Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Drug disposition in the maternal-fetal unit : studies with diphenhydramine and valproic acid in pregnant… Kumar, Sanjeev 1998

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

Item Metadata

Download

Media
831-ubc_1998-345726.pdf [ 16.71MB ]
Metadata
JSON: 831-1.0099371.json
JSON-LD: 831-1.0099371-ld.json
RDF/XML (Pretty): 831-1.0099371-rdf.xml
RDF/JSON: 831-1.0099371-rdf.json
Turtle: 831-1.0099371-turtle.txt
N-Triples: 831-1.0099371-rdf-ntriples.txt
Original Record: 831-1.0099371-source.json
Full Text
831-1.0099371-fulltext.txt
Citation
831-1.0099371.ris

Full Text

Drug Disposition in the Maternal-Fetal Unit: Studies with Diphenhydramine and Valproic Acid in Pregnant Sheep by Sanjeev Kumar B.Pharm., Panjab University, Chandigarh, India, 1991 M.Pharm. (Pharmaceutics), Panjab University, Chandigarh, India, 1993 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 Biopharmaceutics and Pharmaceutics) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September, 1998 © Sanjeev Kumar, 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) ii Abstract Diphenhydramine (DPHM, a high clearance amine) and valproic acid (VPA, a low clearance carboxylic acid) disposition was examined in chronically-instrumented pregnant sheep during late gestation (after 120 d, term 145 d) in order to: i) identify the important pharmacokinetic factors determining fetal drug exposure, and ii) examine the in utero fetal drug metabolism capacity for these two compounds. Both drugs underwent rapid placental transfer after maternal dosing. Although the placental permeability for DPHM was higher, the steady-state fetal exposure relative to the mother was greater for VPA (70% vs. 20%). Significant fetal drug elimination capacity was detectable for both compounds, with DPHM having a much higher clearance. Moreover, there was significant (~44%) fetal hepatic first-pass uptake of the placentally transferred DPHM from the umbilical vein due to the unique geometrical position of the fetal liver. These factors were responsible for the observed lower fetal exposure to DPHM. Maternal and fetal placental and non-placental clearances of both drugs were highly dependent on their plasma protein binding. However, for DPHM, the importance of plasma protein binding in determining fetal drug exposure was overridden by the above more pronounced effect of the fetal liver. Metabolism to diphenylmethoxyacetic acid (DPMA), and DPHM-N-oxide renal excretion accounted for only 1-2% of the DPHM dose in maternal and fetal sheep, as opposed to 50-80% in the monkey, dog and human. Also, DPMA was not secondarily metabolized in sheep. Approximately 95% of the maternal VPA dose was eliminated via glucuronidation and renal excretion, with p-oxidation and cytochrome P-450 pathways iii accounting for the remainder. All these pathways appeared to be functional in the fetus; however, the low fetal VPA clearance could not be accurately measured. Studies conducted in 1 day old lambs revealed -18 fold lower weight-normalized VPA elimination capacity relative to the mother, and this led to a lower newborn VPA clearance and longer half-life, as observed in human newborns. The reduced newborn lamb VPA elimination was mainly attributable to underdeveloped glucuronidation and renal excretion ability; this, combined with a high p-oxidation capacity at birth, resulted in a larger fraction of the VPA dose (~20%) being metabolized via p-oxidation and cytochrome P-450 pathways in lambs. iv Table of Contents Title i Abstract ii Table of Contents iv List of Tables xi List of Figures xv List of Abbreviations xxi Acknowledgments xviii Chapter 1 Introduction 1 1.1 Drug Use in Pregnancy 1 1.2 Factors Affecting Placental Drug Transfer 2 1.3 Study of Placental Drug Transfer and 3 Determinants of Fetal Drug Exposure 1.4 Diphenhydramine 10 1.4.1 Pharmacology and Therapeutic Use 10 1.4.2 Basic pharmacokinetics 11 1.4.3 Metabolism 13 1.4.4 DPHM in Pregnancy 15 1.4.5 DPHM Disposition in Chronically-Instrumented 16 Pregnant Sheep: Earlier Studies 1.5 Valproic Acid 23 1.5.1 Therapeutic Use, Pharmacology and Adverse Effects 23 1.5.2 Basic Pharmacokinetics 25 1.5.3 Metabolism 27 1.5.3.1 Glucuronidation 29 1.5.3.2 Mitochondrial p-oxidation 29 1.5.3.3 Microsomal Metabolism 30 1.5.4 Epilepsy, Pregnancy and Valproic Acid 33 1.5.5 VPA Disposition in Pregnant Sheep: Earlier Studies 36 1.6 Rationale 37 1.7 Objectives 41 Chapter 2 Materials, Instrumentation and Assay Methodologies 42 2.1 Materials 42 2.2 Instrumentation 46 2.2.1 Gas Chromatography-Mass Spectrometry 46 2.2.2 High-Performance Liquid Chromatography-Tandem 47 Mass-Spectrometry V 2.2.3 Spectrophotometer 47 2.2.4 Physiological Monitoring 47 2.2.5 Other Equipment 48 2.3 Analysis of DPHM and [2H10]-DPHM in Biological Fluids 49 2.4 Analysis of DPMA and [2H10]-DPMA in Biological Fluids 50 2.5 Analysis of VPA and its Metabolites in Biological Fluids 51 2.6 Simultaneous LC-MS/MS Analysis of DPHM, [2H10]-DPHM, 53 DPHM-N-Oxide and [2H10]-DPHM-N-Oxide in Biological Fluids 2.6.1 Methods 53 2.6.1.1 Synthesis and Purification of Deuterium-Labeled 53 DPHM-N-Oxide Hydrochloride 2.6.1.2 Standard Stock Solutions 54 2.6.1.3 Sample Extraction 55 2.6.1.4 High Performance Liquid Chromatography 55 2.6.1.5 Electrospray Tandem Mass-Spectrometry (MS/MS) 56 2.6.1.6 Calibration Curves and the Regression Model 57 2.6.1.7 Extraction Recovery 58 2.6.1.8 Analyte Stability in Biological Fluid Samples 58 2.6.1.9 Method Validation 60 2.6.1.10 Application of the Assay to a Sample Study of 60 DPHM, [2H10]-DPHM, DPHMNOX and [2H10]-DPHMNOX Disposition in the Ovine Maternal-Fetal Unit 2.6.2 Results and Discussion 61 2.6.2.1 High Performance Liquid Chromatography 63 of Diphenhydramine and the N-Oxide Metabolite 2.6.2.2 Tandem Mass Spectrometric (MS/MS) 65 Detection of Analytes 2.6.2.3 Extraction Method and Recovery 69 2.6.2.4 Analyte Stability in Biological Fluid Samples 71 2.6.2.5 Calibration Curves and the Regression Model 71 2.6.2.6 Method Validation 72 2.6.2.7 Application of the Assay to a Sample Study of 76 DPHM, [2H10]-DPHM, DPHMNOX and [2H10]-DPHMNOX Disposition in the Ovine Maternal-Fetal Unit Chapter 3 Organs and Metabolic Pathways of Diphenhydramine 79 Clearance in Maternal and Fetal Sheep 3.1 Methods 80 3.1.1 Animals and Surgical Preparation 80 3.1.1.1 Study A: Paired Maternal (DPHM) and Fetal 81 ([2H10]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances vi 3.1.1.2 Study B: Study of Fetal Hepatic First-Pass 82 Uptake of DPHM from the Umbilical Vein 3.1.1.3 Study C: Hepatic Uptake of DPHM in Adult 82 Non-Pregnant Sheep 3.1.1.4 Study D: Gut Uptake of DPHM from the Systemic 82 Circulation in Adult Non-Pregnant Sheep 3.1.1.5 Study E: Contribution of DPMA Formation to DPHM 83 Non-Placental Clearance in Maternal and Fetal Sheep 3.1.1.6 Study F: Disposition of the DPHM-N-oxide 83 Metabolite in the Maternal-Fetal Unit 3.1.2 Experimental Protocols 84 3.1.2.1 Study A: Paired Maternal (DPHM) and Fetal 84 ([2Hi0]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances 3.1.2.2 Study B: Study of Fetal Hepatic First-Pass 85 Uptake of DPHM from the Umbilical Vein 3.1.2.3 Study C: Hepatic Uptake of DPHM in Adult 86 Non-Pregnant Sheep 3.1.2.4 Study D: Gut Uptake of DPHM from the Systemic 87 Circulation in Adult Non-Pregnant Sheep 3.1.2.5 Study E: Contribution of DPMA Formation to DPHM 87 Non-Placental Clearance in Maternal and Fetal Sheep 3.1.2.6 Study F: Disposition of the DPHM-N-oxide 88 Metabolite in the Maternal-Fetal Unit 3.1.3 Physiological Recording and Monitoring Procedures 89 3.1.4 Plasma Protein Binding of DPHM and fH 1 0]-DPHM 90 in the Mother and the Fetus 3.1.5 Drug and Metabolite Analysis 91 3.1.6 Pharmacokinetic Analysis 91 3.1.6.1 Study A: Paired Maternal (DPHM) and Fetal 91 ([2Hio]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances 3.1.6.2 Study B: Study of Fetal Hepatic First-Pass 93 Uptake of DPHM from the Umbilical Vein 3.1.6.3 Study C: Hepatic Uptake of DPHM in Adult 95 Non-Pregnant Sheep 3.1.6.4 Study D: Gut Uptake of DPHM from the Systemic 97 Circulation in Adult Non-Pregnant Sheep 3.1.6.5 Study E: Contribution of DPMA Formation to 97 DPHM Non-Placental Clearance in Maternal and Fetal Sheep 3.1.6.6 Study F: Disposition of the DPHM-N-oxide 100 Metabolite in the Maternal-Fetal Unit 3.1.7 Statistical Analysis 100 3.2 Results 101 vii 3.2.1 Study A: Paired Maternal (DPHM) and Fetal 101 ([2H10]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances 3.2.1.1 Maternal and Fetal Plasma DPHM and 102 [2Hi0]-DPHM Concentrations, and Placental and Non-Placental Clearance Values 3.2.1.2 Fetal and Maternal DPMA and [2H10]-DPMA 104 Plasma Concentrations 3.2.2 Study B: Study of Fetal Hepatic First-Pass 108 Uptake of DPHM from the Umbilical Vein 3.2.3 Study C: Hepatic Uptake of DPHM in Adult 111 Non-Pregnant Sheep 3.2.4 Study D: Gut Uptake of DPHM from the Systemic 115 Circulation in Adult Non-Pregnant Sheep 3.2.5 Study E: Contribution of DPMA Formation to 116 DPHM Non-Placental Clearance in Maternal and Fetal Sheep 3.2.5.1 Maternal-Fetal Steady-State Plasma DPHM 117 Concentrations and Free Fractions, and Placental and Non-placental Clearance Estimates 3.2.5.2 Maternal-Fetal Arterial Plasma AUC Ratios of 118 the Parent Drug, the Preformed Metabolite, and the in vivo Generated Metabolite 3.2.5.3 Comparative Maternal-Fetal Pharmacokinetics of 121 the Parent Drug, the Preformed Metabolite, and the in vivo Generated Metabolite 3.2.5.4 Renal Elimination of the Parent Drug, the Preformed 123 Metabolite, and the in vivo Generated Metabolite in the Mother and the Fetus 3.2.6 Study F: Disposition of the DPHM-N-oxide Metabolite 128 in the Maternal-Fetal Unit 3.3 Discussion 131 3.3.1 Study A: Paired Maternal (DPHM) and Fetal 131 ([2Hi0]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances 3.3.1.1 DPHM and [2H10]-DPHM Plasma Concentrations 132 in the Ewe and Fetus 3.3.1.2 Placental and Non-Placental Clearances 132 of DPHM in Fetal and Maternal Sheep 3.3.1.3 Maternal and Fetal Plasma Concentrations of 133 DPMA and [2H10]-DPMA 3.3.2 * Studies A & B: Fetal Hepatic Uptake and Metabolism of DPHM 134 3.3.2.1 Evidence of Fetal Hepatic First-Pass 134 DPHM Uptake from the Umbilical Vein viii 3.3.2.2 Impact of Fetal Hepatic Drug Uptake on 140 the 2-Compartment Model Estimates of Maternal and Fetal Clearances 3.3.3 Studies C and D: Role of the Liver and Gut in Systemic 144 DPHM Clearance in Adult Non-Pregnant Sheep 3.3.4 Study E: Contribution of DPMA Formation to DPHM 151 Non-Placental Clearance in Maternal and Fetal Sheep 3.3.5 Study F: Disposition of the DPHM-N-oxide 159 Metabolite in the Maternal-Fetal Unit Chapter 4 Inter-relationships between Plasma Drug Protein Binding, 162 Gestational Age, Umbilical Blood Flow, and Diphenhydramine Clearances in the Ovine Maternal-Placental-Fetal Unit 4.1 Methods 163 4.1.1 Animals and Surgical Preparation 163 4.1.2 Experimental Protocols 163 4.1.3 Physiological Recording and Monitoring Procedures 164 4.1.4 Protein Binding of DPHM and [2H10]-DPHM in 164 Fetal and Maternal Plasma 4.1.5 Drug Analysis 165 4.1.6 Pharmacokinetic Analysis 165 4.1.7 Statistical Analysis 166 4.2 Results 166 4.2.1 Relationships of Maternal and Fetal DPHM 167 Clearances with Gestational Age 4.2.2 Plasma Protein Binding Effects on Maternal 168 and Fetal DPHM Clearances 4.2.3 Changes in Maternal and Fetal Plasma Protein 173 Binding with Gestational Age 4.2.4 Inter-relationships between Maternal and Fetal 173 Plasma DPHM Concentrations, Unbound Fractions and the 2-Compartment Model Clearance Estimates 4.2.5 Relationships between the Indices of Fetal Drug 176 Exposure/Placental Transfer, Gestational Age and Plasma Protein Binding 4.3 Discussion 180 4.3.1 Maternal and Fetal DPHM Clearances: Gestational Age, 181 Plasma Protein Binding and Umbilical Blood Flow Effects 4.3.2 Inter-relationships between Maternal and Fetal 189 Plasma DPHM Concentrations, Unbound Fractions and the 2-Compartment Model Clearance Estimates 4.3.3 Relationships between the Indices of Fetal Drug 191 Exposure/Placental Transfer, Gestational Age and Plasma Protein Binding ix Chapter 5 Pharmacokinetics and Metabolism of Valproic Acid in 198 Maternal, Fetal, and Newborn Sheep 5.1 Methods 198 5.1.1 Animals and Surgical Preparation 198 5.1.2 Experimental Protocols 200 5.1.2.1 Pregnant Sheep Experiments 200 5.1.2.2 Newborn Lamb Experiments 201 5.1.3 Physiological Recording and Monitoring Procedures 202 5.1.4 Plasma Protein Binding of VPA and its Metabolites 202 in the Mother and the Fetus 5.1.5 Preparation of Maternal and Fetal Liver Microsomes 203 5.1.6 VPA Glucuronidation in Maternal and Fetal Liver Microsomes 203 5.1.7 Drug and Metabolite Assay 204 5.1.8 Pharmacokinetic Analysis 205 5.1.9 Statistical Analysis 209 5.2 Results 210 5.2.1 Maternal-Fetal VPA Plasma Concentrations, 210 Plasma Protein Binding and Indices of Fetal Drug Exposure 5.2.2 Placental and Non-Placental Clearances of 218 VPA in the Mother and the Fetus 5.2.3 Amniotic and Fetal Tracheal Fluid Concentrations 223 of VPA during Maternal and Fetal VPA Infusions 5.2.4 Overall Total Body Clearance, Terminal 224 Elimination Half-Life, Mean Residence Time, and Steady-State Volume of Distribution of VPA in the Mother and the Fetus 5.2.5 Pharmacokinetics and Plasma Protein Binding of 228 VPA in Newborn Lambs 5.2.6 Kinetics of VPA-Glucuronide Formation in Maternal 234 and Fetal Liver Microsomes 5.2.7 Comparative in vivo Michaelis-Menten 236 Pharmacokinetic Parameters of VPA in the Mother and the Newborn 5.2.8 Pharmacokinetics of Renal Excretion of Unchanged 236 VPA in the Mother, Fetus and the Newborn 5.2.9 Maternal and Fetal Plasma Concentrations of 243 VPA Metabolites 5.2.10 Plasma Concentrations of VPA Metabolites in 251 Newborn Lambs, and Fetal and Newborn VPA Metabolite Exposure Relative to the Mother 5.2.11 Excretion of Unchanged VPA and its Metabolites 259 in Maternal and Fetal Urine 5.2.12 Excretion of Unchanged VPA and its Metabolites 261 in the Newborn Lamb Urine X 5.3 Discussion 265 5.3.1 Maternal-Fetal VPA Plasma Concentrations, 265 Plasma Protein Binding and Indices of Fetal Drug Exposure 5.3.2 Placental and Non-Placental Clearances of 268 VPA in the Mother and the Fetus 5.3.3 Amniotic and Fetal Tracheal Fluid Concentrations 275 of VPA during Maternal and Fetal VPA Infusions 5.3.4 Overall Total Body Clearance, Terminal 277 Elimination Half-Life, Mean Residence Time, and Steady-State Volume of Distribution of VPA in the Mother and the Fetus 5.3.5 Pharmacokinetics and Plasma Protein Binding 281 of VPA in Newborn Lambs 5.3.6 Overall VPA Elimination Capacity in the Newborn 286 Lamb and Maternal Sheep 5.3.7 Pharmacokinetics of the Renal Excretion of Unchanged 290 VPA in the Mother, Fetus and the Newborn Lamb 5.3.8 Maternal, Fetal and Newborn Metabolism of VPA 292 5.3.8.1 Maternal and Fetal Plasma 292 Concentrations of VPA Metabolites 5.3.8.2 Plasma Concentrations of VPA Metabolites 299 in Newborn Lambs, and Fetal and Newborn Metabolite Exposure Relative to the Mother 5.3.8.3 Excretion of Unchanged VPA and its 303 Metabolites in Maternal and Fetal Urine 5.3.8.4 Excretion of Unchanged VPA and its 308 Metabolites in the Newborn Lamb Urine Chapter 6 Global Summary and Conclusions 313 References 324 Appendix I 351 xi List of Tables Table 1.1 -Table 2.1 -Table 2.2 -Table 2.3 -Table 3.1 -Table 3.2 -Table 3.3 -Table 3.4 -Table 3.5-Average values of fetal and maternal placental (CL f m and C L m f , 18 respectively) and non-placental clearances (CL f 0 and C L m o , respectively) of various drugs studied in pregnant sheep during the last part of gestation. Calibration curve concentration range of VPA and its 52 metabolites and the corresponding internal standards that were utilized for their quantitation using the GC-MS assay. Various fragment ions that were monitored during selected ion monitoring are also presented. Intra-assay variability and bias of the LC-MS/MS analytical 74 method for DPHM, [2Hi0]-DPHM, DPHM-N-oxide and [2H10]-DPHM-N-oxide in ovine plasma. Inter-assay variability and bias of the LC-MS/MS analytical 75 method for DPHM, [2H10]-DPHM, DPHM-N-oxide and [2H10]-DPHM-N-oxide in ovine plasma. Weight-normalized estimates of net (CLm(net) and CLf(net)), total 106 ( C L m m and CLff), placental (CLmf and CL f m), and non-placental (CL m o and CL f 0) clearances in the mother and the fetus, respectively, in 5 experiments involving simultaneous i.v. infusion of DPHM to the mother and [2H10]-DPHM to the fetus. Maternal and fetal femoral arterial AUC values for unlabeled 107 and labeled forms of DPHM and DPMA, metabolite/parent drug AUC ratios and estimates of fetal hepatic extraction of the maternally derived DPHM in 5 maternal-fetal paired infusion experiments. Fetal femoral arterial AUC values of labeled and unlabeled 110 DPHM and DPMA after simultaneous but separate umbilical and tarsal venous administration of DPHM and [2H-io]-DPHM, and estimates of fetal hepatic extraction ratio of DPHM in umbilical venous first-pass experiments. Femoral arterial AUC's of the parent drug (DPHM or [2Hi0]- 112 DPHM) and metabolite (DPMA or [2H10]-DPMA), parent drug hepatic first-pass extraction ratio, and fraction of intravenously administered parent drug dose metabolized in the liver during hepatic first-pass uptake studies. Diphenhydramine systemic and intrinsic clearances, estimated 113 hepatic blood flows and percentage of intravenously administered dose eventually delivered to the "hepato-portal system" in 5 adult sheep. XII Table 3.6 - Steady-state femoral arterial and portal venous plasma 116 concentrations, systemic clearance, and gut extraction of DPHM in 4 sheep during gut uptake experiments. Table 3.7 - Gestational age and weight corrected estimates of net (CL m ( n e t ) 119 and CLf(net)), total (CL m m and CLff), placental (CLmf and CL f m), and non-placental (CL m o and CL f 0) DPHM clearances in the mother and the fetus, respectively, obtained using 2-compartment model of the maternal-fetal unit. Table 3.8- Arterial plasma AUC ratios of the parent drug, the preformed 120 metabolite and the in vivo formed metabolite in the mother and the fetus. Table 3.9- Pharmacokinetic parameters of DPHM (or [2H10]-DPHM), 125 preformed DPMA and in vivo formed DPMA in the mother. Table 3.10- Pharmacokinetic parameters of DPHM (or [2H10]-DPHM), 126 preformed DPMA and in vivo formed DPMA in the fetus. Table 3.11 - Pharmacokinetic parameters of renal elimination of DPHM (or 127 [2H10]-DPHM), preformed DPMA and in vivo formed DPMA in maternal and fetal sheep. Table 3.12- Maternal and fetal plasma AUC's of the DPHM-N-oxide (or 130 [2H10]-DPHM-N-oxide) metabolite in 4 pregnant sheep during separate infusion. maternal and fetal 6h DPHM (or fH1 0]-DPHM) Table 3.13-Table 4.1 -Table 5.1 -Table 5.2 -Table 5.3 -Table 5.4 -Pharmacokinetics of the renal elimination of DPHM-N-oxide metabolite in the mother and the fetus. Gestational Age, Maternal and Fetal Plasma Unbound Fractions and DPHM Clearance Data from 18 Pregnant Sheep. Steady-state maternal and fetal total plasma concentrations of VPA during separate maternal and fetal 24 h VPA infusions in 5 pregnant sheep. Steady-state maternal and fetal unbound plasma concentrations of VPA during separate maternal and fetal 24 h VPA infusions in 5 pregnant sheep. Range of unbound fractions of VPA in maternal and fetal plasma over the course of the entire VPA concentration vs. time profile. Steady-state maternal and fetal plasma unbound fractions of VPA during separate maternal and fetal 24 h VPA infusions in 5 pregnant sheep. 131 169 212 213 214 215 xiii Table 5.5- In vivo plasma protein binding parameters of VPA in maternal, 217 fetal and newborn sheep obtained by fitting the data to a 1- or 2-site binding model. Table 5.6 - Weight-normalized estimates of total ( C L m m and CLff), placental 220 (CL m f and CL f m), and non-placental (CL m o and CL f 0) VPA clearances of the total drug in the mother and the fetus, respectively. Clearances were calculated using the measured total plasma VPA concentrations and 2-compartment model of the maternal-fetal unit. Table 5.7 - Weight-normalized estimates of net (CLum(net) and CLVet)), total 221 (CLUmm and CLu f f), placental (CL u m f and CL u f m ) , and non-placental ( C L u m 0 and CLUf0) VPA clearances of the unbound drug in the mother and the fetus, respectively. Clearances were calculated using the measured unbound plasma concentrations and 2-compartment model of the maternal-fetal unit. Table 5.8 - Weight-normalized estimates of net (CLm(net) and CLf(net)), total 222 ( C L m m and CLff), placental (CLmf and CLfm), and non-placental (CLmo and CL f 0) VPA clearances of the total drug in the mother and the fetus, respectively. Clearances were calculated after correcting for the changes in maternal and fetal plasma steady-state unbound fractions between maternal and fetal experiments. Maternal clearances (CLm > CL m f and C L m o ) are presented with respect to steady-state maternal plasma unbound fraction during maternal administration, whereas, fetal clearances (CLff, C L f m and CL f 0) are with respect to steady-state fetal unbound fractions during fetal administration. Table 5.9 - Amniotic and fetal tracheal fluid concentrations of VPA during 223 the steady-state period of maternal and fetal infusion. Table 5.10 - Comparative pharmacokinetic parameters of the unbound and 226 total VPA in the mother during maternal VPA infusion. Table 5.11 - Comparative pharmacokinetic parameters of the unbound and 227 total VPA in the fetus during fetal VPA infusion. Table 5.12- Comparative pharmacokinetic parameters of the unbound and 231 total VPA in newborn lambs. Table 5.13- In vivo Michaelis-Menten pharmacokinetic parameters of the 237 unbound VPA in maternal and newborn sheep. Table 5.14- Pharmacokinetic parameters of renal elimination of VPA in 240 maternal and fetal sheep. Table 5.15- Pharmacokinetic parameters of renal elimination of VPA in 241 newborn lambs. xiv Table 5.16 - Steady-state plasma concentrations of various VPA metabolites 249 in maternal and fetal plasma during the final 4 h period of maternal drug infusion. Table 5.17 - Steady-state plasma concentrations of various VPA metabolites 250 in maternal and fetal plasma during the final 4 h period of fetal drug infusion. Table 5.18- Maximal plasma concentrations (Cm a x) of various VPA 254 metabolites in maternal and fetal plasma during maternal drug infusion. Table 5.19- Maximal plasma concentrations (Cm a x) of various VPA 255 metabolites in maternal and fetal plasma during fetal drug infusion. Table 5.20 - Maximal plasma concentrations (Cm a x) of VPA and its 256 metabolites, and the corresponding times of their occurrence (tmax) in 4 newborn lambs after a 6 h VPA infusion. Table 5.21 - Recovery of unchanged VPA and its metabolites as a 262 percentage of the total dose in maternal urine after maternal steady-state VPA infusion. Table 5.22 - Recovery of unchanged VPA and its metabolites as a 263 percentage of the total dose in maternal urine after fetal steady-state drug infusion. Table 5.23 - Recovery of unchanged VPA and its metabolites as a 264 percentage of the total dose in newborn lamb urine after a 6 h drug infusion. XV List of Figures Figure 1.1 - A representation of various placental and non-placental drug 8 clearances in the 2-compartment pharmacokinetic model of the maternal-fetal unit (CL m o - maternal non-placental clearance; CL f 0 - fetal non-placental clearance; C L m f -placental clearance from the mother to the fetus; C L f m -placental clearance from the fetus to the mother). Figure 1.2 - Chemical Structure of DPHM 10 Figure 1.3- A diagrammatic sketch of the fetal circulation showing the 21 position of fetal liver and sites of drug administration for assessment of fetal first-pass hepatic drug uptake from umbilical vein (Tonn et al., 1996). Figure 1.4 - Chemical structure of VPA. 23 Figure 1.5 - Metabolic pathways of VPA. Dotted arrows indicate pathways 28 for which direct experimental evidence is lacking. Figure 2.1 - Positive ion electrospray daughter ion mass-spectra of DPHM, 67 [2H10]-DPHM, DPHM-N-Oxide, [2H10]-DPHM-N-Oxide, and the internal standard, orphenadrine. Figure 2.2 - LC-MS/MS MRM ion chromatograms of: (A) an extracted 68 blank sheep plasma sample, and (B) an extracted plasma calibration standard containing 0.4 ng/ml each of [2H10]-DPHM-N-oxide and DPHM-N-oxide, and 1 ng/ml each of [2H10]-DPHM and DPHM. The y-axis scales in panel (A) have been magnified to clearly show the baselines at different MRM ion transitions. The HPLC and MS/MS conditions and specifications are described in the text. Figure 2.3 - Representative calibration curves of: (A) DPHM and [2H10]- 73 DPHM, and (B) DPHM-N-oxide and [2H10]-DPHMN-N-oxide, in sheep plasma. Figure 2.4 - Femoral arterial plasma concentration versus time profiles of 77 DPHM, [2H10]-DPHM, DPHM-N-oxide, and [2H10]-DPHM-N-oxide in a fetal lamb after a simultaneous equimolar bolus dose of DPHM and [2H10]-DPHM via the fetal lateral tarsal vein. Figure 3.1 - Average plasma concentrations of DPHM and [2H10]-DPHM in 103 maternal and fetal femoral arterial plasma during and following simultaneous i.v. infusions of DPHM (670 pg/min) to the ewe and [2Hi0]-DPHM (170 pg/min) to the fetus. xvi Figure 3.2 - Average maternal and fetal femoral arterial plasma 105 concentrations of DPMA and [2Hio]-DPMA during and following simultaneous iv. infusion of DPHM (670 uig/min) to the ewe and [2H10]-DPHM (170 ug/min) to the fetus. Figure 3.3- A). Representative fetal femoral arterial plasma 109 concentrations of DPHM, [2Hi0]-DPHM, DPMA and [2H10]-DPMA after simultaneous iv. bolus administration of DPHM and [2H10]-DPHM (5 mg each) to the fetus (E#989). DPHM was given via the fetal tarsal vein and [2H10]-DPHM via the umbilical vein. B). Representative fetal femoral arterial plasma concentrations of DPHM, [ 2Hi 0]-DPHM, DPMA and [2Hio]-DPMA after simultaneous steady-state infusion of DPHM and [2Hi0]-DPHM (60u,g/min each) to the fetus (E#2164). DPHM was infused via the fetal umbilical vein and [2H10]-DPHM via the fetal lateral tarsal vein. Figure 3.4- Representative femoral arterial plasma concentration vs. time 114 profiles of parent drug (DPHM & [2H10]-DPHM; A & C), and metabolite (DPMA and [2H10]-DPMA; B & D) in E1154 (upper panel) and E102 (lower panel). E1154 received 50 mg DPHM via the portal venous route and an equimolar dose of [2Hw]-DPHM via the femoral venous route. The routes of administration of DPHM and [2H10]-DPHM were reversed in E102. Figure 3.5- Average femoral arterial and portal venous plasma 115 concentration vs. time profiles of DPHM in 4 sheep during 6h DPHM infusion. Figure 3.6 - Representative plasma concentration vs. time profiles of the 122 parent drug, the preformed metabolite and the in vivo generated metabolite in maternal and fetal plasma of E303Y. Figures 3.6A and 3.6B are the data from separate maternal and fetal administration experiments, respectively. In both experiments, unlabeled DPMA was administered as the preformed metabolite in combination with [2Hio]-DPHM. The [2Hio]-DPMA is thus the in vivo generated metabolite. Figure 3.7- Representative cumulative renal excretion profiles of the 124 parent drug, the preformed metabolite and the in vivo generated metabolite in maternal and fetal urine of E303Y. Figures 3.7A and 3.7B are the data from separate maternal and fetal administration experiments, respectively. In both experiments, unlabeled DPMA was administered as the preformed metabolite in combination with [2Hi0]-DPHM. The [2H10]-DPMA is thus the in vivo generated metabolite. Figure 3.7B shows that even after fetal administration negligible amounts of preformed as well as in vivo generated metabolite are excreted in fetal urine compared to maternal urine. xvii Figure 3.8- Representative maternal and fetal arterial plasma profiles of 129 the parent drug and the N-oxide metabolite in E4230 after A) 6 h maternal DPHM infusion (670 ug/min), B) after fetal DPHM infusion (170 pg/min). Figure 3.9- Relationship between fetal and maternal placental clearance 144 difference (CL f m -CL m f , ml/min/kg) and fetal non-placental clearance (CL f 0, ml/min/kg) of drugs studied in pregnant sheep. Figure 4 .1- Relationships between fetal DPHM clearances and 170 gestational age (GA). A) CL f f vs. GA; B) CLfm vs. GA; C) C L f o vs. GA. Scatter points are the actual data in 18 pregnant sheep. The regression line (solid) and the 95% confidence interval (dotted) are also shown. CLff: Fetal total clearance; CLf m : Fetal placental clearance; CL f 0 : Fetal non-placental clearance. Figure 4.2- Relationships between fetal and maternal DPHM clearances 171 and corresponding plasma unbound fractions of the drug. A) CLff vs. F-UF; B) C L f m vs. F-UF; C) CL f 0 vs. F-UF; D) C L m m vs. M-UF; E) CLmf vs. M-UF; and F) C L m o vs. M-UF. The CL f 0 , CLmm and C L m o relationships with the corresponding plasma unbound fraction of the drug were analyzed according to the well-stirred model of organ drug clearance. UF: unbound fraction; F and M refer to mother and the fetus, respectively; Clearances are as defined in Table 4.1. Figure 4.3- Alterations in (A) fetal and (B) maternal steady-state plasma 174 unbound fraction of DPHM with increasing gestational age. Actual experimental data (scatter points), regeression line (solid) and 95% confidence interval (dotted) are depicted. Figure 4.4- Relationships between (A) maternal unbound fraction and 175 steady-state plasma concentration of the drug after maternal drug administration and (B) fetal unbound fraction and steady-state plasma concentration of the drug after fetal drug administration. Scatter points are the experimental data in different sheep. The regression line (solid) and the 95% confidence interval (dotted) are also shown. Figure 4.5- Influence of various clearance parameters of the 2- 178 compartment model on different maternal-fetal plasma concentrations. A) CL mo vs. C m ; B) CL m f vs. Cf, C) CLfo vs. Cf; D) CLfm vs. Cf'; E) C L f m vs. C m ' ; and F) C L m o vs. Cm". All relationships except B) and E) were analyzed according to the steady-state clearance model, CL=k 0/C s s; the solid lines represent the best-fit lines determined by this model. Figure 4.6- Relationships between indices of placental drug transfer/fetal 179 drug exposure and their determining factors. A) Cf/C m vs. F-xviii UF/M-UF; B) C f' vs. GA; C) Cf7Cm' vs. GA; D) C f7Cm' vs. CL f 0 ; E) CfVCm' vs. CL f m ; and F) C f7Cm' vs. F-UF. Figure 4.7- Relationship between fetal DPHM placental clearance of the 185 unbound drug and gestational age. Actual experimental data (scatter points), linear regression line (solid) and 95% confidence interval are depicted. Figure 4.8- Relationship between fetal acetaminophen placental 186 clearance and gestational age. Data were taken from Wang ef al., 1986. Actual experimental data (scatter points), linear regression line (solid) and 95% confidence interval are shown. Figure 4.9- Weight-normalized umbilical blood flow in 11 pregnant sheep 187 between 128-141 d gestation as a function of gestational age. Blood flow was measured by an electromagnetic flow transducer (n = 4), an ultrasonic flow probe (n = 5), or radioactive microspheres (n=2). Figure 5.1 - Representative concentration vs. time profiles of the total and 211 unbound VPA in maternal and fetal plasma during and after a 24 h steady-state VPA infusion. A) maternal infusion, B) fetal infusion. Figure 5.2 - Maternal and fetal plasma protein binding characteristics of 216 VPA; pooled data from all maternal and fetal VPA infusion experiments. C t, Cb and C u are the maternal or fetal plasma concentrations of the total, bound and unbound VPA. A) & D) are the Rosenthal plots of the data in maternal and fetal plasma, respectively, demonstrating a biphasic relationship. C t vs. C u data in maternal (B) and fetal (E) plasma also exhibit a biphasic relationship, indicating VPA binding at 2 classes of binding sites. C) & F) show the relationship between Cb vs. C u in maternal and fetal plasma, respectively; actual data (scatter points) and model predicted line obtained from fit of the data to a 2-site binding model are depicted. Figure 5.3 - Representative amniotic and fetal tracheal fluid concentration 224 vs. time profiles of VPA in E4241, showing apparent steady-state concentrations in these fluids during the 6-24h infusion period and subsequent rapid decline in concentrations during the post-infusion phase. Figure 5.4 - A) Representative plasma concentration vs. time profile of the 230 total and unbound VPA in the newborn lamb NL0123z. B) Unbound and total VPA plasma concentrations in NL0123z during the 6 h infusion period, showing continuous accumulation of the drug in newborn plasma. C) Typical plasma concentration vs. time profiles of the total and unbound VPA in maternal and fetal plasma during the initial 6 h period of the 24 h maternal VPA infusion in E5108, showing lack of any drug xix accumulation. Profiles similar to C) were also observed in all animals after fetal VPA infusion. Figure 5.5 - Plasma protein binding characteristics of VPA in newborn lamb 232 plasma; pooled data from 4 lambs. C t, C b and C u are the plasma concentrations of the total, bound and unbound VPA, respectively. A) Rosenthal plot of the data. B) Relationship between C t and C u . C) Relationship between C b and C u . Anomalous VPA binding characteristics in the newborn plasma are evident when compared to maternal and fetal data in Figure 5.2. Figure 5.6 - Partitioning of newborn lamb VPA plasma protein binding data 233 into those obtained during day 1 of the experiment (A-C) and those obtained after day 1 (D-F). C t, C b and C u are the plasma concentrations of the total, bound and unbound VPA. A) and D) are the Rosenthal plots B) and E) are the relationships between C t and C u . C) and F) are the relationships between C b and C u . The data in F) were fitted to a 1-site binding model and the model-predicted line is also shown. Striking differences in VPA binding characteristics between the 2 groups of data are evident. Figure 5.7 - In vitro glucuronidation of VPA in pooled maternal sheep liver 235 microsomes. A) Saturation plot of reaction velocity vs. substrate concentration. B) Eadie-Hofstee plot showing monophasic nature of glucuronidation reaction in maternal liver microsomes, v and [S] are the reaction velocity and substrate concentration, respectively. Figure 5.8 - Representative pharmacokinetic fitting of newborn and 238 maternal sheep unbound plasma concentration vs. time data to a one-compartment model with Michaelis-Menten elimination. Actual data (scatter) and model predicted profile (solid line) are shown. A) newborn lamb (NL4124), B) pregnant ewe (E5108). Figure 5.9 - Representative cumulative amount excreted vs. time plots of 242 VPA and VPA-glucuronide in a pregnant ewe (E4241, after maternal administration) and a newborn lamb [NL2243(1)]. Profound differences in the kinetics of urinary excretion of these two compounds relative to the duration of VPA infusion in the ewe and the newborn lamb are evident. Figure 5.10- Average maternal and fetal plasma concentration vs. time 245 profiles of the (E)-2-ene (A and C) and 3-keto (B and D) VPA metabolites in 5 pregnant sheep during and after a 24 h steady-state VPA infusion. Upper panel: maternal VPA infusion; lower panel: fetal VPA infusion. Figure 5.11- Average maternal and fetal plasma concentration vs. time 246 profiles of the (E)-3-ene (A and C) and 4-ene (B and D) VPA XX metabolites in 5 pregnant sheep during and after a 24 h steady-state VPA infusion. Upper panel: maternal VPA infusion; lower panel: fetal VPA infusion. Figure 5.12- Average maternal and fetal plasma concentration vs. time 247 profiles of the 4-OH (A and C) and 4-keto (B and D) VPA metabolites in 5 pregnant sheep during and after a 24 h steady-state VPA infusion. Upper panel: maternal VPA infusion; lower panel: fetal VPA infusion. Figure 5.13 - A) and B) Average maternal and fetal plasma concentration vs. 248 time profiles of 5-OH VPA in 5 pregnant sheep during maternal and fetal VPA infusions, respectively. C) Average maternal plasma concentration vs. time profile of 2-PGA during 24 h maternal steady-state VPA infusion. 2-PGA was below the LOQ in all fetal plasma samples and was measurable in maternal samples from only one ewe during fetal VPA infusion (see Tables 5.16 and 5.17 for details). Figure 5.14- Average plasma concentration vs. time profiles of the VPA 257 metabolites formed via the p-oxidation pathway in 4 newborn lambs during and after a 6 h VPA infusion. A) (E)-2-ene VPA, B) 3-keto VPA, C) (E)-3-ene VPA. Figure 5.15- Average plasma concentration vs. time profiles of the VPA 258 metabolites formed via microsomal oxidation pathways in 4 newborn lambs during and after a 6 h VPA infusion. A) 4-ene VPA, B) 4-keto VPA, C) 4-OH VPA, D) 5-OH VPA. Figure 5.16- Average concentrations of VPA before and after base- 260 hydrolysis in fetal urine samples obtained from 4 pregnant sheep at different times during and after steady-state VPA infusion experiments. A) maternal infusion experiments, B) fetal infusion experiments. xxi List of Abbreviations u. Micron a Alpha, an exponential rate constant (apparent rate of distribution) (3 Beta, an exponential rate constant (apparent rate of elimination) x Duration of infusion u,g Microgram uJ Microliter # Ewe or lamb identification number Approximately [2Hi0]-DPHM Stable isotope (deuterium) labeled diphenhydramine [2H10]-DPMA Stable isotope (deuterium) labeled diphenylmethoxyacetic acid [2H10]-DPHMNOX Stable isotope (deuterium) labeled diphenhydramine-N-oxide 2 H Deuterium ACS American Chemical Society AMN Amniotic AUC Area under the plasma concentration vs. time curve AUMC Area under the first moment curve Bmax Maximal binding capacity °C Degree Celsius ca. Approximate CA Fetal carotid artery Cb or Cbound Plasma concentration of the protein bound drug C u or Cunbound Plasma concentrations of the unbound drug CLf(net) Net clearance from the fetal compartment based on total drug concentrations CLff Total drug clearance from the fetal compartment based on total drug concentrations CLfm Placental clearance of drug from the fetal to maternal compartment based on total drug concentrations CLfo Non-placental clearance of drug from the fetal compartment based XXI I on total drug concentrations CLm(net) Net clearance from the maternal compartment based on total drug concentrations C L m m Total clearance from the maternal compartment based on total drug concentrations Cl_mf Placental clearance from the maternal to fetal compartment based on total drug concentrations C L m o Non-placental clearance from the maternal compartment based on total drug concentrations u CL f(net) Net clearance from the fetal compartment based on unbound drug concentrations u CL ff Total clearance from the fetal compartment based on unbound drug concentrations u CL f m Placental clearance from the fetal to maternal compartment based on unbound drug concentrations u CL f0 Non-placental clearance from the fetal compartment based on unbound drug concentrations u CL m(net) Net clearance from the maternal compartment based on unbound drug concentrations u CL m m Total clearance from the maternal compartment based on unbound drug concentrations u CL m f Placental clearance from the maternal to fetal compartment based on unbound drug concentrations u CL mo Non-placental clearance from the maternal compartment based on unbound drug concentrations CL i n t Intrinsic clearance based on total drug concentrations u CL Intrinsic clearance of the drug based on unbound drug concentrations CL r Renal Clearance of the total drug u CL r Renal Clearance of the unbound drug CLtb Total body clearance based on total drug concentrations u CL tb Total body clearance based on unbound drug concentrations Ciast Plasma concentration at the last sampling time point C m Maternal plasma steady-state total drug concentration after maternal administration XXII I Cm cf C f ' c, C m c", cm Cmax cP Css CV d Dbolus dl DOS DPAA DPHM DPMA DPHMNOX EDTA E H Eg El Maternal plasma steady-state total drug concentration after fetal administration Fetal plasma steady-state total drug concentration after maternal administration Fetal plasma steady-state total drug concentration after fetal administration Total (bound + unbound) plasma concentration of the drug Maternal plasma steady-state unbound drug concentration after maternal administration Maternal plasma steady-state unbound drug concentration after fetal administration Fetal plasma steady-state unbound drug concentration after maternal administration Fetal plasma steady-state unbound drug concentration after fetal administration Centimeter Maximal plasma concentration Plasma concentration Steady state plasma concentration Coefficient of variation Day of gestation or gestational age in days Bolus loading dose Deciliter Disc Operating System Diphenylacetic acid Diphenhydramine Diphenylmethoxyacetic acid Diphenhydramine-N-oxide Ethylenediaminetetraacetic acid Hepatic extraction ratio Gut extraction ratio Electron impact xxiv eV Electron volts F Bioavailability FA Fetal femoral arterial plasma fn Fraction of intravenous dose metabolized in liver F m Fraction of drug converted to metabolite or formation clearance of the metabolite as a fraction of total body clearance of the drug F m ' Formation clearance of the metabolite as a fraction of non-placental clearance of the drug fp Area weighted free fraction of the drug (=AUCunbound/AUCt0tai) F r Fraction of total dose excreted unchanged in urine F-UF Unbound fraction of the drug in fetal plasma at steady-state g Gram GA Gestational age GC Gas chromatography GFR Glomerular filtration rate h Hour H 2 Hydrogen gas HCI Hydrochloric acid HP Hewlett Packard HPLC High performance liquid chromatograph i.d. Internal diameter i.e., id est; that is i.v. Intravenous IVC Inferior vena cava KCI Potassium chloride K d Dissociation constant for drug - plasma protein interaction Kf Apparent first-order rate constant describing the formation of metabolite from parent drug kg Kilogram K m A Michaelis-Menten parameter for enzymatic reactions; substrate concentration at which the reaction velocity is at half-maximal. XXV k0 Drug infusion rate to the mother k0' Drug infusion rate to the fetus kPa Kilopascals LC Liquid chromatograph LOQ Limit of quantitation of the assay M Molar (moles/litre) m Meter m/z Mass to charge ratio MA Maternal arterial plasma mg Milligram MHz Megahertz min Minute mm Millimeter mM Millimolar MRM Multiple reaction monitoring MRT Mean residence time of the total drug u MRT Mean residence time of the unbound drug MS Mass spectrometry MSD Mass selective detector MS/MS Tandem mass spectrometry msec Millisecond MTBSTFA N-Methyl-N-(terf-butyldimethylsilyl)trifluoroacetamide M-UF Unbound fraction of drug in maternal plasma at steady-state MV Maternal femoral vein n Number of subjects or animals NADPH Reduced p-nicotinamide-adenine dinucleotide tetrasodium salt NaOH Sodium hydroxide NCI Negative chemical ionization ng Nanogram xxvi NPD Nitrogen-phosphorous detector o.d. Outer diameter peg Picogram Po 2 Partial pressure of oxygen in the blood Pco2 Partial pressure of carbon dioxide in blood pH Negative logarithm of hydrogen ion concentration P.S.I. Pounds per square inch PTFE Polytetrafluoroethylene p.v. Portal venous Q H Liver blood flow Q u m Umbilical blood flow r Pearson correlation coefficient r2 Coefficient of determination RBC Red blood cells S.D. Standard deviation SIM Selected ion monitoring SIR Single ion recording t Time t.1/2 Half-life in a 1-compartment model based on total drug concentrations ti/2U Half-life in a 1-compartment model based on unbound drug concentrations ti/ 2 p Terminal elimination half-life in a 2-compartment model based on total drug concentrations t1/2Up Terminal elimination half-life in a 2-compartment model based on unbound drug concentrations tiast Time of the last sample tm a x Time of occurrence of maximal plasma concentration TEA Triethylamine TLC Thin layer chromatography TR Tracheal fluid xxvii TRIS U U S P U V V Vc Vd V d s s Vdss' V d U s s V P A U D P G A Vmax Tris(hydroxymethyl) aminomethane International units (insulin dosages) United States Pharmacopeia Umbilical vein Volts Apparent volume of distribution of the central compartment Volume of distribution Apparent steady-state volume of distribution of the total drug Apparent steady-state volume of distribution of the total drug corrected using area weighed free fraction Apparent steady-state volume of distribution of the unbound drug Valproic Acid Times gravity (centrifugal force) Uridine-5'-diphosphoglucuronic acid Maximal velocity of an enzymatic reaction; a Michaelis-Menten parameter xxviii Acknowledgments I would like to express my sincere appreciation and thanks to my research advisors, Drs. K. Wayne Riggs and Dan W. Rurak, for their support, friendship and thoughtfulness throughout my graduate training. The multifaceted, free thinking and independent learning atmosphere that they provided in the lab has been the most fruitful and enjoyable experience of my Ph.D. program. Dr. Dan W. Rurak's excellent surgical expertise with the chronic pregnant sheep preparation lies at the very foundation of many studies conducted as part of this project. Many thanks are due to the members of my graduate research committee, Drs. Frank S. Abbott, James E. Axelson, Guenter Eigendorf and Don M. Lyster, for their valuable time and input. My sincere thanks to Dr. George R. Tonn for laying the ground work for many aspects of this project and for his support and guidance during the initial stages of my training. A very special thanks to my friends, Mr. Harvey Wong and Mr. John Kim, for their selfless help with many aspects of this work, and most of all for the good times that we shared together over these years. Excellent technical assistance of Ms. Caroline Hall and Ms. Nancy Gruber with sheep studies, and of Mr. Roland Burton with LC-MS/MS is very much appreciated. I would also like to thank the undergraduate students, Mr. Sam Au Yeung and Mr. Randy Dumont, for their assistance with biological sample analysis as part of their directed studies program. Thanks to a number of my colleagues, Dr. Wei Tang, Dr. Anthony G. Borel, Mr. Eddie Kwan, Mr. Caly Chien, Ms. Janna Morrison, Dr. Weiping Tan and Mr. Ahmad Doroudian for their timely help and friendship. Financial support in the form of a University of British Columbia Graduate Fellowship and Berlex Award is gratefully acknowledged. These studies were funded by the Medical Research Council of Canada. This thesis is dedicated to my wife, Harsimrat, for always being there for me with her love, assurance and strength, and to my mom, dad and brother for their unconditional love, patience and belief in me. 1 Chapter 1 Introduction 1.1 Drug Use in Pregnancy Although drug use in pregnancy is generally not recommended, many surveys have shown that it does occur in a majority of pregnancies as a result of legitimate medical problems. Rurak et al., (1991) reviewed the available epidemiological data on drug use during pregnancy prior to 1990 and concluded that the incidence of drug use in pregnant women reported in various surveys ranged from 35-100%, with an average of -2-4 drugs taken by each woman during gestation (Rurak et al., 1991). More recent surveys of drug use in pregnancy as well as reviews by other investigators also present similar statistics (Bonati et al., 1990; Collaborative Group on Drug Use in Pregnancy, 1992; de Jong-van den Berg et al., 1993; Irl et al., 1997). These epidemiological data come from all parts of the world indicating that a huge proportion of human fetuses are exposed to drug(s) at some point during gestation. The most commonly used agents include vitamin and mineral supplements, antibiotics, vaginal and urinary tract antiinfectives/antiseptics, analgesics, cough and cold remedies, laxatives, drugs for the control of nausea and vomiting, and antiallergic medications. In addition, drugs are also used to treat certain pregnancy-associated complications (e.g., preeclampsia, Kyle and Redman, 1992; Hauth etal., 1993; preterm labour, Viamontes, 1996) and other existing medical conditions (e.g., epilepsy, Yerby et al., 1992; heart disease, Mitani et al., 1987). Recently, there has also been increased interest in maternal drug administration for the treatment of fetal disorders via transplacental therapy (Ward, 1992). Finally, there is illicit drug use during pregnancy that exposes the fetus to agents, which may have profound effects on fetal neurological, cardiovascular and metabolic functions (Rurak, 1992; Chiriboga, 1993). Hence, in spite 2 of societal efforts to limit drug consumption during pregnancy, the human fetus will likely continue to be exposed to a wide range of non-prescription, prescription and illicit drugs. Continued study of fetal toxicology and adverse effects of drugs is thus warranted, and since these are in large part determined by the extent of fetal drug exposure, investigation of the factors affecting fetal drug disposition is also important. 1.2 Factors Affecting Placental Drug Transfer It has long been recognized that most drugs can cross the placenta from the mother to the fetus and vice versa. Hence, there has been extensive study of the factors affecting placental transfer of the drugs and the resulting fetal drug exposure. Although significant anatomical differences exist among the placentas of different species, the limiting barrier to the movement of drugs and other compounds is essentially a single or multi-layered lipid membrane in all species. Thus, substances can move across the placenta by the usual mechanisms of biological membrane transport, i.e., passive diffusion, active transport, carrier-mediated transfer or facilitated diffusion, pinocytosis, and via passage through the paracellular aqueous pores (Reynolds and Knott, 1989). However, it appears that the major mechanism of placental transfer for most drugs is passive diffusion and other processes are relatively unimportant (Rurak et al., 1991). Thus, it has been suggested that the rate and extent of placental drug transfer is dependent on the physicochemical properties of the drug and is governed by the principles of simple lipid diffusion (Reynolds and Knott, 1989). In agreement with this, low molecular weight lipid soluble compounds (molecular weight < 600) exhibit almost unimpeded placental transfer, whereas there is a significant barrier to the transport of hydrophilic molecules (Reynolds and Knott, 1989; Rurak et al., 1991). Factors 3 influencing placental transfer of drugs have been extensively reviewed (Mirkin, 1973; Levy and Hayton, 1973; Green et al., 1979; Waddell and Marlowe, 1981; Szeto, 1982; Mihaly and Morgan, 1984; Reynolds and Knott, 1989; Rurak et al., 1991; Nau, 1992). These include: 1) physicochemical properties of the drug such as its molecular weight, lipophilicity, and degree of ionization at physiological pH, 2) plasma protein binding of the drug in maternal and fetal circulation, 3) maternal and fetal placental blood flows, 4) stage of placental development, 5) concentration gradient of the free drug across the placenta, and 6) maternal-fetal blood pH gradient (Rurak etal., 1991). 1.3 Study of Placental Drug Transfer and Determinants of Fetal Drug Exposure Due to the obvious implications of the extent of fetal drug exposure in fetal adverse drug effects, the quantitative study of fetal drug exposure has been the subject of numerous scientific investigations in animals as well as humans. Because of practical and ethical restraints, detailed placental transfer and fetal exposure studies cannot be conducted in humans during pregnancy. Hence, most human data on placental drug transfer and fetal exposure are based on the measurement of relative drug concentrations in maternal and umbilical cord blood at birth. Intuitively, these data, although clinically relevant in terms of providing evidence of fetal drug exposure in utero in humans, are highly dependent upon the time of sampling after the last dose of the drug. In addition, at the time of birth, the placenta undergoes considerable trauma and there are alterations in uterine and umbilical blood flow patterns (Hamshaw-Thomas et al., 1984). These factors may influence the maternal-fetal drug concentration relationships at birth and hence the data obtained may not be quantitatively representative of the situation during pregnancy. With the advent of ultrasound assisted fetal blood sampling 4 techniques, maternal-fetal drug concentration ratios have also been determined during, pregnancy in humans (Brown et al., 1990; Moise Jr. et al., 1990; Pons et al., 1991). However, these techniques can also provide maternal and fetal drug concentration data only at a single time point, and unless the drug is at steady-state, the quantitative interpretation of these concentrations with respect to overall fetal drug exposure is difficult. Because of the difficulties in obtaining reliable and detailed in utero placental drug transfer data in humans, the in vitro perfused term human placental preparation has been extensively utilized to obtain such data (Krishna et al., 1993; Johnson et al., 1995; Bassily et al., 1995; Bourget et al., 1995; Bloom et al., 1996). The perfused human term placenta has a number of advantages. These include: 1) ease of availability, 2) this is likely the only way to obtain detailed placental transfer data in humans (Rurak et al., 1991), 3) it can be used to examine the placental transfer of toxic substances such as environmental toxins, dioxins etc. (Bourget et al., 1995), 4) the effects of variables such as changes in plasma protein binding and placental perfusion rate on placental transfer can be studied (Krishna et al., 1993; Bassily et al., 1995). However, this model has a number of serious limitations as well. These include: 1) difficulty in extrapolating the results to the in vivo situation, 2) it provides no information on the extent of fetal exposure resulting from placental transfer of the drug, 3) it lacks maternal and fetal physiological and metabolic components which could themselves have significant influence on the extent of fetal drug exposure (see below), 4) the viability and integrity of the placental membranes may be questionable because of trauma during delivery and deterioration of the tissue after a few hours. Due to the fact that limited information can be obtained from human studies, a number of pregnant animal models have been utilized to examine placental transfer and fetal 5 drug exposure. Small animals such as rats, guinea-pigs and rabbits possess hemochorial placentas, similar to the human. Due to the small size of their fetuses, studies in these species are conducted by measuring drug concentrations in whole conceptuses after maternal drug administration (DeVane and Simpson, 1985; Laishley et al., 1989; Huang et al., 1996). A fetal drug concentration vs. time profile can be constructed by removing different fetuses from the same animal at different time points or by conducting studies in different animals and sampling at different times (Laishley et al., 1989; Huang et al., 1996). The potential limitations of studies in these species include the inability to perform serial sampling of maternal and fetal blood and hence clearly define maternal and fetal pharmacokinetic relationships. In addition, the majority of these studies are conducted in acute anesthetized animal preparations; this could affect maternal and fetal drug disposition and the degree of placental drug transfer (Rurak et* al., 1991). Similar to the human, perfused placental preparations from these species have also been utilized to examine placental transfer of drugs and these studies share the advantages and limitations of the human placental perfusion model. Perhaps the most detailed studies on placental drug transfer and fetal drug exposure have been conducted in two large chronically-catheterized animal models, pregnant sheep and the pregnant mecaque (Macaca nemestrina and M. mulatta). Such animal models allow serial sampling of maternal and fetal blood and also of other fluid compartments (e.g., amniotic and allantoic fluids, fetal tracheal fluid) so that detailed maternal-fetal drug disposition studies can be conducted. Fetal drug exposure indices after single dose administration (fetal-to-maternal plasma AUC ratio) as well as those at steady-state (fetal-to-maternal steady-state plasma concentration ratio) can be determined. In terms of the validity of these data to the human situation, the mecaque 6 model has a hemochorial placenta similar to the human, and also being a primate, it is likely to be more similar to the human in terms of its physiologic and drug metabolism characteristics. On the other hand, the epitheliochorial placenta of sheep has a significantly lower permeability to hydrophilic endogenous compounds as well as polar drugs and their metabolites (Faber and Thornburg, 1983; Olsen et al., 1988; Rurak et al., 1991). Thus, placental transfer of such compounds in sheep may not correlate with that in the human (Reynolds and Knott, 1989). However, it can be safely said that placental transfer of most compounds in the human will be at least equal to or greater than that in sheep, and in the case of lipophilic compounds, the sheep and the human may be fairly similar. Examples of drugs studied in sheep include acetylsalicylic acid (Anderson etal., 1980b), indomethacin (Anderson etal., 1980a), meperidine (Szeto et al., 1978), morphine (Szeto et al., 1982b; Olsen etal., 1988), methadone (Szeto etal., 1982b), acetaminophen (Wang et al., 1986a), omeprazole (Ching et al., 1986), propranolol (Czuba et al., 1988; Morgan et al., 1988), cimetidine (Mihaly et al., 1983), ranitidine (Mihaly et al., 1982a), metoclopramide (Riggs et al., 1988), diphenhydramine (Yoo 'etal., 1986a), quinine and quinidine (Czuba et al., 1991), nifedipine (Nugent et al., 1991), labetalol (Yeleswaram et al., 1992), ritodrine (Wright, 1992) and valproic acid (Gordon et al., 1995). The pregnant mecaque models have been essentially utilized by two groups of investigators for the study of anti-HIV nucleosides or cocaine (Binienda et al., 1993; Pereira etal., 1994; Pereira etal., 1995; Sandberg etal., 1995; Odinecs etal., 1996a; Odinecs etal., 1996b; Odinecs etal., 1996c; Tuntland etal., 1996; Patterson et al., 1997; Tuntland et al., 1998). Thus, it appears that until now, the chronically-catheterized pregnant sheep has been by far the most commonly used model in the study of placental transfer and fetal exposure of drugs. This is likely related to its generally lower cost and ease of handling as compared to non-human primate models. 7 Sheep have also been the most widely used species for the study of fetal physiology during late gestation. This is primarily because of the fact that the physiologic and behavioral parameters in the fetal lamb during the last 3rd of gestation are similar to those of the human fetus under normal conditions and also in response to cardiovascular and CNS drugs (Rurak, 1992; Thornburg and Morton, 1994; Nijhuis and van de Pas, 1992; Szeto, 1992). Many of the above referenced studies in pregnant sheep and the mecaque model have quantified the extent of fetal drug exposure relative to the mother by measuring the fetal-to-maternal plasma AUC or steady-state concentration ratios after maternal drug administration. However, in some of these studies, separate maternal and fetal administration of the drug has been carried out and maternal and fetal placental and non-placental clearances of the drug have been determined using a 2-compartment pharmacokinetic model of the maternal-fetal unit (Figure 1.1) (Szeto et al., 1982a). Drugs studied in pregnant sheep in this manner include morphine (Szeto et al., 1982b), methadone (Szeto etal., 1982b), acetaminophen (Wang etal., 1986a), metoclopramide (Riggs et al., 1990), diphenhydramine (Yoo et al., 1993) and labetalol (Yeleswaram et al., 1993). Similarly, in the mecaque model, maternal and fetal placental and non-placental elimination of the drug has been studied for dideoxyinosine (Pereira et al., 1994), zalcitabine (2',3'-dideoxycytidine, Tuntland et al., 1996), stavudine (2'3'-didehydro-3'deoxythymidine, Odinecs et al., 1996a) and zidovudine (Tuntland et al., 1998). This approach, in addition to providing information on the bi-directional placental transfer rates of the drug, also provides evidence of possible fetal drug elimination in the form of a fetal non-placental clearance parameter. The relative magnitudes of fetal and maternal non-placental clearance can be used as a measure of the extent of 8 development of fetal drug elimination capacity via routes other than the placenta (e.g., metabolism, renal excretion etc.) as compared to the mother. k ' MATERNAL COMPARTMENT CL, mf CL n CL •fm FETAL COMPARTMENT Placenta CL f o Figure 1.1 - A representation of various placental and non-placental drug clearances in the 2-compartment pharmacokinetic model of the maternal-fetal unit ( C L m o - maternal non-placental clearance; CLf 0 - fetal non-placental clearance; C L m f - placental clearance from the mother to the fetus; CLf m - placental clearance from the fetus to the mother). Although the kinetics of placental transfer of most drugs appear to follow the principles of simple diffusion across biological membranes, the extent and kinetics of fetal drug exposure are not related solely to the ease of placental drug transfer. Instead these are the result of a complex interplay between the kinetics of placental drug transfer as well as many other factors related to maternal and fetal components of the pregnant system. These include the relative extent of maternal and fetal plasma protein binding of the drug, the efficiency of maternal and fetal drug elimination via metabolism or renal excretion, and recirculation of the drug between amniotic and allantoic fluid compartments and the fetal circulation. The measurement of fetal plasma AUC or 9 steady-state concentration after maternal drug administration, although a clinically useful index of the extent of fetal drug exposure, does not provide any information about the role of different factors in determining its magnitude. The computation of maternal and fetal placental and non-placental clearances partitions the complex array of these pharmacokinetic factors into 3 main categories, i.e., factors related to the placenta (maternal and fetal placental clearance), the mother (maternal non-placental clearance) and the fetus (fetal non-placental clearance). Thus, it is possible to separately examine the effect of various physicochemical (e.g., drug lipophilicity and pKa efc.^and maternal and fetal biological variables (e.g., plasma protein binding, placental blood flows, drug metabolism capacity) on these 3 classes of pharmacokinetic factors and the resultant effects on fetal drug exposure. This makes it feasible to determine the relative importance of each pharmacokinetic variable in determining fetal exposure to a particular drug, and to make comparisons among different drugs in terms of the most important factor(s). In recent years there has been an increased interest in in utero fetal drug therapy via maternal, intraamniotic, fetal intraperitoneal or intravascular drug administration for the treatment of a number of medical conditions such as arrhythmias, congestive heart failure, pulmonary immaturity, bacterial infections and maternal-fetal HIV transmission (Hamamoto etal., 1990; Ward, 1995; Gilbert etal., 1995; Rayburn, 1997). The detailed studies of drug disposition in the maternal-fetal unit and surrounding fluid compartments (amniotic and allantoic fluids), that are possible with large chronically-catheterized animals, may provide important data for pharmacokinetic rationalization of fetal drug therapy in humans. The study of fetal drug elimination capacity yields important information regarding the ontogenetic development of drug metabolism pathways and 10 renal drug excretion capacity. Such information also has obvious importance in devising better therapeutic strategies for pregnant women and also in the immediate newborn and infant period. 1.4 Diphenhydramine Diphenhydramine [2-(diphenylmethoxy)-N,N-dimethylethylamine, DPHM] is a classical first-generation ^-receptor antagonist of the ethanolamine class (Garrison, 1991). The drug is a low molecular weight (255.4 Da) lipophilic (octanol/water partition coefficient 1862) weakly basic amine with a pKa of 9.0 (Figure 1.2) (de Roose et al., 1970) and is marketed as its hydrochloride salt. Figure 1.2- Chemical Structure of DPHM 1.4.1 Pharmacology and Therapeutic Use 11 The pharmacology of DPHM has been extensively reviewed (Melville, 1973; Hahn, 1978; Drouin, 1985; Garrison, 1991). It is a competitive Hrhistamine receptor antagonist with little or no effect on H2-receptors (Cooper et al., 1990). Thus, DPHM can inhibit the action of histamine mediated allergic and anaphylactic responses such as smooth muscle contraction, wheal formation, edema and increased capillary permeability (Hahn, 1978). It is used therapeutically for the relief of allergic symptoms such as hay fever, allergic rhinitis, cough, urticaria, dermatoses and pruritis (Garrison, 1991). It is also effective in the management of motion sickness, post-operative nausea and vomiting, emesis due to antineoplastic drug use and as a hypnotic (Garrison, 1991). The most common side effects with normal DPHM therapy include sedation or drowsiness and anticholinergic effects, and at higher doses it may result in convulsions and death (Garrison, 1991; Koppel et al., 1987). The newer second generation antihistamines (e.g., astemizole, terfenadine, loratidine) lack the undesirable drowsiness and impaired performance caused by first-generation ^-antagonists primarily because of their much reduced blood-brain barrier permeability (Estelle et al., 1991). 1.4.2 Basic Pharmacokinetics In humans, DPHM is rapidly absorbed after oral administration with peak plasma concentrations occurring 2-4 h after administration (Carruthers et al., 1978; Blyden et al., 1986; Luna et al., 1989; Simons et al., 1990). In healthy adult humans, the peak plasma concentration after a 50 mg therapeutic dose is 40-100 ng/ml (Albert et al., 1975; Carruthers et al., 1978; Blyden et al., 1986; Luna et al., 1989). The drug undergoes substantial first-pass metabolism after oral administration and systemic 12 bioavailability ranges from 40-70% (Albert et al., 1975; Carruthers et al., 1978; Berlinger etal., 1982; Blyden etal., 1986). In human plasma, DPHM is ~70-85% protein bound, with unbound fractions being significantly higher in orientals compared to Caucasians (Spector et al., 1980; Meredith et al., 1984; Zhou et al., 1990). The degree of DPHM binding to human serum albumin is low (Drach et al., 1970), and therefore as with other basic amine drugs, DPHM may bind to a-i-acid glycoprotein in plasma. The apparent volume of distribution of DPHM is large (~3-7 L/kg) and suggests extensive distribution within the body (Carruthers et al., 1978; Spector et al., 1980; Berlinger et al., 1982; Blyden et al., 1986). The classical tissue distribution studies of Glazko and Dill (1949) in rats and guinea pigs showed that the highest concentrations of drug, in decreasing order, were present in lung, spleen, brain, liver and kidney after oral, subcutaneous, intraperitoneal or intravenous administration. The human systemic total body clearance of DPHM is in the range of 6-15 ml/min/kg (Carruthers etal., 1978; Spector et al., 1980; Meredith etal., 1984; Blyden etal., 1986). DPHM clearance appears to show ethnic variation, likely due to inter-racial differences in plasma protein binding (Spector et al., 1980; Zhou et al., 1990). The adult human oral clearance of the drug is considerably higher (~20-30 ml/min/kg) due to a substantial first-pass effect (Luna et al., 1989; Simons et al., 1990). DPHM clearance also exhibits age-dependency, being ~2 fold lower in the elderly (-70 years old) and -2 fold higher in children (-10 year old) as compared to adult humans (-30 years old) (Simons et al., 1990). The terminal elimination half-life of DPHM in humans ranges from 3-9 h (Albert etal., 1975; Carruthers etal., 1978; Spector etal., 1980; Berlinger et al., 1982; Blyden 13 et al., 1986; Luna et al., 1989). The excretion pattern of the drug shows considerable species variation. In rats, after the administration of a 10 mg/kg subcutaneous dose, ~4-6% is excreted unchanged in the urine (Glazko and Dill, 1949), whereas in rabbits this figure is -21% (Hald, 1947). In man, urinary DPHM excretion accounts for only 2-4% of the administered dose (Hald, 1947; Albert et al., 1975). The sum of urinary excretion of unchanged DPHM and its metabolites identified thus far can account for -50-60% of the oral dose in humans (Glazko etal., 1974). In contrast, only -33% of the total administered radioactivity was excreted in urine following subcutaneous administration of 1 4 C-DPHM to rats, with the remainder being recovered in feces, suggesting the possibility of biliary excretion (Glazko et al., 1949). 1.4.3 Metabolism The liver of the rat, guinea pig, and rabbit is highly active in metabolizing DPHM and appears to be the primary site for DPHM biotransformation, although the lung and kidney also have some metabolic activity (Glazko et al., 1949). A reduced clearance of DPHM in patients with cirrhotic liver disease suggests the role of hepatic metabolism in human DPHM clearance as well (Meredith et al., 1984). Liver microsomes from rats, mice, guinea-pigs and rabbits appear to rapidly N-demethylate DPHM (Roozemond et al., 1965; Kataoka and Takabatake, 1971). Incubation of DPHM with liver microsomes from rat, guinea-pig and rabbit results in the formation of methylamine but not dimethylamine (Yamada et al., 1993). This indicates that biotransformation of DPHM proceeds via an N-demethylation and subsequent deamination, rather than direct oxidative deamination of the dimethylamino group (Yamada et al., 1993). It is, however, not clear whether the direct oxidative deamination of N-demethyl DPHM or its further N-14 demethylation to N,N-didemethyl DPHM and subsequent deamination is quantitatively the more important pathway. In rhesus monkeys, the major urinary DPHM metabolites include unchanged DPHM (2-10%), N-demethyl (5-11%) and N,N-didemethyl (3-13%) analogs of DPHM, DPHM N-oxide (7-15%), an acid metabolite (diphenylmethoxyacetic acid, DPMA) (4-20%), a trace of benzhydrol (1-2%), a glutamine conjugate of DPMA (35-59%), and an uncharacterized glucuronide (0-14%) (Drach and Howell, 1968; Drach .et al., 1970). Thus, in this species, N-demethylation followed by oxidative deamination to DPMA and subsequent conjugation of DPMA with glutamine appears to be the major metabolic pathway. Dogs appear to have a similar sequence of metabolic reactions, except that DPMA is conjugated to glycine and this pathway accounts for -40% of the dose (Drach et al., 1970). Drach et al., (1970) also demonstrated the presence of unchanged DPHM, N-demethyl DPHM, N,N-didemethyl DPHM, and DPHM-N-oxide in rat urine. However, in contrast to the rhesus monkey and dog, neither the free nor the conjugated form of DPMA was detectable in rat urine (Drach et al., 1970); thus, the exact nature of the majority of DPHM metabolites in this species still remains to be determined. The metabolic pathways of DPHM in man have also not been fully established; however, the urinary metabolites identified thus far include small amounts of unchanged DPHM and N,N-didemethyl DPHM, and relatively large amounts of N-demethyl DPHM and DPMA, the latter being the major metabolite (Chang et al., 1974). DPMA is excreted both in its free and conjugated form in humans but the exact nature of the conjugate(s) is not known (Chang et al., 1974). Recently, a quaternary ammonium glucuronide conjugate of DPHM has been identified in human urine and it may account for -2-15% of the total administered dose (Luo et al., 1991; Luo et al., 1992; Fischer and Breyer-Pfaff, 1997). 15 Human liver microsomes have also been shown to form this quaternary ammonium glucuronide in significant amounts (Breyer-Pfaff et al., 1997). DPHM has been found to interact by forming a complex with the enzyme systems, cytochrome P-450 (Bast et al., 1990) and monoamine oxidase (Yoshida et al., 1989; 1990). Both these enzyme systems are known to metabolize a number of xenobiotics; however, it is not yet clear which enzyme system or isoform thereof, is responsible for the metabolism of DPHM. In the rat, it has been demonstrated that DPHM and its structural analogues (e.g., orphenadrine) form a metabolic-intermediate complex with cytochrome P-450 2B1/2B2 and 2C6, and thus may act as inhibitors of these isozymes (Rekka et al., 1989; Reidy et al., 1989; Bast ef al., 1990). In this regard, DPHM has been shown to significantly inhibit the clearance of diltiazam in the isolated perfused rat liver (Hussain etal., 1994). 1.4.4 DPHM in Pregnancy A number of surveys of drug use during pregnancy indicate that on average 10-20% of women take antihistamine containing preparations at some point during pregnancy (Peckham and King, 1963; Forfar and Nelson, 1973; Doering and Stewart, 1978; Piper et al., 1987). DPHM is one of the commonly used antihistamines and is taken by 3-7% of pregnant women (Piper et al., 1987; Briggs et al., 1990; Smith et al., 1994; Bologa et al., 1994). The high incidence of DPHM use during pregnancy is evident from the fact that it appears on the lists of 'Drugs and Chemicals Most Commonly Used by Pregnant Women' and 'Drugs of Choice for Pregnant Women' compiled during 1994 at the Hospital for Sick Children in Toronto, Canada (Smith et al., 1994; Bologa ef al., 1994). 16 The predominant therapeutic applications of DPHM during pregnancy are similar to its uses in the normal population. These include the symptomatic treatment of allergic conditions such as rhinitis and contact dermatitis, motion sickness, pruritis, as a hypnotic, and as a cough and cold remedy (Forfar and Nelson, 1973; Ahmed and Kaplan, 1981; Smith et al., 1994). In addition, it is also used in the treatment of certain pregnancy-specific conditions such as nausea and vomiting during the first trimester and a pregnancy-related urticarial condition in late gestation (Nagoette et al., 1996; Ahmad and Kaplan, 1981; The Drugs and Pregnancy Study Group, 1994). Common coughs, colds, nausea and vomiting are among the most common complaints during pregnancy (de Jong-van den Berg et al., 1993; Irl et al., 1997). In one survey, DPHM was administered to pregnant women for an average duration of -19 days over the entire gestation (Forfar and Nelson, 1973). DPHM, despite initial reports of possible association with an increased incidence of oral clefts, is now considered to have no significant teratogenic potential (Saxen, 1974; Briggs et al., 1990; Bologa et al., 1994). There appear to be no systematic studies on the placental transfer and fetal effects of DPHM in the human during the later part of gestation. 1.4.5 DPHM Disposit ion in Chronically-Instrumented Pregnant Sheep: Earlier Studies During the course of the last few years, a series of studies have been conducted in this lab to examine the placental transfer, comparative maternal-fetal pharmacokinetics and metabolism, and fetal effects of DPHM in chronically-instrumented pregnant sheep (Yoo, 1989; Tonn, 1995). Rapid placental transfer of DPHM across the sheep placenta was demonstrated after i.v. bolus administration to the mother, with peak fetal plasma concentrations occurring within 5 min after injection. The fetal-to-maternal AUC ratio of 17 the drug averaged 0.85 indicating significant fetal DPHM exposure after maternal drug administration (Yoo et al., 1986a). Also, the drug did not appear to accumulate or persist to any significant extent in the fetal circulation and the observed maternal and fetal elimination half-lives were similar (40-50 min). Maternal and fetal placental and non-placental clearances of DPHM were determined after separate maternal and fetal steady-state drug administration using a 2-compartment pharmacokinetic model of the maternal-fetal unit (Figure 1.1) (Yoo et al., 1993; Szeto et al., 1982a). The estimates of maternal and fetal placental and non-placental clearances of DPHM along with those of other drugs studied in pregnant sheep are given in Table 1.1. The most significant features of these data are the generally higher magnitudes of fetal weight-normalized placental and non-placental clearances compared to the mother. Also, the fetal non-placental drug elimination capacities for most drugs are remarkable in comparison to the mother. The magnitude of C L f m compared to C L m f is also higher for all other drugs where estimates are available, except acetaminophen. One property of the 2-compartment model is that if maternal and fetal plasma protein binding of the drug is similar and the mode of its placental transfer is via passive diffusion, the magnitude of its placental clearance in the two directions should be equal. This point is discussed at further length in chapter 3. Some of the drugs listed in Table 1.1 do not bind significantly to maternal and fetal plasma proteins (e.g., morphine) and yet exhibit a higher Cl_fm compared to C L m f . For others, this CLf m -CL m f difference still remains after the differences in maternal and fetal plasma protein binding of the drug are taken into account (e.g., DPHM, methadone) (Yoo et al., 1993; Szeto et al., 1982b). These findings are surprising considering the fact that all these drugs appear to undergo placental transfer via passive diffusion. 18 Table 1.1 - Average values of fetal and maternal placental (CL f m and CL m f , respectively) and non-placental clearances (CL f 0 and CL m o , respectively) of various drugs studied in pregnant sheep during the last part of gestation. Clearance ml/min/kg) Drug C L f m a C L m f a C L f o a C L m o b DPHMC 116.8 37.2 88.8 43.5 Morphined 19.4 8.3 42.0 39.7 Methadone6 101.1 32.2 70.9 26.3 Labetalol' 23.4 7.3 27.1 30.5 Acetaminophen9 30.5 31.1 10.8 14.6 Metoclopramideh 103.9 72.0 27.8 46.1 Ritodrine' 9.2 - 52.8 -lndomethacinj 2.0 - 4.0 _ clearances are ml/min/kg estimated fetal body weight; - clearance is ml/min/kg maternal body weight; c - from Yoo ef a/., (1993); d and e - clearance data from Szeto et al. (1982b) and are weight-normalized to maternal and fetal body weights of 70 kg and 3 kg, respectively; f- from Yeleswaram et al. (1993) and clearances were determined using separate maternal and fetal bolus drug administration; 9- from Wang ef al. (1986a); h - from Riggs et al. (1990); '- from Wright ef al. (1991), determined using the Fick principle and only fetal steady-state drug infusion; j-Krishna et al. (1995), determined using the Fick principle and only fetal steady-state drug infusion. Similar to other amine drugs, DPHM accumulates in fetal tracheal and amniotic fluids, with drug concentrations in fetal tracheal fluid being 4-5 times higher compared to fetal plasma (Riggs etal., 1987; Yoo, 1989; Rurak etal., 1991). The concentration of DPHM in amniotic fluid increases progressively during infusion and subsequently declines more slowly than fetal and maternal plasma (Yoo, 1989). Drug appearance in the amniotic fluid may result from fetal renal and tracheal fluid excretion of the drug into this compartment. In addition, there could be direct drug transfer from the fetal blood into the amniotic fluid via the fetal chorioallantoic membranes (Rurak et al., 1991). After injection of DPHM into the amniotic fluid, the drug is preferentially taken up by the fetus 19 via fetal swallowing and absorption, and exchange between amniotic fluid and fetal circulation via the fetal membranes (Yoo, 1989). Thus, the drug present in amniotic fluid may be recirculated in the fetal lamb via these mechanisms (Yoo, 1989). The above results indicate that the fetus is capable of eliminating a substantial amount of DPHM via non-placental routes. Subsequent studies have attempted to examine the components of this high CL f 0 . Since DPHM was found to accumulate in the tracheal fluid to a high extent, Yoo (1989) examined the fetal pulmonary clearance of the drug and found this to be a minor component of CL f 0 (-8% of CL f 0). An extension of the assessment of the contribution of various fetal organs to DPHM elimination are the recent studies employing simultaneous administration of stable-isotope labeled ([2Hi0]-DPHM, with deuterium labels on the two aromatic rings) and unlabeled DPHM (Tonn, 1995). Tonn et al., (1996) examined the in vivo first-pass uptake/metabolism of DPHM by the fetal and adult sheep liver. After randomized simultaneous but separate administration of [2H10]-DPHM and DPHM via the portal and i.v. routes, it was demonstrated that hepatic first-pass extraction of DPHM in adult sheep was ~95%. This indicates that hepatic uptake and/or metabolism may be a major component of DPHM clearance in adult sheep. The study of fetal hepatic uptake of DPHM was much more complicated due to the unique geometry of the fetal hepatic circulation and multiple number of blood flow inputs into the fetal liver. The fetal liver receives ~75% of its blood supply from the umbilical vein, -20% from the portal vein and -5% from the hepatic artery (Holzman, 1984). Also, -50% of the umbilical venous blood flow returning from the placenta passes through the fetal liver before reaching the fetal circulation and the rest bypasses the fetal liver via the ductus venosus (Figure 1.3) (Edelstone et al., 1978). Since drugs transfer across the placenta and reach the fetal circulation via the umbilical 20 vein, the fetal liver may exert a 'partial first-pass effect' on the drug present in the umbilical vein (Figure 1.3). This fetal hepatic first-pass uptake of the drug from the umbilical vein, if present, may be the most significant factor in minimizing fetal drug exposure. Thus, fetal hepatic first-pass extraction of the drug from the umbilical vein was examined by simultaneous bolus or infusion administration of stable-isotope labeled and unlabeled forms of the drug at umbilical and inferior vena caval sites (Figure 1.3). In contrast to the adult sheep, no fetal hepatic first-pass extraction of DPHM was observed after umbilical venous administration (Tonn etal., 1996). The previous studies of Yoo etal., (1993) involved maternal and fetal drug infusions on different days separated by a washout period in order to calculate maternal and fetal placental and non-placental clearances. It has been suggested that any time-related developmental changes occurring in the rapidly growing fetus at this stage of gestation and inter-occasion variability in drug kinetics can bias the results from such experiments (Rurak et al., 1991). In order to eliminate this potential bias in calculated pharmacokinetic parameters, Tonn (1995) employed a simultaneous steady-state infusion of DPHM and [2Hio]-DPHM to the mother and the fetus, respectively, to re-examine the placental and non-placental clearances. In general, results similar to those of Yoo et al., (1993) were obtained; the higher magnitudes of C L f m and CL f 0 as compared to corresponding maternal clearance parameters were still observed (Tonn, 1995). In addition, it was demonstrated in these studies that fetal renal clearance of DPHM accounts for -2% of CL f 0 . 21 Ascending Aorta ing Figure 1.3 - A diagrammatic sketch of the fetal circulation showing the position of the fetal liver and sites of drug administration for the assessment of fetal first-pass hepatic drug uptake from the umbilical vein. 22 During the above simultaneous maternal-fetal steady-state infusion studies, maternal and fetal plasma concentrations of diphenylmethoxyacetic acid (DPMA or [2H10]-DPMA), a major DPHM metabolite in many species, were also measured. After the administration of [2H10]-DPHM to the fetus, much higher concentrations of [2Hi0]-DPMA were detected in the fetal arterial plasma compared to the mother, indicating that fetus can form this metabolite in utero. In contrast to the parent drug, the [2Hio]-DPMA metabolite was not detectable in fetal tracheal or amniotic fluids. In addition, weight-normalized fetal renal clearance of this metabolite was much lower compared to the mother (Tonn, 1995). A similar phenomenon has been observed in this lab with another carboxylic acid drug, indomethacin (Krishna et al., 1995), and may reflect a limited ability of the fetal lamb kidney to excrete organic acids (Elbourne et al., 1990). The likelihood of maternal and fetal secondary metabolism of DPMA to its amino acid or other conjugates, as in other species, was not investigated in these studies. Other possible metabolites of DPHM such as N-demethyl DPHM and N,N-didemethyl DPHM were not detected in maternal or fetal plasma and urine in any significant amounts. Thus, over the years a considerable amount of information has been gathered on the disposition and metabolism of DPHM in the ovine maternal-fetal unit. However, the study of the components of high fetal CLf0 has not yielded quantitatively significant information and major routes of CL f 0 are still not clear. The results obtained also raise the question as to whether the estimated CL f o's for DPHM and other drugs are in fact related to intrinsic fetal drug clearance capacity, or are some feature of the intrauterine environment, or are perhaps just the product of a pharmacokinetic modeling exercise. In addition, the metabolic routes of DPHM in the mother and the fetus have not been fully elucidated. Thus, a meaningful comparison of maternal and fetal drug 23 elimination/metabolism capacity and its role in determining fetal drug exposure is not possible. 1.5 Valproic Acid Valproic acid (2-propylpentanoic acid, VPA) is a low molecular weight (144.2 Da) antiepileptic drug with a unique branched-chain fatty acid structure (Figure 1.4) (Davis et al., 1994; Baillie and Sheffels, 1995). It is considerably less lipophilic compared to DPHM (octanol/water partition coefficient 398) and has a pKa of 4.8. It is available for clinical use as the parent compound, its sodium salt, its amide derivative and as a combination of the parent compound and its sodium salt (Davis et al., 1994). COOH Figure 1.4 - Chemical structure of VPA. 1.5.1 Therapeutic Use, Pharmacology and Adverse Effects VPA possesses a broad spectrum of clinical efficacy against several types of epilepsy including generalized seizures (such as tonic-clonic, myoclonic and absence), some partial seizures (such as simple, complex and secondarily generalized), and 24 combination seizures including those that are refractive to other anticonvulsant drugs (Davis etal., 1994; Bourgeois, 1995). The precise mechanism(s) of action of VPA remains poorly understood. A mechanism involving potentiation of y-aminobutyric acid (GABA) neurotransmission has been the focus of major scientific investigation. Based on the results from numerous animal studies, it has been suggested that VPA increases brain concentrations of the inhibitory neurotransmitter, GABA, via its effects on the enzymes involved in GABA production or degradation. In this context, VPA has been shown to result in increased synthesis of GABA via activation of glutamic acid decarboxylase (GAD) (Loscher, 1981; Loscher, 1989). Blockade of GABA degradation by inhibition of GABA-transaminase (GABA-T) and succinic semialdehyde dehydrogenase (SSADH), two enzymes involved in successive degradation steps of GABA, has also been demonstrated (Loscher and Vetter, 1985; Loscher, 1993; Zeise et al., 1991). Lastly, an enhancement of potassium-induced GABA release into the synapse has also been suggested as one of the possible mechanisms (Gram et al., 1988). The other proposed mechanisms of action include, inhibition of y-hydroxybutyric acid (GHB) (a minor but epileptogenic metabolic product of GABA itself) release (Vayer et al., 1988), inhibition of NMDA receptor-mediated excitation by the actions of aspartate (Zeise et al., 1991), and a nonspecific membrane action which reduces high frequency repititive firing of neurons thorugh effects on sodium and/or potassium channels (Slater and Johnson, 1978; McLean and MacDonald, 1986). There is ample evidence in the literature both supporting and contradicting all of these postulated mechanisms. It is almost certain that more than one molecular mechanism is necessary to explain the broad-spectrum activity of VPA against a variety of epileptic disorders. 25 The most common adverse effects of VPA therapy can be broadly classified into two categories: dose-related side effects and idiosyncratic reactions (Davis et al., 1994; Dreifuss, 1995). The dose-related adverse effects include gastrointestinal side effects (nausea, vomiting and gastrointestinal distress), excessive weight gain, hair loss, CNS effects (tremor, drowsiness, acute confusional states and irritability), and metabolic effects (hypocarnitinemia, hyperammonemia). Many of these side effects can be minimized/eliminated by lowering the dose or by a more gradual increase in dose or by administering the drug in its enteric-coated form. The most serious side effect of VPA therapy is a rare but fatal idiosyncratic hepatotoxicity characterized by microvesicular steatosis and necrosis of the liver (Dreifuss, 1995). In one survey, the overall incidence of valproate induced hepatic failure was 1 in 49,000, and patients on polytherapy with VPA and other antiepileptic drugs appeared to be more susceptible (Dreifuss et al., 1989). A higher incidence of fatalities due to hepatic failure appears to occur in patients <2 years of age, including those on monotherapy (1 in 7000) and especially on polytherapy (1 in 800) (Dreifuss et al., 1989; Bryant and Dreifuss, 1996). However, the overall incidence of VPA-induced hepatoxicity and especially that in <2 years old population appears to have decreased over the years due to the recognition of susceptibility and subsequent changes in prescribing patterns (Dreifuss et al., 1987; Dreifuss etal., 1989; Dreifuss, 1995). 1.5.2 Basic Pharmacokinetics VPA is rapidly and completely absorbed after oral administration in humans with an absolute bioavailability of 90-100% (Klotz and Antonin, 1977; Perucca et al., 1978a). 26 Peak plasma concentrations occur within 1-3 hours for rapid-release dosage forms and within 3-8 h for enteric-coated formulations (Klotz and Antonin, 1977, Gugler et al., 1977; Davis et al., 1994). The drug has a relatively small volume of distribution (0.1-0.2 L/kg in adults and 0.15-0.4 L/kg in neonates, infants and children), indicating that it is likely confined to the vascular space and extracellular fluids (Gugler and von Unruh, 1980; Gugler et al., 1977; Perucca etal., 1978a & b; Gal etal., 1988; Irvine-Meek etal., 1982; Hall et al., 1985; Herngren et al., 1991; Cloyd et al., 1993). This appears to be primarily related to its high degree of ionization at physiological pH (pKa 4.8) and also to its extensive serum/plasma protein binding (~90% at therapeutic concentrations) (Levy and Shen, 1995; Davis et al., 1994). The major binding protein for VPA in plasma/serum appears to be albumin (Kober et al., 1980). The serum/plasma protein binding of VPA is nonlinear and exhibits a concentration dependency, leading to increased unbound fractions at higher total plasma concentrations (Gugler and Mueller, 1978; Yu, 1984; Cramer et al., 1986; Scheyer et al., 1990; Cloyd et al., 1993; Levy and Shen, 1995). VPA has been shown to easily enter the CNS in spite of its high degree of ionization in the blood. Brain and CSF concentrations in humans, as well as animals, range from 60-100% of the serum unbound concentrations and are to some extent related to the serum protein binding of the drug (Levy and Shen, 1995). Carrier-mediated, probenecid-sensitive transport processes have been suggested to play a significant role in CNS transport of VPA (Levy and Shen, 1995). The elimination half-life of VPA ranges from 9-18 h in humans (Nau et al., 1982; Zaccara et al., 1988; Davis et al., 1994; Perucca et al., 1978a & b; Gugler et al., 1977; Bowdle etal., 1980; Bialeref al., 1985; Schapel etal., 1980; Hoffman etal., 1981). VPA is a low clearance drug with plasma clearances in the range of 1.0-3.0 ml/min/kg and 27 0.1-0.3 ml/min/kg based on unbound and total drug concentrations, respectively. VPA clearance is generally higher in children and epileptic patients on polytherapy as compared to adults on monotherapy (Perucca et al., 1978b; Gugler et al., 1977; Bowdle et al., 1980; Bialer et al., 1985; Schapel et al., 1980; Hoffman et al., 1981). The clearance of VPA is dose-dependent, with possible nonlinearities due to alterations in the unbound fraction of the drug and/or saturable metabolism with increasing dose (Bowdle etal., 1980; Gomez Bellver etal., 1993). The major route of VPA elimination in humans is via hepatic metabolism and on average only 1-3% of the total dose is excreted unchanged in urine (Gugler et al., 1977; Gugler and von Unruh, 1980; Dickinson etal., 1989). 1.5.3 Metabolism In spite of the deceptively simple branched short-chain fatty acid structure of VPA, its metabolic fate is extremely complex. Because of its chemical structure, VPA enters into metabolic pathways normally reserved for fatty-acid oxidation (e.g., mitochondrial p-oxidation). In addition, it is also metabolized via the common pathways of xenobiotic metabolism (e.g., glucuronidation, cytochrome P-450 mediated oxidation). The level of complexity in VPA metabolism is evident from the fact that there are approximately 50 known VPA metabolites, with at least 16 being observed consistently in humans (Kassahun et al., 1990; Baillie and Sheffels, 1995). The complex and ever-changing subject of VPA metabolism has been extensively reviewed (Gugler and von Unruh, 1980; Granneman et al., 1984a; Zaccara et al., 1988; Baillie and Rettenmeier, 1989; Baillie and Levy, 1991; Ponchaut and Veitch, 1993; Baillie and Scheffels, 1995) and only a very brief overview will be presented here. The major routes of VPA metabolism 28 can be broadly divided into 3 categories: glucuronidation, mitochondrial p-oxidation and microsomal oxidative metabolism including desaturation and co- and co-1 oxidations (Figure 1.5) (Baillie and Scheffels, 1995). 3-keto VPA 2,4-diene VPA (E)- and (Z)-isomers Figure 1.5 - Metabolic pathways of VPA. Dotted arrows indicate pathways for which direct experimental evidence is lacking. 29 1.5.3.1 Glucuronidation Glucuronidation is the major route of VPA metabolism and results in the formation of 1-O-acyl-p-D-ester linked glucuronide (Dickinson etal., 1979a; Dickinson et al., 1984). In different studies, on average 20-40% (individual subject range 10-70%) of the total dose has been recovered in urine as VPA-glucuronide in humans (Gugler et al., 1977; Dickinson et al., 1989; Levy et al., 1990). Also, it has been suggested that the contribution of the glucuronide pathway to total VPA metabolism may increase as a function of increasing dose due to saturation of the p-oxidation pathway (Granneman et al., 1984a; Pollack etal., 1986; Dickinson etal., 1989). 1.5.3.2 Mitochondrial p-Oxidation A significant fraction of the VPA dose is also metabolized via the p-oxidation pathway in humans. The VPA metabolites formed via the mitochondrial p-oxidation pathway include 2-n-propyl-2-pentenoic acid (2-ene VPA) (formed predominantly as the E-isomer), 2-n-propyl-3-pentenoic acid (3-ene VPA) (predominantly the E-isomer), 2-n-propyl-3-hydroxypentanoic acid (3-OH VPA) and 2-n-propyl-3-oxopentanoic acid (3-keto VPA) (Bjorge and Baillie, 1991; Li et al., 1991). The 3-keto VPA metabolite is a prominent urinary metabolic product of VPA in humans and may account for 10-60% of the total administered dose (Dickinson et al., 1989; Levy et al., 1990; Sugimoto et al., 1996). Small amounts (-1-3% of dose) of other p-oxidation metabolites, (E)-2-ene and 3-OH VPA are also detectable in human urine. Structurally, all these metabolites resemble the intermediates of fatty acid p-oxidation. The fact that these metabolites are 30 excreted in urine in appreciable quantities, while similar intermediates from the p-oxidation of straight-chain endogenous fatty-acids are not, indicates that the branched-chain nature of VPA renders it a poor substrate for its complete oxidation via this pathway (Li et al., 1991). Further experiments conducted via selective deuterium labeling of VPA have demonstrated that 3-OH VPA is not an exclusive product of the p-oxidation pathway (Rettenmeier et al., 1987). Instead, it has a dual origin and may also be formed via cytochrome P-450 mediated microsomal hydroxylation (Prickett and Baillie, 1984; Rettenmeier et al., 1987). Also, 3-keto VPA originates via direct oxidation of 2-ene VPA rather than from the oxidation of 3-OH VPA (Rettenmeier et al., 1987). The (E)-3-ene VPA metabolite arises from the isomerization of (E)-2-ene VPA, and may be further metabolized to (E,E)-2,3'-diene VPA via p-oxidation (Bjorge and Baillie, 1991). Also, (E)-2-ene VPA, (E)-3-ene VPA and (E,E)-2,3'-diene VPA are interconvertible via isomerization and reduction processes, and thus all may serve as precursors of 3-keto VPA (Bjorge and Baillie, 1991). Presumably because of its entry into the p-oxidation pathways, VPA has been found to be an inhibitor of fatty acid p-oxidation via competition with p-oxidation enzymes (Bjorge and Baillie, 1985). 1.5.3.3 Microsomal Metabolism The products of microsomal co- and co-1 hydroxylation of VPA are 2-n-propyl-5-hydroxypentanoic acid (5-OH VPA) and 2-n-propyl-4-hydroxypentanoic acid (4-OH VPA), respectively (Prickett and Baillie, 1984). Further oxidation of 5-OH VPA leads to 2-propylglutaric acid (2-PGA) while that of 4-OH VPA results in the formation of 2-n-31 propyl-4-oxopentanoic acid (4-keto VPA) and 2-propylsuccinic acid (2-PSA) (Granneman etal., 1984b). In addition to 2-ene and 3-ene VPA above, another monounsaturated metabolite of VPA is 2-n-propyl-4-pentenoic acid (4-ene VPA). It has been shown that in contrast to 2-ene and 3-ene VPA metabolites, 4-ene VPA is formed via a distinct microsomal cytochrome P-450 mediated desaturation reaction (Rettie et al., 1987; Rettie et al., 1988). More recently, studies with recombinant human cytochrome P-450 enzymes have demonstrated that CYP2C9 and CYP2A6 enzymes catalyze the formation of 4-ene VPA (Sadeque et al., 1997). The study of 4-ene-VPA metabolite formation and its subsequent fate has received considerable attention because of its possible involvement in VPA-induced idiosyncratic hepatotoxicity. This metabolite structurally resembles a known hepatotoxin, 4-pentenoic acid, and has been shown to cause cytochrome P-450 destruction (Prickett and Baillie, 1986), inhibition of p-oxidation (Bjorge and Baillie, 1985), and microvesicular steatosis in rats (Kesterson et al., 1984; Granneman et al., 1984b). In addition, the formation of this metabolite is enhanced in epileptic patients on polytherapy with other antiepileptic drugs, the population more susceptible to VPA-induced hepatoxicity (Levy et al., 1990). The 4-ene VPA metabolite can be subsequently metabolized via mitochondrial p-oxidation to form the diunsaturated metabolite (E)-2,4,-diene VPA; although the latter species can also arise via microsomal desaturation of (E)-2-ene VPA (Kassahun and Baillie, 1993; Kassahun et al., 1994). One hypothesis of VPA-induced hepatotoxicity is that subsequent oxidative metabolism of 4-ene VPA, possibly via (E)-2,4-diene VPA formation, leads to the generation of chemically reactive and potentially toxic intermediates that are capable of reacting with and depleting mitochondrial glutathione stores (Rettenmeier et 32 al., 1985; Kassahun et al., 1991; Kassahun and Abbott, 1993; Kassahun et al., 1994; Tang and Abbott, 1996). This may eventually result in mitochondrial alterations and inhibition of (3-oxidation, two characteristic features of VPA-induced hepatotoxicity (Kesterson etal., 1984). The plasma and urine concentrations of VPA metabolites exhibit extremely high inter-individual variability. However, the major VPA metabolites in plasma are generally (E)-2-ene VPA, (E,E)-2,3'-diene VPA and 3-keto VPA (Rettenmeier et al., 1989; Kassahun et al., 1990). At therapeutic VPA plasma concentrations (40-100 pg/ml), the plasma concentrations of these metabolites are usually within the range of 1-10 pg/ml. Other metabolites such as 3-ene VPA, 4-keto VPA, 3-OH VPA, 4-OH VPA and 5-OH VPA are also present in significant concentrations (0.5-2 pg/ml) (Rettenmeier et al., 1989; Granneman et al., 1984a). The 4-ene VPA, (E)-2,4-diene VPA, 2-PSA and 2-PGA metabolites are usually present only in trace amounts in plasma (Rettenmeier et al., 1989; Kassahun et al., 1990); however, plasma concentrations of 4-ene VPA may be elevated in patients on polytherapy (Levy et al., 1990). The major urinary metabolites of VPA are VPA-glucuronide (10-70% of the dose) and 3-keto VPA (10-60% of the dose) (Dickinson et al., 1989; Levy et al., 1990; Sugimoto et al., 1996). Other metabolites such as (E,E)-2,3'-diene VPA, 3-OH VPA, 4-OH VPA, 5-OH VPA, 4-keto VPA and 2-PGA may account for 1-5% of the administered VPA dose (Dickinson et al., 1989; Levy et al., 1990). The 4-ene VPA, (E)-2,4-diene VPA, (E)-2-ene VPA, 3-ene VPA, and 2-PSA metabolites are minor urinary products and each accounts for 0-0.5% of the total VPA dose (Dickinson etal., 1989; Levy etal., 1990). 33 In addition to the above metabolites, a number of other minor VPA metabolites have also been identified. These include carnitine and glycine conjugates of VPA, a VPA coenzyme A thioester, 2-OH VPA and a 4,5-epoxide metabolite of 4-ene VPA (Baillie and Sheffels, 1995). The conjugation of VPA with carnitine may be significant in light of the fact that VPA therapy is associated with a secondary carnitine deficiency (Coulter, 1991). 1.5.4 Epilepsy, Pregnancy and Valproic Acid Epilepsy occurs in 0.6-1% of the normal population and 0.5% of all pregnancies occur in epileptic women (Martin and Millac, 1993). Due to its prevalence and high risk, epilepsy is considered the most common major neurologic complication of pregnancy (Rochester and Kirchner, 1997). There are approximately 12,000 births per year to epileptic women in United States alone, and approximately 95% of these women are on antiepileptic therapy during pregnancy (Vorhees etal., 1988). VPA is one of the 4 major drugs (phenytoin, phenobarbital, carbamazepine and VPA) used to treat epilepsy in the pregnant population (Lindhout and Omtzigt, 1994; Malone and D'Alton, 1997). Unfortunately, the incidence of major birth defects in children born to epileptic women is 2-3 times greater compared to the normal population (Kelly, 1984; Yerby, 1994). The occurrence of fetal malformations has been associated with the use of all 4 major antiepileptic drugs. However, causal relationships have not been established because the possible confounding effects of the disease itself on organogenesis are not clear (Lindhout and Omtzigt, 1994; Malone and D'Alton, 1997). All of the above anticonvulsants are associated with major cardiovascular defects, orofacial clefts, genitourinary defects and dysmorphic syndromes (Malone and D'Alton, 1997). In 34 addition, the use of VPA appears to be associated with a 20-fold higher risk of neural tube defects (spina bifida aperta) compared to the general population (Lindhout and Omtzigt, 1994; Yerby, 1991). A fetal valproate syndrome consisting of dysmorphic features such as flat nasal bridge, up-turned nasal tip, down-turned mouth, low-set ears, microencephaly, thin overlapping fingers or toes, and flat orbits has also been reported (DiLiberti et al., 1984). In addition to the teratogenic effects associated with the use of VPA and other antiepileptic drugs, there are also reports that prenatal exposure to these compounds may result in alterations in cognitive function and behaviour during postnatal life (Trimble, 1990). At least in one study, children exposed prenatally to VPA showed poorer motor performance and impaired neurological function at 6 years of age as compared to controls (Koch etal., 1996). VPA has been shown to undergo extensive placental transfer in animals as well as humans (Nau, 1986; Kondo etal., 1987; Nau etal., 1984; Nau etal., 1981; Ishizaki etal., 1981; Dickinson et al., 1979b; Nau and Krauer, 1986; Dickinson et al., 1980). In one study in a single pregnant rhesus monkey (Dickinson et al., 1980), VPA appeared rapidly in fetal blood after maternal i.v. bolus administration and reached concentrations comparable to those in maternal blood within 15 min. Also, during the terminal phase, the average fetal-to-maternal plasma concentration ratio was 1.3, indicating a high degree of fetal exposure (Dickinson et al., 1980). Similar to the monkey, cord-to-maternal blood VPA concentration ratios in humans at birth range from 0.5 to 4.6 (Kondo et al., 1987; Nau etal., 1984; Nau etal., 1981; Ishizaki etal., 1981; Dickinson etal., 1979b; Nau and Krauer, 1986; Nau et al., 1982). It has been demonstrated that the plasma protein binding of VPA gradually decreases in the mother over the course of gestation, whereas that in the fetus gradually increases, such that at birth fetal VPA plasma protein binding 35 exceeds that in the mother (Nau and Krauer, 1986). Also, there is a further reduction in maternal plasma protein binding of the drug at birth due to elevated plasma free fatty acids (Nau et al., 1984; Nau and Krauer, 1986). These phenomena result in fetal accumulation of the drug and a greater than unity cord-to-maternal blood VPA ratio at birth (Nau et al., 1984). The decrease in maternal plasma protein binding of VPA and other anticonvulsants with advancing gestation also results in a significant increase in their clearance and a fall in steady-state plasma concentrations (Yerby et al., 1992). Similar to VPA, the 2-ene and 3-keto VPA metabolites have also been measured in higher concentrations in cord and newborn blood compared to maternal blood (Nau et al., 1981; Nau et al., 1984; Kondo et al., 1987). A mechanism related to maternal-fetal plasma protein binding differences may be operative in this case also; however, this remains to be demonstrated. Rapid placental transfer of VPA has also been demonstrated in the perfused human placental model where placental clearance of VPA was -95% of that of antipyrine (a flow limited marker for placental transfer) (Fowler et al., 1989). In contrast, the placental clearance rate of VPA-glucuronide was only -13% of that of antipyrine, possibly due to its extremely hydrophilic nature (Fowler et al., 1989). In some of the above studies in pregnant women at birth, the elimination of VPA from neonates was also followed (Nau et al., 1981; Nau et al., 1984; Kondo et al., 1987; Ishizaki et al., 1981; Dickinson et al., 1979b). In addition, there are 2 more studies of VPA pharmacokinetics in the immediate newborn period (Irvine-Meek et al., 1982; Gal et al., 1988). In these studies, elimination half-lives of VPA in newborns were in the range of 36 15.1-80 h, and are approximately 2-8 fold longer compared to adults. Similar findings have been reported for newborns of other species such as rats and guinea-pigs (Haberer and Pollack, 1994; Yu et al., 1985; Yu et al., 1987). This indicates a much reduced elimination capacity for the drug during the immediate newborn period in all species studied till date. 1.5.5 VPA Disposition in Pregnant Sheep: Earlier Studies A study was conducted in this lab to examine the placental transfer of VPA in chronically-catheterized pregnant sheep during late gestation (Gordon et al., 1995). After maternal i.v. bolus administration, VPA appeared rapidly in plasma (within 2 min) and the average fetal drug exposure index based on fetal-to-maternal AUC ratio was 0.41 (range 0.29 -0.61). Similarly, after separate fetal i.v. bolus administration, VPA was rapidly detectable in maternal plasma (within 2 min). During both maternal as well as fetal administration, maternal and fetal plasma concentrations of the drug declined in parallel with each other with an apparent elimination half-life of -2-4 h, indicating that the drug does not persist in the fetal circulation. VPA did not appear to accumulate in amniotic and fetal tracheal fluids in contrast to a number of amine drugs studied earlier in this model. A number of human VPA metabolites, such as (E)-2-ene VPA, (Z)-2-ene VPA, (E)-3-ene VPA, (Z)-3-ene VPA, 4-ene VPA, 3-keto VPA, 4-keto VPA, 3-OH VPA, 4-OH VPA, 5-OH VPA and 2-PGA were also detectable in maternal sheep serum. Some metabolites (e.g., (E)-2-ene VPA, 4-ene VPA, 3-keto VPA, 4-OH VPA and 5-OH VPA) were also detected in fetal serum, albeit at lower concentrations compared to the mother. In contrast to the human, however, the diunsaturated VPA metabolites (e.g., (E,E)-2,3'-diene VPA and (E)-2,4-37 diene VPA) were not detected in sheep, indicating a possible species difference (Gordon etal., 1995). 1.6 Rationale The continued use of drugs during pregnancy to treat maternal and fetal disease states and the problem of maternal illicit substance abuse necessitate a better understanding of the factors determining maternal-fetal drug disposition and fetal drug exposure. In the above discussion, a point has been made that such studies cannot be conducted in humans due to practical and ethical constraints, and small animal models do not provide detailed data on maternal-fetal concentration relationships due to technical difficulties related to their small size and blood volume. The chronically-instrumented pregnant sheep is the most commonly used model to study maternal-fetal drug disposition and fetal physiologic functions during late gestation due to the ability to study the fetal lamb for days or weeks in its normal intrauterine environment. Also, the size of the sheep fetus permits serial sampling of blood and other fluids, thereby allowing detailed study of drug disposition in the fetus, surrounding fluid compartments and the mother (Rurak et al., 1991). In comparison to non-human primate models, sheep are much more economical, and are safer and easier to work with. In certain aspects, there are clear differences between the sheep and the human, such as the difference in placental permeability to polar drugs due to differences in placental structure. However, we feel that these differences are not likely to be very great for the lipophilic compound, DPHM, and the relatively polar but low molecular weight, VPA. Also, as with all other animal models, there are likely to be differences in ovine drug metabolizing enzymes and different drug clearance rates compared to the human. However, the overall advantages of sheep for 38 the study of maternal-fetal disposition of lipophilic and low molecular weight compounds far outweigh any disadvantages. The extent of fetal drug exposure after maternal administration, although a clinically useful index, is the result of a complex interaction of a host of pharmacokinetic factors related to the placenta, the mother and the fetus. The determination of maternal and fetal placental and non-placental clearances partitions these factors into their individual contributions in determining fetal drug exposure, and also provides valuable information on in utero fetal drug elimination capacity compared to the mother. For all drugs studied in pregnant sheep, the fetal ability to eliminate drugs via non-placental pathways is remarkable (Table 1.1). However, for none of these drugs have the exact routes of fetal non-placental clearance been fully elucidated, and for many of them, the routes responsible for the majority of maternal non-placental clearance do not appear to fully account for fetal non-placental clearance. This includes studies with DPHM in this lab where there were large apparent differences in hepatic uptake of the drug in adult and fetal sheep (Tonn era/., 1996). Also, for acetaminophen, glucuronidation and sulfation of the drug appear to be responsible for ~97% of maternal non-placental clearance but account for only -33% of fetal non-placental clearance (Wang et al., 1986a). Similar findings have been observed with ritodrine where -35% of the fetal dose could be accounted for by placental transfer to the mother and fetal glucuronidation of the drug (Wright et al., 1991). Thus, it appears important to fully elucidate the components of fetal non-placental clearance of at least one compound so as to validate that this parameter is attributable to actual fetal drug elimination capacity and is not a feature of the intra-uterine environment or a by-product of the pharmacokinetic modeling. This 39 includes the identification of fetal organs responsible for drug elimination as well as of the exact metabolic pathways contributing to this drug elimination. At the outset of this project, we felt that there could be some confounding factors associated with the geometry of the fetal circulation and fetal blood flow patterns that may have led to an apparent lack of any fetal hepatic DPHM uptake from the umbilical vein in previous studies in this lab (Tonn et al., 1996). Thus, we decided to re-examine this fetal hepatic DPHM uptake from the umbilical vein and its contribution to fetal non-placental DPHM clearance. Also, it was felt essential to examine and compare the metabolic pathways responsible for DPHM non-placental clearance and their relative capacity in the mother and the fetus. A detailed analysis of the importance of various placental, maternal and fetal pharmacokinetic factors in determining fetal drug exposure has not been performed for any drug. Also, the last 3rd of gestation is a very dynamic period in terms of fetal development and numerous changes in fetal physiological variables occur during this time. These include possible changes in maternal and fetal plasma protein binding of drugs, development of fetal drug metabolism capacity and renal excretion, and alterations in fetal circulatory and hemodynamic processes with advancing gestation. All these variables could differentially affect the individual placental, maternal and fetal components of drug disposition in the maternal-placental-fetal unit and hence affect fetal drug exposure. Hence, we examined the inter-relationships between maternal and fetal placental and non-placental clearances of DPHM, plasma protein binding, umbilical blood flow and placental transfer and fetal drug exposure as a function of gestational age. 40 VPA is a drug with several contrasting features compared to DPHM. These include a relatively greater polarity, acidic nature, much lower clearance and possible nonlinearities in its pharmacokinetics due to the phenomena of saturable plasma protein binding and metabolism. Thus, the comparative study of DPHM and VPA may provide information on the relative importance of different factors in determining fetal exposure to these two extreme classes of drugs (e.g., low vs. high polarity; low vs. high clearance). In addition, in many species, both these drugs are metabolized via many distinct drug metabolism pathways. Thus, the study of the maternal-fetal metabolism of DPHM and VPA may provide information on the ontogenetic development of fetal xenobiotic metabolism in general. Lastly, both DPHM and VPA appear to be relatively widely used during pregnancy. VPA undergoes extensive placental transfer in humans and is also occasionally administered to epileptic newborns to treat seizures that are refractory to other drugs (Gal et al., 1988). Although data on placental transfer of DPHM in humans are not available, it is likely as rapid and extensive as in sheep. Thus, a detailed study of the maternal-fetal pharmacokinetics and metabolism, placental transfer and fetal exposure of these two drugs is relevant in its own right. 41 1.7 Objectives The major objectives of the research presented in this thesis were: 1. To identify the components of maternal and fetal DPHM non-placental clearance in sheep, including the organs and metabolic pathways responsible for drug elimination. 2. To examine the inter-relationships between maternal and fetal placental and non-placental clearances, plasma protein binding, placental transfer and fetal exposure of DPHM as a function of advancing gestation. 3. To study steady-state placental transfer, fetal exposure, and pharmacokinetics and metabolism of VPA in the sheep maternal-fetal unit. 4. To study VPA disposition during the immediate newborn period in comparison to that in the mother and the fetus. 5. To draw parallels and comparisons between DPHM and VPA in terms of the factors affecting their placental transfer and fetal exposure. 6. To compare the development of fetal (and also newborn in case of VPA) drug metabolism and renal excretion capacity in comparison to the adult for these two drugs. 42 Chapter 2 Materials, Instrumentation and Assay Methodologies During the studies described in subsequent chapters, a number of parent drug and metabolite assay methods were employed. The analysis of DPHM and [2Hi0]-DPHM, and DPMA and [2Hi0]-DPMA was accomplished using two gas-chromatographic-mass spectrometric (GC-MS) methods developed previously in this lab (Tonn et al., 1993; Tonn et al., 1995). In addition, VPA and its metabolites were also quantified using a previously developed GC-MS method (Yu et al., 1995). During the course of these studies, a new high-performance liquid chromatographic-tandem mass spectrometric (LC-MS/MS) analysis method was developed for the simultaneous quantitation of DPHM, [2H10]-DPHM, and their corresponding N-oxide metabolites, DPHM-N-oxide (DPHMNOX) and [2H10]-DPHM-N-oxide ([2H10]-DPHMNOX), respectively. A brief summary of the procedures and specifications of the previously developed methods, and a detailed account of the LC-MS/MS method is presented in this chapter. 2.1 Materials Reference standards, chemicals, reagents and other materials used during the studies described in this thesis, along with the information on their purity (where applicable) and source, are listed below. Unless otherwise specified, the materials were used without further purification or modification. Diphenhydramine hydrochloride [2-(diphenylmethoxy)-N,N-dimethylethylamine] (>99% purity), orphenadrine hydrochloride [N,N-dimethyl-2-[(2-methylphenyl)phenyl]ethylamine] (>99% purity), diphenylacetic acid (>98% purity), ethylenediaminetetraacetic acid (EDTA), sucrose, 43 tris[hydroxymethyl]aminomethane (Trizma® Base), p-nicotinamide-adenine dinucleotide phosphate (reduced) tetrasodium salt ca. 98% (NADPH), uridine-5'-diphosphoglucuronic acid (UDPGA), p-glucuronidase (Glucurase®), sodium 2-propylpentanoate (sodium valproate) and cellophane dialysis membrane "sacks" (molecular weight cutoff = 12,000 daltons) were purchased from Sigma Chemical Co. (St. Louis, MO., U.S.A.). The DPMA metabolite was synthesized and purified using a previously described method (Tonn et al., 1995). Deuterated analogues of DPHM hydrochloride {i.e., [2H10]-DPHM hydrochloride), and DPMA {i.e., [2H10]-DPMA) were also synthesized and purified as described previously (Tonn et al., 1993; Tonn et al., 1995). DPHM-N-oxide hydrochloride (>98% purity) was generously provided by Parke-Davis Pharmaceutical Research Division (Ann Arbor, Ml, U.S.A.). Deuterium-labeled DPHM-N-oxide hydrochloride ([2H10]-DPHM-N-oxide.HCI) was synthesized and purified as described later in this chapter. The metabolites of valproic acid used for the standard calibration curves were synthesized and purified as reported elsewhere (Acheampong et al., 1983; Acheampong and Abbott, 1985; Lee et al., 1989). These metabolites include: (E)-2-propyl-2-pentenoic acid [(E)-2-ene VPA], (E)-2-propyl-3-pentenoic acid [(3)-ene VPA)], 2-propyl-4-pentenoic acid (4-ene VPA), 2-propyl-3-hydroxypentanoic acid (3-OH-VPA), 2-propyl-4-hydroxypentanoic acid (4-OH-VPA), 2-[(E)-1'-propenyl]-(E)-2-pentenoic acid [(E,E)-2,3'-diene VPA], (E)-2-propyl-2,4-pentadienoic acid [(E)-2,4-diene VPA], 2-propylglutaric acid (2-PGA) and 2-propylsuccinic acid (2-PSA). The metabolites 2-propyl-3-oxopentanoic acid (3-keto-VPA) (Kassahun et al., 1990), 2-propyl-4-44 oxopentanoic acid (4-keto-VPA) (Kassahun et al., 1990) and 2-propyl-5-hydroxypentanoic acid (5-OH-VPA) (Rettenmeier et al., 1985) were also synthesized according to previously reported procedures. The following heptadeuterated compounds (utilized as internal standards for VPA and VPA-metabolite analysis) were synthesized as described elsewhere (Zheng, 1993): 2-[2H7]propylpentanoic acid ([2H7]VPA) , 2-[2H7]propyl-2-pentenoic acid (2-ene[2H7]VPA), 2-[2H7]propyl-4-pentenoic acid (4-ene[2H7]VPA), 2-[2H7]propyl-3-oxopentanoic acid (3-keto[2H7]VPA), 2-[2H7]propyl-4-oxopentanoic acid (4-keto[2H7]VPA), 2-[2H7]propyl-3-hydroxypentanoic acid (3-OH[2H7]VPA), and 2-[2H7]propyl-5-hydroxypentanoic acid (5-OH[2H7]VPA). Deuterated benzene ([2H6]-benzene, 99.96% isotopic purity) for the synthesis of deuterated DPHM and its metabolites was obtained from Cambridge Isotope Laboratories (Andover, MA, U.S.A.). Ammonium acetate, sodium carbonate, glacial acetic acid, disodium hydrogen orthophosphate (dibasic), potassium dihydrogen orthophosphate (monobasic), potassium chloride, sodium hydroxide pellets, magnesium chloride, and hydrochloric acid were obtained from BDH Chemicals (Toronto, Ontario, Canada) and were of analytical reagent grade. Anhydrous aluminum chloride, anhydrous sodium sulfate, bromoacetic acid, carbon tetrachloride, magnesium chloride, petroleum ether, sodium metal, and para-toluene sulfonic acid utilized during synthesis procedures were also purchased from BDH Chemical Co. (Toronto, Ontario, Canada). Chloroperoxybenzoic acid, 2-methylglutaric acid (2-MGA), triton-X 100, and deuterium oxide were obtained from Aldrich Chemical Co. (Milwaukee, Wl, USA). Sequanal grade 45 triethylamine (TEA) and N-methyl-N-(i'eAt-butyldimethylsilyl)trifluoroacetamide (MTBSTFA) were purchased from Pierce Chemical Co. (Rockville, IL, U.S.A.). Acetonitrile, ethylacetate, methanol, methylene chloride, toluene, ethanol, isopropanol, acetone, diethyl ether, ethyl acetate and n-hexane were purchased from Caledon Laboratories (Georgetown, Ontario, Canada) and were of distilled in glass HPLC or GC grade. The Bradford protein assay kit was purchased from Bio-Rad Laboratories (Mississauga, Ontario, Canada). Centrifree® micropartition devices were from Amicon (Amicon, Inc., Danver, MA, U.S.A.). Ultra-high purity grade helium, hydrogen, and nitrogen were utilized (Praxair, Vancouver, BC, Canada). Deionized high purity water (referred to as 'distilled or deionized water' in text) was produced on-site by reverse osmosis and subsequent filtration using a Milli-Q® water system (Millipore, Bedford, MA, U.S.A.). Materials used during sheep experiments were as follows. Veramix Sheep Sponges (Tuco Products Co., Orangeville, Ontario, Canada); Pregnant Mares' Serum Gonadotropin (Ayerst Laboratories, Montreal, Quebec, Canada); thiopental sodium injectable 1 g/vial; sodium chloride for injection USP (Abbott Laboratories, Montreal, Quebec, Canada); injectable ampicillin (250 mg/vial) (Novopharm, Toronto, Ontario, Canada); injectable atropine sulfate (0.6 mg/mL) (Glaxo Laboratories, Montreal, Quebec, Canada); heparin 1000 units/mL (Organon Canada Ltd., West Hill, Ontario, Canada); halothane (Ayerst Laboratories, Montreal, Quebec, Canada); and lidocaine 46 2% (Astra Pharma Inc., Mississauga, Ontario, Canada). All injectable drug formulations were purchased from BC Women's Hospital Pharmacy, Vancouver, BC, Canada. Also used were: syringe needles and plastic disposable Luer-Lok® Syringes for drug administration and sample collection (Becton-Dickinson Canada, Mississauga, Ontario, 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.); heparinized Vacutainer® tubes (Vacutainer Systems, Rutherford, NJ, U.S.A.); 15 ml Pyrex® disposable culture tubes (Corning Glass Works, Corning, NY, U.S.A.); polytetrafluoroethylene (PTFE) lined screw caps (Canlab, Vancouver, BC, Canada.); silicone or polyvinyl rubber tubing for catheter preparation (Dow Corning, Midland, Ml, U.S.A.); and cryovials and closures (Nalgene Company, Rochester, NY, 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 quadrapole mass selective detector, and a Vectra 486 25T Computer with MS DOS® HP Model G1030A workstation software was utilized for all GC-MS assays (Hewlett-Packard, Avondale, PA, U.S.A.). The gas-chromatograph was equipped with either a DB1701 (30 m x 0.25 mm i.d.; film thickness 0.25 pm; 5% phenylmethylsilicone and 7% cyanopropylsilicone; J&W Scientific, Folsom, CA, U.S.A; DPHM/[2H10]-DPHM and VPA+VPA metabolites assay methods) or an HP Ultra-2 (25 m x 0.25 mm i.d.; film 47 thickness 0.25 pm; 5% phenylmethylsilicone; DPMA/[ 2Hi 0]DPMA assay method) cross-linked fused silica capillary column. Samples were injected into the GC-MS via a 4 x 78 ® ® mm deactivated Pyrex glass inlet liner and a Thermogreen LB-2 silicone rubber septum. 2.2.2 High-Performance Liquid Chromatography-Tandem Mass-Spectrometry The LC-MS/MS instrumentation consisted of a HP 1090 II liquid chromatograph (Hewlett-Packard, Avondale, PA, U.S.A.) interfaced to a Fisons VG Quattro I Triple Quad Tandem-Mass Spectrometer (Micromass, Cheshire, UK). The operation of both instruments and mass-spectrometric data acquisition were controlled with a Windows-NT based Pentium Pro 200 MHz personal computer using the mass-spectrometry data ® handling software, MassLynx (MicroMass, Cheshire, UK). Chromatographic separations were carried out on a YMC propyl amino (NH2, 100 mm x 2.0 mm ID, 5 urn) column (YMC, Inc., Wilmington, NC, U.S.A.) at ambient temperature. The HPLC autoinjector syringe and sample loop volumes were 25 and 250 uJ, respectively. 2.2.3 Spectrophotometer ® A HP 8452A diode array spectrophotometer equipped with a Vectra computer interface was used for spectrophotometric measurements for the microsomal protein assay. 2.2.4 Physiological Monitoring 48 Physiological monitoring of the animals was performed using a Beckman R-711 Dynograph Recorder (Beckman Instruments, Inc., Palo Alto, CA, U.S.A.), disposable DTX pressure transducers (Spectramed, Oxnard, CA, U.S.A.), cardiotachometer (Model 9857, Sensormedics, Anaheim, CA, U.S.A.), and transit-time blood flow transducers (Transonic Systems Inc., Ithaca, NY, U.S.A.). An Apple lie computer and data acquisition system consisting of an interactive systems analog to digital converter (Daisy Electronics, Newton Square, PA, U.S.A.), and a clock card (Mountain Software, Scott's Valley, CA, U.S.A.) were used for online data acquisition from the polygraph recorder. Blood pH, Po2> and Pco2 were measured using an IL 1306 pH/Blood gas analyzer (Allied Instrumentation Laboratory, Milan, Italy). Blood 02-saturation and hemoglobin concentration were determined using a Hemoximeter (Radiometer, Copenhagen, Denmark). Blood glucose and lactate concentrations were measured with a 2300 STAT plus glucose/lactate analyzer (Y.S.I. Inc., Yellow Springs, OH, U.S.A.). 2.2.5 Other Equipment Also utilized were: a vortex-type mixer and incubation oven (Isotemp model 350) (Fisher Scientific Industries, Springfield, MA, U.S.A.); IEC model 2K centrifuge (Damon/IEC division, Needham Hts., MA, U.S.A.); rotating-type mixer (Labquake model 415-110, Lab Industries, Berkeley, CA, U.S.A.); infusion pumps (Harvard model 944, Harvard Apparatus, Millis, MA, U.S.A.); DIAS Roller pump (DIAS, Ex154, DIAS Inc. Kalamazoo, Ml, U.S.A.); high speed centrifuge model J2-21, ultra-centrifuge model L8-60M or L5-50, JA-17 fixed angle rotor, Ti 50.2 fixed angle rotor (Beckman Instruments, Inc., Palo Alto, CA, U.S.A.). 49 2.3 Analysis of DPHM and [ 2 H 1 0 ] -DPHM in Biological Fluids This was accomplished using the GC-MS assay method developed previously in this lab, using orphenadrine as the internal standard (Tonn etal., 1993). Briefly, appropriate volumes of biological fluids (e.g., plasma, urine etc.) were alkalinized with 0.5 ml of 1M NaOH, and extracted for 20 min with 7 ml of a hexane-2% isopropanol mixture containing 0.05M TEA on a rotary mixer. The organic extract was separated and dried under a gentle stream of nitrogen at room temperature. The residue was reconstituted in 150 uJ of dry toluene containing 0.05M TEA and 2 pi was injected into the GC-MS using the splitless mode of sample introduction (purge time 1.5 min). Chromatographic separations were performed on a DB1701 (30 m x 0.25 mm i.d., 0.25 pm film thickness) fused silica capillary column with helium as the carrier gas at a 12.5 psi column head pressure. The GC operating conditions were as follows. The injection port temperature was 180°C. The oven temperature program consisted of an initial temperature of 140°C for 1 min, a 30°C/min ramp to 200°C, and another 17.5°C/min ramp to 265°C where it was held for 5.0 min. This resulted in a total run time of 12.7 min. The mass spectrometer was operated in electron impact ionization mode (electron ionization energy 70eV) with selected ion monitoring (EI-SIM) at transfer line and ion source temperatures of 280°C and 180°C, respectively. Ion fragment m/z 165 was used to monitor both DPHM and orphenadrine, whereas [2Hi0]-DPHM was monitored using m/z 173. The calibration curve concentration range for this assay is 2.0-200.0 ng/ml for both DPHM and [2Hio]-DPHM. During earlier validation, the inter- and intra-day variability and bias of this assay were <20% at the limit of quantitation (LOQ) and <10% at all other concentrations (Tonn et al., 1993). In addition, the extraction recoveries of 50 DPHM and [2Hi0]-DPHM are nearly complete with the above procedure and analyte stability under the conditions of the assay has been established (Tonn etal., 1993). 2.4 Analysis of DPMA and [ 2 H 1 0 ] -DPMA in Biological Fluids The biological fluid concentrations of the DPHM metabolite, DPMA, and its deuterated analog, [2H-|0]-DPMA, were also measured using the previously developed GC-MS analytical method (Tonn et al., 1995). Briefly, appropriate volumes of biological fluids (e.g., plasma, urine etc.) were acidified with 0.4 ml of 1M HCI, and extracted for 20 min with 5 ml of toluene on a rotary mixer. The organic extract was separated and dried under a gentle stream of nitrogen at 40°C, and the residue was reconstituted in 200 pl of dry toluene. The reconstituted residue was derivatized with 25 pl MTBSTFA (N-methyl-N-(terf-butyldimethylsilyl)trifluoroacetamide) at 60°C for 1 h in order to form tert-butyldimethylsilyl (t-BDMS) derivatives of DPMA and [2H10]-DPMA. A 1 pl aliquot of the derivatized extract was injected into the GC-MS using the splitless mode of sample introduction (purge time 1.5 min). Chromatographic separation was achieved using a HP Ultra-2 (25 m x 0.25 mm i.d., 0.25 pm film thickness) fused silica capillary column with helium as the carrier gas at a 15 psi column head pressure. The GC operating conditions were as follows. The injection port temperature was 280°C. The oven temperature program consisted of an initial temperature of 125°C for 1 min, and a 12.5°C/min ramp to 280°C where it was held for 4.0 min. This resulted in a total run time of 17.4 min. The mass spectrometer was operated in electron impact ionization mode (electron ionization energy 70eV) with selected ion monitoring (EI-SIM) at transfer line and ion source temperatures of 285°C and 180°C, respectively. Ion fragments m/z 165, 183, 177 were used to monitor diphenylacetic acid (DPAA; internal standard), 51 DPMA and [2H10]-DPMA, respectively. The calibration curve concentration range for this assay is 2.5-250.0 ng/ml for both DPMA and [2Hi0]-DPMA. During earlier validation, the inter- and intra-day variability and bias of this assay were <20% at the LOQ and <10% at all other concentrations (Tonn et al., 1995). The extraction recoveries of DPMA and [2Hi0]-DPMA using the above procedure are in the range of ~75-80%, and the analytes are stable under the conditions of the assay (Tonn ef al., 1995). 2.5 Analysis of VPA and its Metabolites in Biological Fluids The biological fluid concentrations of VPA and 12 of its metabolites [(E)-2-ene VPA, (E)-3-ene VPA, 4-ene VPA, (E,E)-2,3'-diene VPA, (E)-2,4-diene VPA, 3-keto VPA, 4-keto VPA, 3-OH VPA, 4-OH VPA, 5-OH VPA, 2-PSA and 2-PGA] were measured during our studies of VPA disposition in maternal, fetal and newborn sheep. A previously developed GC-MS analytical method was utilized for these measurements (Yu ef al., 1995). Briefly, the procedure involves acidification of biological fluid samples to pH 3.0-3.5 with 1M HCI. The samples are then extracted twice with 3 ml ethyl acetate on a rotary mixer for 30 min each. The absorbed water from the combined ethyl acetate extract is removed by vortexing with anhydrous sodium sulfate, and the dry extract is concentrated to -100 uJ under a gentle stream of nitrogen. A 50 pi aliquot of MTBSTFA is then added, samples are derivatized by heating at 60°C for 1 h, and 1 uJ is injected into the GC-MS in splitless mode. Chromatographic separations were performed on a DB1701 (30 m x 0.25 mm i.d., 0.25 pm film thickness) fused silica capillary column with helium as the carrier gas at a 15 psi column head pressure. The GC operating conditions were as follows. The injection port temperature was 250°C. The oven 52 temperature program consisted of an initial temperature of 80°C (0.1 min hold time), a 10°C/min ramp to 100°C (0.1 min hold time), a 2°C/min ramp to 130°C (0.1 min hold time), and a 30°C/min ramp to 260°C (8 min hold time). This resulted in a total run time of 29.5 min. The mass spectrometer was operated in electron impact ionization mode (electron ionization energy 70eV) with selected ion monitoring (EI-SIM) at transfer line and ion source temperatures of 280°C and 180°C, respectively. The calibration curve concentration range for VPA and various metabolites, the internal standards utilized for each compound, and the ions monitored are presented in Table 2.1. Table 2.1 - Calibration curve concentration range of VPA and its metabolites and the corresponding internal standards that were util ized for their quantitation using the GC-MS assay. Various fragment ions that were monitored during selected ion monitor ing are also presented. Analytes to be quantitated Internal Standards Analyte Calibration curve range (pg/ml) Ion monitored Compound Ion Monitored VPA 0.025-20.0 201 [2H7]VPA 208 (E)-2-ene VPA 0.01-8.0 199 (E)-2-ene[2H7]VPA 206 (E)-3-ene VPA 0.0025-2.0 199 (E)-2-ene[2H7]VPA 206 4-ene VPA 0.0025-2.0 199 (E)-2-ene[2H7]VPA 206 (E,E)-2,3'-diene VPA 0.0025-2.0 197 (E)-2-ene[2H7]VPA 206 (E)-2,4-diene VPA 0.0075-6.0 197 (E)-2-ene[2H7]VPA 206 3-keto VPA 0.02-4.0 329 3-keto[2H7]-VPA 336 4-keto VPA 0.0025-2.0 215 4-keto[2H7]- VPA 222 3-OH VPA 0.02-2.0 217 3-OH[2H7]- VPA 224 4-OH VPA 0.04-4.0 100 [2H7]-VPA 208 5-OH VPA 0.02-2.0 331 5-OH [2H7]-VPA 338 2-PSA 0.01-2.0 331 2-MGA 317 2-PGA 0.01-2.0 345 2-MGA 317 53 Previous assay validation studies have established that the variability and bias of this assay for all the compounds within these concentration ranges does not exceed 15% (Yu etal., 1995). 2.6 Simultaneous LC-MS/MS Analysis of DPHM, [ 2 H 1 0 ] -DPHM, DPHM-N-Oxide and [ 2H 1 0 ]-DPHM-N-Oxide in Biological Fluids As mentioned above, a LC-MS/MS analytical method was developed for the simultaneous quantitation of DPHM, [2H10]-DPHM, DPHM-N-oxide (DPHMNOX) and [2Hi0]-DPHM-N-oxide ([2Hi0]-DPHMNOX) in biological fluids obtained from pregnant sheep. This was essential for the study of the DPHM-N-oxide pathway of DPHM metabolism in the mother and the fetus. Development and validation of this assay is described below in detail. 2.6.1 Methods 2.6.1.1 Synthesis and Purification of Deuterium-Labeled DPHM-N-oxide Hydrochloride The [2Hio]-DPHMNOX metabolite was synthesized from [2H10]-DPHM by its oxidation with 3-chloroperoxybenzoic acid using a slight modification of the method described for the synthesis of S- and N-oxides of phenothiazine antipsychotics (Jaworski et al., 1993). For this purpose, [2Hio]-DPHM hydrochloride was converted to its free base by alkalinization of an aqueous solution of the hydrochloride salt with sodium hydroxide. The free [2H10]-DPHM base was extracted with diethyl ether and the solvent was evaporated under vacuum. The [2Hio]-DPHM base (10 mmol) was then dissolved in 30 ml of dry dichloromethane. To this solution, 3-chloroperoxybenzoic acid (12 mmol) was 54 added and the mixture was stirred in an ice bath for 1 h. At the end of the reaction, unreacted 3-chloroperoxybenzoic acid was consumed by addition of an excess of TEA (12 mmol). The crude product was purified by flash column chromatography over silica gel using a benzene and methanol (85:15) solvent mixture as the eluant. The fraction containing [2Hi0]-DPHMNOX was collected, and the solvent was removed under vacuum. The residue was washed repeatedly with hexane and cold methanol (-20°C). The washed residue was dissolved in dry acetone, cooled to -20°C, and [2H10]-DPHMNOX hydrochloride was precipitated by the addition of isopropanol saturated with hydrogen chloride gas. The precipitate was recrystallized from acetone to give a white crystalline powder. The final product gave a single spot on TLC and only one peak on a number of HPLC columns under a variety of elution conditions, indicating acceptable purity for the synthesized metabolite. Also, the HPLC retention time and daughter ion mass spectrum of the purified metabolite were similar to that of the Parke-Davis DPHMNOX standard, except with an expected 10 a.m.u difference in certain fragment masses (Figure 2.1; also see below). 2.6.1.2 Standard Stock Solutions An aqueous stock solution of analytes containing 2.5 p.g/ml of DPHM, 2.6 u.g/ml of [2Hi0]-DPHM (to account for the mass of the deuterium labels), 1.0 ng/ml of DPHMNOX and 1.04 u.g/ml of [2H10]-DPHMNOX (to account for the mass difference of the deuterium labels) was prepared by dissolving appropriate amounts of the analytes (based on free base) in deionized water. Two additional solutions were prepared that were 10 and 25 fold dilutions of the above standard stock. The internal standard (I.S.) 55 solution containing 250.0 ng/ml orphenadrine was prepared by dissolving an accurately weighed amount of orphenadrine hydrochloride in deionized water. 2.6.1.3 Sample Extraction The analytes of interest were extracted from the biological fluid samples using a single step liquid-liquid extraction procedure. Sheep plasma or urine samples (up to 1.0 ml) or the spiked standards were pipetted into clean borosilicate glass tubes with polytetrafluoroethylene (PTFE) lined caps. The sample volume was adjusted to 1.0 ml with deionized water. A 100 pl aliquot of the I.S. solution (containing 25.0 ng orphenadrine) was added to each sample and the samples were alkalinized (pH 11.0) by adding 0.5 ml of a saturated sodium carbonate solution. Ethyl acetate (6 ml) containing 0.05 M TEA was then added to each sample and the tubes were capped. Samples were vortex mixed for 10 sec, extracted with a slow rotary motion on a rotary shaker for 20 min, cooled to -20°C for 10 min (to break any emulsion formed during mixing), and then centrifuged at 3000 g for 10 min. The top organic ethyl acetate layer was separated, transferred to a clean set of tubes, and evaporated to dryness under a gentle stream of nitrogen at 25°C using a Zymark Turbo Vap® LV Evaporator (Zymark Corporation, Hopkinton, MA, USA). The residue was reconstituted in 200 pl of a 9:1 mixture of acetonitrile:water and the tubes were vortex mixed for 30 sec. Samples were transferred to HPLC autosampler vials with 0.35 ml disposable glass inserts and a 10 pl volume was injected into the HPLC. 2.6.1.4 High Performance Liquid Chromatography 56 The samples were chromatographed on a HP 1090 II LC using a 100 mm x 2.0 mm ID, 5 pm YMC amino (-NH2) column (YMC, Inc., Wilmington, NC, U.S.A.) employing normal-phase chromatography. Precolumn filters, with replaceable 2 pm frits, were installed in the LC between the sample loop and the column. Gradient elution was used to achieve a quick run time as well as optimal retention of the compounds on the HPLC column. The chromatographic run began with a 95:5 mixture of acetonitrile: 2 mM ammonium acetate buffer containing 1% glacial acetic acid (pH 3.0). The proportion of the aqueous buffer was increased to 25% in a 6 min linear gradient, held for 0.5 min, brought back to the initial 95:5 proportion at 7.0 min, and again held for 3 min before the next injection. The mobile phase flow rate was 0.4 ml/min with a 50:50 split to the mass-spectrometer and waste. This HPLC procedure resulted in a total run time of 10 min for all 5 compounds of interest. 2.6.1.5 Electrospray Tandem Mass-Spectrometry (MS/MS) The effluent from the HPLC column was split, and 50% (0.2 ml/min) was introduced into the Fisons VG Quattro I Triple Quad Tandem-Mass Spectrometer for detection of the analytes. Nitrogen was used as the nebulizing and bath gas. The compounds were ionized in the positive ion electrospray mode and detected using multiple reaction monitoring (MRM). The ion transitions monitored were m/z 256 ->• m/z 167 (DPHM), m/z 266 m/z 177 ([2H10]-DPHM), m/z 272 -> m/z 167 (DPHMNOX), m/z 282 -> m/z 177 ([2H10]-DPHMNOX), and m/z 270 -> m/z 181 (orphenadrine). These transitions were selected based on the predominant fragmentation pattern of various compounds in their daughter ion spectra (Figure 2.1). The dwell time for each transition was set at 0.2 57 sec with an inter-channel delay of 20 milliseconds to provide optimal sampling of each peak of interest (12-15 scans/peak). Collision induced dissociation (CID) was achieved with argon at a pressure of 3 X 10"4 mbar in the collision cell. For maximal sensitivity, the collision energy, ion source temperature and cone voltage of the mass-spectrometer were optimized at 70 eV, 110 °C, and 30 V, respectively. 2.6.1.6 Calibration Curves and the Regression Model Calibration standards (at concentrations of 0.2, 0.5, 1.0, 2.5, 5.0, 10.0, 25.0, 50.0, 125.0 and 250.0 ng/ml for both amines, and at 0.4, 1.0, 2.0, 4.0, 10.0, 20.0, 50.0 and 100.0 ng/ml for both N-oxides) were prepared by adding appropriate amounts of the prepared standard stock solutions to 1 ml of blank ovine plasma or urine. The I.S. (25.0 ng orphenadrine) was then added to each sample and the samples were extracted and analyzed using LC-MS/MS as described above. Weighted linear regression (weighting factor = 1/y2) was performed between the ratio of peak area of each analyte to that of the I.S. vs. the corresponding spiked concentration in order to reduce bias at the lower concentrations. Linearity of calibration curves was demonstrated by calculating the regression bias. This was accomplished by analyzing 6 sets of calibration curve samples and back calculating the concentration of each standard from the obtained slope, intercept and the peak area ratios. The bias (%) was calculated as: n . B a c k Calculated Concentration-Nominal Concentration A n n % Bias = x 100 Nominal Concentration A bias of < ±15% at each concentration was considered evidence of linearity of the calibration curves. 58 2.6.1.7 Extraction Recovery Absolute recoveries of all analytes in plasma and urine were determined at four different concentrations (2.0, 5.0, 50.0, and 250.0 ng/ml for both amines, and 2.0, 5.0, 50.0 and 100.0 ng/ml for both N-oxides) representing the entire range of the calibration curves. Two sets of samples, the control group and the recovery group, were prepared. Both sets of samples were prepared by spiking 1.0 ml of blank ovine plasma or urine with 25.0 ng of I.S. In addition, the samples in the recovery group were also spiked with known amounts of analytes (as above) at this point. All the samples were then subjected to the extraction procedure described earlier. After drying of the ethyl acetate organic extract, the samples in the control group were also spiked with the above amounts of analytes from an aqueous analyte stock solution. These control group samples were again dried under a gentle stream of nitrogen at 25°C using the Zymark Turbo Vap® LV Evaporator. All samples were then reconstituted as described above (Section 2.6.1.3) and injected into the LC-MS/MS system. The concentrations of the analytes in the control and recovery group samples were measured against the extracted duplicate standard curves prepared in the corresponding biological matrix. The absolute recovery was calculated as the ratio of measured concentration of recovery samples to that of the corresponding control samples at each different analyte concentration. 2.6.1.8 Analyte Stability in Biological Fluid Samples A number of tests were carried out in order to establish the stability of the analytes under the routine sample handling conditions in the lab. This included the following: 59 Bench-Top Stability: Blank plasma was spiked with 250.0 ng/ml of each amine and 100.0 ng/ml of each N-oxide. The samples were left on the bench-top overnight (12 h at ~25 °C) and processed the next day. Freeze-Thaw Stability: Blank plasma was spiked with the analytes as above. The samples were repeatedly frozen (at -20°C) and thawed (on bench-top at ~28 °C) for a total of three cycles, and then analyzed for drug and metabolite concentration. Stability in Saturated Sodium Carbonate Solution: Stability of the analytes in the saturated sodium carbonate solution was evaluated because some tertiary amine N-oxides are known to decompose in plasma after alkalinization (Hubbard et al., 1985; Midha et al., 1993; Lin et al., 1994). Plasma samples spiked at the above concentrations were mixed with 0.5 ml of saturated sodium carbonate solution and left on the bench-top for 1 h. Samples were then processed using the described procedure. Autosampler Stability: Processed samples on the autosampler tray were injected repeatedly 3 times during a 48 h period after extraction. The area counts of peaks and their ratios to those of the I.S. were evaluated. Freezer Stability: Plasma samples were spiked with known concentrations of analytes as above, and processed after a 3 month storage at -20°C. In all stability tests, the concentration of analytes in the samples was measured after appropriate treatment and compared to the nominal values. The analytes were considered "stable" if the measured concentration after the treatment was within ±10% of the nominal value. 60 2.6.1.9 Method Validation Method validation was performed by evaluating the intra- and inter-assay variance and bias (inaccuracy) in the quantitation of quality control samples (QC's). The QC samples were prepared by spiking blank plasma or urine with analytes at concentrations representing the limit of quantitation, and low, medium and high range of the standard curve. Thus, for amines, QC's were prepared at concentrations of 0.2, 1.0, 5.0, 50.0 and 250.0 ng/ml (QC's 1-5, Tables 2.2 and 2.3), whereas those for N-oxides had concentrations of 0.4, 2.0, 20.0 and 100.0 ng/ml (QC's 2-5, Tables 2.2 and 2.3). Intra-assay variance and bias were estimated by quantitating six QC's at each concentration using a duplicate standard curve in one batch. For inter-assay variance and bias, six batches of samples, each consisting of six QC's at each concentration and a duplicate standard curve were analyzed on six separate days. The assay method was also independently cross-validated for the quantitation of DPHM and [2Hio]-DPHM with our earlier GC-MS assay (Tonn et al., 1993). This was accomplished by comparing the results obtained from the two methods for the analysis of plasma samples spiked with 3 different concentrations (5.0, 50.0 and 250.0 ng/ml) of DPHM and [2H10]-DPHM. The cross-validation of the quantitation of N-oxide metabolites could not be performed because to our knowledge no methods have been reported for the measurement of these compounds in biological fluids. 2.6.1.10 Appl icat ion of the Assay to a Sample Study of DPHM, [ 2 H 1 0 ] -DPHM, DPHMNOX and [ 2H 1 0 ]-DPHMNOX Disposit ion in the Ovine Maternal-Fetal Unit 61 A pregnant sheep (125 d gestation, term 145 d) was surgically prepared under halothane anesthesia by placing fluid sampling polyvinyl catheters and other monitoring devices (e.g., ultrasonic blood flow probe) in maternal and fetal blood vessels as described earlier (Rurak et al., 1988). After a recovery period (4 days), an equimolar dose of DPHM (2.5 mg) and [2H10]-DPHM (2.6 mg) was simultaneously administered as a bolus via the fetal lateral tarsal vein. Serial fetal (~2 ml each) and maternal (~ 4 ml each) femoral arterial blood samples were collected at 5, 10, 15, 20, 30, 40, 50, 60, 75, 90, 105, 120, 140, 160, 180, 210, 240, 300 and 360 min after drug administration. Plasma was separated by centrifugation and the plasma samples were stored in borosilicate glass tubes at -20°C until analysis. Blank plasma samples were also collected just before drug administration (-5 min) for use in calibration curve sample preparation. Maternal and fetal plasma samples were then analyzed for DPHM, [2Hio]-DPHM, DPHMNOX and [2H10]-DPHMNOX concentrations using the LC-MS/MS method described above. 2.6.2 Results and Discussion The ability to simultaneously administer unlabeled and stable-isotope labeled drugs and metabolites has significantly improved our ability to study maternal-fetal drug disposition in a scientifically unbiased, efficient and cost-effective way (Tonn et al., 1996; Kumar et al., 1997; also see Chapter 3). The focus of studies described in this thesis was to study the factors affecting fetal drug exposure and to elucidate the in-utero fetal development and functional capacity of various drug-metabolizing enzyme systems as compared to the adult. We planned to achieve this by studying in vivo maternal-fetal drug pharmacokinetics and metabolite formation using a combination of stable-isotope 62 labeled compounds and mass-spectrometry. As mentioned previously, the GC-MS assays to quantiate DPHM, DPMA (major DPHM metabolite in many species), and their corresponding deuterated anlogs ([ 2Hi 0]-DPHM and [2Hi0]-DPMA) were developed earlier in this lab. DPHM-N-oxide is also a major metabolite of DPHM in many species, accounting for ~10-20% of the administered dose (rat, dog, rhesus monkey) (Drach and Howell, 1968; Drach et al, 1970). Hence, we wished to examine the relative capacity of this pathway in maternal and fetal sheep. The rationale for studying DPHM-N-oxide also lies with the fact that N-oxides are commonly formed via microsomal flavin-containing-monooxygenases (Poulsen and Ziegler, 1995; Cashman et al., 1995). In contrast to the cytochrome P450 and phase II conjugation enzymes, there is almost no information in the literature on the extent of development of these enzymes and the pharmacokinetics of tertiary amine N-oxides in the fetus of any species. The study of the DPHM N-oxide metabolite in pregnant sheep required a rapid, sensitive and selective assay method capable of analyzing low concentrations (in the range of ng/ml) of DPHMNOX and its stable-isotope labeled analog, [2Hi0]-DPHMNOX, in biological fluids (e.g., plasma, urine). Since the volume of fluids (e.g., blood) that can be sampled from the fetus is limited, we felt it would be advantageous to include the parent drug as well as its deuterated analog in this assay so as to be able to analyze all these compounds simultaneously in a single run. The tertiary-amine-N-oxide metabolites are generally unstable at the high temperatures encountered in gas-chromatography (Midha et al., 1991). The LC-MS/MS technique offers an obvious choice for the analysis of such compounds in biological matrices due to its sensitivity, selectivity, and chromatography and ionization at a relatively lower temperature. Hence, we developed and validated an LC-MS/MS method for simultaneous 63 quantitation of DPHM, DPHMNOX and their deuterium-labeled analogues in plasma and urine samples obtained from chronically-instrumented pregnant sheep. 2.6.2.1 High Performance Liquid Chromatography of Diphenhydramine and the N-Oxide Metabolite Simultaneous analysis of the tertiary amine, diphenhydramine (DPHM and [2Hi0]-DPHM), and its N-oxide metabolite (DPHMNOX and [2H10]-DPHMNOX), in a single run using LC-MS/MS, presented an interesting analytical challenge. The parent amines and their N-oxide metabolites have widely differing polarities. In general, tertiary amine N-oxides are some of the most polar drug metabolites, whereas the parent amines have a predominantly lipophilic character (Midha et al., 1991). This presented us with some difficulties with respect to the choice of a single stationary phase for the optimal chromatographic retention of both the parent drug and the metabolites. The N-oxide metabolites exhibited little retention on many conventional reversed phase columns such as C18, C8, C2, Phenyl, and Cyano. The extreme reversed-phase conditions (>99% aqueous content in the mobile phase) necessary for only a minimal retention of the N-oxides led to severe "sticking" of the parent amines to the column and subsequent slow elution. This resulted in extremely long run times, severe peak tailing for the parent amines, and also "carry-over" problems. Since, reversed phase chromatography utilizes interactions between the lipophilic moieties of the analytes with the nonpolar stationary phase for analyte retention on the HPLC column, we concluded that the N-oxide metabolites do not exhibit sufficient lipophilic character for optimal interaction with these nonpolar stationary phases. 64 Some investigators have used "ion-pairing" as an approach to improve the retention of tertiary amine N-oxides on reversed phase columns in HPLC analytical methods with ultraviolet detection (Koyoma et al., 1993). This, however, was not possible with our method because electrospray mass-spectrometry prohibits the use of non-volatile additives such as ion-pairing agents. This led us to consider the use of normal phase chromatography in order to utilize the polar component of the analyte molecules for their interaction with and retention on a relatively polar column. For normal phase chromatography, the propyl amino (-NH2) phase provides a useful alternative to silica. In contrast to silica, the -NH 2 phase is compatible with the aqueous components of the mobile phase. This stationary phase essentially allows the use of the same solvents as in traditional reversed phase chromatography (in different proportions compared to reversed phase mode, with water being the stronger solvent) without any restriction to purely organic mobile phases (e.g., hexane, dichloromethane) as is the case with silica. This is important because a complete lack of an aqueous or buffer component in the mobile phase could lead to inadequate ionization of the analytes during the electrospray ionization process in the mass spectrometer. In our experiments, both the parent amines and their N-oxide metabolites exhibited excellent retention on a relatively short (column length 10 cm) -NH 2 column (Figure 2.2B). However, due to greatly different polarities of the parent drug and metabolites, and in order to achieve a balance between adequate retention of all analytes and a rapid analysis time (~ 10-15 min), gradient elution was found to be necessary. In general, low mobile phase flow rates (~50-100 pl/min) provide maximal sensitivity in LC-MS/MS assays. However, it is difficult to use gradient elution at these low flow rates because of a relatively large void volume between the HPLC pump and the column (~1 65 ml in the HP 1090 series II chromatograph). Hence, we decided to use a higher mobile phase flow rate (0.4 ml/min) with a 50:50 split to the mass-spectrometer and waste. This provided sufficient sensitivity for all analytes for our purposes and a rapid run time (10 min) for all five compounds of interest (Figure 2.2B). Using normal phase chromatography and the described gradient time-table, the relatively lipophilic amines, orphenadrine, DPHM and [2H10]-DPHM, eluted earlier (at 3.06, 3.12, and 3.12 min, respectively) compared to the more polar N-oxides (at 5.27 min). The variations in retention times of different analytes within a single run were <5%. Although the two groups of compounds (tertiary amines and N-oxides) were well resolved from each other, it was neither possible nor necessary to chromatographically resolve different compounds within the amine or the N-oxide group (Figure 2.2B; also see below). Tertiary amine drugs may exhibit severe peak-tailing on many reversed phase columns. However, the use of an -NH 2 column, in addition to providing optimal retention for all the compounds, also resulted in excellent symmetrical peak shape (Figure 2.2B). Use of the acidic ammonium acetate buffer, pH 3.0 (2 mM ammonium acetate and 1% acetic acid) instead of pure water led to an improvement in peak shape for all the compounds and significantly reduced peak tailing for the parent amines. 2.6.2.2 Tandem Mass Spectrometric (MS/MS) Detection of Analytes Figure 2.1 shows daughter ion mass spectra of all the analytes and the internal standard, orphenadrine, in the positive ion electrospray mode. All analytes appear to predominantly fragment at the ether linkage on the aliphatic side chain of the molecules, thus forming analogous daughter ions. Some additional fragmentation also takes place 66 and the corresponding m/z assignments are depicted in Figure 2.1. Our initial attempts to quantitate these compounds involved the use of single ion recording (SIR) by monitoring the [M+H]+ ion for each analyte in the first quadrupole of the mass spectrometer. However, for some ions, this resulted in significant cross-over of signal from one channel to the other (e.g., m/z 266 and m/z 270 for [2H10]-DPHM and orphenadrine, respectively), especially at higher analyte concentrations. Mass-spectrometric resolution of these interferences was important because it was not possible to chromatographically resolve many of these compounds from one another due to very subtle differences in their chemical structure (Figures 2.1 and 2.2). The use of multiple reaction monitoring mode (MRM) helped to eliminate these interferences by utilizing the enhanced selectivity of MS/MS detection as compared to the single mass spectrometer configuration in SIR. Also, the MRM mode provided much cleaner baselines compared to SIR, leading to an increase in signal to noise ratio. The molecular ion [M+H]+ and the daughter ion formed by fragmentation at the ether linkage are the major ions in the mass spectra of all the analytes (Figure 2.1) and would be expected to provide maximal sensitivity for the analysis of these compounds. Based on this, the mass transitions of m/z 256 -» m/z 167 (DPHM), m/z 266 -> m/z 177 ([2H10]-DPHM), m/z 272 -> m/z 167 (DPHMNOX), m/z 282 -> m/z 177 ([2H10]-DPHMNOX), and m/z 270 -» m/z 181 (orphenadrine) were selected for the detection of the compounds of interest (Figures 2.1 and 2.2). A number of mass-spectrometric parameters such as ion source temperature, argon pressure in the collision cell, collision energy of the collision gas, and cone voltage were optimized to achieve maximal sensitivity. In general, lower ion source temperatures would be safer in terms of the stability of the N-oxide metabolites. In our experiments, 67 [ 2H 1 0]-DPHM-N-Oxide 2 H 5 " 193 j L | C H 2 - C H 2 - N — O - H4 ] \H3 L 60 80 100 120 H O 160 180 200 220 240 260 280 3d Figure 2.1 - Positive ion electrospray daughter ion mass-spectra of DPHM, [ 2 Hi 0 ] -DPHM, DPHM-N-Oxide, [2H10]-DPHM-N-Oxide, and the internal standard, orphenadrine. 68 B 1 0 6 % m/z 282 H> 177 1.50 ?_62 [2H10]-DPHM-N-Oxide 10J m/z 272 -> 167 " ' I n q i i i i | i 5.27 A DPHM -N-, , 6 | i 0 2 Oxide 10E % oi m/z 270 -> 181 10E| % 0* 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 m/z 266-> 177 0.04 TOO m/z 256-* 167 0.04 ^ 1 2 Orphenadrine (I.S.) f H 1 0]- DPHM r" 1 1 5.32 DPHM i 1 1 1 1 1 1 | 1 1 1 1 1 1 1 1 1 1 1 i i ^ ^ ^ p - " 1 1 1 1 1 1 1 11 i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 i 1 1 T i m e 1.00 2.00 3.00 4.00 5.00 6.00 Figure 2.2 - LC-MS/MS MRM ion chromatograms of: (A) an extracted blank sheep plasma sample, and (B) an extracted plasma calibration standard containing 0.4 ng/ml each of [2Hi0]-DPHM-N-oxide and DPHM-N-oxide, and 1 ng/ml each of [2Hi0]-DPHM and DPHM. The y-axis scales in panel (A) have been magnified to clearly show the baselines at different MRM ion transitions. The HPLC and MS/MS conditions and specifications are described in the text. 69 lower temperatures (80 - 90°C), however, resulted in a loss of sensitivity and deposition of material in the mass spectrometer, thus requiring frequent cleaning of the ion source. An increase in ion source temperature beyond 110°C did not offer any significant increase in sensitivity and the N-oxide metabolites appeared to decompose at temperatures higher than 150°C. Thus, an ion source temperature of 110°C was considered optimal for final analysis. Similarly, the argon pressure in the collision cell, the collision energy of the argon molecules and the cone voltage were optimized for maximal sensitivity at 3 X 10"4 mbar, 70 eV and 30 V, respectively. 2.6.2.3 Extraction Method and Recovery A number of organic solvents and solvent mixtures such as dichloromethane, dichloromethane containing 2% isopropanol, hexane containing 2% isopropanol, toluene, and ethyl acetate were evaluated for maximizing the extraction recovery of the analytes. Triethylamine (TEA; 0.05 M) was included in all the solvent systems in order to prevent nonspecific binding of the analytes to glassware (Tonn et al., 1993) (also see below). All solvent systems were very efficient in terms of extracting the parent amines. However, due to the extreme polarity of the N-oxide metabolites, their recoveries were low (ranging from <20% for the hexane-2% isopropanol mixture to ~50% for the dichloromethane-2% isopropanol mixture). Extraction with an ethyl acetate-0.05 M TEA mixture provided maximal recoveries for the N-oxide metabolites and also resulted in clean extracts devoid of any chromatographic or mass-spectrometric interference from the biological matrix (Figure 2.2A). Thus, the ethyl acetate-0.05 M TEA mixture was chosen as the final extraction solvent. The recoveries for all the analytes were consistent (coefficient of variation <10% for the parent amines and <15% for the N-70 oxides) and independent of the analyte concentration; hence an overall mean recovery was calculated. The mean recoveries for DPHM, [2Hi0]-DPHM, DPHMNOX and [2H10]-DPHMNOX from ovine plasma were 79.7 ± 6.6 %, 76.6 ± 6.6 %, 71.6 ± 9.4 % and 69.0 ± 8.3 %, respectively. Similar recoveries were obtained for the extraction of urine samples. Earlier we observed a dramatic improvement in DPHM and [2H10]-DPHM extraction recovery by the addition of 0.05M TEA to a hexane and 2% isopropanol solvent mixture, possibly due to a reduction in nonspecific binding of the tertiary amines to active sites on glassware (Tonn et al., 1993). A similar phenomenon was observed with ethyl acetate extraction; the recovery of all analytes was reduced by ~20-30% when 0.05M TEA was not included in the extraction solvent. Some investigators have demonstrated decomposition of the N-oxide metabolites in plasma on alkalinization with sodium hydroxide (Hubbard et al., 1985; Midha et al., 1991; Lin et al., 1994). These metabolites, however, appear to be stable with the use of saturated sodium carbonate solution (Hubbard et al., 1985; Lin et al., 1994). Hence, we decided to use a saturated sodium carbonate solution for alkalinization of samples in this method. The DPHM N-oxide metabolites were found to be stable in sodium carbonate (see below). Also, the use of a saturated sodium carbonate solution resulted in cleaner extracts as compared to sodium hydroxide as evidenced by a reduction in baseline noise. 71 2.6.2.4 Analyte Stability in Biological Fluid Samples Analyte stability in biological samples under conditions encountered in their routine handling is of paramount importance in order to accurately study drug disposition. Thus, a number of studies were conducted to evaluate stability of the analytes in biological fluid samples under conditions simulated to match routine sample handling. The mean analyte concentrations measured in samples subjected to a 12 h bench-top stability test, 3 freeze-thaw cycles, 1 h exposure to saturated sodium carbonate solution, and a 3 month freezer storage at -20 °C were found to be within ±10% of the nominal concentrations with an acceptable coefficient of variation (<10%; according to acceptability standards established for bioanalytical assays by Shah et al., 1992). There was no change in analyte to I.S. peak area ratios of various compounds during repeated injections of the processed samples on the autosampler tray for up to 48 h after extraction. Based on these results, it was concluded that the analytes would be sufficiently stable in biological samples during actual freezer storage and analysis conditions. 2.6.2.5 Calibration Curves and the Regression Model The calibration curves for all the analytes showed good linearity in the concentration range tested (0.2 - 250.0 ng/ml for parent amines and 0.4 - 100.0 ng/ml for N-oxides). Representative calibration curves of all 4 analytes are shown in Figure 2.3. Weighted linear regression (weighting factor = 1/y2) was performed on all the calibration curve data in order to reduce bias at the lower concentrations. This weighting function resulted in an acceptable regression bias (Shah et al., 1992) for all analytes at the lower 72 as well as the upper limits of the calibration curves (-11.6 % and +1.5 % for DPHM; -11.3% and -1.5% for [2Hi0]-DPHM; -7.1% and +3.0% for DPHMNOX; +9.4% and -8.6% for [2Hio]-DPHMNOX at the lower and upper end of the calibration curve, respectively), again indicating good linearity. The peak response for the N-oxides was nonlinear at concentrations greater than 100.0 ng/ml. This could be due to fact that at higher concentrations, partial dimerization of the N-oxides was observed. The slopes of the [2Hio]-DPHM and [2Hi0]-DPHMNOX standard curves were consistently higher compared to those of DPHM and DPHMNOX, respectively. This indicates that the fragmentation patterns of the compounds may have been quantitatively altered by the presence of deuterium labels. The LOQ's of 0.2 ng/ml and 0.4 ng/ml were established for the parent amines and N-oxides, respectively, based on a signal to noise ratio of at least 20 and an acceptable variability and bias (Shah et al., 1992) at this concentration (see below). The LOQ of 0.2 ng/ml for DPHM and [2H10]-DPHM achieved in this assay method is an order of magnitude lower than that of our earlier GC-MS method (2.0 ng/ml) (Tonn et al., 1993) indicating a significant improvement in sensitivity. 2.6.2.6 Method Validation The validation of the assay involved estimation of intra- and inter-assay variability and bias in ovine plasma and urine. The results from plasma validation are presented in Tables 2.2 and 2.3. Variability and bias data in urine were similar and are not presented here. The intra- and inter-assay variabilities (CVs) for all analytes were <15% below a concentration of 2.0 ng/ml and < 10% at all other concentrations. The mean intra- and 73 Concentration (ng/ml) 0 30 60 90 120 150 180 Concentration (ng/ml) re 2.3 - Representative calibration curves of: (A) DPHM and [2H10]-DPHM, and (B) DPHM-N-oxide and [2H10]-DPHMN-N-oxide, in sheep plasma. 74 inter-assay bias (inaccuracies) ranged from -6% to +12% of the nominal concentration for the amines and from -10% to +14% for the N-oxides over the calibration curve concentration range (Tables 2.2 and 2.3). Table 2.2 - Intra-assay variability and bias of the LC-MS/MS analytical method for DPHM, [ 2 H 1 0 ] -DPHM, DPHM-N-oxide and [2Hi 0 ]-DPHM-N-oxide in ovine plasma. Analyte QC-1 QC-2 QC-3 QC-4 QC-5 Nominal Concentration 3 0.20 1.0 5.0 50.0 250.0 Measured Concentration 3 0.20 1.04 5.0 55.3 239.9 DPHM S D . 0.02 0.05 0.32 1.8 6.8 C V . (%) 9.7 5.1 6.4 3.3 2.8 Bias (%) -0.5 + 3.8 + 0.2 + 10.6 -4.0 Nominal Concentration 3 0.20 1.0 5.0 50.0 250.0 Measured Concentration 3 0.19 1.05 5.0 56.0 241.4 [2H10]-DPHM S.D. 0.02 0.05 0.25 1.3 8.7 C V . (%) 13.2 4.4 5.0 2.3 3.6 Bias (%) -5.4 + 5.1 + 0.8 + 12.0 -3.4 Nominal Concentration 3 0.40 2.0 20.0 100.0 Measured Concentration 3 0.45 2.48 21.8 94.7 DPHM-N- S.D. 0.06 0.22 2.1 6.6 oxide C V . (%) 13.3 9.0 9.6 7.0 Bias (%) + 12.8 + 14.0 + 9.2 -5.3 Nominal Concentration 3 0.4 2.0 20.0 100.0 Measured Concentration 3 0.41 2.09 21.7 94.9 [2H10]-DPHM- S.D. 0.06 0.28 1.8 4.8 N-oxide C V . (%) 15.0 13.4 8.3 5.0 Bias (%) + 3.0 + 4.4 + 8.6 -5.1 - All concentrations are in ng/ml. The cross-validation of the LC-MS/MS assay with our earlier GC/MS method yielded excellent agreement between the two methods. The concentrations of DPHM and [2H10]-DPHM measured in 1.0 ml aliquots of the plasma samples spiked at 5.0, 50.0 and 250.0 ng/ml by the two analysis methods were highly correlated (Pearson correlation coefficient, r = 1.0 at all the concentrations) and were not significantly different from each other (unpaired t-test, p>0.05). This indicates that the plasma concentrations of both the parent amines can be measured with a high degree of confidence by either of these two methods. However, as discussed above, the current method offers the Table 2.3 - Inter-assay variability and bias of the LC-MS/MS analytical method for DPHM, [ 2 H 1 0 ] -DPHM, DPHM-N-oxide and [2H 1 0 ]-DPHM-N-oxide in ovine plasma. Analyte QC-1 QC-2 QC-3 QC-4 QC-5 Nominal Concentration" 0.20 1.0 5.0 50.0 250.0 Measured Concentration 8 0.21 1.06 5.5 54.5 237.7 DPHM S.D. 0.01 0.04 0.3 0.9 4.3 C V . (%) 5.6 3.5 5.9 1.7 1.8 Bias (%) + 5.3 + 6.3 + 11.0 + 9.0 -4.9 . Nominal Concentration 8 0.20 1.0 5.0 50.0 250.0 Measured Concentration 8 0.20 1.09 5.5 54.8 238.2 [2H10]-DPHM S.D. 0.02 0.05 0.3 1.1 4.5 C V . (%) 10.1 4.6 5.6 2.0 1.9 Bias (%) -1.1 + 8.8 + 10.1 + 9.6 -4.7 Nominal Concentration" 0.40 2.0 20.0 100.0 Measured Concentration" 0.40 2.24 22.1 90.9 DPHM-N- S.D. 0.05 0.14 0.6 3.3 oxide C V . (%) 12.9 6.1 2.8 3.6 Bias (%) + 0.6 + 12.0 + 10.4 -9.1 Nominal Concentration 8 0.4 2.0 20.0 100.0 Measured Concentration 8 0.41 2.02 22.1 92.0 [2H10]-DPHM- S.D. 0.03 0.25 0.6 2.8 N-oxide C V . (%) 6.7 12.2 2.8 3.0 Bias (%) + 2.6 + 0.9 + 10.4 -8.0 a - All concen trations are in ng/ml. 76 advantage of a lower LOQ (0.2 ng/ml vs. 2.0 ng/ml) compared to the earlier method. In addition, this method can also simultaneously quantitate the N-oxide metabolite of the drug with good sensitivity and selectivity. 2.6.2.7 Appl icat ion of the Assay to a Sample Study of DPHM, [ 2 H 1 0 ] -DPHM, DPHMNOX and [ 2H 1 0 ]-DPHMNOX Disposit ion in the Ovine Maternal-Fetal Unit The developed assay was applied to a pharmacokinetic study in one chronically-catheterized pregnant sheep, designed to test the bioequivalency of DPHM and [2H-|0]-DPHM in the fetal lamb in terms of the parent amine disposition, formation of DPHMNOX and [2H10]-DPHMNOX from the two compounds, and the clearance of DPHMNOX and [2H10]-DPHMNOX. This bioequivalency is extremely important in order to obtain meaningful pharmacokinetic data by utilizing the stable-isotope labeled drug. Figure 2.4 shows the fetal femoral arterial plasma profiles of DPHM, [2Hi0]-DPHM, DPHMNOX, and [2Hi0]-DPHMNOX after simultaneous equimolar fetal intravenous bolus administration of DPHM (2.5 mg) and [2Hi0]-DPHM (2.6 mg). These data show that plasma concentrations of DPHM and [2Hi0]-DPHM decline rapidly after drug administration and are virtually superimposable. This indicates approximately equal rates of clearance of these two compounds from the fetal circulation. Also, the corresponding N-oxide metabolites (DPHMNOX and [2H10]-DPHMNOX) were detectable in fetal plasma (Figure 2.4). The concentration of metabolites increased gradually and then declined rapidly in parallel with the parent drug. Similar to the parent drugs, the concentrations of the two metabolites were virtually identical indicating a lack of any effect of the deuterium labels on this metabolic pathway. Thus, these data demonstrate the absence of any isotope-effect in the disposition of [2Hio]-DPHM compared to DPHM, 77 1000 -q 0 50 100 150 200 Time (h) Figure 2.4 - Femoral arterial plasma concentration vs. time profiles of DPHM, [2H10]-DPHM, DPHM-N-oxide, and [2H10]-DPHM-N-oxide in a fetal lamb after a simultaneous equimolar bolus dose of DPHM and [2Hi0]-DPHM via the fetal lateral tarsal vein. and also for [2H10]-DPHMNOX compared to DPHMNOX. In addition to fetal plasma, the maternal femoral arterial plasma samples collected at the corresponding time points were also analyzed for the labeled and unlabeled parent drug and the metabolite. However, all compounds were present at much lower concentrations (near LOQ) compared to the fetal plasma. The higher plasma concentrations of the N-oxide metabolites in fetal plasma compared to maternal plasma indicate the ability of the fetus to form this metabolite after DPHM and [2H10]-DPHM administration (also see Chapter 3). 78 In conclusion, we have developed and validated an LC-MS/MS assay for the simultaneous quantitation of DPHM, [2H10]-DPHM, DPHMNOX and [2H10]-DPHMNOX in plasma and urine samples obtained from chronically-instrumented pregnant sheep. The assay is rapid (fast sample processing, a 10 min run time), sensitive and selective (no interference from biological matrices, LOQ's of 0.2 and 0.4 ng/ml for amines and N-oxides, respectively) and robust (acceptable variability and bias, sample stability). This assay and three previously developed GC-MS analytical methods (for DPHM and [2Hi0]-DPHM, DPMA and [2H10]-DPMA, and VPA and its metabolites) were utilized in the studies of DPHM and VPA disposition in pregnant sheep presented in Chapters 3-5 of this thesis. 79 Chapter 3 Organs and Metabolic Pathways of Diphenhydramine Clearance in Maternal and Fetal Sheep A number of studies were conducted to elucidate the role of different organs and metabolic pathways in maternal and fetal DPHM clearance. Firstly, as discussed in the Introduction section (Chapter 1), we felt that there could be some confounding factors associated with the geometry of the fetal circulation and its blood flow patterns that might have led to the earlier conclusion of no apparent DPHM uptake by the fetal liver. In order to reexamine this issue, additional data were obtained from the fetal plasma samples collected during the previous study of fetal hepatic first-pass uptake of DPHM from the umbilical venous blood (Tonn et al., 1996). Subsequent to the demonstration of significant fetal hepatic uptake of DPHM, the impact of this phenomenon on the estimation of maternal and fetal placental and non-placental clearances of DPHM, and also of other drugs in general, was assessed. Hence, maternal and fetal DPHM clearance and metabolite (DPMA) formation data obtained during an earlier study, where DPHM and [2H10]-DPHM were infused simultaneously but separately to the mother and the fetus, respectively (Tonn, 1995), are also presented in the context of fetal hepatic uptake of the drug from the umbilical vein. In the previous study (Tonn et al., 1996), a high hepatic first-pass extraction of DPHM was demonstrated in adult non-pregnant sheep after simultaneous but separate administration of DPHM and [2H10]-DPHM via the portal venous (p.v.) and i.v. routes. During the current studies, the adult sheep plasma samples collected during the previous study were used to obtain additional metabolite (DPMA and [2Hi0]-DPMA) 80 plasma concentration data. A detailed theoretical analysis of these metabolite data in combination with the previous parent drug concentration data are presented later in this chapter to provide evidence for a significant role of the gut in DPHM systemic clearance in adult sheep. Additional studies were subsequently conducted to directly confirm this gut uptake of DPHM in adult sheep. In terms of the pathways of DPHM clearance, studies were conducted to examine the role of DPMA and DPHM-N-oxide metabolites (two major pathways of DPHM clearance in many species) in maternal and fetal DPHM elimination. 3.1 Methods 3.1.1 Animals and Surgical Preparation All studies described in this thesis were approved by the University of British Columbia Animal Care Committee, and the procedures performed on sheep conformed to the guidelines of the Canadian Council on Animal Care. In order to obtain time-dated pregnant sheep, ewes' estrus cycle was synchronized by inhibition of their spontaneous ovulation. This was accomplished by the administration of medroxyprogesterone acetate for two weeks via an intravaginal pessary (Veramix Sheep Sponge®). At the end of this two-week period, ovulation was induced by an intramuscular injection of 500 I.U. pregnant mares' serum gonadotropin. The ewes were then placed with a ram for 1-2 days to result in time-dated pregnancies. Animals were typically brought into the research facility at least 3-4 days prior to surgery for their acclimatization and were kept indoors in large pens in the company of other sheep. A standard diet and free access to water was provided to all 81 sheep. The surgical preparation of animals employed for different studies presented in this chapter is described below: 3.1.1.1 Study A: Paired Maternal (DPHM) and Fetal ([ 2H 1 0]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances Five pregnant Dorset Suffolk cross-bred ewes, with a maternal body weight of 70.5 ± 7.7 kg (mean ± S.D.), were surgically prepared between 119-127 d gestation (122 ± 2 d, term -145 d). Food was withheld for -18 h prior to surgery, but free access to water was provided. Aseptic techniques were used throughout the surgical procedure. Approximately 20-30 min before surgery, a 6 mg i.v. dose of atropine was administered via the jugular vein to control salivation. Anesthesia was induced by i.v. administration of 1 g pentothal sodium via the jugular vein. The ewe was immediately intubated and maintained on halothane (1-2%) and nitrous oxide (70%) in oxygen anesthesia. An i.v. infusion of 500 ml 5% dextrose solution containing 500 mg ampicillin was administered over -45-60 min via the jugular vein. A midline abdominal incision was made after cleansing and disinfecting the ewe's abdomen with 10% povidone-iodine solution, and the uterus was exposed. Access to the head and hind quarters of the fetus was gained via two separate uterine incisions made carefully in the areas devoid of major blood vessels and placental cotyledons. Polyvinyl or silicone rubber catheters (Dow Corning, Midland, Ml) were implanted in both fetal femoral arteries and lateral tarsal veins, a fetal carotid artery, fetal trachea, fetal urinary bladder (via a suprapubic incision), and the amniotic cavity (catheter i.d. 1.02 mm and o.d. 2.16 mm). Electrodes (Cooper Corporation, Chatsworth, CA) were implanted biparietally on the dura to record the fetal electrocorticogram (ECoG). In four animals, a transit-time 4SB blood flow transducer (Transonic Systems, Inc., Ithaca, NY) was placed around the common umbilical artery to measure umbilical blood flow. The 82 amniotic fluid lost during surgery was replaced with warm sterile irrigation saline and the uterine and abdominal incisions were closed in layers. Catheters were also implanted in a maternal femoral artery and vein (catheter i.d. 1.02 mm and o.d. 2.16 mm). 3.1.1.2 Study B: Study of Fetal Hepatic First-Pass Uptake of DPHM from the Umbilical Vein Additional data were obtained from the fetal plasma samples from eight animals out of a total of 11 employed in the previous study (Tonn etal., 1996). The surgical preparation for these animals was similar to that described above, with the exception that fetal and maternal bladder catheters and electrodes for monitoring fetal behavior were not implanted. 3.1.1.3 Study C: Hepatic Uptake of DPHM in Adult Non-Pregnant Sheep A total of five adult non-pregnant sheep (body weight 70.1 ± 10.5 kg) were employed in these studies. Aseptic surgical procedures, similar to those described above for pregnant sheep, were also employed for non-pregnant sheep. Femoral arterial and venous catheters were implanted in all five sheep. In addition, a catheter was implanted in a branch of one of the mesenteric veins with the catheter tip advanced to the main mesenteric vein in the direction of the hepatic portal vein. 3.1.1.4 Study D: Gut Uptake of DPHM from the Systemic Circulation in Adult Non-Pregnant Sheep A total of four adult non-pregnant sheep (body weight 64.0 ± 7 . 9 kg) were used for these studies. Femoral arterial and venous catheters were implanted as above. In addition, a 83 catheter was implanted in the main hepatic portal venous trunk just prior to its entry into the liver, as described below. A longitudinal abdominal incision was made in order to gain access to various compartments of the ruminant sheep stomach. A prominent branch of the main gastric vein was identified either on the surface of the rumen or the omasum, and a segment of the intact vessel was then carefully isolated from the surface of the stomach compartment. An -12-18 inch length of a sterile polyvinyl catheter was then advanced into this vessel towards the direction of liver, via the main gastric vein and into the hepatic portal vein. The catheter was secured in place using sterile silk sutures and was anchored to the surface of the rumen or the omasum. The position of the portal venous catheter was verified in all animals by autopsy at the end of each experiment. 3.1.1.5 Study E: Contribution of DPMA Formation to DPHM Non-Placental Clearance in Maternal and Fetal Sheep A total of five pregnant Dorset-Suffolk cross-bred ewes, with a maternal body weight of 82.4 ± 14.1 kg (mean ± S.D.), were surgically prepared between 118-129 days gestation for the purpose of these studies. Surgical procedure was similar to that describe above for Study A. The sites where sterile polyvinyl catheters were implanted include both fetal femoral arteries and lateral tarsal veins, and a maternal femoral artery and vein. Catheters were also implanted in the fetal trachea, amniotic cavity and in three animals in the fetal urinary bladder (via a suprapubic incision). Other physiological monitoring devices were not implanted in these animals. Study F: Disposition of the DPHM-N-oxide Metabolite in the Maternal-Fetal Unit 84 Samples from four out of the five animals above in Study E were utilized for this part of the studies. Maternal and fetal plasma and urine concentrations of the DPHM-N-oxide metabolite were measured in these four animals. In all animals, the catheters and/or blood flow probe cables and electrode cables for monitoring fetal behavior were tunneled subcutaneously and exteriorized via a small incision on the flank of the ewe, and were stored in a denim pouch when not in use. All catheters were flushed daily with approximately 3 ml of sterile 0.9% sodium chloride containing 12 units of heparin per ml to maintain their patency. Intramuscular injections of ampicillin 500 mg were given to the ewe on the day of surgery and for 3 days post-operatively. In pregnant animals, ampicillin (500 mg) was also given into the amniotic cavity immediately following surgery and daily thereafter for the duration of the preparation. Following surgery, animals were kept in holding pens with other sheep and were given free access to food and water. The sheep were allowed to recover for 3-8 days prior to experimentation. Following the recovery period, the sheep were moved to a monitoring pen adjacent to and in full view of the holding pen for experimentation purposes. In experiments requiring collection of maternal urine, a Foley® bladder catheter was inserted via the urethra of the ewe on the morning of the experiment and attached to a sterile polyvinyl bag for cumulative urine collection. 3.1.2 Experimental Protocols 3.1.2.1 Study A: Paired Maternal (DPHM) and Fetal ([ 2H 1 0]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances 85 Experiments were conducted at 125-133 d gestation (128.8 ± 3.2 d) (term -145 d). Simultaneous infusions of DPHM and [2Hi0]-DPHM to the ewe and fetus, respectively, were administered to all five sheep. DPHM was given as a 20 mg i.v. bolus loading dose over 1.0 min, followed immediately by an infusion of 670 pg/min via the maternal femoral vein. Simultaneously, a 5.0 mg i.v. bolus loading dose of [2Hi0]-DPHM was given via the fetal lateral tarsal vein over 1.0 min, followed by an infusion of the compound at 170 pg/min. Simultaneous blood samples were collected from the fetal (1.5 ml) and maternal (3.0 ml) femoral arterial catheters at 30, 60, 90, 120, 150, 180, 210, 240, 270, 300, 330, and 360 min during the infusion and at 30, 60, 120, 180, 240, 360 min and 8, 12, 18, 24, 30 and 40 h post-infusion. Fetal femoral arterial samples (0.6 ml) were also collected at the same time intervals for blood gas analysis and measurement of glucose and lactate concentrations. Fetal carotid and umbilical venous blood samples (1.5 ml) were collected at 30, 60, 90, 120, 150, 180, 210, 240, 270, 300, 330, and 360 min during the infusion period. All fetal blood removed for sampling was replaced at intervals during the experiment by an equal volume of maternal blood obtained prior to the start of the experiment. Serial amniotic and tracheal fluid (3.0 ml) and cumulative maternal urine samples were obtained at 60, 120, 180, 240, 300, and 360 min during the infusion and at 60, 120, 180, 240, 360 min and 8, 12, 18, 24, 30 and 40 h post-infusion. 3.1.2.2 Study B: Study of Fetal Hepatic First-Pass Uptake of DPHM from the Umbilical Vein These experiments were conducted between 124-138 d gestation (128.1 ± 4.9 d). In order to assess the fetal first-pass hepatic uptake for DPHM from the umbilical vein, two different types of administration protocols were employed: 86 1. In four animals, simultaneous but separate randomized bolus injections of [2H10]-DPHM and DPHM (5.0 mg each) were administered via the common umbilical vein and fetal lateral tarsal vein (which drains directly into the inferior vena cava) (see Figure 1.3 for position of the administration routes relative to the fetal liver). 2. In the other four animals, 90 min i.v. infusions of the compounds (60 p.g/min each, preceded by a 2.0 mg each i.v. bolus) via the same routes were employed (see Figure 1.3 for position of the administration routes relative to the fetal liver). During bolus experiments, samples of fetal arterial blood (~2 ml) were collected at 5, 10, 15, 20, 30, 40, 50, 60, 75, 90, 105 min and 2, 2.5, 3, 3.5, 4, 5, 6, 8, 10 and 12 h after the drug administration. During infusion studies, fetal femoral arterial blood samples were collected at 5, 15, 30, 45, 60, 75 and 90 min after the beginning of infusion. In the present study, fetal plasma samples from the above fetal bolus (n=4) and infusion (n=4) experiments were used for the measurement of [2H10]-DPMA and DPMA concentrations. The parent drug plasma concentration data were taken from the previous study (Tonn et al., 1996). 3.1.2.3 Study C: Hepatic Uptake of DPHM in Adult Non-Pregnant Sheep Equimolar amounts of DPHM and [ 2Hi 0]-DPHM (equivalent to 50 mg DPHM) were administered simultaneously but separately via the femoral (intravenous or i.v. route) and mesenteric vein (portal venous or p.v. route) catheters in a randomized manner. Serial samples of femoral arterial plasma (~ 3 ml) were collected at 5, 10, 20, 30, and 40 min, and 1, 1.5, 2, 2.5, 3, 4, 5, 6, 8, 10 and 12 h after drug injection. 87 3.1.2.4 Study D: Gut Uptake of DPHM from the Systemic Circulation in Adult Non-Pregnant Sheep DPHM gut uptake was measured under steady-state conditions in four adult sheep with implanted portal venous catheters. A 20 mg i.v. bolus loading dose of DPHM was administered at the beginning of the experiment via the femoral venous catheter. This was followed immediately by an infusion of unlabeled DPHM at a rate of 670 pg/min for 6 h via the same route. Simultaneous femoral arterial (before the gut) and portal venous (after the gut) blood samples (~3 ml each) were collected every hour for the entire duration of DPHM infusion (6 h) in order to estimate the steady-state gut extraction ratio of the drug. 3.1.2.5 Study E: Contribution of DPMA Formation to DPHM Non-Placental Clearance in Maternal and Fetal Sheep All experiments were conducted between 124-140 d gestation. Two sets of experiments were carried out on all five pregnant sheep. 1. Maternal Administrat ion Experiments: In three animals (E2174, E4227 and E4230), a 6 h steady-state DPHM infusion at 670 pg/min, combined with initial 20.0 mg DPHM and 2.5 mg [2H-io]-DPMA boluses, was administered via the maternal femoral vein catheter. In the other two animals (E1225A and E303Y), the isotope labels on the compounds were reversed, i.e., [2H10]-DPHM infusion (at 670 pg/min) was administered in combination with [2H10]-DPHM (20.0 mg) and DPMA (2.5 mg) boluses. 2. Fetal Administrat ion Experiments: Similar to the maternal experiments above, in three animals (E2174, E4227 and E4230), a 6 h steady-state DPHM infusion at 170 88 u.g/min, combined with initial 5.0 mg DPHM and 0.85 mg [2Hio]-DPMA boluses, was administered via the fetal lateral tarsal vein catheter. In the other two animals (E1225A and E303Y), [2H10]-DPHM infusion (at 170 ug/min) was administered in combination with [2Hi0]-DPHM (5.0 mg) and DPMA (0.85 mg) boluses. In both maternal and fetal experiments, simultaneous serial blood samples were collected from the fetal (1.5 ml) and maternal (3.0 ml) femoral arterial catheters at 5, 10, 20, 30, 45, 60, 90, 120, 180, 240, 300 and 360 min during the infusion, and at 5, 15, 30, 60, 120, 180, 240, 360 min and 9, 12, 18, 30, 42, 54, 66, 78, 90 h post-infusion. Fetal femoral arterial samples (0.5 ml) were also collected at the same time intervals for blood gas analysis and measurement of glucose and lactate concentrations. All fetal blood removed for sampling during the experiment was replaced, at intervals, by an equal volume of blood obtained from the mother prior to the start of the experiment (during the first day) or from another ewe (after the first day). Cumulative samples of maternal urine were also collected every hour for the first 8 h, and along with each blood sample beyond 10 h. Fetal urine was allowed to drain by gravity into a sterile bag, an aliquot (3.0 ml) was sampled for drug analysis, and the rest was returned to the amniotic cavity via the amniotic catheter after recording the total volume. Fetal urine samples were collected at the same intervals as the maternal urine samples above but only during the first 24 h of the experimental protocol. 3.1.2.6 Study F: Disposition of the DPHM-N-oxide Metabolite in the Maternal-Fetal Unit As indicated above, maternal and fetal plasma and urine concentrations of DPHMNOX (or [2Hio]-DPHMNOX) were measured in four out of five animals employed for study E. 89 All maternal, fetal and adult non-pregnant sheep blood samples collected for drug and metabolite analysis were placed into heparinized Vacutainer® tubes (Becton-Dickinson, Rutherford, NJ) and gently mixed. The blood samples were then centrifuged at 2000 g for 10 min. The plasma supernatant was removed and placed into clean borosilicate test tubes with polytetrafluoroethylene (PTFE)-lined caps. Urine samples were also placed into clean borosilicate test tubes. All samples were stored frozen at -20°C until the time of analysis. All doses were prepared by dissolving accurately weighed amounts of the drug (and the DPMA or [2Hio]-DPMA metabolite in study E) in sterile water for injection and were sterilized by filtering through a 0.22 pm nylon syringe filter into a capped empty sterile injection vial. The doses of [2Hi0]-DPHM and [2H10]-DPMA were corrected for the mass-difference due to the presence of deuterium labels as compared to the unlabeled drug and metabolite. 3.1.3 Physiological Recording and Monitoring Procedures During study A, from at least 24 h before to at least 24 h after the infusion period, fetal amniotic, tracheal and femoral arterial pressures, heart rate, electrocortical activity and urine production rate were continuously monitored. In the animals with an implanted umbilical flow transducer, umbilical blood flow was measured with a Transonic model T201 transit-time flow meter (Transonic Systems, Inc., Ithaca, NY). Fetal urine flow rate was estimated using a computer controlled roller pump assembly developed in our laboratory. The fetal bladder catheter was allowed to drain by gravity into a sterile reservoir (10 ml 90 syringe barrel) to which a disposable DTX transducer was connected. When, due to urine collection, the pressure in the reservoir increased above a preset level (usually 3 mm Hg), the computer activated a roller pump (DIAS, Ex154, DIAS Inc., Kalamazoo, Ml) which pumped a calibrated volume of urine from the reservoir back to the amniotic cavity (via the amniotic catheter) during control periods. During the experimental period, the urine was collected into a sterile sample collection syringe, at appropriate sampling intervals a 5 ml aliquot was taken and the rest returned to the amniotic cavity after recording the total volume. The cumulative volume pumped per min, which equals fetal urine production rate per min, was stored on diskette. All of these physiological data as well as the pharmacokinetic data on DPHM (or [2Hi0]-DPHM) disposition in amniotic and tracheal fluids and maternal and fetal urine from Study A have been reported previously (Tonn, 1995) and will not be presented here. In all pregnant sheep experiments, the fetal blood pH, Po2, and Pco2, 02-saturation, and hemoglobin, glucose and lactate concentrations were measured as described in Chapter 2 (Section 2.2.4). 3.1.4 Plasma Protein Binding of DPHM and [ 2H 1 0 ]-DPHM in the Mother and the Fetus The plasma protein binding/unbound fraction of DPHM (or [2Hio]-DPHM) was measured ex vivo in pooled fetal and maternal steady-state plasma samples using an equilibrium dialysis procedure, as described by Yoo et al. (1993). 91 3.1.5 Drug and Metabolite Analysis The concentrations of DPHM, [2H10]-DPHM, DPMA and [2H10]-DPMA in all biological fluids were measured using previously developed GC-MS analytical methods (Tonn etal., 1993; Tonn et al., 1995) (see Chapter 2, Sections 2.3 and 2.4). The plasma and urine concentrations of DPHMNOX and [2H10]-DPHMNOX were measured using the LC-MS/MS analytical method developed during the course of these studies (see Chapter 2, Section 2.6). 3.1.6 Pharmacokinetic Analysis 3.1.6.1 Study A: Paired Maternal (DPHM) and Fetal ([ 2H 1 0]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances The maternal and fetal steady-state arterial plasma DPHM and [2Hi0]-DPHM concentration data obtained from Study A were treated according to a 2-compartment open model in order to estimate the placental and non-placental clearance parameters of DPHM and [2Hio]-DPHM in the ewe and fetus, respectively. This model assumes steady-state plasma concentrations and drug elimination from both the maternal and fetal compartments (Szeto etal., 1982a). Clearances were calculated from the following equations (1-6) as described previously (Szeto etal., 1982a). CLmm -[ C m - C f * ( C m 7 C f ' ) ] 0 ) CLff = [Cf ' -Cm'*(Cf /Cm)] (2) 92 CLmf = CLff (3) CLfm = C L i .mm (4) CLmo — CLmm — CLmf (5) (6) CLmm and CLff are maternal and fetal total body clearances, respectively. C L m f is the maternal-to-fetal placental clearance, whereas C L f m is the fetal-to-maternal placental clearance. C L m o and C L f 0 are maternal and fetal non-placental clearances, respectively. The symbols k0 and k0' denote the drug infusion rates to the mother (DPHM) and the fetus ([2H10]-DPHM), respectively. C m and Cf are the steady-state plasma DPHM concentrations in the mother and the fetus after maternal DPHM administration, and C m ' and C / are the steady-state maternal and fetal plasma [2H-io]-DPHM concentrations after fetal [2H10]-DPHM administration. The net maternal and fetal steady-state clearances of the drug were calculated by dividing the appropriate infusion rate by the respective steady-state plasma concentration (i.e., CLm(net)= ko/Cm; and CLf(net) = k07Cf'). The CLm(net) and CLf(net) are related to 2-compartment clearances as: CLm(net)*Cm = C L m m * C m - C L f m * C f ; and CLf(net)*Cf' = CL f f*Cf - C L m f * C m ' . 93 The fetal first-pass hepatic extraction ratio (EH) for the maternally derived DPHM in the umbilical vein was indirectly estimated using equation (7). This equation assumes sole hepatic formation of the DPMA and [2Hi0]-DPMA metabolites in the fetus and also no placental transfer of the DPMA metabolite formed in the mother (see Appendix I for theoretical basis and derivation): AUC O-oo *\ D P M A E H = A I i p O - M . D P H M J AI i r ° ~ w A A U , V - ' [ 2 H 1 0 ] - D P M A / V W V ' [ 2 H 1 0 ] - D P H M J ( Al IP 0 - 0 0 ^ D P M A Al I P 0 - " M U V ^ D P H M J (7) A U C ^ D P H M , A U C ^ H I O J - D P H M , A U C ^ D P M A , A U C ^ H I O I - D P M A are the area under the fetal femoral arterial plasma concentration vs. time profiles (from time 0 to oo) of DPHM, [2Hio]-DPHM, DPMA and [2Hi0]-DPMA, respectively. DPHM was infused to the mother and hence it reaches the fetus via the umbilical vein, whereas [2H10]-DPHM was directly infused to the fetus via the lateral tarsal vein. 3.1.6.2 Study B: Study of Fetal Hepatic First-Pass Uptake of DPHM from the Umbilical Vein For these fetal hepatic first-pass experiments, fetal systemic availability of DPHM (or [2H10]-DPHM) after umbilical venous administration was calculated from the parent drug data as before (Tonn et al., 1996): AMP A A U O p a r e n t ( T V ) 94 where, respective AUC values refer to fetal arterial plasma AUC of the parent drug after umbilical (UV) or tarsal venous (TV) administration. Fetal hepatic extraction ratio (EH) of DPHM (or [ 2Hi 0]-DPHM) after umbilical venous administration was calculated as: E H = 1 - F (9) In these experiments, fetal hepatic extraction ratio of DPHM (or [ 2H 1 0]-DPHM) was also calculated from the parent drug and metabolite (DPMA and [ 2H 1 0]-DPMA) data in an analogous fashion to equation (7) as: AUCmetabolitefUv)^ (AUCmetabolite(Tv) AUCparent(UV) AUCparent(TV) AUCmetabolite(Uv) AUCparent(UV) (10) where, AUCmetaboiite(uv) and AUCmetaboiite(TV) refer to fetal arterial plasma AUC's of the metabolite after umbilical and tarsal venous administration of the parent drug, respectively. In this study (Study B), all AUC's were only available from time 0 to the time of the last sampling point. Fetal systemic availability using this extraction ratio data was also calculated using equation (9). 95 3.1.6.3 Study C: Hepatic Uptake of DPHM in Adult Non-Pregnant Sheep In the following equations, subscripts 'parent' and 'metabolite' refer to the 'parent drug' and 'metabolite', respectively. The systemic total body clearance of drug was calculated as: C L , B = ( D ° t ! ) - (11) (AUC P A R E N T ) J V where, i.v. refers to intravenous administration of the parent drug, and the A U C term is the femoral arterial plasma AUC. The hepatic first-pass extraction ratio (EH) of drug is given by: E H = 1 _ ( A U C ^ 1 V ( 1 2 ) (AUC P A R E N L ) I V where, p.v. refers to the portal venous administration of the parent drug. The fraction of the intravenously administered drug metabolized in the liver was calculated from the arterial AUC's of the DPMA metabolite after p.v. and i.v. administration of the parent drug, as: 96 f _ \r^\J\jmetabolite li.v. (AUC m e t a b 0 | j t e ) p v This equation assumes linear pharmacokinetics, sole hepatic formation of the metabolite, and complete metabolism of the p.v. dose in the liver. Total hepatic blood flow (portal venous + hepatic arterial) was estimated from the relationships of the well-stirred model of hepatic elimination (Wilkinson and Shand, 1975; Pang and Gillette, 1978): Q H = ~Auc;;ent" Auc;; n t" Dose i.v. Dose p.v. (14) where, i.v. and p.v. refer to the doses and respective arterial AUC's during intravenous or portal venous administration of the parent drug. The amount of intravenously administered drug eventually delivered to the 'hepato-portal' system (gut + liver) was calculated from the estimated hepatic blood flow and its systemic arterial AUC, as: Amount delivered to the" hepato - portal" system = Q H * (AUC p a" e n t )jv (15) This was subsequently converted to the percentage of administered dose delivered to the 'hepato-portal' system. 97 3.1.6.4 Study D: Gut Uptake of DPHM from the Systemic Circulation in Adult Non-Pregnant Sheep Steady-state systemic total body clearance of the drug is calculated as: ( C U ) S S = - ^ — (16) where, k0 is the DPHM infusion rate, and ( C s s ) M A is the steady-state femoral arterial plasma DPHM concentration. The steady-state gut extraction ratio of the drug (Eg) is calculated as: E g = 1_i2ff}py_ ( 1 7 ) ( U S S ) M A where, ( C s s ) M A is the steady-state DPHM concentration in the femoral arterial plasma (before the gut), and ( C s s ) P V is the steady-state DPHM concentration in the portal venous plasma (after the gut). 3.1.6.5 Study E: Contribution of DPMA Formation to DPHM Non-Placental Clearance in Maternal and Fetal Sheep The maternal and fetal net, total, placental and non-placental clearances of DPHM (or [2Hio]-DPHM) were calculated as above in study A. Other pharmacokinetic parameters were calculated by the equations described below (Kaplan etal., 1973; Pang etal., 1979; Wagner, 1993). All AUC and AUMC (area under 98 the first-moment curve) terms in these equations refer to those in femoral arterial plasma. Fraction of the total parent drug dose converted to the metabolite (DPMA or [2H-|0]-DPMA) in vivo (in the mother or the fetus) or the formation clearance of the metabolite as a fraction of the total body clearance of the parent drug: AUC 0 "" /DOSeparent drug f o r m e d m e t a b o l i t e ^ A U C p r e f o r m e d metabolite / DOSe Preformed metabolite where, 'formed metabolite' refers to the metabolite generated in vivo from the parent drug and 'preformed metabolite' refers to the synthesized metabolite administered per se. The formation clearance of the metabolite as a fraction of maternal or fetal non-placental clearance (Fm') was calculated by dividing the F m value obtained above by the fractional contribution of the maternal or fetal non-placental clearance to the corresponding total body clearance. Mean residence time of the preformed metabolite: AUMC 0"" 3 »yi | - )T preformed metabolite M O \ IVIK I preformed metabolite = ( I C 7 ) A i j r 0 - e o " preformed metabolite Mean residence time of the metabolite formed in vivo: 99 MRT, formed metabolite = A U M C 0 - ^ A U C O-cc formed metabolite ^AUMC 0" 0 0^ v AUC°-M , (20) parent drug Mean residence time of the parent drug: MRTparent drug ^AUMC°-W ^  v AUC 0"" j ko-x 2 parent drug 2(ko « X + Dbolus) (21) where, k 0 , x, and Dboius are the infusion rate, infusion duration, and initial bolus loading dose of the parent drug, respectively. Total body clearance (CLtb) of the preformed metabolite or parent drug: CLtb = Total i.v. Dose AUC O-OO (22) Steady-state volume of distribution (Vdss) of the preformed metabolite: (Vdss)preformed metabolite — (CLtb)preformed metabolite * MRTpreformed metabolite (23) And, V d s s of the parent drug: (Vdss)parent drug — (CLtb)parent drug MRTp a r e n t drug (24) 100 The terminal elimination half-life (t1/2p) of the parent drug as well as the preformed metabolite was obtained from a 2-compartment model fitting of the data using nonlinear least-squares regression software WinNonlin (Scientific Consulting, Inc., Apex, NC). Maternal renal clearance values for the parent drug, the preformed metabolite and the formed metabolite were calculated by dividing the total amount of each compound excreted in maternal urine by the respective maternal plasma AUC0"1". Fetal renal clearances for all these compounds were calculated by dividing the total amount excreted in fetal urine during the 24 h sampling period by the respective fetal plasma AUC°~ 2 4 h. 3.1.6.6 Study F: Disposition of the DPHM-N-oxide Metabolite in the Maternal-Fetal Unit The maternal and fetal renal clearances of DPHMNOX (or [2H10]-DPHMNOX) were calculated in a fashion similar to DPMA (or [2Hi0]-DPMA) above. Maternal and fetal plasma AUC's and AUMC's for the parent drug and metabolites in all studies (A - F) from time 0 to the last sampling point were calculated using the linear trapezoidal rule (Gibaldi and Perrier, 1982). The AUC was then extrapolated to time infinity by adding the factor, Ciast/K; Ciast is the plasma concentration of the drug or metabolite at the last sampling time (tiast), and K is the terminal elimination rate constant. The factor for the extrapolation of AUMC to time infinity was (C|ast*t|ast/K+C|ast/K2) (Gibaldi and Perrier, 1982). 3.1.7 Statistical Analysis 101 All values are reported as mean ± S.D. In all pregnant sheep studies, the fetal weight in utero at the time of experimentation was estimated from the weight at birth and the time interval between the experiment and birth (Koong et al., 1975). The achievement of steady-state in maternal, fetal or adult non-pregnant sheep plasma was established according to two criteria: i) the slope of the plasma concentration vs. time curve should not be significantly different from zero, and, ii) the coefficient of variation of the measured concentrations should be <15%. Maternal and fetal pharmacokinetic parameters in studies A and E were compared against each other using an unpaired t-test. In study C, the adult sheep femoral arterial plasma AUC's of the parent drug or metabolite after i.v. and p.v. drug administration, were compared using a paired t-test. The steady-state DPHM concentrations in femoral arterial and portal venous plasma during gut uptake studies (study D) were also compared against each other using a paired t-test. The significance level was p < 0.05 in all cases. 3.2 Results 3.2.1 Study A: Paired Maternal (DPHM) and Fetal ([2H10]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances The five experiments involving simultaneous 6 h infusions of DPHM and [2Hi0]-DPHM to ewe and fetus, respectively, were carried out at 125-133 days gestation (128.8 ± 3.2 d). Estimated fetal weight on the day of experiment was 2.46 ± 0.21 kg. During the control period, the fetal femoral arterial values for pH, P02, Pco2, 02-saturation, hemoglobin, glucose and lactate concentrations were 7.36 ± 0.02, 22.6 ± 1.65 mm Hg, 47.3 ± 0.5 mm Hg, 55.3 ± 23.8%, 10.0 ± 0.3 g/dl, 0.98 ± 0.09 mM and 0.70 ± 0 . 1 1 mM, respectively. There were no consistent changes in any of these variables during or after the infusion 102 period. Likewise umbilical blood flow (281 ± 39 ml/min/kg, n=3) was not consistently altered during the experiment. 3.2.1.1 Maternal and Fetal Plasma DPHM and [2H10]-DPHM Concentrations, and Placental and Non-Placental Clearance Values The average plasma concentrations of DPHM and [2Hi0]-DPHM in maternal and fetal femoral arterial plasma are illustrated in Figure 3.1. Steady-state was achieved in maternal as well as fetal plasma after 120 min according to the established criteria. Thus, maternal and fetal steady-state plasma concentrations were calculated as the average concentration during the 150-360 min period. The mean steady-state concentrations of DPHM were 260.8 ± 42.3 and 45.7 ± 38.9 ng/ml in maternal and fetal femoral arterial plasma, respectively, while the mean concentrations of [2Hio]-DPHM in the same vessels were 44.6 ± 12.9 and 244.0 ± 94.7 ng/ml, respectively. The total maternal and fetal femoral arterial steady-state concentrations of DPHM (i.e., labeled and unlabeled DPHM) were 305.4 ± 54.7 and 289.6 ± 128.9 ng/ml. As in previous studies (Yoo, 1989), there was accumulation of DPHM (both labeled and unlabeled) in fetal tracheal and amniotic fluids. The average lung fluid to FA plasma drug concentration ratio was 4.0 ± 1.7 for DPHM and 4.5 ± 1.6 for [2H10]-DPHM. The corresponding ratios in amniotic fluid (i.e., amniotic fluid/FA) were 0.6 ± 0.2 and 0.8 ± 0.2 for labeled and unlabeled drug, respectively. Following the infusion, the concentrations of DPHM and [2H10]-DPHM in all fluids declined rapidly with terminal plasma elimination half-lives of 70.5 ± 6.9 and 51.8 ± 7.2 min in the ewe and fetus, respectively. Details of these data have been presented previously (Tonn, 1995). 103 Figure 3.1 - Average plasma concentrations of DPHM and [2Hio]-DPHM in maternal and fetal femoral arterial plasma during and following simultaneous i.v. infusions of DPHM (670 pg/min) to the ewe and [2Hi0]-DPHM (170 pg/min) to the fetus (n=5). The calculated net, total, non-placental and placental plasma clearances in the mother and the fetus are presented in Table 3.1. The weight-normalized estimates of CLf(net) (314.8 ± 101.5 ml/min/kg), CLff (324.2 ± 104.3 ml/min/kg), C L f m (214.4 ± 68.8 ml/min/kg) and C L f 0 (109.8 ± 49.8 ml/min/kg) were all significantly higher than the corresponding maternal values for CLm(net) (37.2 ± 4.6 ml/min/kg), CLmm (38.3 ± 5.1 ml/min/kg), C L m f (50.3 ± 29.6 ml/min/kg) and C L m o (36.6 ± 4.2 ml/min/kg) (unpaired t-test, p < 0.05 in all cases). The non-placental contribution to net DPHM clearance averaged 98.5 ± 0.9% and 34.4 ± 9.4% in ewe and fetus, respectively, and again these were significantly different. In ewes 122z and 2181, the presence of a functional umbilical venous catheter also allowed calculation 104 of the fetal DPHM extraction across the placenta. The umbilical extraction ratios for [ 2 Hi 0 ]-DPHM were 0.64 and 0.51 in E122z and E2181, respectively, indicating net drug transfer from the fetus to the mother. 3.2.1.2 Fetal and Maternal DPMA and [2H10]-DPMA Plasma Concentrations A mean concentration vs. time plot of DPMA and [2H10]-DPMA in maternal and fetal femoral arterial plasma is shown in Figure 3.2. Although concentrations of DPHM and [2Hio]-DPHM reached steady-state at ~120 minutes from the start of the infusion (Figure 3.1), the plasma levels of DPMA and [2H10]-DPMA did not reach steady-state during the whole duration of infusion and continued to increase for 30-120 minutes post-infusion. At all time points during the infusion period, the concentration of [ 2Hi 0]-DPMA was higher in the fetus than in the mother, whereas for the unlabeled metabolite the situation was reversed. The peak concentrations of DPMA in maternal and fetal plasma averaged 137.4 ± 42.5 and 92.8 ± 42.6 ng/ml, respectively, while the peak maternal and fetal plasma concentrations of [2Hi0]-DPMA were 28.7 ± 9.7 and 135.0 ± 49.7 ng/ml, respectively. The time at which the peak levels occurred post-infusion was 18.0 ± 16.4 min for both labeled and unlabeled DPMA in the ewe, whereas in the fetus it was 87.0 ± 40.2 min. Following the peak, the fetal metabolite levels declined much more slowly than in the ewe. The elimination half-life of the metabolite in the fetus (15.2 ± 5.6 h) was significantly longer compared to the mother (3.0 ± 0.5 h) (unpaired t-test, p < 0.01). In the 2 animals with functional umbilical venous catheters, the extraction ratio of [2H-|0]-DPMA across the fetal side of the placenta averaged -0.06 ± 0.04. This value is not significantly different from zero, but is different from the umbilical extraction ratio for [2H10]-DPHM given above. Finally, DPMA or [2H10]-DPMA metabolites were not detected in amniotic or fetal lung fluid. 105 [ 2 H 1 0 ] - D P M A - F A — • — D P M A - F A -e— [ 2 H 1 0 ] - D P M A - M A —H— D P M A - M A 0 360 720 1080 1440 1800 2160 2520 2880 Time (min) Figure 3.2 - Average maternal and fetal femoral arterial plasma concentrations of DPMA and [ 2 H i 0 ] - D P M A during and following simultaneous i.v. infusion of DPHM (670 pg/min) to the ewe and [2H10]-DPHM (170 pg/min) to the fetus (n=5). Table 3.2 gives the A U C values for DPHM, [2H10]-DPHM, DPMA and [2H10]-DPMA in maternal and fetal femoral arterial plasma. The fetal A U C D P M A / A U C D P H M ratio (8.20 ± 3.62) was significantly higher than the corresponding A U C ratio for [ 2Hi 0]-DPMA/[ 2H 1 0]-DPHM (2.24 ± 1.19) (paired t-test, p < 0.02). In the ewe, however, there was the opposite situation. The A U C D P M A / A U C D P H M ratio (0.62 ± 0.15) was significantly less than the corresponding ratio for [2H10]-DPMA/[2H10]-DPHM (0.96 ± 0.26) (paired t-test, p < 0.02), although the magnitude of the difference is much smaller than that in the fetus. Table 3.2 also gives the estimates of the fetal hepatic first-pass extraction ratio for maternally derived DPHM calculated using equation 7 (section 3.1.6.1), with the mean value for this parameter being 0.71 ±0.16 . 106 E O a _i o | a) + J a) E S ro CL <D O C 2 « O i t _i O o E — I o o E — I o E E _l O 0) c ¥ -J o 0) LU CD L O I s - CD co 00 o i CT> CO TT+- CO Is- CO CO CO CO CD CN CM CNI +l o CM Is-CO cb CO o CM L O CN CD CO 00 CD CO CD CO •Sl-LO' CD L O CO I s - T-CM cb o> CO T— L O CD o CO CO CM CN o CD CO CO CO L O CO CO o CO CO CO CO CO CN o CO "id-CD L O 00 CO CD CN co CO CO CO L O o Is-co CN L O 5-o CO CO 00 CO CD o CO CN o 00 Is-' CO CO CO CN CO CO Is-CM 00 CM CO L O CM N I s -o CM I s - CO ^ -CM CM CM T ~ CM CM CM CO oo CO co ° +l CO o +i CD O CD CT> CN +l CD CD CO CO T -00 L O CO + I CN CD CO + l * co S +i E CD L _ CD CO CU _3 CD > CD O £Z CD s CO o L O o o V Q. CD 3 CO i-r > J Z — CO CD CD E T3 E £ |> TO ^ P O CL ™ CD a) "co E _ o > 8 1 J? E J= -I—* - B CD ro CD 9? CD ^ C co o HE ' c D ) CO * CD CD -i_r o -c cn ' 'CD Q)" o o , 0 ) ro "a to a) CD CO ?> E < '+= CD c c o o c c " r o ro E CD LL CO " 6 .. O o d§ • - 2 9 CO o CD CO o s ro 0) o ai — o ro ro c CL a) _ o ro ro Q - L L . ro ' E E Jf a) o ro . -o d s a> ° o ro ro o CD 3 o C CD CD C ro 2 S CD ' E O O f O CD C LL CD o3 0) jg. O _ l *- O CD . . CO c CD CD O O C C CD CO CD CO CO ro ® ^ o o .. " r o " r o —^< -*—' a) C C £ CD CD E O O I CO CO O Q.Q. 107 < s > Q_ Q de pu ns re c i_ X iJ Q. re O E O the rms <*-O a on T3 O U Oi re i . lab ext T3 o C re pa TJ 0) CD CD lab eta c 3 L- o $ o (0 c a) CD CO re E ue im eri ns est Q. > est X <u O T3 c c o < re "to re so;; infu rtei re pa re o rai AU re Q. I femo drug al-feta c nd fe /parei mate re U) c re o ern iS x Mat Mat E Q o O J = + 3 • r 2 .P C O -)—» 0) LU C O O < CD C CO CD _3 CO > ? o O Z) < 0) CO CO LU -ft 99 ""o o x j f 2 X CL Q CL Q CL Q X CL Q X CL Q i "~o X X CL Q oo oo ci co LO o 00 o CO in c\i CN LO Mr co CM o o o M -00 CD o CN CD 00 CO CO T — LO CO LO Mr CD CN 00 CO 00 CO I s -00 oo IS-CN IS-00 o IS-CN IS-< LL CN w 00 CD o CD oo o M-co CO M -d I s -d Mr M -CM CO LO M -o CO IS-CN CD CO o Mr o M " oo LO CD M -00 o Mr CD 00 o LO CO CD CO CO LO CO Mr oc> 00 00 o N CN CN LU CD Mr d M" CD OO CO d co O CD LO LO d cb M " CO o IS-M -CD IS-' CN CO CO IS-CM I s -oo CN o Mr Mr CD oo CN o CN CD LO IS-CN CN I s -o CO 00 Mr CO LO oo CO CO LO o CO CD < LL IS-IS-CNI LU CD CD d CO CM CO CN CN CO d CD LO CD Mr LO CD o CN CN LO CO CD CN CO CO 00 d IS-00 CD IS-00 X — o CD O0 CO CN 00 CD 00 CO LO CO LO CO CO Mr CN d 1^  M" C0 M -CO CD LO M -00 CM LU LO oo d LO CM CD CD d d LO CD d CO o M J CD CO 00 CD IS-00 CD M -cci M -o Mr co 00 d o CD o CD oo LO CD M -CD CN CD CO o oo o Mr CD M " oo CN CN M -00 Mr Mr CN CN L U LO T — d +1 CN d CO CN CO +1 M -IS-LO CO CN CN CO +1 00 CD o M -co IS-LO CM M " LO +1 I s -cb I s -00 00 CO IS-d CD d +1 CD * CD M -CO CO CM CN T*~* d d +1 CM +1 o CN CO * CN CD CO +1 M -d M " CD CO O0 M -LO CD I s -CD O +1 CD M " Mr M -CD CO CO ib CD o CD CO LO oo Mr CO CD CO +1 cq LO d CM o o M " CO CD CO LO +1 " r ~ +1 ^ cb ^ + i 108 3.2.2 Study B: Study of Fetal Hepatic First-Pass Uptake of DPHM from the Umbilical Vein Figure 3.3 illustrates the representative fetal plasma concentration vs. time plots for unlabeled and labeled forms of DPHM and DPMA measured from samples obtained during the previous study of fetal hepatic first-pass DPHM uptake after umbilical venous administration (Tonn et al., 1996). Figure 3.3A shows the data from a bolus experiment where [2Hi0]-DPHM was administered via the umbilical vein, whereas Figure 3.3B shows the data from an infusion study in which unlabeled DPHM was infused via the umbilical route. In both experiments, there were no consistent differences in the fetal arterial plasma concentrations of DPHM and [2Hi0]-DPHM. In contrast, the plasma concentration of the form of DPMA derived from the drug administered via the umbilical vein was consistently higher than that derived from the drug given via the lateral tarsal vein. Table 3.3 gives the fetal femoral arterial plasma AUC values for labeled and unlabeled DPHM and DPMA. The estimates of fetal first-pass hepatic extraction ratio based upon intact drug concentrations were obtained using equations 8 and 9 (section 3.1.6.2), and the calculated mean value of -0.06 ± 0.17 is not significantly different from zero (unpaired t-test, p > 0.1), as reported previously (Tonn et al., 1996). In contrast, the estimates of fetal hepatic extraction ratio obtained using the AUC values for DPMA and [2Hi0]-DPMA in combination with those of DPHM and [2H10]-DPHM and equation 10 (section 3.1.6.2) indicate significant drug uptake by the fetal liver from the umbilical vein. The mean fetal hepatic extraction ratio was 0.44 ± 0.14, and this was significantly lower than the value of 0.71 ± 0 . 1 6 obtained above in the paired maternal-fetal infusion experiments (unpaired t-test, p < 0.005) (Study A; see Table 3.2). 109 Figure 3.3 - A). Representative fetal femoral arterial plasma concentrations of DPHM, [2H10]-DPHM, DPMA and [2H10]-DPMA after simultaneous i.v. bolus administration of DPHM and [2Hi0]-DPHM (5 mg each) to the fetus (E#989). DPHM was given via the fetal tarsal vein and [2H10]-DPHM via the umbilical vein. B). Representative fetal femoral arterial plasma concentrations of DPHM, [2H10]-DPHM, DPMA and [2H i 0]-DPMA after simultaneous steady-state infusion of DPHM and [2H10]-DPHM (60pg/min each) to the fetus (E#2164). DPHM was infused via the fetal umbilical vein and [2H10]-DPHM via the fetal lateral tarsal vein. 110 >. sa i5 -S •» CD > < CD CD •55 O § i | :8 -2 2 & CD CD CD CD i Q LU ^ ro is " r = CO r n rrs n ^ J > D -5 ^ ra o i-. CD CD CD m ™ a: cx 5 LU Q Q I I X c "E cb c co CD l l o o < 0_ Q X CL a X CL Q Is-LO d co o o d 00 Is-d o co 0 co co d 05 o co d d co CD •sr co d d o 10 CM CM d o Is-d Is-co LO d o d CM CD o CO CN CO CO CO T— O Is- O T -co o CO o d co co CM CO o d d o CM CN d o d O CO T - O d d 0 CO 00 Ci T— CM d d CO CO Is- CO CN CN 0 co CN CO LO T— CD h- CN CO CN LO LO CO CO d ^—-CO LO 0 00 co 00 CN 00 Is-LO LO CM LO CM LO CD CO cb Is-CD ^ — 0 CD CM T— LO T~ 00 co CO co CM CN CN LO Is- O LO CN CO LO CN 00 ^—• LO CO 00 cb O CN T— 00 CN CD CO CO CO O Is- CO CN "vT CT> •ST CN CN CO CO CO LO 0 LO ^_ ^_ 00 00 CO T— r^' 00 CM d cb 06 LO CD CO CD CO CN CD 00 CO CN •sr CM CO CO M- co CO CO CO CO co CO co £ LO ^ +1 * co fL. P ! o +1 CO Is-o • O O 1 +1 c c CD CD > > " r o " r o CO (0 -—* *- LO CD CD o ro ro 0 1— 1— A CD CD " ro ro —-CD CD " T — xi JZ c: .co .co 2 2 S 1 x x £ 0- CL C " CD X c CO ro « S .2 co = CD ^ > O CD ~ * JZ -Q . -= " A CD A > .co a. S3 c 2 CD CD CD $ > N ^ O ) g I « O r - L , ^ CD W JO >> ro ro +± E E § c c 0 ro ro a= CD CD g , CO CO • — CD CD CO r r +; •+-* o c c c o o o ' o ICL C O C O C O C O Z J Z J Z5 Z J 0 O O 0 CO ca CQ CQ cz o C O Z J 4 — c c o C O Z J H— c cz o C O cz c o C O cz CD l £ ro (13 CD CO CT) CO 0 CD O) CN d^" LU LU LU ro ni x i x i CO CN CM O LO CD T— T— CN CM T— ^— T— T— CN LU LU LU LU LU CD Q ^ +1 111 3.2.3 Study C: Hepatic Uptake of DPHM in Adult Non-Pregnant Sheep The average body weight of sheep utilized during the DPHM hepatic first-pass uptake experiments was 70.1 ± 10.5 kg. Figure 3.4 shows the typical femoral arterial plasma concentration vs. time profiles of the parent drug and metabolite after separate but simultaneous administration of DPHM and [2Hio]-DPHM via intravenous and portal venous routes in two sheep. In E1154, DPHM was given via the portal venous route and [2Hi0]-DPHM via the intravenous route. In E102, the order of administration was reversed, i.e., DPHM was administered intravenously and [2H10]-DPHM via the portal route. The average maximal plasma concentrations (C-max) of the metabolite generated from the form of drug administered via the portal venous route were significantly higher compared to the C m a x of the metabolite generated from the form of the drug given intravenously (447.8 ± 175.9 vs. 90.7 ± 75.8 ng/ml; paired t-test, p < 0.005). The times of occurrence of plasma C m a x of the metabolite ( tmax) after portal venous administration ranged from 5-20 min compared to 10-90 min after intravenous administration. Table 3.4 presents the calculated femoral arterial AUC's of the parent drug and metabolite, hepatic first-pass extraction ratios and the fraction of intravenously administered dose metabolized in the liver. The AUC of the form of parent drug administered via the portal venous route was much smaller compared to that of the form administered intravenously (894.8 ± 767.2 vs. 13286.1 + 5042.8 ng.min/ml; paired t-test, p < 0.005). The AUC of the metabolite generated from the form of parent drug administered via the portal venous route, however, was significantly larger compared to that generated from the form administered via the intravenous route (65042.8 ± 47634.9 112 vs. 24218.7 + 27973.8 ng.min/ml; paired t-test, p < 0.01). The hepatic first-pass extraction ratio of the drug was high and ranged from 90.4 - 99% (mean 94.2 ± 3.7%; Table 3.4). The fraction of the i.v. parent drug dose metabolized in the liver ranged from 18.2 - 50.4% (mean 32.5 ± 14.0%; Table 3.4). Thus, the fraction of the i.v. dose metabolized/eliminated by extrahepatic tissues will be 49.6 - 81.8 % (mean 67.5 ± 14.0%). Table 3.4 - Femoral arterial AUC's of the parent drug (DPHM or [ 2 H 1 0 ]-DPHM) and the metabolite (DPMA or [ 2 Hi 0 ] -DPMA), parent drug hepatic f irst-pass extraction ratio, and the fraction of intravenously administered parent drug dose metabolized in the liver during hepatic first-pass uptake studies. Ewe# AUC ng.min/ml) Hepatic Fraction of Parent (i.v.f Parent (p.v.)a Metabolite (i.v.)a Metabolite (p.v.)a First-Pass Extraction Ratio (E H) i.v. Dose Metabolized in the Liver (f H) 1158 b 20437.5 1952.8 73888.5 146681.7 0.904 0.504 1154 14728.5 1411.5 6811.4 37415.4 0.904 0.182 1 3 9 b 14140.9 595.3 11518.8 26096.7 0.958 0.441 1 0 2 b 9812.6 437.8 16360.9 60500.2 0.955 0.270 989 7311.1 76.6 12513.7 54520.1 0.990 0.230 Mean 13286.1 894.8 24218.7 65042.8 0.942 0.325 ±S.D. ± 5042.8 ± 767.2* ± 27973.8 ± 47634.9** ± 0.037 ± 0 . 1 4 a - i.v. and p.v. refer to intravenous and portal venous administration of the parent drug. b - In these ewes, [2Hi0]-DPHM was given via the portal venous route and DPHM via the femoral venous route. In the other animals, the route of administration for the two forms of drug was reversed. *- significantly lower than the parent drug AUC after intravenous administration (p < 0.005) **-significantly larger than metabolite AUC after intravenous administration of the drug (p < 0.01). 113 Table 3.5 presents the estimates of systemic clearance, total hepatic blood flow and percentage of intravenously administered dose that is eventually delivered to the 'hepato-portal' system (gut and liver). The average percentage of i.v. dose delivered to the 'hepato-portal' system was not significantly different from 100% (unpaired t-test, p > 0.3). Table 3.5 - Diphenhydramine systemic clearances, estimated hepatic blood f lows and percentage of intravenously administered dose eventually delivered to the "hepato-portal sys tem" in five adult sheep. E w e # Systemic Clearance (ml/min/kg) Estimated Total Hepatic Blood Flow (ml/min/kg) % of i.v. Dose Delivered to the Gut and Liver 3 1158 b 45.1 41.4 91.8 1154 54.1 59.5 109.9 1 3 9 b 56.8 59.0 103.8 1 0 2 b 85.1 74.0 86.9 989 78.5 78.9 100.5 Mean ± S.D. 63.9 ± 17.0 62.6 ± 14.7 98.6 ± 9 . 2 a - i.v. refers to intravenous administration of the parent drug. b - In these ewes, [2Hi0]-DPHM was given via the portal venous route and DPHM via the femoral venous route. In the other animals, the route of administration for the two forms of drug was reversed. 114 (|w/6u) uoiJBJiuaouoo e iusBid a j ! |oqe ja | / \ | t - CM | III 11 I I I I Mil I I I I I III 11 I I I o o o O O 1-O i -(|w/6u) uonej )uaouoo e i i i s e y BnjQ -.uejBd E (|iu/6u) uo j ie j iueouoo Eiuseid a)]|oqBi9|A| o c cu > ro •c o Q . c o o w =3 CO o .fc C CO > 1 ro E c lo j l l l I 1 I I I M U M ! I j l l l l I I i—i 1— o o o o o o 1-o T-(|LU/6U) UOIIBJJU90UOQ BUJSBId 6njQ }U8JBd g X LU C D CL > 9 ~"co X co £ C N Q . . C D T3 CD CO § CD 3 & H — ' c CD X C N S • co - 5 0 c •—• c • 5 to •*- — C O Q . O -D !_ m C D C D C D C O J2 £ C L O C D i<1S Q. LU CT - > g o3 c ^ c CQ ro X O . . ^ ^ 11 S -8x s |S E -o IS 0 — C O 5 Q C L . ° - + - > -= < a> 2 " 1 - S co E — C D -55 " O C O C O 0 0 FT n c _Q CL O (D C c 3 > c t O c ro C D . " O w ^ C D C D O ~ CD CD c or ro I M-co 33 CO CD CD O CD 33 o 115 3.2.4 Study D: Gut Uptake of DPHM from the Systemic Circulation in Adult Non-Pregnant Sheep The average body weight of sheep utilized for DPHM gut uptake studies was 64.0 ± 7.9 kg. Figure 3.5 shows the average concentration vs. time profiles of DPHM in femoral arterial and portal venous plasma in four sheep during a 6 h DPHM infusion. Based on the previously mentioned criteria, DPHM plasma concentrations were at steady-state during the 2-6 h infusion period in both femoral arterial as well as portal venous plasma. The steady-state femoral arterial and portal venous plasma concentrations, systemic clearances, and the estimates of gut extraction ratio of DPHM in four sheep are presented in Table 3.6. Figure 3.5 - Average femoral arterial and portal venous plasma concentration vs. time profiles of DPHM in four sheep during 6 h DPHM infusion. 116 The portal venous plasma concentrations of DPHM were significantly lower compared to its femoral arterial plasma concentrations throughout the experimental period in all animals (paired t-test, p < 0.005 in all cases). The gut extraction ratio of the drug in individual animals ranged from 46.3 - 53.4% (mean 49.0 ± 3.0%). Table 3.6 - Steady-state femoral arterial and portal venous plasma concentrations, systemic clearance, and the gut extraction ratio of DPHM in four sheep during gut uptake experiments. Ewe# Steady-State Plasma Concentration (ng/ml) Systemic Clearance (ml/min/kg) Gut Extraction (E a) Femoral Arterial Portal Venous E0224 296.7 148.2 40.3 0.501 E4140 197.3 105.9 50.0 0.463 E1225A 451.7 210.5 20.3 0.534 E6216 323.3 173.7 35.1 0.463 Mean ± S.D. 317.2 ± 104.8 159.6 ±44 .0* 36.4 ± 1 2 . 4 0.490 ± 0.030 * - significantly lower compared to femoral arterial plasma concentrations. 3.2.5 Study E: Contribution of DPMA Formation to DPHM Non-Placental Clearance in Maternal and Fetal Sheep In these studies, the average maternal body weight was 82.4 ± 14.1 kg and estimated fetal body weights on the day of maternal and fetal DPHM (or [2Hi0]-DPHM) infusion were 3.16 ± 0.41 and 2.92 ± 0.22 kg, respectively. During the maternal experiments, the control period fetal femoral arterial pH, Po2, Pco2, 02.saturation, and hemoglobin, glucose and lactate concentrations were 7.371 ± 0.017, 20.6 + 1.8 mm Hg, 50.2 ± 3.3 mm Hg, 42.6 ± 117 7.0%, 11.3 ± 0.6 g/dl, 0.80 ± 0.21 mM and 1.15 ± 0.18 mM, respectively. Likewise during fetal administration, the control values for these variables were 7.344 ± 0.031, 24.8 ± 5.6 mm Hg, 53.1 ± 2.9 mm Hg, 55.0 ± 8.4%, 10.3 ± 1.3 g/dl, 0.78 ± 0.29 mM and 1.56 ± 0.65 mM, respectively. There were no consistent changes in any of these variables during the course of maternal or fetal administration experiments. 3.2.5.1 Maternal-Fetal Steady-State Plasma Drug Concentrations and Unbound Fractions, and Placental and Non-placental Clearance Estimates Table 3.7 presents gestational age of the animals on the day of experiment, and maternal and fetal DPHM clearance data (net, total, placental and non-placental clearances). The mean gestational age on the day of maternal and fetal experiments was 132.2 ± 5.9 and 129.4 ± 3.8 days, respectively, and these were not statistically different (paired t-test, p > 0.05). The average maternal and fetal steady-state plasma DPHM (or [2H10-DPHM) concentrations in these animals after maternal drug administration (Cm and Cf, respectively) were 220.9 ± 40.3 (range 179.1 - 268.1) and 51.5 ± 45.4 (range 3.5 - 124.1) ng/ml, respectively, whereas those after fetal drug infusion were 32.7 ± 6.5 (Cm': range 24.5 - 40.2) and 231.9 ± 91.9 (Cf': range 132.5 - 374.7) ng/ml, respectively. The steady-state maternal (on the day of maternal experiment) and fetal (on the day of fetal experiment) plasma DPHM unbound fractions were 0.135 ± 0.069 (range 0.032 - 0.211) and 0.347 ± 0.114 (range 0.242 - 0.527), respectively. The mean maternal plasma unbound fraction was significantly lower compared to the mean fetal plasma unbound fraction (unpaired t-test, p < 0.005). All fetal weight-normalized clearances (net, total body, placental and non-placental clearances) were significantly higher compared to the 118 corresponding maternal clearances (unpaired t-test, p < 0.02 in all cases). However, the contribution of CL f 0 to CLf(net) (41.8 ± 12.9%) was significantly lower compared to that of C L m o to CLm(net) (97.8 + 1.7 %) (unpaired t-test, p < 0.001). 3.2.5.2 Maternal-Fetal Arterial Plasma AUC Ratios of the Parent Drug, the Preformed Metabolite, and the in vivo Generated Metabolite Table 3.8 presents the different A U C ratios for the parent drug, the preformed metabolite and the in vivo formed metabolite in maternal and the fetal arterial plasma. The FA/MA ratio of the formed metabolite AUC's after maternal drug administration (2.97 ± 0.82) was significantly higher compared to the FA/MA ratio of the preformed metabolite AUC's after maternal metabolite administration (0.41 ± 0.21; paired t-test, p < 0.005). Although, the MA/FA ratio of the formed metabolite AUC's after fetal drug administration (0.25 ± 0.35) was higher compared to the MA/FA ratio of the preformed metabolite AUC's after fetal metabolite administration (0.02 ± 0.02) in all the individual animals, the difference between the means was not statistically significant. Also, the formed metabolite to parent drug AUC ratio in FA after maternal drug administration (21.5 ± 26.4) was higher compared to the corresponding ratio after fetal drug administration (2.32 ± 1.16) in all the individual animals; however, the difference between means was only near statistical significance (paired t-test, p = 0.08). Similarly, the formed metabolite to parent drug AUC ratio in MA after fetal drug administration (2.69 ± 3.87) was higher compared to the corresponding ratio after maternal drug administration (1.13 ± 1.44) in all the individual animals; however, again the difference between means was not statistically significant (paired t-test, p > 0.05). 119 I l_ <D + J CD E as as a. a> u c as as 0) O co >» as 2, CD Ui < "as c o CO CD CD LU -J O o —I o ifc —I o o E _ l o o £ _ l O E E - I o cu E _J o L O I s - M-O) c\i co CN L O CO T — _^ L O o L O CO T— CO CO T ~ T— c CD S i -Si X LU CO c\i L O M-as CD E E CD Z 2 X LU CD IS-' C N L O M-L O L O M-CO L O iq CO CO CO L O CN CO M-L O CO CNJ L O L O M-CO CD C N L O CN CO L O co CN L O L O CD d CO CD CN CO L O CO co CN M-CN CN CD L O CO oo CN L O co CO CD C N L O CO o M-M-CN CO CO L O I s -CN CO M-CD CN CO L O r - I s - M- L O O d c d L O L O CN L O T— I s - M- M- CO M- CO CN CN CN T — +l o < LO CM CM T -LU •a I S-CM LU o >-co o CO LU IS-CM CM TS-U I L O CM CN CO co I S -0 0 M-o CD CN CO •a O co CM M" LU CO I s -* I s -•sr' CO +l CO CD cb d M-CD ^ co d CD CO CM L O M-M-+l CO CM +l CO CM +l S 3 co w +l M; co d co CM T - +l <M CO T~ + l g q CD CO § + l T3 CD E a CD i Q_ . Q _ £ O U i CN CO " -*—> a o5 TD E a> - a CO +± E § co co CD Q. 2 g x E °-s a E i .g-o CD m CD CO ± ± co "o CD -2 CO 1 E 2 E CO CD O CD CL . C 2 > Q _ | Q "co 3 E E a> -a co +± E c 2> ro C L l l CD ° ro X c o ^ CD CO > CD O c ro ro L O o d v a . CD ro > ro c i CD -I—' CO E C D _ C T3 C o C L CO S> o o CD CD \ ro CL E o o L — CD - C C D C O -<= .12 o co c CC ^ TO J - -Jr; C CD ^ CD a . < E w I I o * ro ro •+-* +-> c c 8 8 jo _ro C L C L i i c c o o c c ro lo B LU ro _ i . -O CD ® ro £ ro ro _CD ro o J D _ o ro ro CD CD ro ro C L co ^ E « JT ? ° . . CO CD i= o ro ro o i ro T = CD iS T5 o ro •= •5 a * s CD ro ^ E •*= CD — I ro O CD O c ro 0 ro - J> CD o o (— -*—< 1 g ro •3J ro 0 "CD 0 L L c ro I a d CO E ro ro — I CD CD O o o 120 Q CO + l c ro cu S O CO CM TT X I CM CM 6 re z >-CD > CO o > co LU n Tr h-T — CM re < LO CM CM O ro Q£ T— CM T T CM CO CD d d T ' CM + l +i + l + l T— Is- co LO T T Oi T— ' d CM CM c o ro w 'E E < "ro c >_ 0) <*-• ro CO o CD T T CO CD d CM CO CM LO LO LO CO CM d CM d o c o CO co CO CO CD CO CO d CO d T— T— CO LO T f o CO d CM d CD Tr' CD CO CO CM O Is- T T Is-' d co d CD CM LO Is- CD o CO CO d d CO +1 +1 +1 +1 CM LO CT) CM o CM CD CO d d CM CM LO Is- CM o co CD Is-d d d T— CD T f CD o o Is- CO d d d n o 2 w 'E E < ro 0) T— LO CM CD o CM CO d d T—' CM CM CO CD CO O o Tr d d d T— CM O Is-o o T— d d T—" Tf ' < < < d CJ o 3 3 < < < o o 3 3 < < 121 3.2.5.3 Comparative Maternal-Fetal Pharmacokinetics of the Parent Drug, the Preformed Metabolite, and the in vivo Generated Metabolite Figure 3.6 shows the representative plasma concentration vs. time profiles of the parent drug, the preformed metabolite, and the in vivo generated metabolite in maternal and fetal arterial plasma during separate maternal (Figure 3.6A) and fetal (Figure 3.6B) administration experiments in E303Y. The peak plasma concentrations (Cmax) of the preformed DPMA in MA and FA after maternal metabolite administration were 602.9 ± 105.9 and 17.3 + 5.5 ng/ml and occurred at 5 (first sampling time) and 336 ± 50 min (tmax) after injection, respectively. Similarly, the Cmax's of the preformed metabolite in MA and FA after fetal metabolite administration were 17.0 ± 9.3 and 1579.9 ± 274.3 ng/ml corresponding to a tm a x of 276 ± 54 and 5 min, respectively. The Cmax's of the in vivo generated metabolite in MA and FA after maternal drug administration were 201.5 ± 281.4 ng/ml (at 389 ± 51 min) and 138.9 ± 1 1 5 . 8 (at 518 ± 137 min), respectively. The Cm ax's of the in vivo generated metabolite in MA and FA after fetal drug administration were 79.8 ± 130.3 and 130.8 ± 37.5 ng/ml at U ' s of 371 ± 12 and 430 ± 50 min, respectively. Tables 3.9 and 3.10 present the comparative pharmacokinetic parameters of the parent drug, the preformed DPMA metabolite, and the in vivo formed DPMA metabolite in the ewe and the fetus, respectively. The total body clearance (CU, = total i.v. dose/AUC) and steady-state volume of distribution (Vdss) of the parent drug were significantly higher in the fetus compared to the ewe (unpaired t-test, p < 0.01 in both cases). 122 0 20 40 60 Time (h) Figure 3.6 - Representative plasma concentration vs. time profiles of the parent drug, the preformed metabolite and the in vivo generated metabolite in maternal and fetal plasma (E303Y). Figures 3.6A and 3.6B are the data from separate maternal and fetal administration experiments, respectively. In both experiments, unlabeled DPMA was administered as the preformed metabolite in combination with [2H10]-DPHM. The [2Hi0]-DPMA is thus the in vivo generated metabolite. 123 However, the elimination half-life (t1/2p) and mean residence time (MRT) of the parent drug in the mother and the fetus were not significantly different (unpaired t-test, p > 0.05 in both cases). The total body clearance of the preformed metabolite in the fetus was not significantly different from that in the mother (unpaired t-test, p > 0.05). The fetal Vd s s , ti/2p and MRT of the preformed metabolite were, however, significantly higher compared to those in the mother (unpaired t-test, p < 0.005 in all cases). Also, in both the fetus and the mother, the C U , and V d s s of the parent drug were significantly higher compared to those of the preformed metabolite; the preformed metabolite ti/2 p and MRT were, however, longer than those of the parent drug. In both the ewe and the fetus, the MRT of the in vivo formed metabolite was significantly longer compared to that of the preformed metabolite (paired t-test, p < 0.05). The percent of total parent drug dose converted to the DPMA metabolite in vivo in the mother (1.72 ± 2.01%) tended to be greater compared to the fetus (0.32 + 0.11%) (but not statistically different, unpaired t-test, p = 0.09). However, the formation clearance of the metabolite as a percent fraction of non-placental clearance in the ewe (1.78 ± 2.12 %) and the fetus (0.87 ± 0.56 %) was not statistically different (unpaired t-test, p > 0.05). 3.2.5.4 Renal Elimination of the Parent Drug, the Preformed Metabolite, and the in vivo Generated Metabolite in the Mother and the Fetus Figure 3.7 shows the representative cumulative excretion profiles of the parent drug, the preformed metabolite and the in vivo generated metabolite in maternal and fetal urine during separate maternal (Figure 3.7A) and fetal (Figure 3.7B) administration experiments in E303Y. 124 0 10 20 30 40 50 60 70 80 Time (h) 00 Time (h) Figure 3.7 - Representative cumulative renal excretion profiles of the parent drug, the preformed metabolite and the in vivo generated metabolite in maternal and fetal urine (E303Y). Figures 3.7A and 3.7B are the data from separate maternal and fetal administration experiments, respectively. In both experiments, unlabeled DPMA was administered as the preformed metabolite in combination with [2H10]-DPHM. The [2H10]-DPMA is thus the in vivo generated metabolite. Figure 3.7B shows that even after fetal administration, negligible amounts of the preformed as well as the in vivo generated metabolite are excreted in fetal urine in comparison to maternal urine. 125 Mean ± S.D. 40.9 ± 14.0 2.1 ± 1.1 57.2 ± 18.2 60.8 ± 19.1 0.55 ±0.18* 0.10 ±0.02* 174.9 ±94.0** 192.1 ±55.7** 1.72 ±2.01 1.78±2.12 298.2 ±71.9*** "D o CO CM TT LU O „ CD O CO -^j CO T -CO CO CD (N O) t « LO O CO LO ° ' ° ' ? £ O CO °? CO LO oo Lb LO £ •o Is-CM CM Tf LU ^ o ^ °? v - • CO CT> CO CO CO OO CD "^  T T f O O LO d d °2 o r- q CO CM o >-CO o r o LU c o L O r - I s -CO ^ 00 00 LO CO Tj" co co °> I s-. T f T— CD CM Q Q CO OO CO CM T f CO <? 00 CO CO o o % "D TT Is-CM LU CD LO I s -T f • cb CM CM LO T f CO 00 P T f T f O CO I s -d d £ °2 T f CD ^ CD CD CO ° ° CM o < i n CM CM LU ^ CD 9 ^ T f O i I s -LO T f CO CO T - CXI T -00 T - CM O d d °. £ CM CM n . I s - I s - O Pharmacokinetic Parameter jz E "o) ^  ^ j r §- a: o > S cn c E '55 ^ - v c" 1 s 11 x i u> 1— _ j T ? §• DC O > ^ 2 'c' ^ ^ E ° ^ r -E ~ E * LL LL 2 cn 3 Q c 2 ro 0 . T J CD o E o •e l Q L S In vivo Formed Metabolite CD CO CO £ J O C T J CD CD JO E CL "° 6 c o CD C CO o CL CD § r *S CO CO CO CO CD o O v -8 O m m CD 2 E CO CD CD O £ T J O _ g CD _ ° 3 ro 2 co M _ CD O O .IS i= E co o CO § CD S o 5 l o I +—' 0 ) - ° E § o -CD CD O = c o CO - Q co iS CD CD E CD Q X C N CD z> c CO CL X CL Q CD "o - Q CO + J CD E XD CD E i _ 2 V. CL CD CL C ' T J a> c ~ o o C L J D CO CO 2> "£ o E ° - D CD CD £ E C O S I r C L CD CD C D J Z C ^ O > * i _ T = CD o E co c CD CO ^ ' C L CO _ ' c? * — * T3 . - C CD O o CD ro C L CD _ c ro o £ P "5 9> Tr> CD O •C c CD CO > ro CL Q CD ZJ c CO CL X Q_ Q I i.-S J D iz O CD "to E .CO d ) E a; "CD E co CO C L CD CD CZ _o "rz co CD o c 4= 1 0 ^ C CD O CO o o CD CD ZJ czo o o ® £ o C L c c ro CD i= ro • ^ °- OT CU CD CD $ J Z o co — M - O ° = i — CO S CD ^ E E g c co q ? ro ° W CL V c co CD .CD CO J Z c o » co c J CD "ro tz 'E L — CD c o " Z J J 3 CO CD E _ZJ o > CD -I—• ro +-» CO I >» CO CD -J—' CO I in in T J > CD" o c ro i _ ro J D o >. T J 0 CD f i 2 O 1 CD _J co o 2? 126 Mean ± S.D. 285.6 + 122.2 13.1 ±3.1 33.1 ±21.6 51.3 ± 18.9 0.37 ±0.11* 0.40 ± 0.06* 866.0 ± 203.6** 1151.1 ±318.1** 0.32 ±0.11 0.87 ±0.56 1401.8 ±423.7*** u O CO CM M" LU ^ O C O C O CO C O ^ L O C M - ~ ^ w O Is- ° > M- 00 •«- Is-Q o C O C O 00 C O C O C O 00 ^ d d oo •D I V CM CM M" LU ^ co M - o S 6 in c \ i T - co oo T - L O 2 3 O C N . . OO C O O O T - C D L O C M C O ~ C M L O g d d T -C M o >-c o o CO LU T M" L O Is-H C O T - L O S - C N M-C O 00 ^ ^ « ^ § CO O O g C N co oo rj °? °°. o o o Mr •a M-Is-CM LU CP L O v - L O Q C O C O M" 3 - - C O r o C D 0 0 ° • d C O M - C M 2 d d fe o Is-O C O Q <N °? ° O O T -o < If) CM CM LU 9 T - c n m c o ^ c o < ° <r- T - 00 o co ^ L O M - o oo o o C O C D co oo 00 C D O ^ oo o O T - O Pharmacokinetic Parameter o5 E '55 "cr" O > J S "55 c E '55 _ 'c? mi _ J ~o S Cd O > ^ 1> 'cT ° ^ — H E ~E * L L i i ^ CO zz Q +J c CD ro Q. "C CD CD . t i E o 2% CD Q. 5 /n vivo Formed Metabolite o c T3 d) CO ro JZ CL " ° fZ C o CD C CO o CL c g co CO CO 4 = (1) o o a -£ o is m CD S E CO CD CD o £ T3 O O ro o CD o rz ro I— CO JD O c g ro E p 0_ Q X CM CO ZJ rz ro CL 0 . Q 0 ~o J3 ro -*—< CD E T3 CD E CL E < O -i= S I -I—* (D-Q E § CD a sz ^ ~ .c o -(D ^ O — c o ro .a ro iS CD CD O E 2. «> E +-O " O CD + - CD O ^ ^ CO Q_ Q CO ZJ ro CL CO Q. 11 o o CL _Q w ro 2 -55 o E ° -a CD CD £ E c o 5*55 CD CD co _c c *-° o 5= CD rz ro 2 .2 E ro . i | w _ * ^ • - c CO o " CD C -CD ° C O CO CL CD CD SI •4—* c ro B & CD £Z E -2 2 ^ ro •£ °- ro CO o fZ ^ T3 C rz go o CO CL i CO CD o ® o o X Q_ Q i o ' ^ - ^ -Z- CD X ± i ^ O -Q rf CO O CD II CO CD .2 E E J5 CL 8 CO CD ZJ c c ro OJ i= ro • CL CO ^ D V o — o >, o = i_ ro E E g c ro R .25 ro ° CO CL V c ro CD E .CD ro c g ro c | CD ro c E i— CD g -•—* ZJ .g i_ ^—» CO T J CD E _ZJ o > CD -*-» CO CO I •D CO CD CO I u> Q) O C CO I— ro _CD o > . 0 CD - E Is a 1 CD -o:2 T co CD o 127 CD o Si CD E XJ CD E £ 2 a o o 3 C £ ro o. . £ • = 2 O CO +3 >*-ro XJ c c E ™ =3 § 1-2 £ E CD C JZ — +- CD ° o CD 4 - i gl ro TJ ro c °- E co p I I 0 C ro <i) 1 = Q- ro co ro CM CM T t UJ flj O 3 C CD k_ to Q . T J •4—< a? o X CD CD co O T J O T J C O E o E ^ O J O T 5 0) -*—' CD i U X CD CZ 3 • o E < T J CD E o o> L L ~ •p. TO >» *•• £ s CD O CM LO CO 0 0 O d d CD d T - ' + 1 +1 + i + 1 +i + 1 0 0 o O J T t o CO 0 0 L O o CD o O d 00 d T — L O T— 0 0 0 0 0 0 T f in 0 0 T f CM cn o CD d cn d T f o cn o CD T f i^ -T f L O 0 0 o 0 0 d d d d d CO CM CM OO o C3J LO 0 0 CO o c i d d d 0 0 o v - CD 0 0 LO 0 0 0 0 T f T f CD o c i d 0 0 d d i^ -o T f cn Is-T f CD 0 0 CM Is-c i c i d 0 ) d d urine urine 1 urine ternal ternal aterna jse) 0 0 CD o 0 0 T f o o o d d d CM + 1 + 1 +1 + 1 T f 0 0 cn T f o Is-d d o T — T f + 1 CM cn oo o d d T f 0 0 0 0 v— o cb O d T f cn d o CM O T f o T f o o d + 1 i^ -o o CM o o d +1 T f O o <c <: CM O O O T f d CM 0 0 <: < 2 2 o o CM CM d CD cz 'l_ ZJ TO c CD O X CD CD cn O T J L O o o T f CM O O o o d d oo T f CM cn r< O 0 0 0 0 o CD LL. D> 3 C CD L_ re a. CD cz ZJ ro -t—> .<D T J CD -I—» CD s O X CD CD CO O •a ro cz a> ro E c T J CD « s o X CD CD CO O T J <c <: Is- T f o o o o O T J CD co a E o °$ £ a> o- 5 O CO jo CD C * L _ ZJ ro $ a? CO E o T J 3 §> 8 c CD 0 i ° T J cu E O £ L L ~ — - re •5 S CD O X i 03 -*—» CD E "a 0) e Q_ < 0_ Q O J ZJ i _ c : CD s— ro C L 0_ Q I a? O X3 co ' CD E T J CD E CD C L CL a CD 3 CZ CO oS -S ! * E a. CD ro^ . 0 > c ro cz 2 "5 .2 co Q..E 1 ™ I l l LO o d v C L T J c ZJ O C L E o o CD E CO CO CD £ b CD O CO s 03 J D O ro cz CD T J — CO c a> ro E CD J Z co x; ^—' CD o cz co o cz cp to I 128 Table 3.11 presents the comparative pharmacokinetic parameters of the renal elimination of the parent drug, the preformed metabolite and the metabolite generated in vivo in the mother and the fetus. Fetal renal clearance (CLr) of the parent drug was higher (but not significantly so) compared to that of the mother in the three animals where fetal urine was collected. In contrast, maternal renal clearance of the preformed as well as the in vivo formed metabolite was significantly greater than that of the fetus in these three animals (unpaired t-test, p < 0.05). However, there was no significant difference between CL r of the preformed and the in vivo generated metabolite in the mother or the fetus (paired t-test, p > 0.05). Renal clearance of the preformed metabolite accounted for 88.8 ± 6.5 % of its total body clearance in the mother, and only 3.0 ± 3.8 % in the fetus. Consequently, a significantly greater percentage of the maternal i.v. dose of the preformed metabolite was excreted in maternal urine (88.0 ± 6.5 %) as compared to the excretion of the fetal dose into fetal urine (1.79 ± 2.08%). Instead the majority (92.1 ± 7.4%) of the fetal i.v. dose of the preformed metabolite was eventually recovered in maternal urine (Table 3.11). 3.2.6 Study F: Disposition of the DPHM-N-oxide Metabolite in the Maternal-Fetal Unit As mentioned earlier, maternal and fetal plasma and urine concentrations of DPHMNOX (or [2Hi0]-DPHMNOX) were measured in four out of five animals (E1225A, E2174, E303Y and E4230) employed above in study E. Figure 3.8 shows the representative maternal and fetal plasma concentration vs. time profiles of the N-oxide metabolite in relation to those of the parent drug in E4230 after maternal (Figure 3.8A) and fetal (Figure 3.8B) drug administration. Table 3.12 presents the maternal and fetal arterial plasma AUC's of the N-oxide metabolite and also the corresponding FA/MA AUC ratios during maternal as well as fetal drug administration. 129 Figure 3.8 - Representative maternal and fetal arterial plasma profiles of the parent drug and the N-oxide metabolite in E4230 after A) 6 h maternal DPHM infusion (670 pg/min), B) 6 h fetal DPHM infusion (170 pg/min). 130 Renal clearances of the N-oxide metabolite, and the percentages of the total administered parent drug dose excreted in maternal and fetal urine as the N-oxide metabolite are presented in Table 3.13. Table 3.12 - Maternal and fetal plasma AUC's of the DPHM-N-oxide (or [ 2 H 1 0 ] -DPHM-N-oxide) metabolite in four pregnant sheep during separate maternal and fetal 6 h DPHM (or [ 2 H 1 0 ]-DPHM) infusion. Ewe# Arterial Plasma AUC (ng.min/ml) FA/MA Ratio of the Metabolite Maternal Plasma (MA) Fetal Plasma (FA) Maternal Administrat ion E1225A a 1762.2 3069.2 1.74 E2174 b 2107.0 3073.3 1.46 E303Y 3 2885.4 5767.3 2.00 E4230 b 16150.7 11137.5 0.69 Mean 5726.3 5761.8 1.47 ±S.D. ± 6965.4 ± 3 8 0 2 . 5 ± 0 . 5 7 Fetal Administrat ion E1225A 3 384.5 8112.9 21.10 E2174 b 795.0 15356.2 19.32 E303Y 3 616.5 20447.5 33.17 E4230 b 1846.1 34494.6 18.69 Mean 910.5 19602.8 23.07 ±S.D. ± 646.0 .—- :—z—1 ± 1 1 1 4 3 . 5 ± 6 . 8 1 * a - in these ewes [^H10]-DPHM was infused. b - in these ewes unlabeled DPHM was infused. * - significantly greater than the corresponding ratio after maternal administration (p < 0.05). 131 Table 3.13 - Pharmacokinetics of the renal elimination of the DPHM-N-oxide metabolite in the mother and the fetus. Ewe# E1225A 3 E2174 b E303Y a E4230 b Mean ± S.D. Maternal Renal Clearance (ml/min/kg) c 5.1 1.4 2.3 1.7 2.6 ± 1.7 % Dose Excreted as the N-oxide Metabolite 0.31 0.14 0.16 0.74 0.34 ± 0.28 Fetal Renal Clearance (ml/min/kg) c 2.3 N/Ad 1.6 N/Ad 2.0 % Dose Excreted as the N-oxide Metabolite 0.11 N/Ad 0.12 N/Ad 0.11 a - in these ewes, fH^j-DPHM was infused. b - in these ewes, unlabeled DPHM was infused. c - maternal renal clearances were obtained from the data during maternal drug administration, whereas fetal renal clearances were obtained from the data during fetal drug administration. d - samples not available. 3.3 Discussion 3.3.1 Study A: Paired Maternal (DPHM) and Fetal ([ 2H 1 0]-DPHM) Infusions for the Determination of Placental and Non-Placental Clearances This study utilized simultaneous administration of unlabeled (DPHM) and deuterium labeled DPHM ([2Hi0]-DPHM) for the determination of maternal and fetal placental and non-placental clearances in a single experiment. The primary reason for employing this methodology in this study was to eliminate the potential confounding effects of fetal growth and maturation that may occur over the period between time-separated maternal and fetal drug infusions. This approach also reduces the overall duration of the experiment. This is an important factor in studies involving chronically-instrumented pregnant animals where there is a finite time window available for each preparation, and hence shorter experiments 132 allow for additional studies to be conducted on the same animal. However, it is important that the labeled and unlabeled forms of the drug be biologically equivalent (Baillie, 1981). If this is not so, the labeled drug could display different dispositional characteristics compared to the unlabeled drug, and thus be of limited use in pharmacokinetic studies (Baillie, 1981). Bioequivalence of DPHM and [2H10]-DPHM in terms of their pharmacokinetic characteristics and also in their metabolism to DPMA and [2H10]-DPMA metabolites has previously been demonstrated in this lab (Tonn, 1995). 3.3.1.1 DPHM and [ 2Hi 0]-DPHM Plasma Concentrations in the Ewe and Fetus The average total fetal plasma concentration of diphenhydramine achieved in the current experiments (i.e., DPHM + [2H10]-DPHM = -289 ng/ml) lies between the two plasma concentrations achieved in the time-separated maternal (-36 ng/ml) and fetal infusions (-450 ng/ml) in an earlier study in this lab (Yoo et al., 1993). In the ewe, the total drug concentration in the current study (305 ng/ml) exceeds the levels achieved in the previous maternal (212 ng/ml) and fetal (31 ng/ml) infusions. 3.3.1.2 Placental and Non-Placental Clearances of DPHM in Fetal and Maternal Sheep The maternal and fetal clearance values obtained for DPHM in the current study are similar to those determined previously (Yoo et al., 1993). In particular, it has been confirmed that fetal placental clearance of the drug is much higher than the maternal placental clearance. Similarly, the weight-normalized fetal non-placental clearance is also greater than that in the ewe. In the current study, CLfm was 5.4 fold higher than CLmf, whereas previous estimates were 3.7 times higher based upon total drug concentrations, 133 and 1.6 times higher based upon unbound drug concentrations (Yoo et al., 1993). The fetal placental and non-placental clearance values for DPHM are the highest of any drug yet examined in pregnant sheep, and this is also the case with the CLfm-CLmf difference. Given the overall agreement between the results of two studies, we conclude that this latter difference is not an artefactual result of the time-separated maternal and fetal infusions in the previous studies that have employed the 2-compartment open model. 3.3.1.3 Maternal and Fetal Plasma Concentrations of DPMA and [ 2H 1 0 ]-DPMA In humans, monkeys, and dogs, DPHM appears to be metabolized via two sequential N-demethylation steps followed by deamination to DPMA (see Chapter 1, Section 1.4.3). This DPMA metabolite and its conjugates are the major urinary metabolites of DPHM in these species (Chang et al., 1974; Drach and Howell; 1968, Drach et al., 1970; Glazko et al., 1974). DPMA is also present in the urine and plasma of non-pregnant ewes following DPHM administration (Tonn etal., 1995). In the present study, DPMA and [2H10]-DPMA were detected in both maternal and fetal plasma during and following the simultaneous infusions of DPHM and [2Hio]-DPHM to the ewe and fetus, respectively. The consistently higher concentrations of the labeled metabolite in fetal plasma compared to those in the mother during the fetal [2Hio]-DPHM infusion provide strong evidence for its formation in the fetus. The presence of DPMA in the fetus and [2H10]-DPMA in the ewe could be the result of two processes: 1) placental transfer of DPHM to the fetus and [2Hi0]-DPHM to the ewe, with subsequent formation of the unlabeled and labeled metabolites in fetal and maternal compartments, respectively; and 2) fetal to maternal transfer of [2Hio]-DPMA and maternal to fetal transfer of DPMA. However, it seems unlikely that the latter process could be of much importance, since the minimal umbilical extraction ratio of [2H10]-DPMA 134 (~0, see above) and the long fetal elimination half-lives of the labeled and unlabeled metabolite (-15 h) suggest its limited transfer across the ovine placenta. This is perhaps due to a greater polarity of the metabolite compared to DPHM and its high degree of plasma protein binding (-99%) (Tonn, 1995). However, a larger maternal plasma ratio for A U C ^ H ^ J - D R M A / A U C ^ H ^ J - D P H M (0.96±0.26) compared to A U C D P M A / A U C D P H M (0.62±0.15) suggests that at least a portion of the labeled metabolite that is formed in the fetus is transferred to the mother. A long half-life of the metabolite in the fetus also suggests that the elimination pathways for this metabolite may not be fully developed in the fetus as compared to the ewe. Our subsequent studies, involving fetal bolus administration of DPMA, have in fact indicated that virtually all of the administered dose ultimately appears in maternal urine over the ensuing 96 h (see Section 3.2.5.4 in this Chapter). This confirms that DPMA can cross the sheep placenta, albeit at a very slow rate. It also suggests that both the fetus and the ewe have no detectable ability to secondarily metabolize DPMA (see Section 3.3.4 later in this chapter for more details). 3.3.2 Studies A and B: Fetal Hepatic Uptake and Metabolism of DPHM 3.3.2.1 Evidence of Fetal Hepatic First-Pass DPHM Uptake from the Umbilical Vein In a previous study (Tonn et al., 1996), the fetal hepatic first-pass uptake of DPHM following umbilical venous drug administration after both bolus and constant rate i.v. infusion was examined. However, no evidence of fetal hepatic first-pass DPHM uptake was obtained. In contrast, a substantial (>90%) hepatic presystemic elimination of the drug was observed after portal venous administration in adult sheep (see Section 3.2.3). However, the results from the fetal experiments did not completely rule out the 135 involvement of fetal liver in DPHM metabolism/DPMA formation because we subsequently detected the formation of DPMA from DPHM in fetal hepatic microsomes at rates similar to those in maternal liver microsomes (S. Kumar, G.R. Tonn, K.W. Riggs and D.W. Rurak, unpublished data). The data from Study A on DPMA/DPHM (8.20 ± 3.62) and [2H10]-DPMA/[2H10]-DPHM AUC (2.24 ±1 .19) ratios in the fetus (Table 3.2) indicate that more of the maternally-derived form of the drug (reaching the fetus via the umbilical vein and hence undergoing a 'partial' fetal hepatic first-pass) is converted to the metabolite than is the form administered directly to the fetus. These AUC ratios clearly demonstrate that the fetal liver is involved in the metabolism and non-placental clearance of the drug. Using equation 7 and the parent drug and metabolite AUC ratios from this study, the fetal hepatic first-pass extraction ratio of DPHM present in umbilical venous blood averaged 0.71 ± 0.16. However, using the DPHM, [2H10]-DPHM, DPMA and [2Hi0]-DPMA concentrations measured in samples from the previous umbilical hepatic first-pass experiments (Study B) and equation 10 (analogous to equation 7; sections 3.1.6.1 and 3.1.6.2), a mean value of 0.44 + 0.14 was obtained. We feel that the higher estimate of this parameter obtained in the paired infusion study (Study A) is due to maternal to fetal transfer of a portion of the maternally formed DPMA to artefactually increase the fetal A U C D P M A / A U C Q P H M ratio. In the fetal hepatic first-pass study (Study B), labeled and unlabeled DPHM were both administered to the fetus, so that maternal to fetal transfer of intact drug or metabolite was unlikely. Thus the mean value of 0.44 for the fetal hepatic DPHM extraction ratio estimated from direct fetal umbilical venous administration is likely more accurate. The failure to detect a fetal hepatic first-pass effect in the previous study (Tonn et al., 1996), which measured only the concentrations of DPHM and [2H10]-DPHM, may have 136 resulted from the geometry and hemodynamics of the fetal circulation combined with a high placental permeability for the intact drug. In terms of the former factors, a greater portion of the umbilical and fetal hepatic venous return is preferentially distributed to the upper body, with only -20% reaching the placenta in one circulation time (Edelstone and Rudolph, 1979; Reuss and Rudolph, 1980). In contrast, ~50% of inferior vena caval blood reaches the placenta (Edelstone and Rudolph, 1979; Reuss and Rudolph, 1980). This is likely a physiological mechanism for the fetus to preferentially supply highly oxygenated nutrient rich umbilical venous blood to vital organs such as the brain and heart, and channel the deoxygenated inferior vena caval blood to the placenta for reoxygenation. In Study A above, approximately 60% of the drug delivered to the placenta was extracted at this site. Thus, after one pass through the fetal circulation, the average systemic availability of the drug (DPHM or [2Hi0]-DPHM) injected via the umbilical vein will be -50%. This is because on average -50% of the drug injected via the umbilical vein is extracted in a single pass through the circulation (calculated as the fraction removed via hepatic first-pass extraction [~44%], plus the fraction extracted by the placenta [~6%], for a total of ~50%). Similarly, after a single pass through the fetal circulation, the systemic availability of the drug administered at the inferior vena cava (tarsal vein) will be ~70%, since ~30% of the drug will be extracted at the placenta. In addition to this factor, fetal circulatory transit times from the umbilical vein to the placenta (-5.1 sec) and from inferior vena cava to the placenta (-3.7 sec) are different (Power and Longo, 1975). Also, the transit time to the placenta for the -50% of the umbilical blood flow that passes through the fetal liver is even longer (-9.8 sec, Power and Longo, 1975). Due to all the above described factors, the drug administered at the umbilical site likely experiences a 'fetal hepatic first-pass effect', while that administered at the tarsal venous site experiences a 'placental first-pass effect' (see Chapter 1, 137 Figure 1.3). Thus, even though fetal hepatic first-pass DPHM uptake from umbilical venous blood was present, the high placental extraction of drug administered via the tarsal vein may act to nullify the difference between systemic concentrations (AUC's or C s s ) of the two forms of the parent drug. This would result in minimal and inconsistent differences in the systemic arterial levels and AUC's of the two forms of drug and thus to an apparent lack of fetal hepatic DPHM elimination, as was concluded in the previous study (Tonn et al. 1996). However, if we assume that fetal hepatic DPHM elimination is via its metabolism in the fetal liver whereas placental elimination involves simple drug transfer to the maternal circulation, there should be differences in the fetal plasma concentrations of the DPHM metabolite (e.g., DPMA and [2Hi0]-DPMA) formed from the two forms of the drug. Also, these differences are more likely to be maintained over time because of the limited placental permeability of more polar and highly protein bound DPMA (see above) in contrast to the parent drug which can readily cross the placenta. In agreement with this, a consistent concentration difference between the labeled and unlabeled forms of DPMA in fetal arterial plasma (resulting from the hepatic first-pass uptake of umbilically administered drug and subsequent formation of higher amounts of metabolite from this form of the drug) was observed (Study B) (Figure 3.3). The data on fetal plasma DPMA and [2Hi0]-DPMA concentrations allowed an indirect estimation of the fetal hepatic first-pass extraction of the parent drug after umbilical venous administration by using equation 10 (section 3.1.6.2). The estimate of 0.44 for fetal hepatic DPHM extraction is less than that determined in adult sheep (0.94; also see Section 3.2.3). However, -50% of the umbilical venous return bypasses the fetal liver via the ductus venosus (Holzman, 1984), and thus drug present in this blood is not available for hepatic first-pass uptake. When the fetal hepatic extraction estimate is corrected for 138 this, it approaches the adult value (0.88 vs. 0.94). Consequently, it appears that the liver of the fetal lamb in late gestation is quite effective in metabolizing DPHM. Moreover, when this hepatic extraction estimate of 0.88 is multiplied by a published value for total hepatic blood flow in the fetal lamb (137 ml/min/kg fetal weight, Edelstone et al., 1978), the resulting estimate of fetal hepatic clearance is -120 ml/min/kg. This can account for the entire CLf0 of DPHM (Table 3.1), suggesting that the fetal liver is the major organ responsible for fetal non-placental clearance of DPHM. Renal clearance of DPHM contributes only -2% to CL f 0 (Tonn, 1995), and previously we found that fetal pulmonary extraction of DPHM contributes another 8% (Rurak et al., 1991). Thus, the combined average clearances of the liver, kidney and lung are somewhat greater than our current estimate of CL f 0 . However, given that fetal hepatic extraction is estimated by an indirect method and is only an average estimate, and that the other data come from different studies, this difference is not unreasonably great. The metabolic fate of the DPHM taken up by the fetal liver was the subject of our subsequent studies (Studies E and F in this chapter). There have been few other attempts to directly examine the role of the fetal liver in drug elimination/metabolism. In fact, there appears to be only one study where fetal hepatic extraction of propranolol was measured in utero (Mihaly et al., 1982b). This study employed the Fick method and quantified fetal sheep hepatic propranolol extraction by simultaneously measuring drug concentrations in umbilical, portal and right hepatic veins. An average drug extraction of 35 ± 12% was demonstrated from the umbilical venous blood supplying the fetal liver. In contrast, there was no apparent difference between drug concentrations in portal and right hepatic venous plasma. These data were explained by the fact that the left lobe of the fetal liver receives -93% of its total blood supply from the 139 umbilical venous flow whereas this figure is only ~60% for the right lobe, with the rest being supplied by the portal vein (hepatic artery supplies only ~5% of the total fetal hepatic blood flow) (Edelstone et al., 1978; Holzman, 1984). Since the portal venous blood is relatively poorly oxygenated as compared to the umbilical venous blood, the right fetal hepatic lobe may also be less well oxygenated as compared to the left lobe. Thus, apparent differences in fetal hepatic propranolol extraction from the umbilical and portal venous blood were postulated to be due to the differences in oxygenation of the two fetal hepatic lobes (Mihaly et al., 1982b). Although direct evidence for this phenomenon is lacking and an accurate estimation of the fetal hepatic extraction of propranolol by the above approach would require additional sampling from the left fetal hepatic vein, these data in combination with our data on DPHM do indicate additional levels of complexity in the study of fetal hepatic disposition of xenobiotics. If indeed there are differences in the metabolic capacity of the right and left fetal hepatic lobes, the above estimate of a total fetal hepatic DPHM extraction of 88% (2 x 44%) may be an overestimate. More recently, an isolated perfused fetal lamb liver preparation, with perfusion via the umbilical vein and simultaneous measurement of the ductus venosus shunt fraction (see Figure 1.3), has been utilized to study fetal hepatic drug disposition. Using this experimental system, significant uptake and metabolism of propranolol and p-nitrophenol by the late-gestation fetal lamb liver has been demonstrated (Ring etal., 1995; Ring etal., 1996). Also, under these in vitro conditions, where the level of oxygenation of the right and left lobes of the fetal liver was similar, the two lobes were found to be equally effective in propranolol uptake (Ring et al., 1998). Overall, significant fetal lamb hepatic uptake of DPHM appears to be in line with these studies of other compounds and indicates an important role of the fetal liver in drug elimination. 140 3.3.2.2 Impact of Fetal Hepatic Drug Uptake on the 2-Compartment Model Estimates of Maternal and Fetal Clearances The placental clearance values calculated using the 2-compartment model are the "fundamental" clearances of the maternal-placental-fetal system and can be used to estimate the maximum possible rate of drug transfer across the placenta (under given conditions of blood flow and protein binding) assuming sink conditions on the other side. These clearance parameters are thus reflective of the true placental permeability of the drug in question. This is in contrast to the net rate of placental drug flux, which depends only upon the rate of non-placental drug elimination on the other side of the placenta. It is important to realize that the proposed 2-compartment model incorporates both maternal as well as fetal drug elimination. Hence, after maternal or fetal drug administration, this system never reaches a state of equilibrium (defined as equal bi-directional drug fluxes and no net transfer of the drug across the placenta). It does, however, reach a steady-state where the rate of drug flux across the placenta becomes equal to the rate of drug elimination from the other side of the placenta. There could be 2 possible situations: a) Steady-state maternal drug administration: In this case, there is net maternal-to-fetal drug transfer. The rate of this transfer is equal to the rate of drug elimination from the fetus via non-placental pathways, i.e., CLmf * Cmss — CLfm * Cfss = CLfo * Cfss b) Steady-state fetal drug administration: In this case, there is net fetal-to-maternal drug transfer and the rate of this transfer is equal to the rate of non-placental drug elimination from the mother, i.e., CLfm * Cfss'—CLmf * Cmss' = CLmo * Cmss' 141 Hence, after maternal steady-state drug administration, the maternal-to-fetal drug flux always exceeds that in the opposite direction. The situation is reversed after fetal drug administration and fetal-to-maternal flux exceeds that in the opposite direction. Now consider a hypothetical situation where equal steady-state drug concentrations are achieved in the mother and the fetus after separate drug administration. In this case, the maximal rate of maternal-to-fetal drug flux after maternal administration should be equal to the maximal rate of fetal-to-maternal drug flux after fetal administration. This is because of the presence of the same drug diffusion barrier in both directions (i.e., placenta) and similar rate of drug delivery to the placenta (due to relative equality between maternal and fetal placental blood flows and equal drug concentrations). By definition, the maximal placental drug flux rate is a product of placental clearance and steady-state drug concentration (maternal-to-fetal flux = C L m f * C m i and fetal-to-maternal flux = CLfm*Cf'). Since we are assuming equal drug concentrations (Cm and Ci after separate maternal and fetal drug administration, respectively), the placental clearance (CLmf and CL f m) in both directions should be equal. Also, in linear pharmacokinetic systems, clearance is constant at all drug concentrations. Thus, CLmf and CLfm should be equal at all rates of drug infusion as long as the assumption of linearity holds. However, as noted in the Introduction (Chapter 1), with the exception of acetaminophen, all drugs studied in pregnant sheep have values of CLfm that are greater than C L m f (Table 1.1). In the case of DPHM, we believe that the lower value of C L m f compared to CLfm is in part due to fetal hepatic first-pass uptake of the maternally administered form of the drug that reaches the fetus via the umbilical vein. As noted above, we feel that the value of 0.44 is a more accurate estimate of this fetal hepatic extraction of DPHM. Since ~44% of the 142 maternally-derived drug present in umbilical venous blood will not reach the fetal systemic arterial circulation, its steady-state concentration (Cf) and AUC in fetal arterial plasma will be reduced by approximately the same magnitude. Thus, the clearance estimates obtained by using equations 1-6 (section 3.1.6.1) will be biased by this reduction in Cf. When these clearance estimates are corrected for fetal hepatic first-pass extraction (using a systemic availability of 0.56 for maternally-derived DPHM), C L m m (39.4 vs 38.3 ml/min/kg, corrected vs. uncorrected), CLff (332.1 vs. 324.2 ml/min/kg), C L f m (219.7 vs. 214.4 ml/min/kg), C L m o (36.1 vs. 36.6 ml/min/kg) and CL f 0 (112.4 vs. 109.8 ml/min/kg) are only minimally altered. However, the corrected estimate of CLmf (92.6 ml/min/kg) is substantially higher than the uncorrected value (50.3 ml/min/kg). It is still lower than the CLfm value (214.4 ml/min/kg). However, it is generally assumed that only the unbound drug can diffuse across the placenta, and we have previously reported that the free fraction of DPHM in fetal plasma (0.277) is significantly higher than that in maternal plasma (0.141, Yoo et al., 1993). When C L m f is corrected for the difference in maternal and fetal plasma unbound fractions of DPHM (i.e., by a factor of 1.96), the estimate of CLmf (181.9 ml/min/kg) approaches CLfm. Thus, the apparent difference between CLmf and CLfm for DPHM, when the 2-compartment open model is employed to determine the clearance values, appears to be due to two factors: a difference in maternal and fetal plasma protein binding and a significant first-pass uptake of maternally derived drug by the fetal liver. As noted in Chapter 1, the estimates of CLmf are lower than those for C L f m for all drugs studied using the 2-compartment open model, with the exception of acetaminophen. Some of these drugs (e.g., morphine) are not bound to any significant extent in maternal and fetal plasma, and for others (e.g., methadone) the placental clearance difference 143 remains even after the maternal and fetal differences in plasma protein binding of the drug are taken into account. Thus, the CLfm-CLmf clearance difference could largely be due to fetal hepatic first-pass uptake of maternally-derived drug for many of these drugs. The calculated clearance values could thus be in error, but this could only be determined by obtaining an estimate of fetal hepatic extraction from umbilical venous blood for each drug. However, from another viewpoint, the difference in maternal and fetal placental clearances of these drugs may provide evidence for significant fetal hepatic clearance of these compounds. In Figure 3.9, estimates of C L f 0 for the drugs studied in pregnant sheep are plotted against their CLfm-CLm f difference. There is a highly significant linear relationship between the two variables, suggesting that with the exception of acetaminophen, fetal hepatic first-pass uptake is likely a significant factor in fetal non-placental elimination of these compounds. From this relationship, methadone and DPHM will be predicted to exhibit the most pronounced fetal hepatic first-pass effect and in agreement with this, these drugs also have the highest values for CL f 0 . In the case of acetaminophen, the lack of evidence for any fetal first-pass effect from CLfm-CL m f difference is consistent with the low value of CLf 0 for this drug. The above finding of the involvement of fetal hepatic drug uptake in the underestimation of CL m f is also significant in light of the fact that the relative magnitudes of C L m f and C L f m are often used as an indication of the active or passive transport of the drug across the placenta (Wang etal., 1986a; Pereira etal., 1994; Odinecs etal., 1996a; Tuntland etal., 1996; Tuntland et al., 1998). Our data indicate that this approach may not yield accurate conclusions if the fetal liver is significantly active in the uptake of the drug from the umbilical vein, as with DPHM. 144 180 - i 120 -\ 150 4 • DPHM v Morphine O Methadone A Metoclopramide • Acetaminophen o U L fo = 0.60*(CLfm-CLmf) + 20.8 r = 0.9461 p<0.01 0 0 30 60 90 120 150 180 •fm-CLmf (ml/min/kg) Figure 3.9 - Relationship between fetal and maternal placental clearance difference (CL f m -CL m f , ml/min/kg) and fetal non-placental clearance (CL f 0, ml/min/kg) of drugs studied in pregnant sheep. During our hepatic uptake studies (Study C), DPHM was extensively extracted across the sheep liver as evidenced by the much lower AUC's of the form of drug administered via the portal venous route, and an estimated mean hepatic first-pass extraction ratio of 94.2 ± 3.7% (Table 3.4). This is in line with the high systemic clearance of the drug in sheep (Table 3.5). The principles of metabolite kinetics dictate that the overall shape of the metabolite plasma concentration vs. time profile is highly dependent on the route of drug 3.3.3 Studies C and D: Role of the Liver and Gut in Systemic DPHM Clearance in Adult Non-Pregnant Sheep 145 administration (Pang, 1981). In agreement with this, a significantly higher C m a x of the form of metabolite generated from drug administered via the portal venous route (447.8 ± 175.9 ng/ml) was observed compared to that generated from the form of drug administered intravenously (90.7 ± 75.8 ng/ml). Also, the C m a x 's of the former metabolite tended to occur at earlier sampling times compared to those of the latter. This is because portal venous administration results in an almost instantaneous metabolism of a large fraction of the administered dose (-94.2%, as indicated above), thus leading to an apparent "bolus" injection of the large amounts of formed metabolite into the systemic circulation. In contrast, the drug administered intravenously is more gradually metabolized, leading to a slower increase in plasma concentrations of the metabolite. It has been demonstrated by computer simulations as well as theoretical analysis that the systemic arterial AUC's of the metabolite plasma profile after portal and intravenous administration of equal doses of the drug should be relatively equal (Pang, 1981; Houston and Taylor, 1984). This concept is valid if elimination of the drug by the kidney or other peripheral organs is < 10% of its total clearance and linear pharmacokinetics exist (Pang, 1981; Houston and Taylor, 1984). This is true in spite of the widely differing shapes of metabolite plasma profiles obtained after these routes of administration (see above). On the other hand, if renal or peripheral elimination of the drug is >10%, the AUC of the metabolite after p.v. administration becomes larger compared to that after i.v. administration (Pang, 1981; Houston and Taylor, 1984). The underlying reason for this theory is that if renal or peripheral elimination of the drug is negligible (<10%), the majority of the drug will be eliminated via hepatic metabolism. Since the fraction of drug metabolized via a particular metabolic pathway is constant in 146 a linear pharmacokinetic system, similar amounts of the metabolite will be eventually formed from equal doses of the drug administered via the p.v. and i.v. routes. This will in turn lead to similar systemic arterial AUC's of the metabolite after these two routes of administration. However, if renal or peripheral elimination is significant (>10%), a larger fraction of the i.v. dose will be eliminated via these routes and less will be available for metabolism in the liver in comparison to the drug administered via the p.v. route. Thus, a larger amount of metabolite will be formed after p.v. administration resulting in a higher AUC compared to that after i.v. administration. In previous studies in adult sheep, negligible renal and biliary clearances of DPHM were observed (<0.5% of the total body clearance combined) (Tonn, 1995). Also, in earlier dose-ranging studies in adult sheep (Yoo et al., 1990), apparently linear DPHM pharmacokinetics were observed at i.v. bolus doses of 50-250 mg, which covers the dose range employed in this study. Based on this, and the principles of metabolite kinetics above, the AUC's of the DPHM metabolites after i.v. and p.v. administration of the drug should be equal if there is no peripheral elimination of the drug by other organs. However, in our DPHM hepatic first-pass studies (Study C), the AUC of the metabolite after i.v. administration was only 32.5 ± 14.0% (range 18.2 - 50.4%) of the metabolite AUC generated after p.v. administration. This provides clear evidence for significant peripheral elimination of the drug. Also, from the data presented in Table 3.4, it is evident that almost the entire DPHM dose administered via the p.v. route is metabolized in the liver. This mainly results from the metabolism of a large fraction of the dose during its first-pass through the liver (94.2 ± 3.7%); plus at least some additional amount of the remaining drug is metabolized in the liver during subsequent passes. Based on this and relative metabolite AUC's after p.v. and i.v. administration, it 147 appears that only 32.5 ± 14.0% of the intravenously administered drug is metabolized in the liver; the rest is likely eliminated via peripheral mechanisms. This conclusion is based on the assumption of sole hepatic formation of the metabolite; if the*metabolite is also formed in peripheral organs, the percent metabolized in the liver will be overestimated using the above approach (equation 13). The reason for this is that due to the much lower plasma parent drug concentrations after p.v. administration (Figure 3.4), peripheral metabolite formation will contribute more towards the AUC of the metabolite after i.v. than after p.v. administration. It should be noted that the above analysis is valid in spite of the fact that the DPMA metabolite is not the only contributor to DPHM total body clearance; it in fact only accounts for a very small fraction of DPHM elimination (-1%, see above Study E). This is based on the principle that the fraction of the dose metabolized via a particular metabolic pathway is constant in a linear pharmacokinetic system (Houston and Taylor, 1984). Hence, the plasma concentration-time profiles of any other metabolites will also exhibit a behavior similar to that of DPMA. In order to gain further insight into the possible sites of peripheral DPHM elimination, additional pharmacokinetic analysis was performed on the data from the hepatic first-pass uptake studies. Parent drug pharmacokinetic data after i.v. and p.v. administration can be used to estimate total hepatic blood flow ( Q H ) using the principles of a well-stirred model of hepatic drug clearance (Wilkinson and Shand, 1975; Pang and Gillette, 1978). Thus, Q H in the five adult sheep was estimated to be 62.6 ± 14.7 ml/min/kg (Table 3.5). This value is in excellent agreement with the reported literature estimates of Q H in adult non-pregnant sheep obtained using more direct methods (Katz and Bergman, 1969: 56.6 ± 18.0 ml/min/kg; Boxenbaum, 1980: 48.6 ml/min/kg). It should 148 be emphasized that the above estimation of Q H assumes lack of significant DPHM uptake by the gut. However, the extraction of DPHM across the liver is high (-94%). Thus, the presence of any gut uptake of the drug will not significantly alter the total DPHM extraction across the 'hepato-portal' system (liver + gut) (i.e., observed 94% vs. maximum possible 100%), and will not unduly influence the Q H estimates. The estimated values of Q H can be used to calculate the amounts of intravenously administered drug eventually delivered to the 'hepato-portal' system using equation 15. Thus, in these five sheep, almost the entire intravenously administered dose (average 98.6 ± 9.2%, Table 3.5) was eventually delivered to the 'hepato-portal' system. Since the extraction of DPHM even across the hepatic component of the "hepato-portal" system is nearly complete (-94%), only a very small fraction of the drug delivered to this system can escape uptake/metabolism. Thus, the liver and/or gut are likely the main organs responsible for elimination of a major portion of the DPHM i.v. dose. Consequently, these data provide evidence for a lack of any significant first-pass uptake of the drug by the lung. Moreover, since the liver metabolizes only 32.5 ± 14.0% of the intravenously administered dose, the gut is likely responsible for the elimination of the remainder (67.5 ± 14.0%). Thus, the data from hepatic first-pass uptake studies provide strong evidence for significant uptake of DPHM by the gut. This possible gut uptake of DPHM was confirmed in a more direct study (Study D) where steady-state extraction ratio of the drug across the gut was measured by simultaneous sampling of femoral arterial and portal venous blood. Significantly lower DPHM concentrations in portal venous plasma as compared to femoral arterial plasma provide direct evidence for gut uptake of the drug from the systemic circulation. The 149 steady-state gut extraction ratio of DPHM was 49.0 ± 3.0%. As discussed above, it appears that the gut and liver are responsible for almost the entire DPHM systemic clearance in sheep. Since, the gut and liver are anatomically arranged in series and the major contributor to hepatic blood flow is portal venous flow (~80%, Katz and Bergman, 1969), roughly they will account for 49.0 ± 3.0% and 51.0 ± 3.0% of DPHM systemic clearance, respectively. In actual fact, the contribution of the liver will be somewhat greater and that of the gut somewhat lower due to the presence of the hepatic arterial blood flow component, which bypasses the gut. This estimate of the gut contribution to DPHM systemic clearance is near the low end of the range predicted from our hepatic first-pass experiments (i.e., 49.6-81.8%; Study C), whereas, the estimated contribution of the liver appears to be near the high end of the predicted range (18.2 - 50.4%). This may be due to inter-animal variability and the small number of animals used in these two studies. Also, the steady-state systemic clearances of DPHM during the gut uptake study appear to be somewhat lower compared to those estimated during the hepatic first-pass experiments. This may also be related to a low n value. Thus, overall the combined data from both studies indicate that gut uptake of the drug may account for ~50-80% of the DPHM systemic clearance. The exact mechanism of this DPHM gut uptake is not known at present. It may involve simple binding of the drug to tissue components, its secretion into the lumen via specific transporters (e.g., P-glycoprotein), or metabolism via pathways other than the formation of DPMA. It has been known for a number of years that gut uptake/metabolism is an important factor in determining the oral bioavailability of a number of drugs (Krishna and Klotz, 1994). However, detailed study of the underlying mechanisms of this gut uptake/metabolism is relatively more recent. A first-pass intestinal extraction of 43% 150 was demonstrated for midazolam after intraduodenal administration in humans (Paine et al., 1996), and this was subsequently attributed to CYP3A-mediated metabolism of the drug in the gut mucosa (Thummel et al., 1996; Wandel et al., 1998). Oral bioavailability of cyclosporine, verapamil and nifedipine in humans also appears to be related to the activity of CYP3A in the intestinal mucosa (Hebert et al., 1992; Fromm et al., 1996; Holtbecker era/., 1996). More recently P-glycoprotein present in the intestinal mucosal wall has also been demonstrated to be an important factor in determining the bioavailability of some compounds such as cyclosporine and paclitaxel (Lown et al., 1997; Sparreboom et al., 1997; Asperen ef al., 1997). It has been suggested that P-glycoprotein present in the intestinal mucosa causes a polarized basolateral-to-apical transport of these drugs in the intestine, thereby resulting in their reduced absorption and bioavailability. Thus, the bioavailability of paclitaxel is increased in gene knockout mice deficient in P-glycoprotein (mdrla -/-), and also in wild type mice when the drug is co-administered with a P-glycoprotein blocker (Sparreboom et al., 1997; Asperen et al., 1997). It should be noted that most of these studies have assessed the effect of gut uptake/metabolism/secretion on the bioavailability of the drug after oral administration. Apart from a few isolated papers (du Souich et al., 1995), there appear to be few systematic studies on the extent of gut uptake of drugs from the systemic circulation, as was observed in our study. Hence, the role of gut uptake in the systemic clearance of drugs has not been extensively studied. For midazolam, gut uptake of the drug from the systemic circulation was negligible and hence it was concluded that the intestinal CYP3A was not accessible to the drug in the peripheral systemic circulation (Paine ef al., 1996). Other studies have assumed that the gut uptake of the drug from the systemic circulation is negligible (Hebert et al., 1992; Holtbecker et al., 1996). Our data with DPHM show that this assumption may not necessarily be true for all drugs. Thus, 151 pharmacokinetic analyses of the hepato-portal drug disposition based on this assumption may not be entirely accurate, as was previously argued by Lin etal., (1997). 3 . 3 . 4 Study E: Contribution of DPMA Formation to DPHM Non-Placental Clearance in Maternal and Fetal Sheep After demonstration of a significant role of fetal and maternal liver in their respective non-placental DPHM clearance, the logical next step was to examine the relative maternal and fetal capacity for various metabolic pathways for DPHM. The major route of DPHM elimination in many species is its metabolism to DPMA which is subsequently conjugated to amino acids such as glycine and glutamine (Drach and Howell, 1968; Drach et al., 1970; Chang et al., 1974). This study (Study E) examined in detail the comparative kinetics of DPHM metabolism to DPMA in maternal and fetal sheep, and the resulting effects on fetal exposure to this metabolite. The above aim required knowledge of the distribution and elimination characteristics of the parent drug as well as the metabolite in the mother and the fetus. However, the kinetics of the drug metabolites after parent drug administration alone are complex; a number of phenomena such as metabolite formation, distribution and elimination occur concurrently and their kinetics are difficult to resolve. More specifically, after parent drug administration alone, the volume of distribution of the metabolite and the extent of its sequential metabolism is not known, and hence the total amount formed can not be estimated unless all the sequential metabolic pathways are accounted for. Also, the estimation of the total contribution of primary as well as secondary metabolic pathways is much more difficult in the fetus where the formed metabolites may be excreted into a multiple number of fluid compartments (e.g., amniotic and tracheal fluids, fetal urine) which cannot be cumulatively sampled. The metabolites formed in the fetus also cross 152 the placenta and mix with those formed in the maternal circulation. One approach to examine the kinetics of metabolite distribution and elimination, that could be useful in such situations, is the i.v. administration of the preformed (synthesized) metabolite (Kaplan etal., 1970; Kaplan etal., 1973; Boxenbaum and Riegelman, 1976; Patel etal., 1978; Lai et al., 1978; Cobby et al., 1978). These data may then be combined with the drug and metabolite data after parent drug administration and the kinetics of metabolite formation can be estimated. However, this approach can drastically underestimate the amount of the drug converted to a particular metabolite if the metabolite undergoes rapid sequential metabolism in the liver after its formation and before its egress into the circulation (Pang and Gillette, 1979). In such situations, the preformed metabolite must be administered via the portal venous route in order to more closely simulate the secondary hepatic metabolism and subsequent systemic availability of the metabolite formed from the drug (Pang et al., 1979). In our preliminary studies, we found a complete lack of any secondary metabolism of DPMA in either maternal or fetal sheep (also see below). Hence, in the current study, it was acceptable to administer the preformed metabolite via the i.v. route in both the mother and fetus. This considerably simplified our surgical procedure especially in the fetus where implanting and maintaining chronic portal venous catheters is a difficult task. In the maternal-fetal system, maternal as well as fetal drug clearance is comprised of a placental and a non-placental component. A comparison of the exact capacity of a particular metabolic pathway in the mother and the fetus requires the estimation of their non-placental clearance components. This is because the role of a particular metabolite in total maternal and fetal drug clearance is confounded by the different and variable 153 contributions of the placental clearance component (~ 1-10% in the mother vs. ~ 40-75% in the fetus for DPHM; Yoo et al., 1993; also see Sections 3.2.1.1 and 3.2.5.1). The estimation of maternal and fetal placental and non-placental clearances in turn requires separate maternal and fetal steady-state drug administration (Szeto et al., 1982a). This, combined with the essential study of maternal and fetal preformed metabolite kinetics (see above), meant that essentially four experiments had to be conducted in each pregnant sheep preparation in our study. Given the duration of each experiment (96 h), this was difficult to accomplish in the limited time window (~ 2 weeks) available for experimentation in these late-gestational chronically-catheterized animals before labor and delivery. Also, the rapid fetal growth and physiological alterations during this part of gestation may confound the results of experiments conducted over a prolonged period of time. As discussed previously, in such situations, the use of stable-isotope labeled compounds in combination with mass-spectrometric analytical techniques provides a useful solution for minimizing the inter-occasion variability in pharmacokinetics (Baillie, 1981; Browne, 1990). Hence, we utilized a protocol where either a combination of the unlabeled parent drug and the labeled metabolite (DPHM and [2H10]-DPMA) or the labeled drug and the unlabeled metabolite ([2H10]-DPHM and DPMA) was administered to the mother or the fetus in two separate experiments. The assumption made in this protocol is that the preformed metabolite does not alter the pharmacokinetics of the parent drug and vice versa. The parent drug clearances in these animals were similar to those obtained in our earlier studies where the preformed metabolite was not administered (Yoo et* al., 1993; Table 3.1). Also, the amount of formed DPMA recovered in maternal urine after parent drug administration to the mother was similar in this and the previous study (Tonn, 1995). These observations appear to support the above assumptions. 154 We have demonstrated that plasma protein binding of the drug is an important factor determining the magnitude of placental as well as non-placental DPHM clearance in the maternal-fetal unit (see Chapter 4). Thus, a greater fetal plasma free fraction of the drug compared to the mother is at least partly responsible for fetal placental and non-placental clearances being higher than the corresponding maternal clearances. The presence of drug metabolites in the fetal circulation could be the result of two processes: 1) formation of the metabolite in the mother and its subsequent placental transfer, and, 2) formation of the metabolite by the fetus itself. In our study, the evidence of in utero fetal formation of the DPMA metabolite is provided by the various AUC ratios presented in Table 3.8. If the fetal ability to form a particular metabolite is absent or negligible, the FA/MA AUC ratios of the preformed as well as the metabolite formed in the maternal circulation from the parent drug should be similar (because in this case the maternal-to-fetal placental transfer of the metabolite will be the determining factor of this ratio in both situations). A higher FA/MA ratio of the formed metabolite (2.97 ± 0.82) compared to the preformed metabolite (0.41 ± 0.21) after maternal parent drug and preformed metabolite administration suggests the additional fetal formation of this metabolite from the drug transferred to the fetal circulation via the placenta. Analogous to this, after fetal parent drug and preformed metabolite administration, a higher (but not statistically significant due to high inter-animal variability) MA/FA ratio of the in vivo generated metabolite (0.25 ± 0.35) compared to the preformed metabolite (0.02 ± 0.02), indicates additional maternal formation of this metabolite from the drug transferred in the fetal-to-maternal direction. 155 As discussed above, the fetal liver can metabolize a significant but variable (due to the high variability in ductus venosus shunt fraction) proportion of DPHM present in the umbilical vein in a first-pass manner before it reaches the fetal circulation (see Section 3.3.2). This leads to a greater metabolism and fetal formation of metabolites from the drug transferred to the fetus via the placenta and umbilical vein, as compared to when it is directly administered i.v. to the fetus (see Section 3.3.2 and Table 3.2). This factor, combined with some maternal-to-fetal placental transfer of the metabolite formed in the mother, leads to higher (but again not statistically significant due to high inter-animal variability) in vivo generated metabolite to parent drug AUC ratios in fetal arterial plasma after maternal drug administration compared to when the drug is given directly to the fetus (21.5 ± 26.4 vs. 2.32 ± 1.16; Table 3.8). These data also indicate that depending upon the extent of fetal hepatic first-pass metabolism of the drug present in the umbilical vein, fetal exposure to drug metabolites relative to parent drug concentrations after maternal drug administration may actually be higher than that predicted based on direct fetal i.v. administration. Similar to the FA ratios above, but to a lesser extent, the MA ratio of the in vivo generated metabolite to parent drug AUC's after fetal administration was greater in all the individual animals compared to that after maternal drug administration (2.69 ± 3.87 vs. 1.13 ± 1.44; Table 3.8). However the only possible explanation for this phenomenon appears to be some fetal-to-maternal placental transfer of the metabolite formed in the fetal circulation after fetal drug administration, leading to an increase in the former ratio. There are significant differences in the disposition of the parent drug and the preformed DPMA (or [2H10]-DPMA) metabolite both in the mother and the fetus. In particular, the CLtb and V d s s of the metabolite are much lower compared to the parent drug in both the mother and fetus. In fact, the V d s s of the preformed metabolite is only about 2- and 3-156 times the blood volume in the mother and the fetus, respectively. This indicates very limited tissue distribution of this compound which could be partly related to its very high plasma protein binding in both maternal and fetal plasma (> 99%; Tonn, 1995) and its low lipophilicity (octanol/pH 7.4 phosphate buffer partition coefficient = 0.29; S. Kumar, K.W. Riggs and D.W. Rurak, unpublished data). Also, the t i / 2 p and MRT of the metabolite are longer compared to the parent drug in both the mother and particularly in the fetus, indicating its slow overall elimination from the maternal and fetal circulation. A very long fetal ti/2p and MRT of the metabolite compared to the mother suggests that although the fetal lamb has the ability to form DPMA from DPHM, the mechanisms involved in DPMA elimination are only partially developed as compared to the mother. The placental transfer of DPMA appears to be extremely slow and limited as compared to the parent drug, since relatively low C m a x's in FA and MA occur at least 3-5 h after maternal and fetal preformed metabolite i.v. bolus administration, respectively. Slow placental transfer of this polar and highly plasma protein bound metabolite is partly responsible for its slow elimination from the fetus (also see below). At comparable parent drug concentrations, the fetal in vivo generated metabolite exposure based on relative AUC's of the metabolite is approximately 4-fold greater compared to the mother. Thus, at least in this situation, fetal metabolism of the drug actually leads to a considerably increased fetal exposure to the metabolite due to its "trapping" in the fetal circulation. The MRT of the in vivo formed metabolite is longer compared to that of the preformed metabolite in both the mother and the fetus, suggesting differences in the disposition of in vivo generated and exogenously administered metabolite. This is probably related to the fact that in vivo metabolites are generated within the cell (e.g., hepatocytes) and 157 they must diffuse out of the cell into the circulation in order to be eliminated. However, this is not true for the exogenously administered preformed metabolite. If significant diffusional barriers exist to the efflux of the in vivo formed metabolite from the cell (due to their polarity, molecular size or binding to cellular components), its disposition would be different from that of the preformed metabolite. A diffusional barrier has been shown to exist for the entry of enalaprilat (the polar dicarboxylic acid metabolite of angiotensin converting enzyme inhibitor, enalapril) into hepatocytes from the circulating perfusate and its subsequent biliary elimination in an isolated-perfused liver preparation (De Lannoy and Pang, 1986; Schwab et al., 1990). A similar barrier may exist for the egress of the polar DPMA metabolite into the maternal and fetal circulation after its formation in the hepatocytes, and this could be responsible for its longer residence time compared to the preformed metabolite. In such instances, it is important to utilize mass-balance based approaches (such as AUC's and equation 18) rather than the rate-constant based pharmacokinetic modeling in order to calculate the quantitative importance of a particular metabolic pathway in total drug clearance using the preformed metabolite administration. The latter approach will actually estimate the rate constant of metabolite efflux from the hepatocytes rather than the actual metabolite formation rate constant in such cases. Using equation 18 (section 3.1.6.5) and after taking into account the differences in placental clearance contribution to maternal and fetal total body clearance of the drug, the contribution of DPMA formation to maternal and fetal (non-placental) DPHM clearance was not statistically different and was typically -1% in all but one animal (Tables 3.9 and 3.10). In E4230, the maternal DPMA formation accounted for 5.56% of the maternal non-placental clearance. We have also observed this high inter-animal 158 variability in our similar studies in non-pregnant sheep (not presented in this thesis) and reasons for this are not clear. Overall these data indicate that although the DPMA formation pathway is almost equally functional in both the mother and fetus, this is not a major route of DPHM clearance in maternal or fetal sheep. This is in contrast to many other species where DPMA and its amino acid conjugates account for ~ 40-60% of total DPHM metabolites (Drach and Howell, 1968; Drach etal., 1970; Chang etal., 1974). Significant differences also exist in the renal handling of the parent drug and the DPMA metabolite in the mother and the fetus. Renal elimination of the parent drug in the mother is negligible and accounts for <0.5% of the total dose and clearance. In contrast, the CL r of preformed DPMA accounts for 88.8 ± 6.5 % (Tables 3.9 and 3.11) of its C L t o in the mother, and correspondingly almost all of the i.v. dose administered to the mother was ultimately recovered in maternal urine (88.0 ± 6.5 %; Table 3.11). This indicates that DPMA is not secondarily metabolized in maternal sheep and the entire dose is excreted unchanged in urine. This is also a species difference in sheep compared to the monkey, dog and the human, where significant proportions of total DPMA are recovered as its glycine, glutamine or an unidentified conjugate in urine (Drach and Howell, 1968; Drach etal., 1970; Chang etal., 1974). Due to the lack of this sequential metabolism of DPMA, the amount of in vivo formed DPMA recovered in maternal urine as a percentage of the parent drug dose was similar to the percent contribution of this pathway to maternal non-placental DPHM clearance (1.60 ± 1.86 vs. 1.78 ± 2.12 %, respectively; Tables 3.9 and 3.11). Similar to the mother, the excretion of the unchanged DPHM in fetal urine also accounted for < 0.5% of the total fetal parent drug dose in the 3 animals where fetal urine was collected. In contrast to the mother, 159 however, only 1.79 ± 2.08 % of the total fetal i.v. dose of the preformed metabolite was recovered in fetal urine due to a very low fetal renal clearance of this metabolite (Table 3.11). This is presumably related to a lack of any significant renal tubular secretion of many organic acid compounds such as para-aminohippurate (Elbourne et al., 1990), acetaminophen and morphine glucuronides (Wang et al., 1985; Olsen et al., 1988), valproic acid (see Chapter 5) and indomethacin (Krishna et al., 1995) in the late-gestation fetal lamb. Due to this lack of renal elimination of the preformed DPMA metabolite in fetal urine, almost all of the fetal i.v. dose of this metabolite (92.1 ± 7.4%; Table 3.11) underwent fetal-to-maternal placental transfer and was eventually recovered in the maternal urine over the ensuing 96 h sampling period. The lack of any significant fetal renal elimination of this metabolite combined with its slow placental transfer leads to the observed prolonged fetal t1/2p (and MRT) and exposure of the preformed as well as the in vivo generated metabolite (Table 3.10). 3.3.5 Study F: Disposition of the DPHM-N-oxide Metabolite in the Maternal-Fetal Unit The N-oxide metabolite was detectable in both maternal and fetal plasma during maternal as well as fetal drug administration. In contrast to DPMA, the elimination of this metabolite was almost as rapid as that of the parent drug in the mother as well as the fetus (Figure 3.8). The much higher FA/MA AUC ratios of the metabolite after fetal drug administration as compared to those after maternal drug administration provide evidence of the fetal ability to form this metabolite as well (Table 3.12). A fraction of the N-oxide metabolite was excreted unchanged in maternal as well as fetal urine. The renal clearance data presented in Table 3.13 suggest significant fetal ability to 160 excrete the N-oxide metabolite via this pathway in comparison to the mother. However, the percentage of administered DPHM dose excreted in maternal as well as fetal urine as the unchanged N-oxide metabolite is <1%. This is in contrast to other species such as the rat, rhesus monkey, and dog where -10-20% of the total dose is recovered as this metabolite in urine (Drach and Howell, 1968; Drach etal., 1970). The extent and nature of sequential metabolism of the N-oxide metabolite in not known in maternal or fetal sheep or any other species so that the total contribution of this pathway to maternal and fetal clearance can not be determined. In summary, we have demonstrated that the liver and gut are the major organs responsible for DPHM systemic clearance in adult sheep. Fetal liver during late-gestation also appears to exhibit a significant ability for DPHM uptake and appears to be the major site of fetal non-placental clearance of the drug. The role of the gut in fetal DPHM non-placental clearance remains to be investigated; however, it may be less significant compared to the adult because gut blood flow is only -20% of the total hepatic blood flow in the fetus, as compared to -80% in the adult. The fetal first-pass hepatic uptake of the placentally transferred drug from the umbilical vein also appears in part responsible for an underestimation of maternal-to-fetal placental clearance. This may also be the case for other drugs studied in pregnant sheep using the 2-compartment open model that demonstrate higher values of CLfm compared to CLmf. Conversely, the difference between fetal and maternal placental clearance for these drugs may provide evidence for fetal hepatic uptake/metabolism of these compounds. We have also demonstrated that the use of stable-isotope labeled drug and metabolites in combination with mass-spectrometry provides a powerful tool for studying drug and 161 metabolite kinetics within the maternal-fetal unit. Using an approach based on the simultaneous administration of differently labeled parent drug and metabolite to the mother and the fetus, we have shown that the contribution of DPMA formation to DPHM non-placental elimination in the maternal and the fetal sheep is typically -0.5-1%. Hence, this is a minor pathway in overall DPHM elimination in this species. However, the in vivo functional capacity of this metabolic pathway appears to be similar in the mother and the late-gestational fetal lamb. The DPMA metabolite is not sequentially metabolized in fetal or adult sheep. It is eliminated solely via the renal pathway in the mother and via the placenta (and eventually in maternal urine) in the fetus. The impaired renal excretion and slow placental transfer of the DPMA metabolite is responsible for its long half-life in the fetal circulation. The renal elimination of another DPHM metabolite, DPHM-N-oxide, also accounts for <1% of the total DPHM elimination in the mother as well as the fetus. The minor contribution of the DPMA pathway and DPHMNOX renal elimination in sheep DPHM clearance is in contrast to other species (dog, rhesus monkey, man) where these routes account for -40-60% and -10-20% of the total DPHM dose, respectively. Thus, the exact pathways of DPHM metabolism are likely different in sheep compared to other species. Also, the fate and mechanism of the extensive gut uptake of the drug in adult sheep is not known. Overall, the exact routes of a large portion of maternal and fetal sheep DPHM elimination remain to be investigated. However, the pathways that were examined in these studies appear to be almost equally functional in utero in the late-gestational fetal lamb in comparison to the mother. 162 Chapter 4 Inter-relationships between Plasma Drug Protein Binding, Gestational Age, Umbilical Blood Flow, and Diphenhydramine Clearances in the Ovine Maternal-Placental-Fetal Unit Earlier studies in this lab and those in Chapter 3 have utilized DPHM as a model high clearance drug, which undergoes rapid and extensive placental transfer, to examine different aspects of maternal-fetal drug disposition of this class of compounds. This includes the study of comparative maternal-fetal drug clearance (Yoo et al., 1993), in utero fetal hepatic drug uptake and its relationship to fetal drug clearance (see chapter 3, study A), and in utero functional capacity of fetal drug metabolism pathways compared to the mother (see chapter 3, study E). As part of these studies, we have determined DPHM placental and non-placental clearances in several pregnant sheep during the last two weeks of their gestation. In this chapter, we have retrospectively examined the overall inter-relationships between maternal and fetal plasma protein binding, umbilical blood flow, DPHM maternal and fetal clearances, and indices of placental drug transfer using pooled data from the above three studies. Our aim was to better understand the qualitative and quantitative importance of these factors in determining the kinetics of placental drug transport and the extent of fetal drug exposure. Also, the last third of gestation is a very dynamic period in terms of fetal development and is associated with profound changes in numerous physiological variables that may alter maternal-fetal drug disposition. These include changes in fetal plasma protein concentrations and hence the extent of drug binding (e.g. propranolol and methadone, Czuba etal., 1988; Szeto etal., 1982c), changes in the drug metabolism capacity of the 163 fetal liver (Wang et al., 1986a), development of fetal renal function and drug excretion, and alterations in fetal circulatory and hemodynamic processes (Battaglia and Meschia, 1988). However, little has been done experimentally to elucidate the quantitative influence of these gestational age-related factors on drug disposition within the maternal-placental-fetal unit and the resulting alterations in fetal drug exposure. Hence, we have also assessed the inter-relationships between the above variables as a function of advancing gestation during the last 2-week period of pregnancy in sheep. 4.1 Methods 4.1.1 Animals and Surgical Preparation Data from a total of 18 pregnant sheep were employed in these studies. Surgical procedures for 10 out of these 18 animals are described in chapter 3 (studies A and E). The surgical preparation for the other 8 animals was similar and has been described earlier (Yoo etal., 1993). 4.1.2 Experimental Protocols All experiments were conducted between 124-140 days gestation. A total of 31 experiments were carried out on 18 pregnant sheep. Each animal received one of the following: 1) a 90 min separate maternal and fetal steady-state DPHM infusion with an appropriate washout period in between (n=8, experiments from Yoo et al., 1993). 164 2) a 6 h separate maternal and fetal steady-state DPHM infusion with an appropriate washout period in between (n=3, experiments from study E of chapter 3). 3) a 6 h separate maternal and fetal steady-state [2H10]-DPHM infusion with an appropriate washout period in between (n=2, experiments from study E of chapter 3). 4) a 6 h simultaneous steady-state infusion of DPHM to the mother and [2Hi0]-DPHM to the fetus (n=5, experiments from study A of chapter 3). Drug (DPHM or [2Hi0]-DPHM) was administered to the mother in each experiment as a 20 mg i.v. bolus loading dose over 1.0 min, followed immediately by an infusion at 670 pg/min via the maternal femoral vein. In fetal experiments, a 5.0 mg i.v. bolus loading dose of DPHM or [2H10]-DPHM was given via the fetal lateral tarsal vein over 1.0 min, followed by an infusion of the same compound at 170 pg/min. Maternal and fetal sample collection, processing and storage protocols for all these experiments have been described in Chapter 3 of this thesis and in the earlier paper (Yoo etal., 1993). 4.1.3 Physiological Recording and Monitoring Procedures Fetal blood pH, Po2, Pco2, 02-saturation, and hemoglobin, glucose and lactate concentrations were measured in all these animals by the procedures described in Chapter 2. All of these fetal blood gas and metabolite concentrations have been reported earlier and were within the normal range observed in our and other laboratories at this stage of gestation in fetal sheep (Yoo et al., 1993; Chapter 3). 4.1.4 Protein Binding of DPHM and [2Hi0]-DPHM in Fetal and Maternal Plasma 165 The plasma protein binding/unbound fraction of DPHM (or [2H10]-DPHM) was measured ex vivo in pooled fetal and maternal steady-state plasma samples using an equilibrium dialysis procedure as described by Yoo et al. (1993). Maternal plasma protein binding was measured in plasma samples obtained during maternal drug infusion, whereas fetal plasma protein binding was measured in plasma samples obtained during fetal drug infusion. Maternal and fetal plasma unbound concentrations during fetal and maternal drug infusion, respectively, were below the limit of quantitation in many animals and were not used for the sake of uniformity. 4.1.5 Drug Analysis The concentrations of DPHM in all biological fluids collected were measured using either a gas chromatographic-nitrogen phosphorus detection method (Yoo et al., 1986b; studies of Yoo et al., 1993) or by a GC-MS assay (studies in Chapter 3) capable of measuring both DPHM and [2Hi0]-DPHM simultaneously (Tonn et al., 1993). Both these assays have been shown to be comparable to each other with a similar limit of quantitation (2.0 ng/ml; Tonn et al., 1993). 4.1.6 Pharmacokinetic Analysis The maternal and fetal steady-state arterial plasma DPHM and [2Hi0]-DPHM concentration data were treated according to a 2-compartment open model in order to calculate the placental and non-placental clearances of DPHM (or [2Hi0]-DPHM when present) in the ewe and the fetus, as described in Chapter 3. 166 4.1.7 Statistical Analysis All values are reported as mean ± S.D. The significance level was p<0.05 in all cases. Fetal weight in utero at the time of experimentation was estimated from the weight at birth and the time interval between the experiment and birth (Koong et al., 1975). The clearance values obtained in the earlier study of Yoo et al., (1993) were also recalculated in order to obtain weight-normalized estimates with respect to the estimated fetal weight by the method of Koong et al. (1975). 4.2 Results The average maternal body weight was 76.9 ± 12.6 kg and the estimated fetal body weights on the day of maternal and fetal DPHM (or [2H10]-DPHM) infusion were 2.61 ± 0.61 and 2.56 ± 0.54 kg, respectively. Table 4.1 presents the gestational age of the animals on the day of experiment, maternal and fetal steady-state plasma unbound fractions of the drug, and maternal and fetal clearance (total body, placental and non-placental clearances) data calculated using the 2-compartment pharmacokinetic model. The mean gestational age on the day of maternal and fetal steady-state DPHM infusion experiments was 130.9 ± 4 . 1 and 130.4 ± 3.7 days, respectively, and these were not statistically different (paired t-test, p > 0.05). The average maternal and fetal steady-state plasma drug concentrations in these animals after maternal administration ( C m and Cf, 167 respectively) were 228.0 ± 56.1 (range 140.3 - 360.3) and 43.1 ± 31.2 (range 3.5 - 124.1) ng/ml, respectively, whereas those after fetal drug infusion were 35.3 ± 1 1 . 9 (Cm': range 17.9 - 66.3) and 331.1 ± 172.4 (C/: range 132.5 - 697.9) ng/ml, respectively. The steady-state maternal and fetal plasma unbound fractions of the drug were 0.120 ± 0.069 (range 0.032 - 0.293) and 0.301 ± 0.094 (range 0.165 - 0.527), respectively. The average maternal plasma unbound fraction (M-UF) was significantly lower compared to the average fetal plasma unbound fraction (F-UF, unpaired t-test, p < 0.0001). Maternal and fetal steady-state unbound plasma drug concentrations were calculated by multiplying the appropriate total plasma concentration with the corresponding plasma unbound fraction. The mean steady-state unbound plasma concentrations thus obtained were: C m = 25.1 ± 11.4 (range 8.6 - 45.6) ng/ml; C f = 12.0 ± 8.6 (range 1.9 - 40.4) ng/ml; C m ' = 3.9 ± 1.8 (range 0.9 - 7.2) ng/ml; and C f' = 89.3 ± 32.0 (range 46.1 - 166.2) ng/ml. All fetal weight-normalized clearances (total body, placental and non-placental clearance) were significantly higher compared to the corresponding maternal clearance parameters (unpaired t-test, p < 0.0001 in all cases), as reported previously (Yoo et al., 1993; Chapter 3). However, the contribution of CL f 0 to CLff (39.5 ± 10.7%) was significantly lower compared to that of C L m o to C L m m (96.3 ± 2.8 %) (unpaired t-test, p < 0.0001). 4.2.1 Relationships of Maternal and Fetal DPHM Clearances with Gestational Age Figure 4.1 depicts the alterations in maternal and fetal clearances with advancing gestation over the 2-week period during which our experiments were conducted. Fetal total body (CLff) and placental (CLfm) clearances exhibit a highly significant negative linear relationship with gestational age (Figures 4.1 A and 4.1B). The calculated regression 168 equations predict a fall of -59% (from 374.3 to 153.2 ml/min/kg) and -66% (from 247.0 to 83.1 ml/min/kg) for CLff and CLfm, respectively, from 125 to 136 d gestation. Although the CLf0 parameter also exhibits a decreasing trend with gestation, this relationship is only near statistical significance (Figure 4.1C). Also, the percent contribution of C L f 0 to CLff did not change as a function of gestational age (%[CLf0/CLff] vs. GA, r = 0.2892, p > 0.2, data not shown). In contrast to fetal clearances, none of the maternal clearance parameters show any relationship with gestational age ( C L m m vs. GA, r = 0.0113, p > 0.9; CL m f vs. GA, r = -0.0021, p > 0.9; C L m o vs. GA, r = -0.0041, p > 0.9; data not shown). 4.2.2 Plasma Protein Binding Effects on Maternal and Fetal DPHM Clearances Figure 4.2 depicts the underlying relationships between maternal and fetal steady-state plasma unbound fraction and the corresponding clearance parameters. The CLff and C L f m are highly correlated with F-UF (Figures 4.2A and 4.2B). In contrast to CLfm , there was no relationship between CL m f and M-UF (Figure 4.2E). Also, C L f m and C L m f were not significantly related to the extent of drug protein binding on the other side of the placenta, i.e., maternal and fetal plasma protein binding, respectively (data not shown). Although the linear relationships between M-UF and C L m o , and between F-UF and C L f o were statistically significant (M-UF vs. C L m o : C L m o = 107.7*M-UF + 28.8, r = 0.7649, p < 0.0005; F-UF vs. C L f 0 : C L f 0 = 234.7*F-UF + 31.6, r = 0.4749, p < 0.05), a better fit of the data (M-UF vs. C L m o , r = 0.8090; F-UF vs. C L f 0 , r = 0.4678; Figures 2C and 2F) was obtained using an equation of the form (see Discussion of this chapter for further details on this): 169 E _ l o co C M M- CD CD T -C D o i CM O C O C D CN LO CO CO I — C O C O 00 L O C D O T - ^ - C M C D C O T -O O O I ^ C O C O C O C D C M C M t - t - C M C M t - t -L O CD r— L O d I S - C N CN CO L O t- co T -IS-d co o co O CD LO CM CO 00 C O C O T -O O C O L O CM C O -^ r C O C M T - C O d co d co I S - co L O C M O C O I S - O M; cq T - C D LO CM CO s i n ro m LO i - O Is- LO CO 00 1 - CO C O C M o M-CD S 2 S = i5 E a> :=• o E _i o C M ^ O l C M f f l f f l C O f f l f f l n O ^ ^ C O L O C O CM C O O CM CM O CO 00 CM LO CM _ T - C O C O C M C O M - C O C M C M C M CM O) CO CD CO Is-LO 05 O \T CO M-L O C M C D M " L O I S - C M L O T - L O C O C O * C M o C O C O C M 00 E _ l o ^ j . CM CO CD OO K oo ^ ^ 1 T - C M L O C O M" O CO CM T -d co cd L O oo M- M- Is- CM M" C O C 0 t - T ~ : c M C D ( D C O ^ N c d S ^ ^ f - c v i I CO |LO d N" CO C O C M C N 0 0 C D O ) C O L O T - ' ^ - C 0 L O I S - C M C D L O C N I— CO I— LO M" OO M " CM ^ n n m 4 4 L O co C D Is-L O C O C O C O o M- CD CO O CO LO LO CM CO CO CM M- LO C M d M-M-d CMI s -C7)00 0 0 C D C 0 0 0 0 0 L 0 O 0 0 r s - l s - C D C D L 0 C D I I s -a J c o i ^ ' i ^ d d d d d ^ d c o d c o T ^ M - L O i ^ C M f c o c o i n ^ t m n ^ c o t c o ^ o o c M ^ m | M -co d T3 c 3 ° C => o CO J -E £ w Q. W n v ( M t D t N t O ( D C O W ^ C D t D N ( M S N -r- TJ-C D I s - O C M C M C D O C O C D C M L O a ) 0 ) C M ' ? f C D C M I s - O O ) T - ; - r - c O C M C O C O M C M C M M - C M - K - C M C O C N C N L O C O C O O d d d d d d d d d d d d d d d d d d CO £ l_ +* co 5 00 o O T " C M Is-O N 0 0 N t 0 n ( D ( M 0 ) T - O 0 0 C O L O C D C D C O 0 0 C O C D L O O O O O d t- C N O O O 0 0 0 0 0 0 LO O CN CD T - O CD M" CO CO O T - CM CD O O t - t - O t - C M T - O CD CD CD CD CD CD CD Ci O) < O TO +* in a> c LL *-C C D C D C D L O M - O O C D C O h - C O t - L O O L O T - C D CO C O C O C N C O C O C O C O C M 0 0 C N C M C O C N C N 0 O 0 0 C N CM d ^ C O C O CO a) • 5 c M - M - C O C O C M C M O O I ^ O O t ^ O O ^ i - L O L O O M - T - T -C O C O C M C O C O C O C M C M C O C N C N C O C N C O M - C M C O C O C D O C O Q) g LU T - l O O f O O O C N M - O i ; C M C M r O C O C O O O O O _ r r r r r N N T f N Is- T~ T-C M N e O t O N N . . C M T — C M C M C M ^ " T t C M _ * * 3 o CO c CO CO a ai 170 I s-< * CN to o LO CN LO ai o CN + < * oo TT' O o o T]-00 CO o d II v co cb oo oo CN + < o CN O 00 CO LO CO CM CO cx> CM C D C N O O CN TT CM 00 o CD o d d o LO o o o LO (6>|/U!W/|W) * io V i _ CL CQ o o CO o o CM o o (6>|/U!UJ/|LU) ^ H O T f JO LO CD o CD o O d II v i _ Q. O O CO O O LO o o Tl-CD o CO o o CN o o CN CO C/> CD TD, CU CO < ro c o ro co CL) O 00 CO LO — * CO w ^ ro T3 CD CO < ro c CN ^ T— -4-* CO CD o CD CM 00 CO LO CO CM CO CO CN CD CM CO > * ro 0 CO < ro c o ro -*—< CO CD $ CD C —I T3 _ O co ^ O CD CJ 8 LL. £ O LO _ i . co O £ CD ... -C CD CQ . - ~ o < o _ CD ^  co s § § < co iS (0 CD ^ CD LL. CD CD E CD CD CO JZ TO |— CO Q_ .2 CD CO CO O CD" o c CO s CO JD O w CD c ° CO CO — TJ ^ iS 5? o co 8? c I, J ro o co _ TJ _J o Q g c 2> CD .b -Q O co °-C L 1-= CD CO o CO c CO o JO CO CD I ro dj „ o = CD 5 •K CO O CD S o CO CO CO co • £ (6>)/U!UJ/|Lij) uio CD < ct: CD CD L Z! CO L L CD o CO .E Q. 171 o GO i n + CD CO •<- oo co r>~ d II o o o CM I o CD O LO O T f o CO d CN O O CD " O c O X ) c Z) CD E a) to a. ro a5 (B>)/U!LU/|LU) °'-|o uT Z> L L * CO Tf CO "2 1 0 JN S CD CO g 11 Is: 5 E o ° O " Q. CQ CD d LO d T f c i CO c i CN ci o o o o o o o o o o LO TT CO CN T-(6>|/u!iu/|w) " ^ Q CD d LO ci T f O CO c i CNI c i o o o o o o o o o o o o CD LO T f CO CN T-(6>|/U!LU/|LU) c o o CO T3 cz Zl o JD c Z) CO E CO JD 0_ 2 U-c o o CO • D cz ZJ o jQ cz => tf) CO I— C N C N CO CZ c i .2 CO 00 o o o CZ zs o .Q c Z> CO E to ro 0. "co cz I £ ro o O CD O LO O T f O CO O CN (6>)/U!LU/|Uj) ° w 1 0 co ™ t LU CD c i CO O O O O LO O CN O CD O CO O CO (6>|/U!W/|w) ""no 0 0 o o CD o LO o T f o CO o CN (6>|/U!iu/|w) m u j - ] 0 o CO T f CN TJ c ZJ o c Z> CO E in co CL "co cz s Q) -<—1 CD CN CO CZ O .2 ts CO T f £ CN -a ° i o X ! d ro co _co a. d ro cz i CD -*-« o to o 2 + 3 CD XI o § 5 I E 5 CD _ a) o (T-5 — CD CO c o -*—' o CD T J T J C C Z J CO o . -JD LL. §=> CO ^ E co CL E LTJ co 3 CD ' O O CD T J c CO CO CD o c co 1 ro JD o c o ^ O TJ co c «i= CO T J LL c . -Z J cz o o _ Q X 3 c o Z J co E ? co o C L - Q cz C O Z J I s o CL CO CD O O P CD O CD O sz co CO CD o T t JD CO !— c LL 5= CZ T J , • CO CD SJ> £ co cz <? 5 ° Q_ Q "co c L— "CO LL T J ^ CO J= TO O "CD CQ c . . CD LL 1 2 CD -Q co CO "^ C L Kr CO O C L CD TJ CO CD ro — ' m SZ viv CO T J CD c o j= o E 'sz co +z; co CO T J CD "m CD o c t cz ~ CO E ^ CO —I = CD O | Q C CD > CD £ fD _ J C J ) g *r O _ " I " L Z CD CD ' N jz c . .9 < CD rf 0- T J Z ) CO CO i CN T t CD i _ Z J C O 172 P1 * X P2 + — x P2 where, P1 and P2 are the parameters required to describe the relationship. This equation is exactly analogous to the following relationship describing hepatic uptake of compounds based on the well-stirred model of the liver (Wilkinson and Shand, 1975): QH * fub * C L " N T C L H = !5L (2) Q H + fub*CL4t V ' where, Q H , CL H , CL u i n t and fub are the total hepatic blood flow, total hepatic clearance, hepatic intrinsic clearance of the unbound drug and the unbound fraction of the drug, respectively. In our previous studies, we have shown that the liver (liver and gut in case of the mother) is the major organ of non-placental clearance both in the mother as well as the fetus (see Chapter 3). Thus, in the fetus, the terms P1, P2 and P1/P2 of equation (1) correspond to Q H * C L U W , Q H , and CL u i n l , respectively. In the mother, these parameters will describe the combined blood flow and intrinsic clearance of the liver and gut. This analysis produced fits that were statistically at least as good (F-UF vs. CLf0; F-test on sum of squared residuals, p > 0.05) or significantly better (M-UF vs. C L m o , F-test on sum of squared residuals, p < 0.05) compared to the corresponding linear model fits. From the fitting of C L m o vs. maternal plasma unbound fraction data to equation (1), DPHM CL u i n t and Q H in pregnant adult sheep were estimated to be 1242.6 ± 176.8 ml/min/kg (mean ± standard error of estimate) and 60.2 ± 5.7 ml/min/kg, respectively. A similar hyperbolic relationship was apparent between C L m m and M-UF as well (Figure 4.2D). Similarly, the treatment of 173 F-UF and CL f 0 data according to equation (1) produced the fetal C L u i n t and Q H estimates of 517.6 ± 251.2 ml/min/kg (mean ± standard error of estimate) and 318.7 ± 306.6 ml/min/kg, respectively. 4.2.3 Changes in Maternal and Fetal Plasma Protein Binding with Gestational Age Figure 4.3 shows the changes in fetal and maternal steady-state plasma unbound fractions with increasing gestational age during the last two week period of gestation under study. F-UF falls significantly from 125 to 136 d gestation by -46.9% (from 0.383 to 0.203 as predicted by the regression equation, Figure 4.3A). In contrast, the M-UF does not change significantly during this period of gestation (Figure 4.3B). 4.2.4 Inter-relationships between Maternal and Fetal Plasma DPHM Concentrations, Unbound Fractions and the 2-Compartment Model Clearance Estimates Maternal plasma drug concentration (Cm) after maternal drug infusion exhibited a highly significant negative linear relationship with M-UF of the drug (Figure 4.4A). In contrast, fetal plasma concentration after maternal infusion (Cf) was not related significantly to either maternal (r = 0.1751, p = 0.5) or fetal (r = -0.3676 , p > 0.1) plasma unbound fraction (data not shown). Analogous to the maternal situation, the fetal plasma DPHM concentration after fetal drug infusion (Cf') was inversely related to F-UF (Figure 4.4B). Also, maternal plasma concentration after fetal infusion (Cm') was not related to F-UF (r = -0.0966, p > 0.5, data not shown) and its negative relationship with M-UF was only near statistical significance (r = 0.4612, p = 0.05, data not shown). 174 c o o CO TD c o J Q c CO E to ro D_ "co CO 0.6 0.5 0.4 -0.3 -0.2 0.1 F-UF = -0.0163*(GA)+2.42 r = 0.6427 p < 0.005 • • • 124 126 128 130 132 134 136 138 Gestational Age (days) c O o CO TD c Zt O .Q C CO E CO JS D. " r o c CU ro 0.35 - i 0.28 0.21 -A 0.14 0.07 0.00 B r = 0.1612 p > 0.5 124 128 132 136 140 Gestational Age (days) Figure 4.3 - Alterations in (A) fetal and (B) maternal steady-state plasma unbound fraction of DPHM (or [2H10]-DPHM) with increasing gestational age. Actual experimental data (scatter points), regression line (solid) and 95% confidence interval (dotted) are depicted. 175 0.1 0.2 0.3 0.4 0.5 0.6 Fetal Plasma Unbound Fraction Figure 4.4 - Relationships between (A) steady-state unbound fraction and plasma concentration of the drug in the mother after maternal drug administration and (B) steady-state unbound fraction and plasma concentration of the drug in the fetus after fetal drug administration. Scatter points are the experimental data in different sheep. The regression line (solid) and the 95% confidence interval (dotted) are also shown. 176 In order to determine the influence of various 2-compartment clearance terms on maternal and fetal plasma drug concentrations after maternal or fetal drug administration, different concentration vs. clearance relationships were analyzed according to the simple steady-state clearance model of the form: CL=ko/Css- The majority of the inter-animal variability in C m was reflected in the estimated value of total C L m o (not weight-normalized, because the total clearance is the actual determinant of plasma concentrations) as demonstrated by an excellent fit of the concentration vs. clearance data to this model (Figure 4.5A). However, C m was not significantly related to