6.12 Plot of the maternal S/R FX concentration ratio versus the corresponding total FX concentration, a compilation of all data from the nonpregnant and pregnant (following either maternal or fetal i.v. FX infusion) sheep experiments. 179 AUC for total NFX accounting for 30% of the parent drug. The presence of drug metabolites in the fetal circulation may be the result of maternal-to-fetal placental transfer or biotransformation by the fetus itself. As previously discussed, the presence of NFX in fetal plasma during the maternal infusions appears to be the result of metabolite formation in the ewe followed by placental transfer to the fetus. Since the fetal lambs in the present studies were not surgically prepared to allow us to determine the existence of a concentration gradient across the fetal liver for NFX, we are not able to provide direct evidence for the fetus to biotransform FX to NFX. Nevertheless, the low NFX F/M ratios along with the delay in the appearance of NFX in the fetal circulation indirectly suggest a deficiency for the fetal liver to biotransform FX to NFX. Since the majority of this biotransformation occurs through CYP2D6 in humans (see Chapter 5, discussion), these results suggest the lack of orthologous enzyme activity in the fetal lamb. In human fetal liver, CYP2D6 mRNA has been detected however, CYP2D6 protein content was negligible (Treluyer et al., 1991; Hakkola et al., 1994). Using dextrorphan formation, via dextromethorphan (9-demethylation (see Figure 4.10 in chapter 4), as a probe for CYP2D6 activity Treluyer et al, 1991, showed that human fetal liver samples accounted for only 1% of the adult values. A substantial metabolic contribution by the fetal liver to the FX-to-NFX pathway is therefore improbable. Besides the difference in drug biotransformation, significant alterations in other pharmacokinetic properties of FX and NFX were observed between the fetus and the adult. The Vd s s of FX in the fetus was 10-fold greater than that in the nonpregnant adult or the mother. In addition to the inherent high tissue distribution characteristics of FX, it is likely that the large volumes of amniotic fluid surrounding the fetus provide additional extravascular fluid space for FX fetal distribution. Another possibility is that the maternal 180 compartment, which acts as a sink, also contributes additional space for FX distribution from the fetus. Moreover, a remarkable difference in the weight normalized C L T B of FX was observed between the fetus and the adult. Our maternal and fetal infusion studies in the pregnant ewes and their fetal lambs indicate that FX is transferred across the sheep placenta from both sides of the membrane. Placental transfer of FX from the fetus to the mother serves as the dominant removal process for FX from the fetus. Further experiments in neonatal lambs may confirm if the high CLTB in the fetus will be lost at birth and the newborn must then eliminate FX using its own immature metabolic pathways. Protein binding of both FX and NFX was significantly lower (higher /u) in the fetus compared to the adult. This is in accordance with the low level of cii-acid glycoprotein, which is the major binding protein for FX and NFX, in (human) fetal plasma (Wood and Wood, 1981). This implies that the fetus is exposed to more unbound drug (i.e. the pharmacological active moiety) if there are equivalent total drug concentrations between the maternal and fetal circulations. Therefore the fetus may be at a higher risk of experiencing toxicity if drug levels exceed the therapeutic window. Our experiments also showed that the CL,enai of FX and NFX was much higher than the corresponding values in adult sheep (Tables 6.2 and 6.3). However, the contribution of Clonal to the overall elimination of FX was negligible and accounted for less than 1% of CLTB in all animals. Comparable CL r e n a i values have also been observed in fetal lambs for the amine drug diphenhydramine (range 7.8 to 60 ml/h/kg; Kumar et al., 1997) and like FX and NFX they were also higher than in the adult. The average CL i e nai values of two other amine drugs cimetidine and meperidine in fetal lambs are however, considerably higher (117 ml/h/kg and 186 ml/h/kg, respectively) than those of FX and NFX (Mihaly et al., 1983; Szeto et al., 1979). For both these drugs, 181 however, metabolism is negligible and renal excretion is the primary route of elimination. These findings for amine drugs are in contrast to acidic compounds such as valproic acid (Kumar et al., 2000b), diphenylmethoxyacetic acid (Kumar et al., 1999a) and acetaminophen (Wang et al., 1986) all of which exhibit very low C L r e n a i in fetuses. Robillard et al., 1977, reported that the renal clearance of inulin in fetal,sheep ranged from 23.4 to 114.6 ml/h/kg. The CL r e n a i of FX and NFX in our fetal lamb experiments were highly variable, with some animals having values that exceed those reported in the literature for inulin clearance. This suggests that the fetal kidney may be able to eliminate FX and NFX by tubular secretion, which may partly explain the reason why R-FX was preferentially concentrated in fetal urine in higher amounts compared to S-FX. Overall, we have observed that the fetus has 1) limited capability to metabolically biotransform FX to NFX, 2) negligible renal excretion of FX in urine, and 3) insignificant distribution of FX in amniotic fluid and fetal tracheal fluid (based on the results from Chapter 5). Unfortunately, the findings we have gathered so far are still far from explaining the fetal-to-maternal concentration gradient observed in the maternal infusion experiments (0.4 for total drug [bound plus unbound drug] and 0.7 for unbound drug). Apparently, other routes of elimination probably exist in the fetus. To further access the significance of fetal drug elimination by nonplacental routes, we attempted to calculate the placental, and nonplacental clearances of FX in the ewe and the fetus from the maternal and fetal steady-state arterial plasma FX concentrations according to the two-compartment model described by Szeto et al., 1982b (Figure 6.13). This model is based on separate steady-state maternal and fetal drug administration and assumes bidirectional placental transfer and drug elimination from both maternal and fetal compartments. The model is defined by four clearance parameters, 182 k0 I Maternal Compartment CL, mf CL fm Fetal Compartment Placenta , „ T L Lfo CT FIGURE 6.13 A representation of various placental and non-placental drug clearances in the 2-compartment pharmacokinetic model of the maternal-fetal unit (CL m o - maternal nonplacental clearance; CLf0 - fetal nonplacental clearance; CLm, - placental clearance from the mother to the fetus; CLrm - placental clearance from the fetus to the mother; k0 - drug infusion rate to the mother; k0' - drug infusion rate to the fetus). (Adapted from Szeto et al., 1982b). 183 including maternal-to-fetal placental clearance (CLmf), fetal-to-maternal placental clearance (CLfm), maternal nonplacental clearance (CL m o ) , and fetal nonplacental clearance (CLf 0). Maternal and fetal total clearances of the total drug ( C L m m and CL f T , respectively) equal the sum of their respective placental and nonplacental clearances. All these clearance parameters were calculated from the following equations as described previously (Szeto et al., 1982b): k C L m m = [ C m - C f * ( C m 7 C / ) ] . ( 6 ' 1 } k ' C L " = [ c y - c , „ ' . ( c f - c , „ ) ] < 6 ' 2 ) CT = CT * V C m J C L r = C L * — fm mm (6.3) (6.4) v C f ' y C L m 0 = C L m m - C L l l l f (6.5) C L f 0 = C L f f - C L f m (6.6) The symbol k0 and k0' denote the drug infusion rates to the mother and the fetus, respectively. C m and Cf are the steady-state plasma drug concentrations in the mother and the fetus after maternal F X administration, respectively, and C m ' and Cf are the steady-state maternal and fetal plasma drug concentrations, respectively, after fetal F X administration. Ideally, maternal and fetal drug infusions should be performed in the same animal to minimize the variability between animals. However, due to the long ti/2p and nonlinear profile of F X , not only is a much longer time required for the drug to reach steady-state (-4-5 ti/2p), but a long wash-out period is also necessary between infusions (ideally 5 to 7 t|/2p). This makes the experimental period too long to be practically performed in sheep during late 184 gestation. Further, dynamic physiological changes that occur in the fetus during the time interval between the two separate infusions may introduce substantial errors into the data due to considerable intra-animal variability. Experiments conducted on separate groups of animals may be devoid of such problems, but the data may suffer from severe inter-animal variability. For exploratory purposes, we attempted to combine the data from the previous maternal infusion study (Chapter 5) and the present fetal infusion study to calculate the placental and nonplacental clearance values. Because unparallel groups of animals are involved with this approach, naive averaged data (i.e. use the overall mean C s s to fit into the equations) instead of the standard two-stage method (i.e. estimate individual clearances then the mean across the animals) was employed for the calculations. The estimates of maternal and fetal placental and nonplacental clearances of FX along with those of other drugs studied in pregnant sheep are presented in Table 6.5. Overall, FX follows the general trend for a number of other drugs in the maternal-fetal unit in that 1 ) fetal weight-normalized placental and nonplacental clearances are typically higher compared to mother, 2) the capacity of fetal nonplacental drug elimination is substantially higher compared to the mother, 3) CLf m has a higher magnitude compared to C L m f for most drugs where estimates are available. The low CLmo for FX is consistent with its long t\/2 p and slow rate of clearance (CLTB) compared to other drugs. The relatively high fetal and maternal placental clearances of FX are in accordance with its .high lipophilicity as indicated by a high octanol/water partition coefficient (log P = 4.05). Kim, 2000, reported a F/M ratio approximating unity in pregnant women dosed with FX, indicating unrestricted diffusion of FX across the human placenta. The finding of a high CLf 0, which is comparable to CLfm in magnitude, is surprising considering our findings that most of the common fetal drug elimination pathways have only 185 CD o c CD 0 or CD co CD CD "55 cn c CD CO CO co 0 CD E o o CM CD "55 CM o o CM CD CO CO Co CO CO CM 00 CO 0 co .a CO CD o CO CO CO cn CD CM CO CO CD "55 o "55 N CO co co co "55 SZ to o o o CM CO E ZJ 0 > u e -a c E _ i O iri CD CD CD LO CO 1-CD LO 1 c i CD CD CO CO CO CO CM CO CO CD U C D CO IT) CO °°. i n co CD o CO co o CO Csi CM CM o 00 CD 0 CJ c CO 1 ro 0 a o 00 o CO CM o' CD CO LO CD CO CM CM CO CM CO CO CO CM o in d 1 -LO CO O d -cf- CD O CO CO CM CO CM CO CO o Tf CO CM CO CM CD n Q CD c -w CD X o 3 C 0 SZ CL o c E 3 "55 u < 0 tz E CD i _ "D >^  _CZ c 0 sz 0 C "55 X o C CJ CO o C o ro "55 co 0 c o ro sz 0 •g E ro i _ CL o CJ o 0 C SZ CL 0 c O or 15 o < o 'o i_ Q_ CD > U U Q 186 limited contribution to the overall clearance of FX in the fetus. Using a modified approach described for labetalol by Yeleswaram et al., 1993b, a different set of placental and nonplacental clearance values for FX (Table 6.5) have been obtained based on the i.v. bolus data (Kim et al., 2004). C L m o obtained from our infusion experiments is much lower than that resulted from the i.v. bolus experiments (5.6 versus 35.6 ml/min/kg), a finding that is consistent with the nonlinear nature of FX pharmacokinetic where inhibition occurs upon prolonged drug administration. Accumulation of FX occurred in the maternal circulation in a disproportional manner secondary to enzyme inhibition, therefore more drug was available for placental transfer, which resulted in higher C L m f and C L | m compared to those estimated from the i.v. bolus data. Our infusion data yielded a C L f 0 value that is substantially higher than that obtained following i.v. bolus administration (156.9 versus 8.5 ml/min/kg). There is no objective reason for this discrepancy, however it is speculated that a change in C L f 0 might have occurred upon long-term infusion, which was associated with a decrease in C L m o compared to acute FX exposure. However, it should be pointed out that in both i.v. infusion and i.v. bolus experiments, the administration of FX to the mother or the fetus took place on two separate occasions. The concentrations of FX achieved following maternal or fetal FX administrations were not identical (e.g. maternal FX concentrations following maternal FX administration were higher compared to those obtained after fetal FX administration) regardless of whether the same (the case for the i.v. bolus studies) or different (the case for the infusion studies) animals were used. This approach of calculating the placental and nonplacental clearance is problematic especially when clearance of FX is concentration dependent. Further experiments that involved simultaneous maternal and fetal infusions, which eliminate variability due to inter-occasional 187 and inter-subject differences, should allow better measurements of actual fetal and maternal nonplacental clearances. However, the technical difficulties of synthesizing sufficient amounts of stable isotope labeled FX (e.g. deuterated FX) for animal dosing must first be overcome before such experiments are possible. Nevertheless, our data suggest that a significant portion of fetal drug elimination remains unexplored for FX. In the previous chapter it was proposed that a possible route of fetal drug elimination might involve uptake into the lungs, which act as a reservoir for many amine drugs including FX due to lyzosomal trapping (Daniel and Wqjcikowski, 1997). Uptake of FX from the blood circulation to the gut may also be possible. Recently our lab found that the AUC of diphenylmethoxyacetic acid, a metabolite of diphenhydramine, was significantly lower when diphenhydramine was given as i.v. bolus versus portal venous (p.v.) bolus injection in nonpregnant sheep (Kumar et al., 1999b). It was also found that the C s s of diphenhydramine was 50% lower when the drug was given by p.v. infusion versus i.v. infusion of the drug (Kumar et al., 1999b). These data suggest the presence of gut extraction of diphenhydramine from the systemic circulation and that this functions as one of the major elimination pathways of diphenhydramine in sheep. However, the question of whether the fetal gut can function as an elimination organ remains to be elucidated. The mechanism of this gut uptake phenomenon is unknown and may involve active transporter systems such as P-glycoprotein. In the previous chapter we discussed the possibility of FX being a P-glycoprotein substrate and hence it would seem reasonable to speculate that gut uptake could potentially function as an elimination route in the fetus. The presence of alternate pathways of biotransformation of FX to other metabolites apart from NFX (see Figure 1.3 in chapter 1) in the fetus is also possible. In humans, FX 188 undergoes phase II metabolism to form glucuronide conjugates which comprised 7-15% of the total radioactivity excreted in urine. In sheep, however, only low levels (< 0.5 % of total administered dose) of FX glucuronide conjugates were identified in adult sheep urine and virtually none was found in amniotic or fetal tracheal fluids (Kim et al., 2004). The later suggests that glucuronidation of FX does not occur in the fetus. FX also undergoes O-dealkylation to form />-trifluoromethylphenol (TFMP) in both humans and rats (Urichuk et al., 1997). The extent of FX metabolism via this pathway is not well understood compared to NFX. However, an in vitro human microsomal study revealed that a significant portion of FX can be O-dealkylated to TFMP (Liu et al., 2001). Table 6.6 lists the literature values for the kinetic parameters for FX TV-demethylation and O-dealkylation in human liver microsomes. The kinetics of TFMP formation share remarkably similar affinity, capacity and intrinsic clearance characteristics to those of R-NFX. These data indicate that O-dealkylation is equally important to 7V-demethylation in the metabolism of FX. It has been shown that CYP2C19 and CYP3A4 are the major cytochrome P450 isoforms responsible for FX O-dealkylation. Although both isoforms are not reported to be expressed in human fetal liver (Treluyer et al., 1997; Hines et al., 2002), CYP3A7, an isoform of cytochrome P450 which is expressed specifically in human fetal liver, is estimated to account for approximately 50% (30-85%) of total P450 content in human fetal hepatic microsomes (Yang et al., 1994). Various studies have shown that both isoforms share similar catalytic activities and substrate specificities (Shimada et al., 1996; de Wildt et al., 1999a; Chen et al., 2000). Therefore, it is very likely that the fetal liver is also capable of biotransforming FX to TFMP. In fact, we have evidence from our preliminary analysis that TFMP was formed and concentrated in urine collected from fetal lambs. Further experiments are in process to 189 CD O c CD CD M— 0 CC s i X I E > 3. o CO CO Q_ o _o ro "55 o o CN lo "S CO c CM co CM uri o c o 16 >^  .c "S E CD T 3 O O CN co "CD cn c CU CC co T t CO 0 0 b X LL z CO o c o jo >-x: co o o CN CD ZD CD oo T t CO CD T t 0. c o ro ro CD OH o TT PH 6 0 O 6 c d <+H o > c o CJ cu -G 6 0 C a 6 0 E o ~5 o E OH E o cd > '3 CT '5b HC H \ 190 delineate the significance of formation of TFMP and its contribution to the overall elimination of FX in fetal and adult sheep. Results from the maternal FX i.v. infusion experiments demonstrated that FX decreased uterine artery blood flow transiently and altered fetal blood gas status, which resulted in fetal hypoxemia and respiratory academia (Morrison et al., 2002). It was postulated that this effect involves FX-mediated inhibition of serotonin reuptake by platelets, resulting in a transient increase in maternal plasma serotonin following onset of maternal drug administration. The present study, however did not reveal much changes in fetal arterial blood gas status when FX was administered directly to the fetus via i.v. infusion. This could be due to high variability between animals such that a larger number of animals may be necessary to show a statistically significant difference (n=5 for fetal infusion compared to n=13 for maternal infusion). Nevertheless, the overall changes in the blood gas variables tended to be less with the fetal FX infusions and also to occur a bit later. For example, PO2 dropped 2.8 mm Hg in 45 minutes after infusion in fetal infusion compared to 5.5 mm Hg in 5 minutes following maternal infusion; O2 saturation fell 12.9 % in 4 hours after fetal infusion compared to 18.5 % in 5 minutes after maternal infusion; PCO2 increased 3 mm Hg in 4 hours following fetal infusion compared to 4.5 mm Hg in 30 minutes post-dose. The delays in the occurrence of these effects suggest that it was the FX in the maternal circulation that elicited the changes via reductions in uterine blood flow as opposed to a similar effect on umbilical blood flow. In addition, the lower maternal FX and NFX concentrations following fetal FX infusion compared to those following maternal FX infusion (Figure 6.14) seem to correlate well with the milder effects on fetal blood gas status following fetal drug infusion. 191 1000 cn c c o C D C u c o O X u_ 100 10 Maternal Infusion Fetal Infusion ~ i 1 r T r -10 12 14 16 18 20 22 24 Time (hour) 100 £ c o zz re C D c o c o o Maternal Infusion Fetal Infusion 10 12 14 Time (hour) 16 20 22 24 FIGURE 6.14 Mean maternal plasma concentration versus time profiles of (A) FX and (B) NFX during the initial 24 hours following maternal or fetal FX i.v. infusion. 192 6.4 Summary In summary, this study has demonstrated that FX exhibits stereoselective disposition in fetal lambs similar to that observed in adult sheep. The differential steady-state concentrations of the FX enantiomers observed in the fetal circulation is likely the consequence of stereoselective protein binding in fetal plasma and differential maternal-to-fetal placental transfer of the enantiomers preceded by stereoselective metabolism in the mother. It appears that the fetal liver has limited metabolic ability to biotransform FX to NFX. Further excretion of FX and NFX in fetal urine also does not make a significant contribution to their overall clearance (CL T B)- Compared to maternal and nonpregnant adult sheep, weight normalized CLTB, CL^nai, and Vd s s are significantly higher in the fetus. Unbound fractions of FX and NFX are also higher in the fetus owing to the lower concentration of plasma protein in fetal plasma. Preliminary calculations for maternal and fetal placental and nonplacental clearances indicate that both placental and nonplacental routes are equally important in the fetal elimination of FX. Lung uptake via lyzosomal trapping, gut uptake by active transporters, and hepatic biotransformation to TFMP are potential routes of fetal nonplacental elimination of FX. However, this is only speculative at this stage and requires experimental data to be conclusively accepted. Alterations in fetal blood gas status were observed following fetal i.v. infusion of FX, although none of these achieved statistical significance due to variability between animals. Nevertheless, the data reveal that peak effects on fetal blood gas status were delayed and had a milder intensity compared to those observed following maternal i.v. infusion of FX, which supports the hypothesis of FX-elicited changes via reductions in uterine artery blood flow as opposed to an effect on diminishing umbilical blood flow. 193 CHAPTER 7 O V E R A L L SUMMARY AND CONCLUSIONS FX is one of the most popular SSRIs used for the treatment of depression. Its use in treating depressed pregnant woman has increased substantially in recent years (Altshuler, 2002). A considerable amount of attention in the literature regarding the use of FX in pregnancy has been devoted to the assessment of birth outcomes following in utero FX exposure while studies that focus on pharmacokinetics during pregnancy are relatively sparse. In terms of the latter studies, none have addressed the stereoselective pharmacokinetics of FX nor the impact of pregnancy-related physiological changes on stereoselective drug disposition (Pohland et al., 1989; Heikkinen et al., 2003). Our laboratory is the first to initiate the study of stereoselective pharmacokinetics and placental transfer of FX and NFX enantiomers using the pregnant sheep model. Previous experiments were conducted using a single i.v. bolus FX administration, and primarily focused on the acute exposure and effects of the drug in both the mother and fetus late in gestation (Kim et al., 2004). In the present study, experiments were conducted to investigate the stereoselective pharmacokinetics of FX and NFX under steady-state condition in nonpregnant sheep and pregnant sheep at late gestation. The exposure of FX and NFX enantiomers to the fetus at steady-state was assessed by evaluating their placental transfer, and distribution in amniotic and fetal tracheal fluid. The pharmacokinetics of FX and NFX in fetal lambs were also studied in an attempt to explore the intrinsic components of fetal drug elimination via nonplacental routes. A summary of all the animal studies performed and the corresponding pharmacokinetic data are presented in Table 7.1. 194 o • o : 53 o 3 ' .<• s E' O CT c ? E d T f co X L L ct CO CO d CO X LL 00 LO CD CO CN CD CN Z ct CD CN c c the ys CD ys _ ro CD a s; TJ T J CO _l -~ CO .c X ^ =S X CT c • ~ CT c O) E o CD E o E o CN CD infusi CD II E CD LO ^ 'co D d infusi c CO CO c CD CN X LL ct LO d d o CO o co d TT X LL co CO O CD O LO_ cn CD CN LO CO O CO LO cn LO S. LO o X LL z LO II c CN o CN d •"cr X LL ct CO d CO T t LO X LL co cn CD cn oo o X LL z ro ™ -i -~ oo £ < =C= v — o E . 2 jj E CN M •£ rri 3 CO O °> HS CO CN O LO LO T f cn CO d CD o LO X LL ct d cn TT CN T f CD CD CD d X LL co c CD .£= ays Q T J _i CO sz X — CT C CT E o E CM 'co o CD p d CD CN d CO T f CN LO T f X LL z ct m CN LO X LL z co LO II 00 00 CO TT CM CO T f LO CN CD CO TT X LL ct CN r-csi co CM CN CM CD T f CD X LL . CO O) o E -s CM CD d CC CM CD CD O CD CN LO CO CO d CO >, ro TJ £ : ro c CD — 55 " CD CD o3 sq CD - ~ o. ro i i £ ro * o o a> S >• I B Q. CO CD > CD — CO CD . J CO c CL CO CD C CD CJ) x : CD CO Q . ~ - - < co co CL : ro S -CO CD >-•£ = ™ ^ ro 2 55 § ) CD 55 55 CL Q. co CL a . ro . c O i - Tj C -2> TO « " S o ro co 1 / 3 ro 0_ H ro 55 CO k-0) a ra O i,5 a S5 T3 • -.b JO RA T J "§ racOcS CO ^ £ CL LL 3 co 195 In the nonpregnant animal study, enantiomers of FX and NFX were found to exhibit significant pharmacokinetic differences under steady-state conditions, confirming the observations made in previous i.v. bolus experiments (Kim et al, 2004). For FX, the R-enantiomer has a lower C s s level but a longer t|/2p than the S-enantiomer while for NFX, both the R- and S-enantiomers have similar C s s levels but t|/2p is slightly longer for R-NFX compared to S-NFX. FX and NFX also display stereospecific protein binding, with free fractions of the R-enantiomers that are about 2 to 3-fold higher than those of the S-enantiomers. In comparison to humans, quantitative differences exist but qualitatively the overall stereoselective disposition of FX and NFX are similar between sheep and humans. Using the literature reported value for hepatic blood flow in sheep, CLj n t, a measurement of metabolic enzyme activity, was estimated for each FX enantiomers based on the present sheep data. It was found that CLj n t of R-FX is about 50% higher than that of S-FX, which agrees with the findings from in vitro human microsomal metabolism studies (Stevens and Wrighton, 1993; Ring et al., 2001). The difference in CLj m between FX enantiomers indicates that in addition to stereoselective protein binding stereoselective metabolism also plays a pivotal role in the overall stereoselective disposition of FX. Comparison of pharmacokinetic parameters from the present infusion experiments to those from the previous i.v. bolus experiments (Kim et al., 2004) reveals a disproportional change in pharmacokinetics of FX in sheep upon long term administration, characterized by a decrease in CLTB and increase in dose normalized AUC values. Interestingly, stereoselectivity is also evident in this dose-related phenomenon, with S-FX being more affected compared to R-FX. Similar dose-related changes in FX pharmacokinetics has been reported in humans previously (Altamura et al., 1994). It is generally believed this nonlinear nature of FX 196 pharmacokinetics is primarily due to CYP2D6 inhibition (Stevens and Wrighton, 1993; Ring et al, 2001), which infers not only that the source of nonlinearity observed in sheep is due to enzyme inhibition but also that sheep may possess enzyme(s) that are analogous to those in humans. Taken together the data from the i.v. infusion and i.v. bolus experiments (Kim et al., 2004), the stereoselective disposition of FX may involve the following mechanisms: 1) differential protein binding, 2) stereoselective metabolism, and 3) stereoselective enzyme inhibition. The pregnant sheep study demonstrates that the stereoselective pharmacokinetics of FX and NFX under steady-state conditions are similar to those observed in nonpregnant animals. No remarkable changes in the S/R ratio of FX and NFX were observed. From the previous single i.v. bolus experiments (Kim, 2000), clearance of FX was observed to be higher in pregnant sheep compared to nonpregnant animals. Similar observations have been made in humans and CYP2D6 induction is likely the cause of increased drug clearance during pregnancy (Wadelius et al., 1997). However, this increase in clearance was not observed in our long-term infusion experiments. Instead, a decrease in clearance was found. It is speculated that the potent enzyme inhibitory effects of FX and NFX may have abolished any increased enzyme activity during pregnancy under steady-state conditions. The findings of decreased C L T B of FX in pregnant sheep at steady-state together with the increased dextromethorphan/dextrorphan metabolic ratio in CYP2D6 PMs during pregnancy (Wadelius et al., 1997) suggests that the activity of CYP3A4, an enzyme secondary to CYP2D6 that is involved in the metabolism of both FX and dextromethorphan, is possibly attenuated during pregnancy. Future studies focused on the effect of pregnancy on the expression of drug metabolizing enzymes are imperative. Ex vivo equilibrium dialysis studies with sheep 197 plasma showed no difference in protein binding between pregnant and nonpregnant sheep for either FX or NFX. This finding is consistent with other reports regarding the lack of significant change in cti-acid glycoprotein concentration during pregnancy (Wood and Wood, 1981; Krauer et al., 1984). However, the relationship between unbound fraction and concentration of the binding protein does not always correlate as exemplified by the alteration of binding of other ai-acid glycoprotein bound basic compounds like lidocaine and propranolol during pregnancy (Wood and Wood, 1981). Analysis of fetal plasma from the 8-day maternal FX infusion study for FX and NFX concentrations revealed that both the parent and metabolite exhibit a moderate degree of placental transfer upon chronic administration, with a F/M AUC ratio of about 0.40 and 0.33, respectively. The diffusibility of a drug across biological membranes is usually well predicted based on its physicochemical properties. Correlation analyses between F/M ratios and 1) partition coefficient, 2) unionized fraction (estimated from pKa), or 3) molecular weight from 37 different compounds (including FX and NFX), however, fail to reveal any significant relationships, implying that physicochemical properties alone cannot predict the ultimate extent of placental drug transfer. When the concentration is corrected for the difference in free fraction between the maternal and fetal compartment, the F/M ratio still remains less than unity, which suggests that other nonplacental routes of elimination for FX are present in the fetus. Analyses of amniotic and fetal tracheal fluids show that excretion of FX and NFX into these fluids result in levels not exceeding those in fetal plasma and significant accumulations are therefore unlikely. There are possibilities for fetal renal excretion, active placental transport, fetal metabolism, and placental metabolism although previous in vivo and in vitro experiments (Kim, 2000) suggest that the latter elimination 198 mechanism is probably of minor significance. In comparison to humans, the extent of placental drug transfer in sheep is generally less extensive, which probably due to the lower permeability characteristics of the sheep placenta. Attempts were made to draw correlations between animal and human placental drug transfer data. Of the nineteen compounds for which placental transfer has been studied in both sheep and humans, fourteen show a significant correlation for F/M ratios between these two species. While this correlation needs to be validated with more data, and the reasons for the other five compounds not following the correlation requires further elaboration, there appears to be an opportunity to establish a model for animal-to-human extrapolation in terms of placental transfer. Comparison of F/M ratios between enantiomers was also made. The F/M ratio does not differ between R-FX and S-FX, while the ratio is slightly higher (-25%) for S-NFX compared to R-NFX, indicating that NFX distributes between the mother and the fetus in a stereoselective manner. A clinically important consideration for this difference in F/M ratio lies in the fact that S-NFX is 20-times more active than R-NFX in terms of 5-HT reuptake inhibition (Wong et al., 1990), and hence the fetus is actually exposed to more pharmacological active species than one would normally predict based on racemic data. The mechanism behind this difference in S/R ratio between the mother and the fetus remains largely unexplored, but may involve stereoselective fetal drug elimination or stereoselective carrier mediated transfer (e.g. P-glycoprotein). In the last experiment, the effect of FX on fetal blood gas status, as well as the stereoselective pharmacokinetics of FX and NFX enantiomers were studied following direct fetal FX administration via i.v. infusion. Alterations of fetal blood gas status, characterized by decreases in PO2 and O2 saturation, and increases in PCO2 and pH, were noted following 199 fetal i.v. infusion of FX, and persisted for up to 24 hours before returning to pre-infusion levels. Due to variability between animals, the majority of these effects did not reach statistical significance. Nonetheless, the data reveal that the maximal effects on fetal.blood gas status were slightly delayed and less intensive compared to those observed following maternal i.v. infusion of FX, which strongly suggests that it was the maternal circulating FX that elicited these changes via reductions in uterine artery blood flow as opposed to an effect on umbilical blood flow. The stereoselective pharmacokinetics of FX and NFX in fetal lambs were similar to those observed in adult sheep, but the extent of stereoselectivity in terms of the S/R ratio is less. The differential steady-state concentrations of the FX enantiomers observed in the fetal circulation is likely due to stereoselective protein binding in fetal plasma and differential maternal-to-fetal placental transfer of the enantiomers preceded by maternal stereoselective metabolism. Based on the F/M ratios for NFX (-0.62) following fetal FX infusion, it appears that the fetus has limited metabolic capability to biotransform FX to NFX, if any. It is not fully known, however, whether other metabolic pathways such as the formation of TFMP via O-dealkylation are functional in the fetus. The elucidation of the possible formation of TFMP in the fetal liver may provide important new information on the ontogeny of drug metabolizing enzymes during the fetal period. Renal excretion of FX and NFX do not appear to make a significant contribution to the overall clearance in the fetus (<1% of CLTB)- In comparison to maternal and nonpregnant adult sheep, weight normalized C L T B , CL r e nai, and Vd s s are significantly higher in the fetus. Unbound fractions of FX and NFX are also higher in the fetus owing to the lower concentration of plasma protein in fetal plasma. In attempt to quantitatively assess the intrinsic capability of the fetus to eliminate FX, maternal and fetal placental and nonplacental 200 clearances were estimated based on the 2-compartment model for maternal-fetal unit (Szeto et al., 1982b). The results indicate that both placental and nonplacental routes are equally important in fetal elimination of FX. It is apparent that other elimination routes are present in the fetus but as of now these remain poorly understood. Among these, lung uptake via lyzosomal trapping, gut uptake by active transporters, and hepatic biotransformation to TFMP are potential routes of fetal nonplacental elimination of FX, but none of these has been systemically examined. Finally, the importance of stereoselectivity for the study of these nonplacental clearance mechanisms cannot be overemphasized. 201 REFERENCES Abebe-Campino G, Offer D, Stahl B and Merlob P (2002) Cardiac arrhythmia in a newborn infant associated with fluoxetine use during pregnancy. Ann Pharmacother 36:533-534. Aboul-Enein HY and Wainer IW (1997) The impact of stereochemistry on drug development and use. Wiley, New York. Addis A and Koren G (2000) Safety of fluoxetine during the first trimester of pregnancy: a meta-analytical review of epidemiological studies. Psychol Med 30:89-94. Altamura AC, Moro AR and Percudani M (1994) Clinical pharmacokinetics of fluoxetine. Clin Pharmacokinet 26:201-214. Altshuler LL (2002) The use of SSRIs in depressive disorders specific to women. J Clin Psychiatry 63 Suppl 7:3-8. Anderson DF, Phernetton TM and Rankin JH (1980a) The measurement of placental drug clearance in near-term sheep: indomethacin. J Pharmacol Exp Ther 213:100-104. Anderson DF, Phernetton TM and Rankin JH (1980b) Prediction of fetal drug concentrations. Am J Obstet Gynecol 137:735-738. Apatu RS and Barnes RJ (1991) Blood flow to and the metabolism of glucose and lactate by the liver in vivo in fetal, newborn and adult sheep. J Physiol 436:431-447. Ariens EJ (1984) Stereochemistry, a basis for sophisticated nonsense in pharmacokinetics and clinical pharmacology. Eur J Clin Pharmacol 26:663-668. Arvela P, Jouppila R, Kauppila A, Pakarinen A, Pelkonen O and Tuimala R (1983) Placental transfer and hormonal effects of metoclopramide. Eur J Clin Pharmacol 24:345-348. Banzai M, Sato S, Tezuka N, Komiya H, Chimura T and Hiroi M (1995) Placental transfer of lidocaine hydrochloride after prolonged continuous maternal intravenous administration. Can J Anaesth 42:338-340. Beasley CM, Jr., Nilsson ME, Koke SC and Gonzales JS (2000) Efficacy, adverse events, and treatment discontinuations in fluoxetine clinical studies of major depression: a meta-analysis of the 20-mg/day dose. J Clin Psychiatry 61:722-728. Belpaire FM, Wynant P, Van Trappen P, Dhont M, Verstraete A and Bogaert MG (1995) Protein binding of propranolol and verapamil enantiomers in maternal and foetal serum. Br J Clin Pharmacol 39:190-193. 202 Bergstrom RF, Lemberger L, Farid NA and Wolen RL (1988) Clinical pharmacology and pharmacokinetics of fluoxetine: a review. Br J Psychiatry Suppl'Al-50. Bhatia SC and Bhatia SK (1999) Depression in women: diagnostic and treatment considerations. Am Fam Physician 60:225-234, 239-240. Biehl D, Shnider SM, Levinson G and Callender K (1978) Placental transfer of lidocaine: effects of fetal acidosis. Anesthesiology 48:409-412. Bourget P, Roulot C and Fernandez H (1995) Models for placental transfer studies of drugs. Clin Pharmacokinet 28:161-180. Brien JF, Clarke DW, Richardson B and Patrick J (1985) Disposition of ethanol in maternal blood, fetal blood, and amniotic fluid of third-trimester pregnant ewes. Am J Obstet Gynecol 152:583-590. Brosen K and Skjelbo E (1991) Fluoxetine and norfluoxetine are potent inhibitors of P450IID6—the source of the sparteine/debrisoquine oxidation polymorphism. Br J Clin Pharmacol 32:136-137. Burchell B, Coughtrie M, Jackson M, Harding D, Fournel-Gigleux S, Leakey J and Hume R (1989) Development of human liver UDP-glucuronosyltransferases. Dev Pharmacol Ther 13:70-77. Burchfield DJ, Lucas VW, Abrams RM, Miller RL and DeVane CL (1991) Disposition and pharmacodynamics of methamphetamine in pregnant sheep. Jama 265:1968-1973. Burt VK and Stein K (2002) Epidemiology of depression throughout the female life cycle. J Clin Psychiatry 63 Suppl 7:9-15. Campbell DB (1990) Stereoselectivity in clinical pharmacokinetics and drug development. Eur J Drug Metab Pharmacokinet 15:109-125. Capece BP, Perez R, Andaluz A, Perez F, Garcia F, Castells G, Arboix M and Cristofol C (2002) Placental transfer of albendazole sulphoxide enantiomers in sheep. Vet J 163:155-160. Capon DA, Bochner F, Kerry N, Mikus G, Danz C and Somogyi AA (1996) The influence of CYP2D6 polymorphism and quinidine on the disposition and antitussive effect of dextromethorphan in humans. Clin Pharmacol Ther 60:295-307. Catterson ML and Preskorn SH (1996) Pharmacokinetics of selective serotonin reuptake inhibitors: clinical relevance. Pharmacol Toxicol 78:203-208. Chambers CD, Johnson KA, Dick LM, Felix RJ and Jones KL (1996) Birth outcomes in pregnant women taking fluoxetine. N Engl J Med 335:1010-1015. 203 Chen H, Fantel AG and Juchau MR (2000) Catalysis of the 4-hydroxylation of retinoic acids by cyp3a7 in human fetal hepatic tissues. Drug Metab Dispos 28:1051-1057. Cheng HY and Jusko WJ (1991) Mean residence time of drugs in pharmacokinetic systems with linear distribution, linear or nonlinear elimination, and noninstantaneous input. J Pharm Sci 80:1005-1006. Ching MS, Morgan DJ, Mihaly GW, Hardy KJ and Smallwood RA (1986) Placental transfer of omeprazole in maternal and fetal sheep. Dev Pharmacol Ther 9:323-331. Chow AW and Jewesson PJ (1985) Pharmacokinetics and safety of antimicrobial agents during pregnancy. Rev Infect Dis 7:287-313. Cohen LS, Heller VL, Bailey JW, Grush L, Ablon JS and Bouffard SM (2000) Birth outcomes following prenatal exposure to fluoxetine. Biol Psychiatry 48:996-1000. Cook CS, Karim A and Sollman P (1982) Stereoselectivity in the metabolism of disopyramide enantiomers in rat and dog. Drug Metab Dispos 10:116-121. Cooper PJ and Murray L (1998) Postnatal depression. Bmj 316:1884-1886. Cox PJ, Farmer PB, Jarman M, Kinas RW and Stec WJ (1978) Stereoselectivity in the metabolism of the enantiomers of cyclophosphamide in mice, rats, and rabbits. Drug Metab Dispos 6:617-622. Crooks PA and Godin CS (1988) N-methylation of nicotine enantiomers by human liver cytosol. J Pharm Pharmacol 40:153-154. Cumming ME, Ong BY, Wade JG and Sitar DS (1984) Maternal and fetal ethanol pharmacokinetics and cardiovascular responses in near-term pregnant sheep. Can J Physiol Pharmacol 62:1435-1439. Czuba MA, Morgan DJ, Ching MS, Mihaly GW, Ghabrial H, Hardy KJ and Smallwood RA (1991) Disposition of the diastereoisomers quinine and quinidine in the ovine fetus. J Pharm Sci 80:445-448. Daniel WA and Wojcikowski J (1997) Contribution of lysosomal trapping to the total tissue uptake of psychotropic drugs. Pharmacol Toxicol 80:62-68. Davison JM and Noble MC (1981) Serial changes in 24 hour creatinine clearance during normal menstrual cycles and the first trimester of pregnancy. Br J Obstet Gynaecol 88:10-17. de Wildt SN, Kearns GL, Leeder JS and van den Anker JN (1999a) Cytochrome P450 3A: ontogeny and drug disposition. Clin Pharmacokinet 37:485-505. 204 de Wildt SN, Kearns GL, Leeder JS and van den Anker JN (1999b) Glucuronidation in humans. Pharmacogenetic and developmental aspects. Clin Pharmacokinet 36:439-452. DeVane CL (1994) Pharmacokinetics of the newer antidepressants: clinical relevance. Am J Med 97:13S-23S. DeVane CL, Burchfield DJ, Abrams RM, Miller RL and Braun SB (1991) Disposition of cocaine in pregnant sheep. I. Pharmacokinetics. Dev Pharmacol Titer 16:123-129. Doberczak TM, Kandall SR and Friedmann P (1993) Relationship between maternal methadone dosage, maternal-neonatal methadone levels, and neonatal withdrawal. Obstet Gynecol 81:936-940. Eap CB, Bondolfi G, Zullino D, Savary-Cosendai L, Powell-Golay K, Kosel M and Baumann P (2001) Concentrations of the enantiomers of fluoxetine and norfluoxetine after multiple doses of fluoxetine in cytochrome P4502D6 poor and extensive metabolizers. J Clin Psychopharmacol 21:330-334. Eichelbaum M, Mineshita S, Ohnhaus EE and Zekorn C (1986) The influence of enzyme induction on polymorphic sparteine oxidation. Br J Clin Pharmacol 22:49-53. Elliott BD, Langer O, Schenker S and Johnson RF (1991) Insignificant transfer of glyburide occurs across the human placenta. Am J Obstet Gynecol 165:807-812. Elliott BD, Schenker S, Langer O, Johnson R and Prihoda T (1994) Comparative placental transport of oral hypoglycemic agents in humans: a model of human placental drug transfer. Am J Obstet Gynecol 171:653-660. Elliott R (2001) Court dismisses Eli Lilly's case against generic drug makers. Can HIV AIDS Policy Law Rev 6:60. Erkkola R, Lammintausta R, Liukko P and Anttila M (1982) Transfer of propranolol and sotalol across the human placenta. Their effect on maternal and fetal plasma renin activity. Acta Obstet Gynecol Scand 61:31-34. Faber J and Thornburg KL (1983) Placental physiology : structure and function of fetomaternal exchange. Raven Press, New York. Fjordside L, Jeppesen U, Eap CB, Powell K, Baumann P and Brosen K (1999) The stereoselective metabolism of fluoxetine in poor and extensive metabolizers of sparteine. Pharmacogenetics 9:55-60. Florey K, Brewer GA and Academy of Pharmaceutical Sciences. Pharmaceutical Analysis and Control Section (1990) Analytical profiles of drug substances. Academic Press, New York. 205 Fujimoto S, Akahane M and Sakai A (1986) Concentrations of ritodrine hydrochloride in maternal and fetal serum and amniotic fluid following intravenous administration in late pregnancy. Eur J Obstet Gynecol Reprod Biol 23:145-152. Fujimoto S, Akahane M, Uzuki K, Inagawa A, Sakai K and Sakai A (1984) Placental transfer of ritodrine hydrochloride in sheep. Int J Gynaecol Obstet 22:269-274. Fuller RW, Snoddy HD, Krushinski JH and Robertson DW (1992) Comparison of norfluoxetine enantiomers as serotonin uptake inhibitors in vivo. Neuropharmacology 31:997-1000. Ganapathy V, Prasad PD, Ganapathy ME and Leibach FH (2000) Placental transporters relevant to drug distribution across the maternal-fetal interface. J Pharmacol Exp Ther 294:413-420. Garland M (1998) Pharmacology of drug transfer across the placenta. Obstet Gynecol Clin North Am 25:21-42. Gerdin E, Rane A and Lindberg B (1990) Transplacental transfer of morphine in man. J PerinatMed 18:305-312. Gibaldi M and Perrier D (1982) Pharmacokinetics. Dekker, New York. Gilbert WM and Brace RA (1989) The missing link in amniotic fluid volume regulation: intramembranous absorption. Obstet Gynecol 74:748-754. Gilmore DA, Gal J, Gerber JG and Nies AS (1992) Age and gender influence the stereoselective pharmacokinetics of propranolol. J Pharmacol Exp Ther 261:1181-1186. Goldstein DJ, Corbin LA and Sundell KL (1997) Effects of first-trimester fluoxetine exposure on the newborn. Obstet Gynecol 89:713-718. Gomez Bellver MJ, Garcia Sanchez MJ, Alonso Gonzalez AC, Santos Buelga D and Dominguez-Gil A (1993) Plasma protein binding kinetics of valproic acid over a broad dosage range: therapeutic implications. J Clin Pharm Ther 18:191-197. Goncalves PV, Matthes Ado C, Da Cunha SP and Lanchote VL (2002) Enantioselectivity in the steady-state pharmacokinetics and transplacental distribution of pindolol at delivery in pregnancy-induced hypertension. Chirality 14:683-687. Gordon JD, Riggs KW, Rurak DW, Kwan E, Hall C and Abbott FS (1995) The pharmacokinetics of valproic acid in pregnant sheep after maternal and fetal intravenous bolus administration. Drug Metab Dispos 23:1383-1389. Gram L (1994) Fluoxetine. N Engl J Med 331:1354-1361. 206 Greig PC, Massmann GA, Demarest KT, Weglein RC, Holland ML and Figueroa JP (1993) Maternal and fetal cardiovascular effects and placental transfer of the oxytocin antagonist atosiban in late-gestation pregnant sheep. Am J Obstet Gynecol 169:897-902. Gresham EL, Rankin JH, Makowski EL, Meschia G and Battaglia FC (1972) An evaluation of fetal renal function in a chronic sheep preparation. J Clin Invest 51:149-156. Gross AS, Heuer B and Eichelbaum M (1988) Stereoselective protein binding of verapamil enantiomers. Biochem Pharmacol 37:4623-4627. Grush LR and Cohen LS (1998) Treatment of depression during pregnancy: balancing the risks. Harv Rev Psychiatry 6:105-109. Gupta SK, Atkinson L, Tu T and Longstreth JA (1995) Age and gender related changes in stereoselective pharmacokinetics and pharmacodynamics of verapamil and norverapamil. Br J Clin Pharmacol 40:325-331. Hakkola J, Pasanen M, Purkunen R, Saarikoski S, Pelkonen O, Maenpaa J, Rane A and Raunio H (1994) Expression of xenobiotic-metabolizing cytochrome P450 forms in human adult and fetal liver. Biochem Pharmacol 48:59-64. Hamelin BA, Turgeon J, Vallee F, Belanger PM, Paquet F and LeBel M (1996) The disposition of fluoxetine but not sertraline is altered in poor metabolizers of debrisoquin. Clin Pharmacol Ther 60:512-521. Hamshaw-Thomas A, Rogerson N and Reynolds F (1984) Transfer of bupivacaine, lignocaine and pethidine across the rabbit placenta: influence of maternal protein binding and fetal flow. Placenta 5:61-70. Hansch C, Sammes PG and Taylor JB (1990) Comprehensive medicinal chemistry : the rational design, mechanistic study & therapeutic application of chemical compounds. Pergamon Press, Oxford ; New York. Haram K, Augensen K and Elsayed S (1983) Serum protein pattern in normal pregnancy with special reference to acute-phase reactants. Br J Obstet Gynaecol 90:139-145. Harris RZ, Benet LZ and Schwartz JB (1995) Gender effects in pharmacokinetics and pharmacodynamics. Drugs 50:222-239. Harris WH and Van Petten GR (1981) Placental transfer of indomethacin in the rabbit and sheep. Can J Physiol Pharmacol 59:342-346. Harrison LI and Gibaldi M (1976) Influence of cholestasis on drug elimination: pharmacokinetics. JPharm Sci 65:1346-1348. Heikkinen T, Ekblad U , Kero P, Ekblad S and Laine K (2002) Citalopram in pregnancy and lactation. Clin Pharmacol Ther 72:184-191. 207 Heikkinen T, Ekblad U, Palo P and Laine K (2003) Pharmacokinetics of fluoxetine and norfluoxetine in pregnancy and lactation. Clin Pharmacol Ther 73:330-337. Hendrick V and Altshuler L (2002) Management of major depression during pregnancy. Am J Psychiatry 159:1667-1673. Hiemke C and Hartter S (2000) Pharmacokinetics of selective serotonin reuptake inhibitors. Pharmacol Ther 85:11 -28. Hill LM and Malkasian GD, Jr. (1979) The use of quinidine sulfate throughout pregnancy. Obstet Gynecol 54:366-368. Hill MD and Abramson FP (1988) The significance of plasma protein binding on the fetal/maternal distribution of drugs at steady-state. Clin Pharmacokinet 14:156-170. Hines RN and McCarver DG (2002) The ontogeny of human drug-metabolizing enzymes: phase I oxidative enzymes. J Pharmacol Exp Ther 300:355-360. Hoffman S and Hatch MC (1996) Stress, social support and pregnancy outcome: a reassessment based on recent research. Paediatr Perinat Epidemiol 10:380-405. Hogstedt S, Lindberg B, Peng DR, Regardh CG and Rane A (1985) Pregnancy-induced increase in metoprolol metabolism. Clin Pharmacol Ther 37:688-692. Holladay JW, Dewey MJ and Yoo SD (1998) Pharmacokinetics and antidepressant activity of fluoxetine in transgenic mice with elevated serum alpha-1-acid glycoprotein levels. Drug Metab Dispos 26:20-24. Hooper WD and Qing MS (1990) The influence of age and gender on the stereoselective metabolism and pharmacokinetics of mephobarbital in humans. Clin Pharmacol Ther 48:633-640. Hostetter A, Stowe ZN, Strader JR, Jr., McLaughlin E and Llewellyn A (2000) Dose of selective serotonin uptake inhibitors across pregnancy: clinical implications. Depress Anxiety 11:51-57. Houston JB (1981) Drug metabolite kinetics. Pharmacol Ther 15:521-552. Howe JP, McGowan WA, Moore J, McCaughey W and Dundee JW (1981) The placental transfer of cimetidine. Anaesthesia 36:371-375. Hsyu PH and Giacomini KM (1985) Stereoselective renal clearance of pindolol in humans. J Clin Invest 76:1720-1726. Jacqz-Aigrain E, Funck-Brentano C and Cresteil T (1993) CYP2D6- and CYP3A-dependent metabolism of dextromethorphan in humans. Pharmacogenetics 3:197-204. 208 Jamali F (1988) Pharmacokinetics of enantiomers of chiral non-steroidal anti-inflammatory drugs. Eur J Drug Metab Pharmacokinet 13:1 -9. Jamali F, Mehvar R and Pasutto FM (1989) Enantioselective aspects of drug action and disposition: therapeutic pitfalls. J Pharm Sci 78:695-715. Jeffries WS and Bochner F (1988) The effect of pregnancy on drug pharmacokinetics. Med J Aust 149:675-677. Jori A, De Ponte P and Caccia S (1978) Stereoselective disposition of fenfluramine enantiomers in the rat. Xenobiotica 8:583-588. Juchau MR (1980) Drug biotransformation in the placenta. Pharmacol Ther 8:501-524. Juchau MR, Chao ST and Omiecinski CJ (1980) Drug metabolism by the human fetus. Clin Pharmacokinet 5:320-339. Kanto J, Sjovall S, Erkkola R, Himberg JJ and Kangas L (1983) Placental transfer and maternal midazolam kinetics. Clin Pharmacol Ther 33:786-791. Kashuba AD, Nafziger AN, Kearns GL, Leeder JS, Gotschall R, Gaedigk A and Bertino JS, Jr. (1999) Limitations of dextromethorphan N-demethylation as a measure of CYP3A activity. Pharmacogenetics 9:453-462. Kashuba AD, Nafziger AN, Kearns GL, Leeder JS, Shirey CS, Gotschall R, Gaedigk A and Bertino JS, Jr. (1998) Quantification of intraindividual variability and the influence of menstrual cycle phase on CYP2D6 activity as measured by dextromethorphan phenotyping. Pharmacogenetics 8:403-410. Katz ML and Bergman EN (1969) Simultaneous measurements of hepatic and portal venous blood flow in the sheep and dog. Am J Physiol 216:946-952. Kessler RC, McGonagle KA, Swartz M, Blazer DG and Nelson CB (1993) Sex and depression in the National Comorbidity Survey. I: Lifetime prevalence, chronicity and recurrence. J Affect Disord 29:85-96. Kim J (2000) Pharmacokinetics and Pharmacodynamics of selective serotonin reuptake inhibitors, fluoxetine and paroxetine, during pregnancy and the nursing period. Ph.D. Thesis, The University of British Columbia, Vancouver, Canada. Kim J, Riggs KW and Rurak DW (2004) Stereoselective pharmacokinetics of fluoxetine and norfluoxetine enantiomers in pregnant sheep. Drug Metab Dispos 32:1-10. Kochak GM and Rakhit A (1987) Determination of metabolite pharmacokinetics for orally administered prodrugs. Drug Metab Dispos 15:456-459. Krauer B (1984) Drugs and pregnancy : maternal drug handling—fetal drug exposure. Academic Press, London ; Orlando. 209 Krauer B, Dayer P and Anner R (1984) Changes in serum albumin and alpha 1 -acid glycoprotein concentrations during pregnancy: an analysis of fetal-maternal pairs. Br J Obstet Gynaecol 91:875-881. Krauer B, Krauer F and Hytten FE (1980) Drug disposition and pharmacokinetics in the maternal-placental-fetal unit. Pharmacol Ther 10:301-328. Krauer B, Nau H, Dayer P, Bischof P and Anner R (1986) Serum protein binding of diazepam and propranolol in the feto-maternal unit from early to late pregnancy. Br J Obstet Gynaecol 93:322-328. Krishna R, Riggs KW, Kwan E, Wong H, Szeitz A, Walker MP and Rurak DW (2002) Clearance and disposition of indometacin in chronically instrumented fetal lambs following a 3-day continuous intravenous infusion. JPharm Pharmacol 54:801-808. Kumar S, Riggs KW and Rurak DW (1999a) Comparative formation, distribution, and elimination kinetics of diphenylmethoxyacetic acid (a diphenhydramine metabolite) in maternal and fetal sheep. Drug Metab Dispos 27:463-470. Kumar S, Riggs KW and Rurak DW (1999b) Role of the liver and gut in systemic diphenhydramine clearance in adult nonpregnant sheep. Drug Metab Dispos 27:297-302. Kumar S, Tonn GR, Kwan E, Hall C, Riggs KW, Axelson JE and Rurak DW (1997) Estimation of transplacental and nonplacental diphenhydramine clearances in the fetal lamb: the impact of fetal first-pass hepatic drug uptake. J Pharmacol Exp Ther 282:617-632. Kumar S, Wong H, Yeung SA, Riggs KW, Abbott FS and Rurak DW (2000a) Disposition of valproic acid in maternal, fetal, and newborn sheep. I: placental transfer, plasma protein binding, and clearance. Drug Metab Dispos 28:845-856. Kumar S, Wong H, Yeung SA, Riggs KW, Abbott FS and Rurak DW (2000b) Disposition of valproic acid in maternal, fetal, and newborn sheep. II: metabolism and renal elimination. Drug Metab Dispos 28:857-864. Lampela ES, Nuutinen LH, Ala-Kokko TI, Parikka RM, Laitinen RS, Jouppila PI and Vahakangas KH (1999) Placental transfer of sulindac, sulindac sulfide, and indomethacin in a human placental perfusion model. Am J Obstet Gynecol 180:174-180. Lane EA and Levy RH (1981) Metabolite to parent drug concentration ratio as a function of parent drug extraction ratio: cases of nonportal route of administration. J Pharmacokinet Biopharm 9:489-496. Lankas GR, Wise LD, Cartwright ME, Pippert T and Umbenhauer DR (1998) Placental P-glycoprotein deficiency enhances susceptibility to chemically induced birth defects in mice. Reprod Toxicol 12:457-463. 210 Levy DM, Williams OA, Magides AD and Reilly CS (1994) Gastric emptying is delayed at 8-12 weeks' gestation. Br J Anaesth 73:237-238. Levy G, Procknal JA and Garrettson LK (1975) Distribution of salicylate between neonatal and maternal serum at diffusion equilibrium. Clin Pharmacol Ther 18:210-214. Lin JH and Yamazaki M (2003) Role of P-glycoprotein in pharmacokinetics: clinical implications. Clin Pharmacokinet 42:59-98. Lin YS, Lockwood GF, Graham MA, Brian WR, Loi CM, Dobrinska MR, Shen DD, Watkins PB, Wilkinson GR, Kharasch ED and Thummel KE (2001) In-vivo phenotyping for CYP3A by a single-point determination of midazolam plasma concentration. Pharmacogenetics 11:781-791. Lipshitz J, Yau MK, Meyer MC, Ahokas RA, Maduska AL, Whybrew WD, Anderson GD, Morrison JC and Schneider J (1981) Hexoprenaline pharmacokinetics in pregnant and nonpregnant sheep. Res Commun Chem Pathol Pharmacol 34:3-16. Little BB (1999) Pharmacokinetics during pregnancy: evidence-based maternal dose formulation. Obstet Gynecol 93:858-868. Liu HS, Chang JS, Juo JS, Shieh JJ, Shih MT, Ho ST, Hou CC, Wong CS, Chang CL and Peng TY (1985) Ranitidine as an antacid in elective cesarean delivery and its placenta transfer—a preliminary report. Ma Zui Xue Za Zhi 23:18-21. Liu ZQ, Zhu B, Tan YF, Tan ZR, Wang LS, Huang SL, Shu Y and Zhou HH (2001) O-Dealkylation of fluoxetine in relation to CYP2C19 gene dose and involvement of CYP3A4 in human liver microsomes. J Pharmacol Exp Ther 300:105-111. Loebstein R and Koren G (1997) Pregnancy outcome and neurodevelopment of children exposed in utero to psychoactive drugs: the Motherisk experience. J Psychiatry Neurosci 22:192-196. Loebstein R, Lalkin A and Koren G (1997) Pharmacokinetic changes during pregnancy and their clinical relevance. Clin Pharmacokinet 33:328-343. Margolis JM, O'Donnell JP, Mankowski DC, Ekins S and Obach RS (2000) (R)-, (S)-, and racemic fluoxetine N-demethylation by human cytochrome P450 enzymes. Drug Metab Dispos 28:1187-1191. McCarver DG and Hines RN (2002) The ontogeny of human drug-metabolizing enzymes: phase II conjugation enzymes and regulatory mechanisms. J Pharmacol Exp Ther 300:361-366. Mhanna MJ, Bennet JB, 2nd and Izatt SD (1997) Potential fluoxetine chloride (Prozac) toxicity in a newborn. Pediatrics 100:158-159. 211 Mihaly GW, Jones DB, Morgan DJ, Ching MS, Webster LK, Smallwood RA and Hardy KJ (1983) Placental transfer and renal elimination of cimetidine in maternal and fetal sheep. J Pharmacol Exp Ther 227:441-445. Mihaly GW and Morgan DJ (1984) Placental drug transfer: effects of gestational age and species. Pharmacol Ther 23:253-266. Mihaly GW, Morgan DJ, Marshall AW, Smallwood RA, Cockbain S, MacLellan D and Hardy KJ (1982a) Placental transfer of ranitidine during steady-state infusions of maternal and fetal sheep. J Pharm Sci 71:1008-1010. Mihaly GW, Morgan DJ, Smallwood R and Hardy KJ (1982b) The developing liver: the steady-state disposition of propranolol in pregnant sheep. Hepatology 2:344-349. Min DI, Ku YM, Vichiendilokkul A and Fleckenstein LL (1999) A urine metabolic ratio of dextromethorphan and 3-methoxymorphinan as a probe for CYP3A activity and prediction of cyclosporine clearance in healthy volunteers. Pharmacotherapy 19:753-759. Mirkin BL (1975) Perinatal pharmacology: placental transfer, fetal localization, and neonatal disposition of drugs. Anesthesiology 43:156-170. Mohan CG and Moore JJ (2000) Fluoxetine toxicity in a preterm infant. JPerinatol 20:445-446. Morrison JL, Chien C, Gruber N, Rurak D and Riggs W (2001) Fetal behavioural state changes following maternal fluoxetine infusion in sheep. Brain Res Dev Brain Res 131:47-56. Morrison JL, Chien C, Riggs KW, Gruber N and Rurak D (2002) Effect of maternal fluoxetine administration on uterine blood flow, fetal blood gas status, and growth. PediatrRes 51:433-442. Morrison JL, Riggs KW, Chien C, Gruber N, McMillen IC and Rurak DW (2004) Chronic maternal fluoxetine infusion in pregnant sheep: effects on the maternal and fetal hypothalamic-pituitary-adrenal axes. PediatrRes (in press). Murad SH, Conklin KA, Tabsh KM, Brinkman CR, 3rd, Erkkola R and Nuwayhid B (1981) Atropine and glycopyrrolate: hemodynamic effects and placental transfer in the pregnant ewe. Anesth Analg 60:710-714. Murad SH, Tabsh KM, Conklin KA, Shilyanski G, Ziadlourad F, Kapur PA and Flacke WE (1985) Verapamil: placental transfer and effects on maternal and fetal hemodynamics and atrioventricular conduction in the pregnant ewe. Anesthesiology 62:49-53. Nagai N, Kawakubo T, Kaneko F, Ishii M, Shinohara R, Saito Y, Shimamura H, Ohnishi A and Ogata H (1996) Pharmacokinetics and polymorphic oxidation of dextromethorphan in a Japanese population. Biopharm Drug Dispos 17:421-433. 212 Nation RL (1980) Drug kinetics in childbirth. Clin Pharmacokinet 5:340-364. Nau H, Luck W and Kuhnz W (1984) Decreased serum protein binding of diazepam and its major metabolite in the neonate during the first postnatal week relate to increased free fatty acid levels. Br J Clin Pharmacol 17:92-98. Nau H, Rating D, Koch S, Hauser I and Helge H (1981) Valproic acid and its metabolites: placental transfer, neonatal pharmacokinetics, transfer via mother's milk and clinical status in neonates of epileptic mothers. J Pharmacol Exp Ther 219:768-777. Newport DJ, Hostetter A, Arnold A and Stowe ZN (2002) The treatment of postpartum depression: minimizing infant exposures. J Clin Psychiatry 63 Suppl 7:31-44. Nonacs R and Cohen LS (2002) Depression during pregnancy: diagnosis and treatment options. J Clin Psychiatry 63 Suppl 7:24-30. Nordeng H, Lindemann R, Perminov KV and Reikvam A (2001) Neonatal withdrawal syndrome after in utero exposure to selective serotonin reuptake inhibitors. Acta Paediatr 90:288-291. Nugent CE, Hill M, Hursey T and Abramson FP (1991) Pharmacokinetics and pharmacodynamics of nifedipine in pregnant sheep. Drug Metab Dispos 19:754-757. Nulman I, Rovet J, Stewart DE, Wolpin J, Pace-Asciak P, Shuhaiber S and Koren G (2002) Child development following exposure to tricyclic antidepressants or fluoxetine throughout fetal life: a prospective, controlled study. Am J Psychiatry 159:1889-1895. Oberlander TF, Eckstein Grunau R, Fitzgerald C, Ellwood AL, Misri S, Rurak D and Riggs KW (2002) Prolonged prenatal psychotropic medication exposure alters neonatal acute pain response. Pediatr Res 51:443-453. Ogawa K, Hasegawa S, Udaka Y, Nara K, Iwai S and Oguchi K (2003) Individual difference in the pharmacokinetics of a drug, pravastatin, in healthy subjects. J Clin Pharmacol 43:1268-1273. O'Hara MW (1995) Postpartum depression : causes and consequences. Springer-Verlag, New York. Orr ST, James SA and Blackmore Prince C (2002) Maternal prenatal depressive symptoms and spontaneous preterm births among African-American women in Baltimore, Maryland. Am J Epidemiol 156:797-802. Otton SV, Wu D, Joffe RT, Cheung SW and Sellers EM (1993) Inhibition by fluoxetine of cytochrome P450 2D6 activity. Clin Pharmacol Ther 53:401-409. Pabst PL (2001) How Barr managed to kill Eli Lilly's patent on Prozac. Tissue Eng 7:843-844. 213 Pasanen M (1999) The expression and regulation of drug metabolism in human placenta. Adv DrugDeliv Rev 38:81-97. PDR (2002) Physicians' desk reference. Medical Economics Company Inc., Montvale, NJ. Pelkonen O (1980) Environmental influences on human foetal and placental xenobiotic metabolism. Eur J Clin Pharmacol 18:17-24. Phillips RE, Looareesuwan S, White NJ, Silamut K, Kietinun S and Warrell DA (1986) Quinine pharmacokinetics and toxicity in pregnant and lactating women with falciparum malaria. Br J Clin Pharmacol 21:677-683. Pirhonen JP, Erkkola RU, Ekblad UU and Nyman L (1990) Single dose of nifedipine in normotensive pregnancy: nifedipine concentrations, hemodynamic responses, and uterine and fetal flow velocity waveforms. Obstet Gynecol 76:807-811. Pohland RC, Byrd TK, Hamilton M and Koons JR (1989) Placental transfer and fetal distribution of fluoxetine in the rat. Toxicol Appl Pharmacol 98:198-205. Raimundo S, Fischer J, Eichelbaum M, Griese EU, Schwab M and Zanger UM (2000) Elucidation of the genetic basis of the common 'intermediate metabolizer' phenotype for drug oxidation by CYP2D6. Pharmacogenetics 10:577-581. Rane A and Tomson G (1980) Prenatal and neonatal drug metabolism in man. Eur J Clin Pharmacol 18:9-15. Rau T, Heide R, Bergmann K, Wuttke H, Werner U, Feifel N and Eschenhagen T (2002) Effect of the CYP2D6 genotype on metoprolol metabolism persists during long-term treatment. Pharmacogenetics 12:465-472. Reynolds F and Knott C (1989) Pharmacokinetics in pregnancy and placental drug transfer. OxfRev Reprod Biol 11:389-449. Riggs KW, Axelson JE, Gruber NC, McErlane BA, McMorland GH and Rurak DW (1988) Metoclopramide pharmacokinetics in pregnant and nonpregnant sheep. JPharm Sci 77:373-379. Riggs KW, Rurak DW, Taylor SM, McErlane BA, McMorland GH and Axelson JE (1990) Fetal and maternal placental and nonplacental clearances of metoclopramide in chronically instrumented pregnant sheep. J Pharm Sci 79:1056-1061. Riggs KW, Rurak DW, Yoo SD, McErlane BA, Taylor SM, McMorland GH and Axelson JE (1987) Drug accumulation in lung fluid of the fetal lamb after maternal or fetal administration. Am J Obstet Gynecol 157:1286-1291. Ring BJ, Eckstein JA, Gillespie JS, Binkley SN, VandenBranden M and Wrighton SA (2001) Identification of the human cytochromes p450 responsible for in vitro formation of R-and S-norfluoxetine. J Pharmacol Exp Ther 297:1044-1050. 214 Ring JA, Ghabrial H, Ching MS, Smallwood RA and Morgan DJ (1999) Fetal hepatic drug elimination. Pharmacol Ther 84:429-445. Robert E (1996) Treating depression in pregnancy. N Engl J Med 335:1056-1058. Robillard JE, Sessions C, Kennedey RL, Hamel-Robillard L and Smith FG, Jr. (1977) Interrelationship between glomerular filtration rate and renal transport of sodium and chloride during fetal life. Am J Obstet Gynecol 128:727-734. Rogers RC, Sibai BM and Whybrew WD (1990) Labetalol pharmacokinetics in pregnancy-induced hypertension. Am J Obstet Gynecol 162:362-366. Romberg RM, Rieger CH, Hill JH, Danielson J and Matadial L (1978) Cord and maternal serum meperidine concentrations and clinical status of the infant. Biol Neonate 33:80-89. Ruelius HW (1987) Extrapolation from animals to man: predictions, pitfalls and perspectives. Xenobiotica 17:255-265. Rurak DW, Wright MR and Axelson JE (1991) Drug disposition and effects in the fetus. J Dev Physiol 15:33-44. Sasaki M, Tateishi T and Ebihara A (1993) The effects of age and gender on the stereoselective pharmacokinetics of verapamil. Clin Pharmacol Ther 54:278-285. Schadel M, Wu D, Otton SV, Kalow W and Sellers EM (1995) Pharmacokinetics of dextromethorphan and metabolites in humans: influence of the CYP2D6 phenotype and quinidine inhibition. J Clin Psychopharmacol 15:263-269. Schneider TJ, Struijk PC, Lotgering FK and Wallenburg HC (1996) Placental transfer and maternal and fetal hemodynamic effects of ketanserin in the pregnant ewe. Eur J Obstet Gynecol Reprod Biol 68:179-184. Schneider TJ, Struijk PC and Wallenburg HC (1999) Placental transfer of the thromboxane synthetase inhibitor ridogrel in the late-pregnant ewe. Eur J Obstet Gynecol Reprod Biol 86:83-87. Shimada T, Yamazaki H, Mimura M, Wakamiya N, Ueng YF, Guengerich FP and Inui Y (1996) Characterization of microsomal cytochrome P450 enzymes involved in the oxidation of xenobiotic chemicals in human fetal liver and adult lungs. Drug Metab Dispos 24:515-522. Simone C, Derewlany LO and Koren G (1994) Drug transfer across the placenta. Considerations in treatment and research. Clin Perinatol 21:463-481. Simpson KH, Stakes AF and Miller M (1988) Pregnancy delays paracetamol absorption and gastric emptying in patients undergoing surgery. Br J Anaesth 60:24-27. 215 Sisenwine SF, Tio CO, Hadley FV, Liu AL, Kimmel HB and Ruelius HW (1982) Species-related differences in the stereoselective glucuronidation of oxazepam. Drug Metab Dispos 10:605-608. Smit JW, Huisman MT, van Tellingen O, Wiltshire HR and Schinkel AH (1999) Absence or pharmacological blocking of placental P-glycoprotein profoundly increases fetal drug exposure. J Clin Invest 104:1441-1447. Somogyi AA, Bochner F and Sallustio BC (1992) Stereoselective inhibition of pindolol renal clearance by cimetidine in humans. Clin Pharmacol Ther 51:379-387. Spencer MJ (1993) Fluoxetine hydrochloride (Prozac) toxicity in a neonate. Pediatrics 92:721-722. Srinivas NR, Barbhaiya RH and Midha KK (2001) Enantiomeric drug development: issues, considerations, and regulatory requirements. J Pharm Sci 90:1205-1215. Steer RA, Scholl TO, Hediger ML and Fischer RL (1992) Self-reported depression and negative pregnancy outcomes. J Clin Epidemiol 45:1093-1099. Stevens JC and Wrighton SA (1993) Interaction of the enantiomers of fluoxetine and norfluoxetine with human liver cytochromes P450. J Pharmacol Exp Ther 266:964-971. Stowe ZN (2000) Fetal exposure of SSRIs at delivery. Presented at American Psychiatric Association annual meeting, May 2000. Streetman DS, Ellis RE, Nafziger AN, Leeder JS, Gaedigk A, Gotschall R, Kearns GL and Bertino JS, Jr. (1999) Dose dependency of dextromethorphan for cytochrome P450 2D6 (CYP2D6) phenotyping. Clin Pharmacol Ther 66:535-541. Szeto HH (1993) Kinetics of drug transfer to the fetus. Clin Obstet Gynecol 36:246-254. Szeto HH, Clapp JF, 3rd, Larrow RW, Hewitt J, Inturrisi CE and Mann LI (1981) Disposition of methadone in the ovine maternal-fetal unit. Life Sci 28:2111-2117. Szeto HH, Clapp JF, Desiderio DM, Schiller PW, Grigoriants OO, Soong Y, Wu D, Olariu N, Tseng JL and Becklin R (1998) In vivo disposition of dermorphin analog (DALDA) in nonpregnant and pregnant sheep. J Pharmacol Exp Ther 284:61-65. Szeto HH, Kaiko RF, Clapp JF, Larrow RW, Mann LI and Inturrisi CE (1979) Urinary excretion of meperidine by the fetal lamb. J Pharmacol Exp Ther 209:244-248. Szeto HH, Mann LI, Bhakthavathsalan A, Liu M and Inturrisi CE (1978) Meperidine pharmacokinetics in the maternal-fetal unit. J Pharmacol Exp Ther 206:448-459. Szeto HH, Umans JG and McFarland J (1982a) A comparison of morphine and methadone disposition in the maternal-fetal unit. Am J Obstet Gynecol 143:700-706. 216 Szeto HH, Umans JG and Rubinow SI (1982b) The contribution of transplacental clearances and fetal clearance to drug disposition in the ovine maternal-fetal unit. Drug Metab Dispos 10:382-386. Thiessen JJ, Salama RB, Coceani F and Olley PM (1984) Placental drug transfer in near-term ewes: acetylsalicylic and salicylic acid. Can J Physiol Pharmacol 62:441-445. Tolle-Sander S, Rautio J, Wring S, Polli JW and Polli JE (2003) Midazolam exhibits characteristics of a highly permeable P-glycoprotein substrate. Pharm Res 20:757-764. Tonn GR, Abbott FS, Rurak DW and Axelson JE (1995) Simultaneous analysis of diphenylmethoxyacetic acid, a metabolite of diphenhydramine, and its deuterium-labeled stable isotope analog in ovine plasma and urine. J Chromatogr B Biomed Appl 663:67-81. Tracy TS (1995) Stereochemistry in pharmacotherapy: when mirror images are not identical. Ann Pharmacother 29:161-165. Treluyer JM, Gueret G, Cheron G, Sonnier M and Cresteil T (1997) Developmental expression of CYP2C and CYP2C-dependent activities in the human liver: in-vivo/in-vitro correlation and inducibility. Pharmacogenetics 7:441-452. Treluyer JM, Jacqz-Aigrain E, Alvarez F and Cresteil T (1991) Expression of CYP2D6 in developing human liver. Eur J Biochem 202:583-588. Uhr M, Steckler T, Yassouridis A and Holsboer F (2000) Penetration of amitriptyline, but not of fluoxetine, into brain is enhanced in mice with blood-brain barrier deficiency due to mdrla P-glycoprotein gene disruption. Neuropsychopharmacology 22:380-387. Upton RA and Williams RL (1986) The impact of neglecting nonlinear plasma-protein binding on disopyramide bioavailability studies. J Pharmacokinet Biopharm 14:365-379. Urichuk LJ, Aspeslet LJ, Holt A, Silverstone PH, Coutts RT and Baker GB (1997) Determination of p-trifluoromethylphenol, a metabolite of fluoxetine, in tissues and body fluids using an electron-capture gas chromatographic procedure. J Chromatogr B Biomed Sci Appl 698:103-109. van Harten J (1993) Clinical pharmacokinetics of selective serotonin reuptake inhibitors. Clin Pharmacokinet 24:203-220. Veereman-Wauters G, Monbaliu J, Meuldennans W, Woestenborghs R, Verlinden M, Heykants J and Rudolph CD (1991) Study of the placental transfer of cisapride in sheep. Plasma levels in the pregnant ewe, the fetus, and the lamb. Drug Metab Dispos 19:168-172. 217 Veng-Pedersen P and Gillespie WR (1987) A method for evaluating the mean residence times of metabolites in the body, systemic circulation, and the peripheral tissue not requiring separate i.v. administration of metabolite. Biopharm Drug Dispos 8:395-401. von Moltke LL and Greenblatt DJ (2000) Drug transporters in psychopharmacology—are they important? J Clin Psychopharmacol 20:291-294. von Moltke LL, Greenblatt DJ, Duan SX, Schmider J, Wright CE, Harmatz JS and Shader RI (1997) Human cytochromes mediating N-demethylation of fluoxetine in vitro. Psychopharmacology (Bed) 132:402-407. Vree TB, Reekers-Ketting JJ, Fragen RJ and Arts TH (1984) Placental transfer of midazolam and its metabolite 1-hydroxymethylmidazolam in the pregnant ewe. Anesth Analg 63:31-34. Vree TB, Reekers-Ketting JJ, Nouws JF and Arts TH (1983) Placental transfer and renal clearance of sulphamethoxazole and its metabolite N4-acetylsulphamethoxazole in a pregnant ewe. J Vet Pharmacol Ther 6:77-81. Wadelius M, Darj E, Frenne G and Rane A (1997) Induction of CYP2D6 in pregnancy. Clin Pharmacol Ther 62:400-407. Walle T (1985) Stereochemistry of the in vivo disposition and metabolism of propranolol in dog and man using deuterium-labeled pseudoracemates. Drug Metab Dispos 13:279-282. Walle UK, Walle T, Bai SA and Olanoff LS (1983) Stereoselective binding of propranolol to human plasma, alpha 1-acid glycoprotein, and albumin. Clin Pharmacol Ther 34:718-723. Wang LH, Rudolph AM and Benet LZ (1986) Pharmacokinetic studies of the disposition of acetaminophen in the sheep maternal-placental-fetal unit. J Pharmacol Exp Ther 238:198-205. Weiss J, Dormann SM, Martin-Facklam M, Kerpen CJ, Ketabi-Kiyanvash N and Haefeli WE (2003) Inhibition of P-glycoprotein by newer antidepressants. J Pharmacol Exp Ther 305:197-204. Wilkinson GR (1983) Plasma and tissue binding considerations in drug disposition. Drug Metab Rev 14:427-465. Wilkinson GR (1987) Clearance approaches in pharmacology. Pharmacol Rev 39:1-47. Wilkinson GR and Shand DG (1975) Commentary: a physiological approach to hepatic drug clearance. Clin Pharmacol Ther 18:377-390. 218 Wisner KL, Gelenberg AJ, Leonard H, Zarin D and Frank E (1999) Pharmacologic treatment of depression during pregnancy. Jama 282:1264-1269. Wong DT, Bymaster FP, Reid LR, Mayle DA, Krushinski JH and Robertson DW (1993) Norfluoxetine enantiomers as inhibitors of serotonin uptake in rat brain. Neuropsychopharmacology 8:337-344. Wong DT, Fuller RW and Robertson DW (1990) Fluoxetine and its two enantiomers as selective serotonin uptake inhibitors. Acta Pharm Nord 2:171-180. Wong H, Rurak DW, Kumar S, Kwan E, Abbott FS and Riggs KW (2001) Dose-dependent pharmacokinetics and metabolism of valproic acid in newborn lambs and adult sheep. Drug Metab Dispos 29:664-675. Wood M and Wood AJ (1981) Changes in plasma drug binding and alpha 1-acid glycoprotein in mother and newborn infant. Clin Pharmacol Ther 29:522-526. Woolf TF (1999) Handbook of drug metabolism. Dekker, New York. Wright JD, Boudinot FD and Ujhelyi MR (1996) Measurement and analysis of unbound drug concentrations. Clin Pharmacokinet 30:445-462. Wright MR and Jamali F (1993) Methods for the analysis of enantiomers of racemic drugs application to pharmacological and pharmacokinetic studies. J Pharmacol Toxicol Methods 29:1-9. Wright MR, Rurak DW, van der Weyde MP, Taylor SM and Axelson JE (1991) Clearance and disposition of ritodrine in the fluid compartments of the fetal lamb during and after constant rate fetal intravenous infusion. J Pharmacol Exp Ther 258:897-902. Yang HY, Lee QP, Rettie AE and Juchau MR (1994) Functional cytochrome P4503A isoforms in human embryonic tissues: expression during organogenesis. Mol Pharmacol 46:922-928. Yeleswaram K, Rurak DW, Hall C, Wright MR and Axelson JE (1992) Pharmacokinetics and pharmacodynamics of labetalol in the pregnant sheep. J Pharmacol Exp Ther 262:683-691. Yeleswaram K, Rurak DW, Kwan E, Hall C, Doroudian A, Abbott FS and Axelson JE (1993a) Disposition, metabolism, and pharmacodynamics of labetalol in adult sheep. Drug Metab Dispos 21:284-292. Yeleswaram K, Rurak DW, Kwan E, Hall C, Doroudian A, Wright MR, Abbott FS and Axelson JE (1993b) Transplacental and nonplacental clearances, metabolism and pharmacodynamics of labetalol in the fetal lamb after direct intravenous administration. J Pharmacol Exp Ther 267:425-431. 2 1 9 Y o o S D , A x e l s o n J E , T a y l o r S M and Rurak D W (1986) Placental transfer o f d iphenhydramine i n chron ica l ly instrumented pregnant sheep. J Pharm Sci 75:685-687. Y o u n g A M , A l l e n C E and A u d u s K L (2003) E f f l u x transporters o f the human placenta. Adv DrugDeliv Rev 55 :125-132. Y u A and H a i n i n g R L (2001) Compara t ive contr ibution to dextromethorphan metabol ism b y cytochrome P 4 5 0 isoforms i n vi t ro: can dextromethorphan be used as a dual probe for both C Y P 2 D 6 and C Y P 3 A activit ies? Drug Metab Dispos 29 :1514-1520. Z a r J H (1996) Biostatistical analysis. P ren t i ce -Ha l l , E n g l e w o o d C l i f f s , N . J . Z i m m e r m a n C L and Slattery J T (1983) Maintenance-dose predic t ion based on a single determination o f concentration: general appl icab i l i ty to two-compartment drugs w i t h reference to l i th ium. J Pharm Sci 72 :1262-1266. Z u c k e r m a n B , A m a r o H , Bauchner H and Cabra l H (1989) Depress ive symptoms dur ing pregnancy: relationship to poor health behaviors. Am J Obstet Gynecol 160:1107-1111. Zucke rman B , Bauchner H , Parker S and Cabra l H (1990) Mate rna l depressive symptoms dur ing pregnancy, and newborn i r r i tabi l i ty . JDev Behav Pediatr 11:190-194. 220 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0091754/manifest

Comment

Related Items