@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Pharmaceutical Sciences, Faculty of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Gordon, John David"@en ; dcterms:issued "2009-04-17T22:53:50Z"@en, "1995"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """The purpose of these studies was to assess the pharmacokinetics, metabolism and fetal effects of valproic acid (VPA) in pregnant sheep. Following i.v. bolus administration, the disposition of VPA was best described by a biexponential function with a t½β of 2.13 ± 0.49 h in maternal serum and 3.37 ± 1.37 h in fetal serum. The total body clearances (CLT) for the ewe and fetus were 0.083 ± 0.027 Lh⁻¹ kg⁻¹ and 0.529 ± 0.306 Lh⁻¹ kg⁻¹ , respectively. Significant fetal exposure to VPA was observed following maternal dosing (AUC∞FA / AUC∞¹¹MA of 0.410 ± 0.118). VPA appeared rapidly in amniotic and fetal tracheal fluids but did not accumulate extensively in either. 11 VPA metabolites were detected following i.v. bolus dosing. The metabolite levels in maternal serum were consistently higher than fetal levels irrespective of the site of drug administration. Simultaneous i.v. infusion studies were also performed achieving steadystate concentration ratios of VPA and [¹³C₄]VPA in fetal serum to that in maternal serum (i.e., Cfss / Cmss) of 0.55 ± 0.11 and 1.90 ± 0.38, respectively. Maternal (CLm) and fetal (CLf) total body clearances were 3.6 ± 0.9 mL min⁻¹ kg⁻¹ maternal weight and 63.0 ± 22.8 mL min⁻¹ kg⁻¹ fetal weight. Of these values, the maternal nonplacental clearance contributed approximately 69 % of the CLm while the fetal nonplacental clearance (CLfo) apparently made no contribution to CLf. The maternal renal clearance, 0.521 ± 0.071 mL min⁻¹ kg⁻¹ maternal weight, accounts for approximately 20 % of the total maternal nonplacental clearance (CLmo). In contrast, there was essentially no renal excretion of VPA from the fetus. Metabolite data suggests that [¹³C₄]VPA undergoes transfer to the ewe followed by metabolism, with the resulting labelled metabolites being transferred back to the fetus. This data further supports the apparent lack of fetal nonplacental clearance of VPA. The most significant physiological effect of the administration of VPA was a decrease in fetal breathing activity although other transient cardiovascular effects were noted in the infusion studies."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/7367?expand=metadata"@en ; dcterms:extent "3213070 bytes"@en ; dc:format "application/pdf"@en ; skos:note "PHARMACOKINETICS, METABOLISM, PLACENTAL TRANSFER AND FETALEFFECTS OF VALPROIC ACID IN PREGNANT SHEEPbyJOHN DAVID GORDONB.Sc. (Hon.), Daihousie University, 1988M.Sc., The University of British Columbia, 1991A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Faculty of Pharmaceutical Sciences)(Division of Pharmaceutics and Biopharmaceutics)We accept this thesis as conformingto the required standard.THE UNIVERSITY OF BRITISH COLUMBIAJuly, 1995© John David Gordon, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis br scholarly purposes may be ‘granted by the head of mydepartment or by his or her representatives. it is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)_______________________Department of ?ri ( SeicecThe University of British ColumbiaVancouver, Canada—cTh i QC)Date sJ j.. I Q( t IDE.6 (2/88)IIABSTRACTThe purpose of these studies was to assess the pharmacokinetics,metabolism and fetal effects of vaiproic acid (VPA) in pregnant sheep. Followingi.v. bolus administration, the disposition of VPA was best described by abiexponential function with at112 of 2.13 ± 0.49 h in maternal serum and 3.37 ±1.37 h in fetal serum. The total body clearances (CLT) for the ewe and fetus were0.083 ± 0.027 Lh1kg1 and 0.529 ± 0.306 Lh1kg-1, respectively. Significantfetal exposure to VPA was observed following maternal dosing (AUCOOFA /AUCOOMA of 0.410 ± 0.118). VPA appeared rapidly in amniotic and fetal trachealfluids but did not accumulate extensively in either. 11 VPA metabolites weredetected following i.v. bolus dosing. The metabolite levels in maternal serumwere consistently higher than fetal levels irrespective of the site of drugadministration.Simultaneous i.v. infusion studies were also performed achieving steady-state concentration ratios of VPA and [13C4]VPA in fetal serum to that in maternalserum (i.e., Cfss / Cmss) of 0.55 ± 0.11 and 1.90 ± 0.38, respectively. Maternal(CLm) and fetal (CLf) total body clearances were 3.6 ± 0.9 mL mirn1 kg1maternal weight and 63.0 ± 22.8 mL mirn1kg-1 fetal weight. Of these values, thematernal nonplacental clearance contributed approximately 69 % of the CLmwhile the fetal nonplacental clearance (CLf0) apparently made no contribution toCLf. The maternal renal clearance, 0.521 ± 0.071 mL mirn1 kg1 maternal weight,accounts for approximately 20 % of the total maternal nonplacental clearance(CLmo). In contrast, there was essentially no renal excretion of VPA from thefetus.Metabolite data suggests that [13C4]VPA undergoes transfer to the ewefollowed by metabolism, with the resulting labelled metabolites being transferredback to the fetus. This data further supports the apparent lack of fetalnonplacental clearance of VPA.The most significant physiological effect of the administration of VPA was adecrease in fetal breathing activity although other transient cardiovascular effectswere noted in the infusion studies.ivTABLE OF CONTENTSAbstractTable of contents ivList of Figures viiiList of Tables xvList of Abbreviations xxAcknowledgments xxivDedication xxv1. INTRODUCTION 11.1 Valproic acid 11.1.1 Clinical use of valproicacid 11.1.2 Adverse effects of VPA 21.1.3 Valproic acid mechanisms of action 51.1.4 Pharmacokinetics and metabolism of VPA 91.1.4.1 Absorption 91.1.4.2 Distribution 101.1.4.3 Metabolism and elimination 111.1.5 Methods of analysis of VPA and its metabolites 151.1.6 Drugs, epilepsy and pregnancy 171.2 Fetal drug exposure 191.3 Rationale and specific objectives 261.3.1 Rationale 261 .3.2 Objectives 282. EXPERIMENTAL 292.1 Materials 292.2 Preparation of stock solutions 31V2.3 Instrumentation and equipment 332.3.1 Gas chromatography with mass selective detection 332.3.2 Physiological monitoring 332.3.3 General experimental 342.4 Quantitation of VPA and metabolites using GC-MS 342.4.1 GC-MS condition 342.4.2 Preparation of standard samples 352.4.3 Determination of calibration curves 362.4.4 Assessment of GC-MS assay in sheep biofluids 362.4.5 Analysis of experimental samples 372.4.6 Protein binding assessment 372.5 Animal and surgical preparation 392.6 Physiological monitoring 412.7 Experimental protocols 422.8 Data analysis 462.8.1 Calculation of in utero fetal weights 462.8.2 Pharmacokinetic analysis - i.v. bolus studies 462.8.3 Calculation of placental and nonplacental clearances 482.8.4 Renal clearances - simultaneous infusion studies 502.8.5 Statistical analysis 503. RESULTS 513.1 Assessment of GC-MS assay in sheep biofluids 513.2 Nonpregnant animal bolus studies 603.2.1 Comparison of parent drug disposition 603.2.2 Comparison of metabolism of parent compounds 623.3 Maternal i.v. bolus studies 66vi3.4 Fetal i.v. bolus studies 743.5 Simultaneous infusion studies 823.5.1 VPA and [13C4]VPA disposition in serum 823.5.2 Placental and nonpiacental drug clearances 893.5.3 VPA and[13C4]VPA disposition in fetal tracheal andamniotic fluids 903.5.4 VPA and[13C4]VPA in maternal and fetal urine 953.5.5 Physiological effects 1003.5.6 VPA and[13C4]VPA metabolism during simultaneousinfusion 1113.5.6.1 VPA metabolites in maternal and fetal serum 1113.5.6.2 VPA metabolites in maternal and fetal urine 1264. DISCUSSION 1284.1 Assessment of GC-MS assay in sheep biofluids 1284.2 Nonpregnant animal i.v. bolus studies 1294.3 Maternal and fetal i.v. bolus studies 1304.3.1 Pharmacokinetics 1314.3.2 Metabolism 1364.3.3 Physiological effects 1374.4 Simultaneous infusion studies 1 394.4.1 Maternal and fetal clearance of VPA from serum 1434.4.2 Maternal and fetal renal clearance of VPA 1474.4.3 Physiological effects 1504.4.3.1 Breathing activity and respiration 1504.4.3.2 Cardiovascular effects 1554.4.3.3 Modelling of physiological effects 1574.4.4 Metabolism 1584.4.4.1 VPA metabolites in maternal and fetal serum 1584.4.4.2 VPA metabolites in maternal and fetal urine 161vii5. SUMMARY AND CONCLUSIONS 1635.1 Quantitation of VPA and metabolites in sheep biofluids 1635.2 Nonpregnant animal bolus studies 1635.3 Maternal i.v. bolus studies 1645.4 Fetal i.v. bolus studies 1655.5 Simultaneous infusion studies 1666. BIBLIOGRAPHY 172v I ILIST OF FIGURESFigure 1. The chemical structure of valproic acid (V PA).Figure 2. Illustration of VPA and sixteen of its most frequentlyobserved metabolites in humans. 13Figure 3. A two-compartment open model describing the dispositionof a drug in the mother and fetus where M represents thematernal compartment, F represents the fetal compartment,CLmf is the maternal to fetal clearance, CLfm is the fetal tomaternal clearance, CLmo is the maternal nonplacentalclearance and CLf0 is the fetal nonplacental clearance. 25Figure 4a. SIM chromatograms of blank sheep serum (lower trace)and tBDMS derivatives of VPA and VPA metabolites inspiked, control sheep serum (upper trace). Continued inFigure 4b. 53Figure 4b. SlM chromatograms of blank sheep serum (lower trace)and tBDMS derivatives of VPA metabolites in spiked,control sheep serum (upper trace). Continued inFigure 4c. 54Figure 4c. SIM chromatograms of blank sheep serum (lower trace)and tBDMS derivatives of VPA metabolites in spiked,control sheep serum (upper trace). Continued inixFigure 4d. 55Figure 4d. SIM chromatograms of blank sheep serum (lower trace)and tBDMS derivatives of VPA metabolites in spiked,control sheep serum (upper trace). 56Figure 5. Representative concentration versus time profiles for[13C4]VPA and unlabelled VPA in serum from anonpregnant ewe following Lv. bolus administrationof 50% VPA / 50% [13C4]VPA (1000 mg dose). 61Figure 6a. SIM chromatograms of tBDMS derivatives of somerepresentative VPA metabolites collected in urine froma nonpregnant ewe following Lv. bolus administration.SIM chromatograms of tBDMS derivatives of someinternal standards are also shown. Continued inFigure 6b. 64Figure 6b. SIM chromatograms of tBDMS derivatives of somerepresentative VPA metabolites collected in urine froma nonpregnant ewe following i.v. bolus administration.SIM chromatograms of tBDMS derivatives of someinternal standards are also shown. 65Figure. 7. Representative VPA concentration versus time profilesfor maternal and fetal serum as well as amniotic andfetal tracheal fluids from the ewe and fetus (E54)xfollowing maternal i.v. administration of VPA(1000 mg dose). 69Figure 8. Representative VPA concentration versus timeprofiles for maternal and fetal serum as well asamniotic and fetal tracheal fluids from the ewe andfetus (E499) following fetal i.v. administration of VPA(250 mg dose). 76Figure 9. Fetal femoral arterial lactate concentrations (mean ± Sd)monitored during fetal bolus dose administration(250 mg) experiments. 80Figure 10. Representative VPA and[13C4]VPA serum concentrationversus time profiles from ewe 2181 during simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus. 84Figure 11. Representative VPA and [13C4]VPA serum concentrationversus time profiles in the fetal femoral and carotidarteries (E2177) during simultaneous infusion of VPAto the ewe and [13C4]VPA to the fetus. 87Figure 12. Representative VPA and[13C4]VPA serum concentrationversus time profiles for the umbilical vein and fetalfemoral artery (E2181) during simultaneous infusionof VPA to the ewe and [13C4]VPA to the fetus. 88xiFigure 13. VPA and [13C4]VPA concentration versus time profilesin amniotic and fetal tracheal fluid from ewe 2177 duringthe simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus. 93Figure 14. Fetal urine flow (mean ± Sd) for six hours before, duringand six hours following the simultaneous infusion of VPAto the ewe and [13C4]VPA to the fetus. 97Figure 15. Fetal femoral arterial oxygen saturation (mean ± sd)monitored before, during and after the simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus. 103Figure 16. Fetal femoral arterial p02 (mean ± Sd) monitored before,during and after the simultaneous infusion of VPA to theewe and [13C4]VPA to the fetus. 104Figure 17. Fetal femoral arterial lactate concentrations (mean ± Sd)monitored before, during and after the simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus. 105Figure 18. Fetal heart rate (mean ± Sd) before and during thesimultaneous infusion of VPA to the ewe and [13C4]VPAto the fetus. 106Figure 19. Fetal arterial pressure (mean ± Sd) before and duringthe simultaneous infusion of VPA to the ewe andXII[13C4]VPA to the fetus. 107Figure 20. Fetal breathing activity (mean ± Sd) for six hours before,during and six hours following the simultaneous infusionof VPA to the ewe and [13C4]VPA to the fetus. 108Figure 21. EOG activity (mean ± sd) for six hours before, duringand six hours following the simultaneous infusion ofVPA to the ewe and [13C4]VPA to the fetus. 109Figure 22. EC0G activity (mean ± sd) for six hours before, duringand six hours following the simultaneous infusion ofVPA to the ewe and [13C4]VPA to the fetus. 110Figure 23. Mean (± sd) serum concentration versus time profilesfor (E)-2-ene VPA and (E)-2-ene[13C4]VPA duringthe simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus. 113Figure 24. Mean (± sd) serum concentration versus time profilesfor (Z)-2-ene VPA and (Z)-2-ene [13C4]VPA duringthe simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus. 114Figure 25. Mean (± sd) serum concentration versus time profilesfor (E)-3-ene VPA and (E)-3-ene[13C4]VPA duringthe simultaneous infusion of VPA to the ewe andxm[13C4]VPA to the fetus. 115Figure 26. Mean (± Sd) serum concentration versus time profiles for(Z)-3-ene VPA and (Z)-3-ene[13C4]VPA during thesimultaneous infusion of VPA to the ewe and [13C4]VPAtothe fetus. 116Figure 27. Mean (± sd) serum concentration versus time profilesfor 4-ene VPA and 4-ene [13C4]VPA during thesimultaneous infusion of VPA to the ewe and [13C4]VPAto the fetus. 117Figure 28. Mean (± sd) serum concentration versus time profile for3-OH VPA during the simultaneous infusion of VPA tothe ewe and [13C4]VPA to the fetus. 118Figure 29. Mean (± sd) serum concentration versus time profiles for4-OH VPA and 4-OH[13C4]VPA during the simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus. 119Figure 30. Mean (± sd) serum concentration versus time profiles for5-OH VPA and 5-OH [13C4]VPA during the simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus. 120Figure 31. Mean (± sd) serum concentration versus time profiles for3-keto VPA and 3-keto [13C4]VPA during thesimultaneous infusion of VPA to the ewe and [13C4]VPAxivto the fetus. 121Figure 32. Mean (± sd) serum concentration versus time profiles for4-keto VPA and 4-keto[13C4]VPA during the simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus. 122Figure 33. Mean (± sd) serum concentration versus time profile for2-PSA during the simultaneous infusion of VPA to theewe and [13C4]VPA to the fetus. 123Figure 34. Mean (± Sd) serum concentration versus time profilesfor 2-PGA and 2- [13C4]PGA during the simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus. 124Figure 35. SIM chromatograms of the tBDMS derivative of(E)-2-ene VPA (mlz = 199; peak 1) and the mlz = 197profile between 16.50 and 19 minutes, the time rangeover which the diunsaturated metabolites of VPA aredetected, from a maternal serum sample collectedduring the simultaneous infusion of VPA and[13C4]VPA to the ewe and fetus, respectively. 125Figure 36. Cumulative recovery of the most abundant metabolites ofVPA in maternal urine, collected during the simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus. 127xvLIST OF TABLESTable 1. Concentration ranges, diagnostic ions and retentiontimes of tBDMS derivatives of VPA and metabolites fromstandard reference samples. 32Table 2. Diagnostic ions and retention times of the derivatizedinternal standard compounds. 38Table 3. Accuracy and precision of assay with respect to analysisof VPA and metabolites extracted from sheep urine. 57Table 4. Accuracy and precision of assay with respect to analysisof VPA and metabolites extracted from sheep serum. 58Table 5. Assessment of inter-assay variability with respect toanalysis of VPA and metabolites extracted from serum. 59Table 6. Percentage (%) of dose excreted as parent drug in urinefollowing i.v. bolus administration of a 1000 mg dose of50% VPA /50% [13C4]VPA to a nonpregnant ewe. 62Table 7. Time-averaged peak area ratios from serum for the mostprominent metabolites detected following i.v. bolusadministration of 50% VPA / 50% [13C4]VPA(1000 mg dose). 63xviTable 8. Ratio of amount of unlabelled VPA metabolite to amountof [13C4]VPA metabolite recovered from urine for themost prominent metabolites following i.v. bolusadministration of 50% VPA / 50% [13C4]VPA(1000 mg dose). 66Table 9. Information for experimental preparations used in thematernal i.v. bolus administration studies. 67Table 10. Pharmacokinetic parameters in the ewe estimated fromthe maternal i.v. bolus administration experiments. 70Table 11. The VPA metabolites detected in maternal serumfollowing maternal Lv. bolus administration of VPAalong with the range of peak concentrations (C max)typically observed in the five animals and the timefollowing drug administration at which peak levelswere observed (t max). 71Table 12. Fetal blood gas, acid-base, metabolic andcardiovascular parameters (mean ± Sd; n=5) priorto maternal bolus administration of VPA. 72Table 13. Fetal breathing movements monitored during maternalbolus dose administration of VPA (1000 mg) experiments. 73Table 14. Information regarding experimental preparations used inxviithe fetal i.v. bolus administration studies. 75Table 15. Pharmacokinetic parameters in the fetus estimated fromfetal i.v. bolus administration experiments. 77Table 16. Fetal blood gas, acid-base, metabolic and cardiovasculareffects related parameters (mean ± s.d.; n=5) prior to fetalbolus administration of VPA. 79Table 17. Fetal breathing movements monitored during fetal bolusdose administration (250 mg) experiments. 81Table 18. Information regarding experimental preparations usedin the simultaneous infusion studies. 83Table 19. Mean steady-state concentrations for VPA and[13C4]VPA in maternal and fetal femoral arterialserum obtained during the simultaneous infusionof VPA to the ewe and [13C4]VPA to the fetus. 85Table 20. Total body, placental and non-placental clearanceparameters calculated from serum steady-state totaldrug concentrations achieved during the simultaneousinfusion of VPA and [13C4]VPA to the ewe and fetus,respectively. 91Table 21. Total body, placental and non-placental clearancexviiiparameters calculated from serum steady-state levelsof unbound VPA and [13C4]VPA achieved during thesimultaneous infusion of VPA and [13C4]VPA to the eweand fetus, respectively. 92Table 22. Mean concentrations of VPA and[13C4]VPA inamniotic and fetal tracheal fluids during the final threehours of the simultaneous infusion of VPA to the eweand [13C4]VPA to the fetus. 94Table 23. Mean pH and volume of urine produced per hour by theewe and fetus during the six hour simultaneous infusionof VPA to the ewe and [13C4]VPA to the fetus. 96Table 24. Summary of the cumulative amounts of VPA and[13C4]VPA recovered from unhydrolyzed (unhydrol.)and hydrolyzed (hydrol.) fetal and maternal urineduring the six hour simultaneous infusion period. 98Table 25. Maternal and fetal renal clearance values based onunconjugated and total (unconjugated + conjugated)VPA and[13C4]VPA levels in maternal and fetal urine,respectively, collected during the six hour simultaneousinfusion period. 99Table 26. Fetal blood gas, acid-base, metabolic andcardiovascular parameters (mean ± s.d.; n=5) prior tothe simultaneous infusion of VPA and [13C4]VPA. 102Table 27. The mean (± s.d.) ratio of metabolite concentration infetal serum relative to maternal serum in the first twohours and final four hours of the simultaneous infusionof VPA to the ewe and [13C4]VPA to the fetus. 112xixxxLIST OF ABBREVIATIONSA exponential equation coefficient (zero time intercept ofthe alpha phase)exponential rate constantAIC Akaike Information CriterionAMN amniotic fluidANOVA analysis of varianceASA acetylsalicylic acidAUG area under the concentration vs. time curveAUMC area under the first moment curveB exponential equation coefficient (zero time intercept ofthe beta phase)exponential rate constantBE base excessbpm beats per minute°C degrees CelsiusCf fetal concentrationCL clearanceCLf fetal total body clearanceCLfm fetal-to-maternal placental clearanceCLf0 fetal nonplacental clearanceCLm maternal total body clearanceCLmf maternal-to-fetal placental clearanceCLmo maternal nonplacental clearanceCL placental clearanceCm maternal concentrationxxiCmjd concentration at the midpoint of the intervalCSF cerebrospinal fluidCNS central nervous systemCV coefficient of variationdL decilitreDPHM diphenhydramineDPMA diphenylmethoxyacetic acid(E)- trans isomerEC0G electroencephalogramEl electron impact ionizationEOG electro-oculargrameV electron voltsexpt experimentFA fetal arteryFBM fetal breathing movementsFUR fetal urineGABA ‘y-aminobutyric acidGABA-T y-aminobutyric acid transaminaseGAD glutamic acid decarboxylaseGC gas chromatographyGFR glomerular filtration rateGHB ‘y-hydroxybutyrateg gramsh hourHb hemoglobinHz hertzID internal diameterxxiii.v. intravenousJ infusion rateL litreLOQ limit of quantitationM molarMA maternal arteryMCP metoclopramidemeq milliequivalentsmg milligramsmicrogramsmm minutemL millilitremicrolitremM millimolarMRT mean residence timeMS mass spectroscopyMSD mass selective detectionMUR maternal urinem/z mass to charge ration number of samples or subjectsng nanogramsNICI negative ion chemical ionizationCD outer diameterPAH para-aminohippuric acidPFB pentafluorobenzyl bromidepsi pounds per square inchPTFE polytetrafluoroethylenexxiiiQum umbilical venous flowREM rapid eye movementss.d. standard deviationSIM selected ion monitoringSRF sustained repetitive firingss (as a suffix) steady stateSSA succinic semialdehydeSSADH succinic semialdehyde dehydrogenaseSSAR succinic semialdehyde reductasetBDMS tert.-.butyldimethylsilylt112 biological half-lifeTMS trimethylsilylTR fetal tracheal fluidVd volume of distributionVPA valproic acidX amount in urine(Z)- cis isomerxxivAcknowledgmentsI would like to sincerely thank my supervisors Dr. Wayne Riggs and Dr.Frank Abbott for their guidance, patience, friendship and support over the courseof my Ph.D. training. I would like to thank Dr. Dan Rurak for his invaluablesupport and assistance during my studies. Thanks also to my committeemembers, Dr. Jack Diamond, Dr. Keith McErlane and Dr. Jim Orr for their interestand guidance in this work.I would like to express my appreciation to Mr. Eddie Kwan and Ms.Caroline Hall for their exceptional assistance with the experimental and technicalaspects of the sheep studies. I would also like to thank Dr. George Tonn for hishelp in conducting experiments as well as for his continued friendship andsupport.Many thanks to my colleagues in the laboratory, Dr. Andras Szeitz, Mr.Rajesh Krishna, Mr. John Kim, Mr. Ahmad Doroudian, Mr. Sanjeev Kumar, Dr.Anthony Borel, Mr. Wei Tang and Dr. Jan Palaty for their friendship andassistance. Thanks also to my colleagues at the Research Centre, Ms. KathleenStobbs and Mr. Weiping Tan for lending a helping hand when needed.My sincere thanks to Ms. Karen Brodeur and Mr. Jamie Eyers as well asMs. Laurie Chesworth and Mr. Chris Lalonde for providing me with homes awayfrom home during the latter portion of my studies in Vancouver.This work would not have been possible without the love, understandingand patience of my wife, Dr. Darlene Brodeur and my children, Jacob andSamuel. Thank you. I would also like to thank Darlene for her assistance withthe statistical analysis of this work. Thanks also to my parents, Barbara Gordonand Donald Gordon for their support throughout the course of my graduate work.Financial support from the Pharmaceutical Manufacturer’s Association ofCanada (PMAC), the Medical Research Council of Canada (MRC) and theUniversity of British Columbia is gratefully appreciated. This project wassupported by a Medical Research Council of Canada Program Grant.DedicationThis thesis is dedicated to my wife, Darlene, and to my children, Jacob andSamuel, for their love, support and patience.xxv11. INTRODUCTION1.1 Valproic acidSince the first discovery of it’s anticonvulsant activity in 1963 (Meunier eta!., 1963), valproic acid (2-propylpentanoic acid; Depakene®; VPA) has becomewidely used in the treatment of several types of epileptic disorders (Chadwick,1987; Wilder, 1987; Davis etaL, 1994). VPA is a branched-chain fatty acid (Fig. 1)and is structurally unrelated to any other antiepileptic drug. It is available forclinical use as the parent compound, its sodium salt, its amide derivative and as acombination of the parent compound and its sodium salt (Davis et al., 1994).COCHFigure 1. The chemical structure of vaiproic acid (VPA).1.1.1 Clinical Use of valproic acidAs mentioned above, VPA has been demonstrated to be effective in thetreatment of several types of epileptic seizures including generalized convulsive(e.g., tonic-clonic and myoclonic seizures) and non-convulsive seizures (e.g.,absence seizures), several types of partial seizures (including simple, complex andsecondarily generalized), and compound/combination seizures including those thatare refractory to other anticonvulsant compounds (Davis et aL, 1994). The dose ofVPA required to achieve therapeutic concentrations differs markedly between2patients on VPA monotherapy and those on combination therapy because of avariety of pronounced pharmacokinetic interactions. Moreover, the required dose(and concentration) of VPA can be a function of the type of seizure being treated(Lundberg etah, 1982). Patients on monotherapy usually require doses between10 and 20 mg kg1day- in order to achieve an adequate clinical response(Bourgeois, 1989). Higher doses are usually required in order to achieve similardrug levels in patients on polytherapy, with 30 to 60 mg kg-1day-1 being typical.The dosing of VPA is further complicated by a nonlinear relationship between VPAdose and concentration, age-dependent kinetics and the effects of other diseasestates on the kinetics of the drug (Bourgeois, 1989).1.1.2 Adverse effects of VPAThe adverse effects of VPA can be divided into dose-related side effects,non dose-related metabolic side effects and pregnancy-related teratogenic effects.The effects falling in the first of these classes are usually not life threatening(except in the case of acute overdose), however, the more idiosyncratic metaboliceffects can be fatal. Pregnancy-related effects such as morphologicalteratogenicity occur following a normal therapeutic and non-toxic dose to themother which results in fetal exposure to VPA.Dose-related side effectsThis class of side effects includes problems such as gastrointestinaltroubles (e.g., nausea, vomiting and anorexia), excessive weight gain, skin rash,hair changes, and CNS effects such as sedation and tremor (Dreifuss, 1989).Generally these effects occur at higher doses of VPA and will usually subside3when the dose is lowered or, in the case of sedation, after long-term continued useof VPA.Metabolic side effects: HepatotoxicityThe most significant metabolic side effect of VPA therapy is a rare but fatalhepatotoxicity. Fulminating hepatotoxicity associated with VPA use usually occurswithin the first four to six months of therapy, with children under the age of 10years representing 73 % of all fatalities (Dreifuss eta!., 1987). At highest risk werepatients less than two years of age on multiple-anticonvulsant therapy.Although the exact mechanism of VPA hepatotoxicity is not known,similarities to Reye’s syndrome, hypoglycin toxicity and 4-pentenoic acid toxicityhave been noted (Kesterson eta!., 1984). Both 4-pentenoic acid and a metaboliteof hypoglycin, methylenecyclopropylacetic acid, are known steatogenic agents(Zimmerman and Ishak, 1982). Two metabolites of VPA, 2-n-propyl-4-pentenoicacid (4-ene VPA) and 2-n-propyl-2,4-pentadienoic acid (2,4-diene VPA), arestructurally similar to 4-pentenoic acid and methylenecyclopropylacetic acid. GCMS analysis of urine and plasma from patients which have succumbed to VPAinduced hepatotoxicity has detected abnormally high levels of both 4-ene VPA and(E)-2,4-diene VPA as well as decreased levels of 2-n-propyl-3-hydroxypentanoicacid (3-OH VPA) and 2-n-propyl-3-oxopentanoic acid (3-keto VPA) (Scheffner eta!., 1988). There is evidence, such as mitochondrial damage and a decrease inthe 3-oxidation products (i.e., 3-keto VPA and 3-OH VPA) in humans, thatindicates that the 3-oxidation pathway is being inhibited in these cases. Aninhibition of f3-oxidation will result in a shunting of VPA metabolism to otherpathways which may explain the higher than normal levels of 4-ene VPA and 2,4-diene VPA in the urine and plasma of hepatotoxic patients. Thus, the disruption orinterference of fatty acid p-oxidation could be a prelude to VPA hepatotoxicity4which is possibly mediated through a reactive intermediate commonly generatedby VPA metabolism.Pregnancy-related side effects: Teratogenic effectsThe teratogenic effects of VPA include malformations such as facialanomalies of the nose, lips, palate, eyes and ears, digital anomalies such aspolydactylism and hypoplasia of the phalanges and nails, congenital heart defectssuch as ductus arteriosus and ventricular septal defects, and neural tube detectssuch as lumbosacral meningocele and spina bifida (Nau and Loscher, 1982;Kaneko et a!., 1988). Approximately 1 % of fetuses exposed to VPA in uterodevelop spina bifida (Nau and Loscher, 1984). The first trimester is thought to bethe period during which there is the most significant risk for exposure-relatedteratogenesis. Unlike the hepatotoxic side effects of VPA therapy, the teratogeniceffects appear to be related to VPA itself, perhaps in a dose-dependent fashion,and not the products of VPA metabolism (Dreifuss, 1989). Unfortunately, thealternatives are either to use other anticonvulsants, which also have teratogeniceffects associated with them, or to discontinue drug use and risk seizures.Concern about the possible behavioral teratogenicity of VPA has arisen asinformation has accumulated concerning impairments in learning, abnormal activityand delayed behavioral maturation patterns after prenatal exposure (Janz, 1982;Chapman and Cutler, 1984; DiLiberti eta!., 1984; Jager-Roman eta!., 1986;Ardinger eta!., 1988; Christianson eta!., 1994). Neurological manifestations suchas excitability, abnormal tone, feeding problems, sedation and seizures have beenreported in neonates (Yerby et aL, 1992). Exposure to valproate in utero is alsoreported to have caused significant changes in social investigation and exploratoryactivity after birth in animal models (Koch eta!., 1985).51.1.3 Valproic acid mechanisms of actionThe therapeutic diversity of VPA, in conjunction with the variety ofneurochemical and neurophysiological mechanisms that underlie different seizuretypes, suggest that VPA may have several mechanisms of action. Althoughseveral possible mechanisms of action have been proposed for VPA, the mostlikely involve a potentiation of brain synaptic y-aminobutyric acid (GABA) (Loscher,1993). GABA is a neurotransmitter that inhibits the firing of an assortment ofneurons by binding to the GABAA receptor, thereby increasing the permeability ofpostsynaptic membranes to chloride ions. This causes the membrane to becomehyperpolarized, thus increasing the threshold for the initiation of an actionpotential. The proposed modes of VPA’s influence on GABAergicneu rotransmission are as follows:1) GABA levels may be elevated via an activation of glutamic aciddecarboxylase (GAD). Experiments carried out in rats (Godin et aL, 1969) andmice (Taberner et aL, 1980) have indicated that an increase in the production ofGABA occurs following VPA administration. Furthermore, increases in the activityof GAD, the enzyme responsible for the decarboxylation of glutamate required forthe synthesis of GABA, have been demonstrated ex vivo in mice (Loscher, 1981 b)and rats (Phillips and Fowler, 1982) following VPA administration. The timecourse of GAD activation has been found to match that of the GABA increase andthe anticonvulsant effect (Nau and Loscher, 1982). These studies strongly supportthe notion that GAD activation is one of VPA’s modes of GABAergic action.Recent studies on GABA turnover in rat brain have indicated that the mostsignificant increase in GABA synthesis by VPA occurs in the substantia nigra(Loscher, 1989).62) GABA levels may be increased through a reduction in synaptosomalGABA-transaminase (GABA-T) activity in discrete brain areas such as thesubstantia nigra. The initial step in GABA degradation involves itstransamination to succinic semialdehyde (SSA) in nerve terminals, glia cells andpostsynaptic neurons by GABA-T. Following the realization that VPA appears toelevate GABA levels, several groups reported an in vitro inhibition of GABA-T byVPA (e.g., Godin eta!., 1969; Maitre eta!., 1978; Larsson eta!., 1986). However,these inhibitory effects were only observed at very high concentrations of VPA,levels not seen in vivo. This confound has since been rationalized by assumingthat presynaptic (nerve terminal) GABA-T is different from glial GABA-T in terms ofsusceptibility to VPA (Loscher, 1993). Since glial GABA-T predominates in wholetissue homogenates and is less susceptible to VPA, higher concentrations of VPAmust be used in whole tissue homogenates in order to produce an effectcomparable to that observed in synaptosomes. Inhibition of nerve terminal GABAT could explain the increase of presynaptic GABA levels by VPA, although theobserved reductions of synaptosomal GABA-T activity are not large (Loscher,1981b; Loscher, 1993).3) GABA levels may be increased through inhibition of succinic semialdehyde dehydrogenase (SSADH) activity. Succinic semialdehyde (SSA)produced by GABA-T is further biotransformed to either a reduced form, yhydroxybutyrate (GHB), orto succinate via oxidation through SSADH (Vayer eta!.,1988). VPA has been found to exert an in vitro inhibitory effect on SSADH (Maitreeta!., 1976; Whittle and Turner, 1978; Laan eta!., 1979), and thus it has beenpostulated that the accumulation of SSA caused by this inhibition either inhibits theforward reaction of GABA-T or promotes a reverse reaction to that of GABA-T(Loscher, 1993). This does not seem likely to be a mechanism of VPA action,7however, as studies have suggested that it is not possible to raise brain GABAlevels by inhibition of SSADH (Maitre eta!., 1976; Simler eta!., 1981).4) VPA’s GABAergic action may be due to its ability to potentiate thepostsynaptic response to GABA. It has also been proposed that VPAs effecton presynaptic GABA levels could be caused indirectly through direct potentiationof postsynaptic GABAergic function leading to feedback inhibition of GABAturnover and thus to increases in nerve terminal GABA (Macdonald and Bergen,1979). However, the appearance of various reports (including those mentionedabove) demonstrating a primary effect of increases in GABA turnover by VPAsuggest that the hypothesis that VPA’s GABAergic actions are due to a secondaryfeedback inhibition is not valid.It should also be mentioned that obviously, an increase in the presynapticconcentration of GABA would only potentiate GABAergic neurotransmission if therelease of GABA into the synaptic cleft were also increased. There is both directand indirect evidence that suggests that the potassium induced release of GABA isindeed enhanced by VPA (Zimmer eta!., 1980; Loscher and Siemes, 1984;Loscherand Siemes, 1985; Gram eta!., 1988; Ekwuru and Cunningham, 1990).On the other hand, the uptake of GABA from the synaptic cleft is not affected byVPA (Loscher, 1980; Ross and Craig, 1981).Other proposed mechanisms of VPA’s anticonvulsant action include thefollowing:1) An inhibition of the release of the excitatory substance -hydroxybutyricacid (GHB) may result in anticonvulsant effects. GHB, a minor in vivodegradation product of GABA, has been shown to produce epileptogenic effects inseveral species (Snead, 1988). VPA has been shown to be a potent inhibitor of an8NADPH-dependent aldehyde reductase, a nonspecific SSA reductase (SSAR)(Whittle and Turner, 1978; Vayer eta!., 1988) . Unlike the nonspecific SSAR, thespecific SSAR which is thought to reduce SSA to GHB is not affected by VPA(Vayer etah, 1988). However, studies in rat brain homogenates havedemonstrated that VPA inhibits the formation of GHB in vitro, suggesting that thespecific SSAR is not exclusively responsible for GHB formation but that thenonspecific aldehyde reductase may also contribute (Whittle and Turner, 1982).The involvement of nonspecific SSAR in GHB formation may provide VPA with anadditional mode of action. It is more likely, however, that the importance of GHBon VPA’s mechanism(s) of action lies in the fact that VPA also inhibits the synapticrelease of GHB (Vayer eta!., 1988). This reduction in GHB release could play animportant role in the anti-absence seizure action of VPA.2) VPA may cause a nonspecific membrane action which would reduce highfrequency repetitive firing of neurons through indirect effects on sodiumand/or potassium channels. Generally, high concentrations of VPA arereportedly required to inhibit the spontaneous firing of neurons (Chapman et a!.,1982). However, VPA has been shown to reduce high frequency repetitive firing ofaction potentials of central neurons at concentrations lower than those used todepress normal neuronal cell activity (McLean and Macdonald, 1986; Loscher,1993). VPA’s anticonvulsant effect on general tonic-clonic seizures may berelated to this action (McLean and Macdonald, 1986). The reduction of sustainedrepetitive firing (SRF) caused by VPA is similar to that produced by phenytoin andcarbamazepine, hence, a likely explanation for this action of VPA would be a usedependent reduction of inward sodium current (McLean and Macdonald, 1986).However, biochemical studies have failed to find a direct interaction between VPAand Na channels at relevant VPA concentrations. It has been proposed thatperhaps SRF can be reduced through a modulation of potassium channels9involved in action potential repolarization. In support of this, an increase inmembrane conductance to potassium ion has been observed in Aplysia neuronsfollowing administration of VPA at concentrations 15 to 50 times greater thanlevels seen in clinical serum samples (Slater and Johnston, 1978). In work withhippocampal slices, VPA has been found to suppress spontaneous epileptiformactivity and prolonged the after-discharge elicited by antidromic stimulation(Franceschetti et al., 1986). These works and others (see Loscher, 1993) supportthe idea that VPA causes activation of Ca-dependent K conductance (Levyand Shen, 1989), although high concentrations of VPA are necessary to achievethese effects.As noted above, there is evidence supporting several of the proposedmechanisms of VPA action, with some mechanisms being more stronglysubstantiated than others. However, none of these proposals is completelysatisfactory in terms of explaining the anticonvulsant activity of VPA. It is almostcertainly the case that more than one mechanism of action is necessary to affectthe wide range of disorders treatable with VPA. For example, while GABApotentiation and SRF limitation may explain VPA’s action on focal and generalizedmotor seizures, a reduction of GHB release is more likely to be the path involved inVPA’s effect on non-convulsive seizures (Loscher, 1993).1.1.4 Pharmacokinetics and metabolism of VPA1.1 .4.1 AbsorptionIn humans, VPA is rapidly and completely absorbed following oraladministration (Loiseau etal., 1975; Meinardi etal., 1975; Perucca etal., 1978).Peak plasma levels are typically attained within 0.5 - 2 hours for rapid-release10formulations such as the oral syrup (of the sodium salt), capsule (containing thefree acid) and tablet (of either acid or sodium salt). The absolute bioavailability ofVPA is consistently close to 1 for both immediate and delayed-release products.Rectal administration of the VPA syrup appears to produce steady-state plasmalevels in children which are similar to those observed following oral administration,suggesting that this could provide an alternative route of administration whenneeded (Cloyd and Kriel, 1981). VPA appears to be absorbed throughout theintestine with no evident site specificity (Levy and Shen, 1989).1.1 .4.2 DistributionIn humans, VPA is highly bound (approximately 90 %) to plasma albumin,which results in most of the drug remaining within the vascular compartment (Levyand Lai, 1982). Typical volumes of distribution (Vd) range between 0.1 -0.4 L/kgindicating that VPA distribution is limited primarily to the circulation and rapidlyexchangeable extracellular waters (Klotz and Antonin, 1977; Gugler and vonUnruh, 1980). However, a larger Vd for unbound drug in plasma (approximately 1L kg-i) suggests that there is significant distribution of free VPA into theintracellular structures (Gugler eta!., 1977). VPA appears to enter the brain andcerebrospinal fluid (CSF) rapidly in humans, as well as in other species such asmice, rats and rhesus monkeys, as evidenced by the successful treatment ofstatus epilepticus with single-dose oral or rectal administration of VPA (Snead andMiles, 1985; Rosenfeld et al., 1987). This result is rather suprising since VPA (pKa= 4.56) is predominantly ionized at physiological pH, and suggests that a carriermediated transport mechanism may exist (Loscher and Frey, 1984). Further, VPAconcentrations in the brain and CSF correlate reasonably well with total and freedrug concentrations in plasma (Levy and Shen, 1989) and VPA does not appear toconcentrate in brain tissue (Vajda etaL, 1981).111.1.4.3 Metabolism and eliminationThe elimination half-life (t112) for VPA has been reported to range from 9 -18 h in humans and 0.3-6 h in various animals (Dickinson eta!., 1980; Nau,1986a; Hendrickx eta!., 1988; Kobayashi eta!., 1990; Davis eta!., 1994). Plasmaclearance of VPA in healthy volunteers has been reported to range from 0.4 to 0.6L h1 while higher values (1.0 to 1.1 L h-i) have been observed in patients withepilepsy on polytherapy (Schoben eta!., 1975; Klotz and Antonin, 1977; Peruccaeta!., 1978).In humans, as in other mammals, VPA is eliminated primarily by hepaticmetabolism with only 1- 3 % of a dose excreted unchanged in urine (Gugler andvon-Unruh, 1980; Bailer eta!., 1985). The branched short-chain fatty acidstructure of VPA is quite simple, however, its metabolic fate is complex,presumably because of its entry into xenobiotic metabolic pathways, as well aspathways normally reserved for endogenous lipids (e.g., 13-oxidation). In fact,there are approximately 50 known VPA metabolites (Levy and Shen, 1989) with atleast 16 of these being observed consistently in humans (see Fig. 2) (Gugler andvon-Unruh, 1980; Kassähun eta!., 1990). The major routes of VPAbiotransformation are glucuronidation, 13-oxidation in mitochondria and phase Imicrosomal oxidative processes (including Co and Co-i oxidation). In addition, otherroutes of metabolism in humans and other animals include ö and ydehydrogenation, glycine conjugation, carnitine ester formation, coenzyme Aconjugation, epoxidation and isomerization (Granneman et a!., 1984; Levy andShen, 1989).12GlucuronidationIn humans, glucuronidation accounts for approximately 10 - 70 % of a doseof VPA (Bruni and Wilder, 1979; Bailer eta!., 1985; Dickinson eta!., 1989). A netincrease in glucuronidation can be expected with increases in dose of VPA as lessdrug will be shunted toward the 3-oxidation pathway (Granneman eta!., 1984).f-oxidatio nThe metabolites produced by the mitochondrial f3-oxidation pathway are 2-n-propyl-2-pentenoic acid (2-ene VPA) (predominantly as the E-isomer), 2-n-propyl-3-hydroxypentanoic acid (3-OH VPA) and 2-n-propyl-3-oxopentanoic acid(3-keto VPA) (Granneman eta!., 1984). Cumulatively, these metabolites accountfor approximately 12 % of VPA in serum in humans (Abbott eta!., 1986; Dickinsonetah, 1989; Kassahun eta!., 1990). These compounds are analogous tointermediates of fatty acid 13-oxidation and as a result VPA may compete withendogenous lipids for the enzymes of f3-oxidation (Granneman eta!., 1984; Bjorgeand Baillie, 1985). However, since stable isotope experiments in rats haverevealed that 3-OH VPA is not an exclusive product of 13-oxidation but is alsoproduced by direct cytochrome P-450 hydroxylation of VPA (Rettenmeier et al.,1987), measurement of the combined urinary output of these metabolites cannotbe used to monitor the flux of VPA through this metabolic pathway. Furthermore,the mere fact that these f3-oxidative products of VPA, a branched-chain fatty acid,are excreted into urine while the 13-oxidative products of the straight-chainendogenous fatty acids are not excreted to any appreciable extent, suggests thatVPA’s branched structure renders it a poor substrate for further oxidation by the 13-oxidation enzymes (Levy and Shen, 1989). Lastly, it is worth noting that 3-ketoVPA, a major urinary metabolite, is generated mainly by oxidation of 2-ene VPAinstead of 3-OH VPA as one might expect (Rettenmeier et a!., 1987).13Microsomal 0) and (0)-i) oxidationThe principal products of 0)-oxidation of VPA occurring in the endoplasmicreticulum are 2-n-propyl-5-hydroxypentanoic acid (5-OH VPA), 2-propylg lutaricacid (2-PGA) and 2-propylmalonic acid (2-PMA) while (0)-i )-oxidation isresponsible for the production of 2-n-propyl-4-hydroxypentanoic acid (4-OH VPA),2-n-propyl-4-oxopentanoic acid (4-keto VPA) and 2-propylsuccinic acid (2-PSA)(Granneman etaL, 1984).COO-GluVPA glucuronideCOOHHO’\\’5-OH VPAOH COOH4-OH VPAO COOH4-keto VPACOOHHOOC\\/kFigure 2. Illustration of VPA and sixteen of its most frequently observedmetabolites in humans.COOH/%/c/\\3-ene VPA(E)- and (Z)-isomersCOOHWA/ NCOOH2,3-diene VPA COOH(EE)- and (EZ)-isomersOH3-OH VPA \\COOH\\ HOOC2-PGA2,4-diene VPA(E)- and (Z)-isomers /3-keto VPA 2-PSA14The 4-keto VPA and 2-PSA metabolites are formed by further oxidation of4-OH VPA while 2-PGA arises from oxidation of 5-OH VPA (Granneman et al.,1984). 2-PMA may arise via the 5-OH VPA pathway, however, there has beendisagreement as to whether it is a genuine metabolite of VPA (Levy and Shen,1989).DesaturationStudies have identified the following mono-unsaturated metabolites of VPAin serum of epileptic children (Kochen and Scheffner, 1980) and subsequently, inbodily fluids of other animal species (e.g., Granneman eta!., 1984; Rettenmeier etal., 1986): 2-ene VPA (both E and Z-isomers), 2-n-propyl-3-pentenoic acid (3-eneVPA; both E and Z-isomers) and 2-n-propyl-4-pentenoic acid (4-ene VPA). While3-ene VPA appears to be reversibly formed from isomerization of 2-ene VPA(Bjorge and Baillie, 1991), 4-ene VPA has been shown to be formed viacytochrome P-450 enzymes in the liver microsomes (Rettenmeier et aL, 1987;Rettie eta!., 1987). Thus, there are two separate pathways of desaturation forVPA, one in the endoplasmic reticulum that is responsible for 4-ene VPA formationand the other in mitochondria that produces 2-ene VPA which in turn isomerizes toform 3-ene VPA.Di-unsaturated metabolites of VPA have also been detected, with the mostsignificant of those being 2-(1 ‘-propenyl)-2-pentenoic acid (2,3-diene VPA; in both(E,E)- and (E,Z)-forms) and 2-n-propyl-2,4-pentadienoic acid (2,4-diene VPA; inboth (E)- and (Z)-forms) (Kochen eta!., 1984; Acheampong and Abbott, 1985).The origins of 2,3’-diene VPA have been traced to further 3-oxidation of 3-eneVPA (Rettenmeler et al., 1987; Bjorge and Baillie, 1991). The production of 2,4-diene VPA is an example of the interaction of the mitochondrial and microsomalmetabolism of VPA. That is, 2,4-diene VPA is known to be not only a f3-oxidation15product of the microsomally produced 4-ene VPA (Rettenmeier et al., 1985;Rettenmeier et aL, 1986; Singh eta!., 1987), but also a product of the microsomalmetabolism of 2-ene VPA (Kassahun and Baillie, 1993).Other conjugation reactionsOther minor pathways for conjugation of VPA exist aside from theprominent glucuronidation pathway. Carnitine conjugates have been found in theurine of pediatric patients receiving prolonged administration of VPA (Bohan eta!.,1984; Millington etal., 1985). These conjugates probably account for less than 1% of the administered dose of VPA but may be significant in that VPA therapy hasbeen associated with both a secondary carnitine deficiency and a Reye-likesyndrome. Either of these conditions may be connected with the diversion ofcarnitine to VPA conjugation (Levy and Shen, 1989). Glycine conjugates of VPAhave been identified in rat urine, however, the real importance of this pathway maylay in the conjugation of unsaturated VPA metabolites which have been found ingreater quantities (Granneman eta!., 1984). Lastly, a coenzyme A thioesterderivative of VPA has been proposed based on evidence from animal studies(Thurston etaL, 1983; Thurston eta!., 1985). Rigorous structural characterization,however, has not been reported to date.1.1.5 Methods of analysis of VPA and its metabolitesDue to the potential involvement of metabolites in both the efficacy andtoxicity of VPA, there has been considerable interest in the development of asimultaneous, sensitive assay for VPA and its metabolites. Attempts to developsuch an assay have been hampered by the large number of metabolites, the largerange of concentrations over which they are found in biological fluids, and the16presence of endogenous substances similar in structure to VPA. Some successhas been achieved however as several gas chromatographic - mass spectroscopicmethods have been reported for VPA and its metabolites. These methods haveemployed both electron impact ionization (El) of the tert-butyldimethylsilyl (tBDMS)(Phillipou, 1975; deJong eta!., 1980; Woollard, 1983; Abbott eta!., 1986; Dariusand Meyer, 1994; Yu eta!., 1995) and trimethylsilyl (TMS) (Kochen and Scheffner,1980; Nau eta!., 1981; Kochen eta!., 1984; Rettenmeier eta!., 1989; Fisher eta!.,1992b) derivatives as well as negative-ion chemical ionization (NICI) of thepentafluorobenzyl bromide (PFB) derivatives (Kassahun eta!., 1990). Recently, amethod using capillary gas chromatography with ion trap detection was reportedwhich is capable of resolving VPA and all of its mono-unsaturated metabolitesfrom serum in the form of methyl ester derivatives (Gaetani eta!., 1992). Perhapsthe most versatile and useful of the El / tBDMS assays available was recentlyreported by our laboratory (Yu et aL, 1995). This assay provides a convenientmethod for the separation and quantitation of VPA and 16 metabolites in a singlechromatographic run with a time of 29.5 minutes. The assay utilizes the intense(M-57y fragment of the tBDMS derivative which has been found to provide bettersensitivity with fatty acids than the less intense (M-15) fragment of TMSderivatives (Phillipou, 1975; deJong eta!., 1980; Woollard, 1983; Darius andMeyer, 1994). The assay also makes use of stable isotope labelled VPA andmetabolites as internal standards because their structural similarity to thecompounds of interest produces chromatographic characteristics akin to those ofthe compounds being examined (Roncucci eta!., 1975; Murphy and Sullivan,1980). Although the NICI method available (Kassahun eta!., 1990) is alsoadequate for the detection and quantitation of most of the metabolites of interest,analysis work involving large numbers of samples would be best handled using amass selective detector with El detection.171.1.6 Drugs, epilepsy and pregnancyAlthough it is now evident that the fetus is exposed to most drugsadministered to the mother during pregnancy, surveys suggest that a largepercentage of women still take at least one pharmaceutical preparation during theirpregnancy (Rubin eta!., 1986; Piper eta!., 1987; Kasilo eta!., 1988; Simpson eta!., 1989). Aside from vitamins and minerals, the most commonly consumedproducts are predominantly prescribed antibiotics and non-prescription items suchas non-narcotic analgesics, laxatives, antinauseants and antiemetics, andantihistamine containing preparations (Rurak eta!., 1991). Drug consumptionduring pregnancy is likely to continue since societal efforts to date have reducedthe usage of non-prescription items but have not eliminated it. Furthermore, therewill continue to be legitimate medical reasons for maternal (or fetal) drugadministration.Epilepsy in pregnancyEpilepsy has been estimated to occur in approximately 0.33 to 1 % of thegeneral population which would include approximately 2.5 million people in theUnited States alone (Kelly, 1984; Vorhees eta!., 1988). There are approximately12, 000 births per year to epileptic women in the United States, and approximately95 % of these women are on anticonvulsant therapy during pregnancy (Vorhees eta!., 1988). Unfortunately, women on anticonvulsant therapy have been reported togive birth to children with major malformations with a frequency 2 -3 times greaterthan that seen in the general population. It remains unclear as to whether this isdue to the disease-state itself or to the drug therapy (Kaneko et a!., 1988). Themost common malformations include cardiovascular effects, orofacial clefts, and18skeletal deformations as well as gastrointestinal and genitourinary anomalies(Janz, 1982; Bossi, 1983; Kelly, 1984; Ardinger eta!., 1988).A further area of concern, albeit less well studied than the congenitalmalformations mentioned above, is the possible effect of drug therapy duringpregnancy on the behavioral development of the fetus/neonate. The use of drugsduring the second and third trimesters of pregnancy could affect CNS function andresult in possible learning impairment, deficient motor function and patterns ofdelayed behaviour maturation (Chapman and Cutler, 1984; Nau and Loscher,1984; Jager-Roman eta!., 1986; Vorhees, 1987; Ardinger eta!., 1988;Christianson et a!., 1994). It is believed that neurotransmitters are important incontrolling cell proliferation in the nervous system. If this is true, it is possible thatcompounds that affect neurotransmitter activity in the CNS (e.g., antihistaminesand anticonvulsants) may interfere with brain cell differentiation and proliferation,resulting in a permanent change in later functional capacity without obviousstructural malformations (Beeley, 1986; Ardinger eta!., 1988). Studies in this areahave been primarily restricted to research on functional impairments after birth.More in depth studies examining fetal CNS function and behaviour during theactual period of drug exposure would be useful in elucidating the potential hazardsof such drug therapies.VPA in pregnancyThe placental transfer of VPA has been reported in both animals andhumans (Knott and Reynolds, 1986; Nau, 1986b; Nau and Krauer, 1986;Hendrickx eta!., 1988; Fowler eta!., 1989; Kobayashi eta!., 1990; Omtzigt eta!.,1992). Fetal/maternal blood concentration ratios of VPA ranging from 0.5 to 4.6have been reported in humans at term (Nau eta!., 1982; Hendrickx eta!., 1988)VPA was reported to exhibit three-compartment elimination kinetics in the19pregnant rhesus monkey with a terminal elimination t112 of 6.3 h in the mother andterminal phase fetal/maternal blood concentrations ratios of approximately 1.3(Dickinson etal., 1980). In pregnant rats, the terminal ti2 was reported to beapproximately 2.2 h, significantly longer than the t112 in a nonpregnant controlgroup (1.2 h) (Kobayashi etal., 1991). A significantly lower total body clearancewas also reported for the pregnant group (5.4 compared to 7.2 mLmin1kg -1)(Kobayashi etal., 1991).Pregnancy can have a significant effect on the maternal clearance of VPA.The blood level-to-dose ratio for VPA begins to decrease in the second trimesterand continues to decline throughout most of the remaining pregnancy (Plasse etal., 1979). Typically, the lowest ratios are observed late in the third trimester, just2-3 weeks before delivery. Following delivery, VPA levels quickly return to prepregnancy values. The change in this ratio is probably due to a decrease inmaternal serum protein binding of VPA as a result of elevated nonesterified fattyacids and hypoalbuminemia (Nau and Krauer, 1986). Fetal serum protein bindingdisplays a gestational time-dependency with a progressive increase in bindingoccurring as fetal serum albumin levels increase with fetal maturation (Nau andKrauer, 1986). These changes in the serum free fraction of VPA could beimportant as studies of VPA binding in the perinatal period suggest that decreasesin maternal serum binding results in fetal accumulation of the parent drug andpossibly several metabolites (Nau et al., 1984).1.2 Fetal drug exposureMost reports on the degree of fetal/maternal drug transfer are based asingle maternal and umbilical blood sample at the time of delivery following single20dose administration to the mother. Based on these studies, an index of relativefetal drug exposure has been defined as the ratio of the concentration of drug infetal plasma (Cf) to the concentration of drug in maternal plasma (Cm) (Waddelland Marlowe, 1981; Mihaly and Morgan, 1984). It has been clearly demonstrated,however, that the value of this ratio (Cf/Cm) depends largely on the time lapsedbetween administration of drug and collection of the plasma samples (Levy andHayton, 1973). Despite the obvious problems associated with estimating fetalexposure from such single point determinations, some reports continue to makeuse of such measurements. It has been proposed that a more informativeestimate of relative fetal exposure is the ratio of fetal-to-maternal areas under theplasma unbound drug concentration curve from the time of drug administration tothe time when all of the drug has been eliminated (i.e., (AUCI)u / (AUCm)u) (Levyand Hayton, 1973) . This approach has been successfully utilized in studies withvarious drug substances, such as diphenhydramine (Yoo etah, 1986) andmetoclopramide (Riggs etaL, 1990). Following prolonged chronic administrationof a drug, the ratio of the steady-state drug concentration in fetal plasma to that inmaternal plasma may also be used as an index of relative fetal exposure (Szeto,1982). This approach has been applied for drugs such as cimetidine (Ching et aL,1985), morphine (Szeto etal., 1982a), methadone (Szeto eta!., 1982a),indomethacin (Anderson etaL, 1980c) and acetaminophen (Wang etal., 1986).Methods of studyAs mentioned above, single point measurements of fetal / maternal druglevels in humans have commonly been used to estimate fetal drug exposure.While these experiments are of some clinical relevance, the fetal / maternal drugratios can be dependent on sampling time if the drug is not at steady-state. In aneffort to study fetal / maternal drug levels over time (instead of at a single point in21time), some researchers have attempted to create a composite by sampling fromdifferent subjects at different times (e.g., Bray eta!., 1966). Not surprisingly,intersubject variability and the numerous artefacts created by data averagingusually limit the kinetic interpretation of such collections of data (Krauer andKrauer, 1977; Levy, 1981; Waddell and Marlowe, 1981). Due to the limitedinformation that can be gained directly from humans, many researchers haveutilized animal models in order to study fetal-maternal pharmacokinetics. Perhapsthe most successfully used model has been that of chronically-catheterizedpregnant sheep. This model allows for serial sampling of both maternal and fetalblood as well as sampling from other fluid compartments following drugadministration, thus permitting a more extensive study of fetal-maternal drugrelationships (Rurak eta!., 1991). Studies with pregnant sheep have permitted thecalculation of fetal placental, nonplacental and organ clearances of several drugs(e.g., Szeto eta!., 1982b; Wang etal., 1986; Riggs eta!., 1990; Wright eta!., 1991;Yoo eta!., 1993). However, the epitheliochorial placenta of the sheep has asignificantly lower permeability to hydrophilic compounds than does thehemochorial placenta of the human (Faber and Thornburg, 1983) and thus studiescarried out with hydrophilic compounds in the sheep may not be relevant tohumans (Reynolds and Knott, 1989).Some researchers have studied placental drug transfer in several smallanimal species which possess hemochorial placentas such as the guinea pig andrabbit (Faber and Thornburg, 1983; Devane and Simpson, 1985; Laishley eta!.,1989). The small size of these animals (especially the fetus), however, limitsresearchers’ abilities to collect serial fetal samples and thus construct a goodpicture of the fetal-maternal drug concentration relationship (Rurak etal., 1991).Some attempts to overcome this problem have been made by exploiting thepolyctous nature of some species like the rabbit by removing individual fetuses22overtime (Laishley eta!., 1989). This work is compromised, however, due to theunknown effects of the surgery on fetal / maternal disposition (Rurak et aL, 1991).As a result, many studies have resorted to utilizing perfused placentas. This invitro work has provided the only feasible method of obtaining detailed transferdata through human placentas (Rurak et a!., 1991). These techniques are notwithout criticisms, however, as it has been reported that perfused rabbit andguinea pig placental preparations may leak when umbilical venous pressures areraised (Faber and Thornburg, 1983).The movement of substances (both endogenous and xenobiotic) across theplacenta may occur through any of the following processes: passive diffusion,facilitated diffusion, active transport, pinocytosis and/or via bulk flow (Reynoldsand Knott, 1989). For drug substances, the most important of mechanism oftransfer between mother and fetus is passive diffusion (Rurak eta!., 1991). Thus,despite the anatomic differences between the placentas of various species (e.g.,human vs. sheep), the limiting layer of tissue is essentially a continuous lipidmembrane and movement across this membrane is controlled by simple lipiddiffusional processes (Reynolds and Knott, 1989). If this is the case, thephysicochemical properties of the drug, such as polarity, molecular weight, degreeof ionization and protein binding, are of primary importance in determining theextent of drug transfer. However, other factors such as the concentration gradientacross the placenta and the characteristics of the maternal and fetal blood flowsare also important considerations (Reynolds and Knott, 1989; Rurak eta!., 1991).Approaches to the calculation of placental drug clearanceThe placental clearance of drugs has been determined in vivo using threedifferent approaches. The first of these, a compartmental approach to thesemeasurements, has been developed by Anderson and co-workers (Anderson et23aL, 1980a; Anderson etaL, 1980b; Anderson eta!., 1980c). This techniqueinvolves the paired administration of a maternal intravenous drug infusion and afetal bolus dose of labelled drug (either a radio- or stable-isotope of the drug).Following the completion of post-administration blood collections, fetal andmaternal blood flows are measured using a radioactively labelled microspheretechnique. The bolus administration is followed by loss of drug from the fetalcompartment due to both distribution from the fetal vascular compartment into theextravascular compartment and total fetal clearance. The total fetal clearance (CIf)is the summation of clearance due to placental flux (CIp) and fetal tissue usage bysequestration and / or metabolism (Clnp) and may be calculated from the slope ofthe fetal drug concentration (Cf) versus time plot. This method has been appliedto the determination of both placental and nonplacental clearances in sheep forboth acetylsalicylic acid (Anderson eta!., 1980b) and indomethacin (Anderson etaL, 1980a).The most notable criticisms and limitations of this technique are thefollowing: 1) The method is only valuable if the protein binding of drug is equal inthe fetus and mother (Anderson etaL, 1980a); 2) It must be assumed that thebidirectional placental clearances are necessarily equal and that diffusion is theonly mechanism of transplacental transfer (Szeto, 1982); and 3) That thecompartmental model upon which this method is based may not hold if the systemis not at steady-state because of the subsequent changes in the pharmacokinetics(Szeto, 1982). Therefore, the use of an intravenous bolus dose may not beappropriate for the calculation of placental and nonplacental clearances (Szeto,1982). The most obvious practical difficulties involved in utilizing this method arethe requirement for labelled drug plus the ability to detect and quantitate both thelabelled and unlabelled drug in the presence of one another.24Szeto and co-workers have developed an alternative approach to thecalculation of placental clearances utilizing steady-state compartmental techniquesand the model illustrated in Fig. 3 (Szeto, 1982; Szeto eta!., 1982a; Szeto etal.,1 982b). This method makes use of the administration of drug infusions to steady-state in the fetus and the mother accompanied by the sampling of both fetal andmaternal blood during both infusions (Szeto, 1982). With a knowledge of the ratesof infusion and the steady-state drug concentrations resulting from each infusionboth the placental and nonplacental clearances of the fetus and mother may bedetermined.This approach to the examination of the placental clearance of drugs hasbeen frequently employed in conjunction with the pregnant sheep model.Placental and nonplacental clearances in the mother and fetus have beencalculated for morphine (Szeto eta!., 1982a), methadone (Szeto eta!., 1982a),acetaminophen (Wang et aL, 1986), diphenhydramine (Yoo et a!., 1986) andmetoclopramide (Riggs et al., 1990) using this approach. Unlike the technique ofAnderson and co-workers, this model does not make any assumptions withrespect to the mechanisms or extent of drug transfer between mother and fetus.In order to make use of this model it is necessary to either employ alabelled (radio- or stable-isotope) species of the drug so that simultaneousinfusions may be carried out or, perform the infusions on separate days whileminimizing changes to any factors which might affect placental transfer (Szeto,1982). Most drug studies with this model in the past have utilized time-separatedinfusions to the ewe and fetus in order to collect the appropriate data, due to theinability to obtain or analyze for an isotope-labelled form of the compound. Afurther limitation associated with this technique is that each respective infusionmust provide a measurable drug concentration on the opposite side of theplacenta. This may be a problem if the fraction of the drug clearing the placenta25following fetal infusion becomes unmeasurable once diluted in the maternalvolume of distribution.CLmf___potenalsite M F potenalseCLfm I CLfoCLmoFigure 3: A two-compartment open model describing the disposition of a drug inthe mother and fetus where M represents the maternal compartment, F representsthe fetal compartment, CLmf is the maternal to fetal clearance, CLfm is the fetal tomaternal clearance, CLmo is the maternal nonplacental clearance and CLf0 is thefetal nonplacental clearance (Szeto etal., 1982b).The third approach used to monitor placental and nonplacental clearancesis based on Fick’s Law of Diffusion and has been utilized for such studies since the1960’s (e.g., Meschia et aL, 1967; Stacey et aL, 1978). The application of the Fickapproach involves a fetal infusion to steady-state followed by either simultaneoussampling of maternal and uterine blood plus determination of uterine blood flow orfetal arterial and umbilical venous blood plus determination of umbilical blood flow26(Wright et al., 1991). Recently the Fick approach has been used to estimate theplacental clearance of ritodrine in the pregnant sheep model (Wright et al., 1991).One of the limitations to this approach is that only the fetal-to-maternalplacental clearance or the maternal-to-fetal placental clearance may bedetermined at any one time. Also, one should be aware that uterine or umbilicalvenous blood flows may not be totally representative of blood at the sites oftransfer (Wright etah, 1991). In addition, the Fick approach should only be appliedto studies involving non-protein bound substances which are not metabolized inthe placenta and are rapidly cleared through the placenta (Anderson eta!., 1 980a).Other practical difficulties encountered using this method are the extensive surgeryrequired for cannulation, the need for a measurement technique to determine therate of blood flow, and the need for measurable differences in arterial-venous drugconcentrations.1.3 Rationale and specific objectives1.3.1 RationaleThe continued administration of anticonvulsants to epileptic women duringpregnancy and the popularity of VPA as an anticonvulsant of choice because of itsversatility and effectiveness, highlights the need to better examine the placentaltransfer and pharmacokinetic disposition of this drug in pregnancy. Work withhumans and in animal models indicates that significant placental transfer of VPAcan occur. Also, the observed morphological and behavioral teratogenicity causedby maternal VPA therapy and VPA’s potential effect on fetal behaviour through itseffects on central GABA levels, indicate the need to examine the fetal CNS,cardiovascular and metabolic effects of VPA in pregnancy in detail. Furthermore,27considering VPA’s propensity to undergo extensive hepatic metabolism, anexamination of a fetus’ ability to metabolize and excrete this compound will notonly add to our knowledge of the disposition of this compound in pregnancy butalso possibly to our understanding of fetal mechanisms of elimination ofxenobiotics in general.The use of the pregnant sheep model, despite the structural differences inthe human and ovine placentas, will permit a detailed study of the fetal andmaternal pharmacokinetics and fetal effects of VPA in a chronic preparation. Thefetal pharmacokinetic and phamacologic monitoring that can be accomplished inthis model cannot be duplicated in either humans, due to ethical and technicalconstraints, or in smaller animal models due to difficulties introduced by sizelimitations (e.g., impractical surgical preparation) and by limited blood volumes ofboth mother and fetus. Our ability to collect data from this model will be greatlyenhanced by the capability to administer and quantitate a stable isotope labelledform of VPA, [13C4]VPA. Possession of this compound will permit simultaneousinfusions to steady state in both the ewe (VPA) and fetus ([13C4]VPA), therebyeliminating any confounds introduced by performing infusions on separate days.Furthermore, our ability to monitor both labelled and unlabelled metabolites of VPAin various biological fluids should enhance our understanding of the metabolism ofVPA in the ewe and fetus.While several basic compounds have been examined in detail in this model,acidic compounds, such as VPA, have not been studied to date. Therefore,comparisons will be made between the data collected with VPA and the availabledata obtained with basic compounds such as diphenhydramine, metoclopramideand ritodrine.281.3.2 ObjectivesThe specific objectives of the proposed work are the following:1. To verify the pharmacokinetic bioequivalence of [13C4]VPA with unlabelledVPA in nonpregnant sheep.2. To assess the pharmacokinetics, protein binding, placental transfer andmetabolism (metabolite profiles) of VPA in the pregnant sheep using threemodes of drug administration: maternal i.v. bolus dosing, fetal i.v. bolusdosing and maternal and fetal i.v. infusions to steady-state.3. To assess the CNS, cardiovascular and metabolic effects of VPA on thefetus.4. To compare the results of these pharmacokinetic studies with thoseobtained from similar studies with basic compounds such asdiphenhydramine, metoclopramide and ritodrine.292. EXPERIMENTAL2.1 MaterialsThe following is a list of reference standards, chemicals, reagents and othermaterials used: 2-Propylpentanoic acid (di-n-propylacetic acid, vaiproic acid) (K &K Fine Chemicals Inc., Plainview, NY, U.S.A. & Aldrich Chemical Co., Milwaukee,WI, U.S.A.; redistilled priorto use); 2-(propyl-1-13C)-pentanoic-1,2,3- acid([13C4]VPA) (MSD Isotopes, Montreal, PQ, Canada); ethyl acetate and methanol(distilled-in-glass grade) (Caledon, Georgetown, ON, Canada); anhydrous sodiumsulfate, sodium hydroxide and hydrochloric acid (reagent grade) (BDH Chemicals,Toronto, ON, Canada); N- tert.-butyldi methylsi lyl-N-methyltrifluoroacetamide(MTBSTFA) (Aldrich Chemical Co., Milwaukee, WI, U.S.A.); injectable Atropinesulphate (0.6 mg/mL) (Glaxo Laboratories, Montreal, PQ, Canada); injectablethiopental sodium (1 g/viaI) (Pentothal®); halothane (Fluothane®) (AyerstLaboratories, Montreal, PQ, Canada); injectable ampicillin (250 mg/vial)(Penbritin®); injectable gentamicin sulphate (40 mg/vial) (Garamycin®) (ScheringCanada Inc., Pointe Claire, PQ, Canada); injectable heparin (1000 U/mL)(Hepalean®) (Organon Canada Ltd., West Hill, ON, Canada); sodium chloride forinjection USP (Abbott Laboratories, Montreal, PQ, Canada); ultra-pure gradehelium (Matheson Gas Products Canada Ltd., Edmonton, AB, Canada); deionized,high purity water was produced on-site by reverse osmosis, followed by filtrationthrough ion-exchange cartridges and a 0.45 pm membrane filter (MiIIi-Q(Millipore, Mississauga, ON, Canada).Needles and plastic disposable LuerLok® syringes for drug administrationand sample collection (Beckton-Dickinson Canada, Mississauga, ON, Canada);nylon membrane filters for dose filtration (0.22 jim) (MSI, Westboro, MA, U.S.A.);30Foley® bladder catheters (l6fr), (Bard, Ltd., Sunderland, U.K.); heparinized bloodgas syringes (Marquest Medical Products Inc., Englewood, CO, U.S.A.); additive-free Vacutainer® tubes (Vacutainer Systems, Rutherford, NJ, U.S.A.);polytetrafluoroethylene (PTFE) lined screw caps (Canlab, Vancouver, BC,Canada); borosilicate glass culture tubes with screw top (John Scientific, TorontoON, Canada); disposable plastic pipet tips (National Scientific, San Rafael, CA);borosilicate glass pasteur pipets (John Scientific, Toronto, ON, Canada); siliconerubber tubing for catheter preparation (Dow Corning, Midland, Ml, U.S.A.); PTFEcoated stainless steel wire for electrode preparation (Cooper Corp., Chatsworth,CA, U.S.A.).VPA metabolitesThe metabolites used for the standard calibration curves were synthesizedas reported elsewhere (Acheampong etaL, 1983; Acheampong and Abbott, 1985;Lee eta!., 1989). These metabolites were namely (E)-2-propyl-2-pentenoic acid[(E)-2-ene VPA], (Z)-2-propyl-2-pentenoic acid [(Z)-2-ene VPA)], (E) and (Z)isomers of 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 as y-lactone isomers (4-OH-VPA), 2-[(E)-1 ‘-propenyl]-(E)-2-pentenoic acid [(E,E)-2,3’-diene VPA], 2-[(E)-1 ‘-propenyl]-(Z)-2-pentenoic acid[(E,Z)-2,3’-diene VPA], (E)-2-propyl-2,4-pentadienoic acid [(E)-2,4-diene VPA], (Z)2-propyl-2,4-pentadienoic acid [(Z)-2,4-diene VPA], 2-propylglutaric acid (2-PGA)and 2-propylsucci nic acid (2- PSA). The metabolites 2-propyl-3-oxopentanoic acid(3-keto-VPA) (Kassahun eta!., 1990), 2-propyl-4-oxopentanoic acid (4-keto-VPA)(Kassahun eta!., 1990) and 2-propyl-5-hydroxypentanoic acid (5-OH-VPA)(Rettenmeier eta!., 1985) were also synthesized according to previously reportedmethods.31Internal standards2-Methylgiutaric acid (2-MGA) (Aldrich Chemical Co., Milwaukee, WI,U.S.A.). The following heptadeuterated compounds were prepared as describedelsewhere (Zheng, 1993):2-[H7]propylpentanoic acid([2H7]VPA) , 2-[H7]propyl-2-pentenoic acid (2-ene[H7]VPA),2-[H7]propyl-4-pentenoic acid (4-ene[2H7]VPA), 2[2 H7]propyl-3-oxopentanoic acid (3-keto[2H7]VPA), 2-[2H7]propyl-4-oxopentanoic acid (4-keto[2H7]VPA), 2[2 H7]propyl-3-hydroxypentanoic acid (3-OH[2H 7]VPA), 2[2 H7]propyl-5-hydroxypentanoic acid(5-OH[2H7]VPA).2.2 Preparation of stock solutionsAlkaline solutions of the internal standards were diluted with distilled waterand combined to obtain a stock solution containing 4 p,g/mL of 3-OH[2H7]VPA and5-OH[2H7]VPA, 2 .tg/mL of[2H7]VPA, 3-keto[2H7]VPA, 4-keto[2H7]VPA and 2-MGA and, 1.7 p.g/mL of (E)-2-ene[H7]VPA and 1 ig/mL of 4-ene[H7]VPA.Individual solutions of VPA and the sixteen metabolites listed aboveprepared using 0.1 M NaOH were combined and diluted with distilled water toprovide a standard stock solution with metabolite concentrations equivalent to theupper end of the ranges listed in Table 1. Subsequent serial dilutions of the stocksolution with distilled water provided a series of seven standard stock solutionswith the concentration ranges of VPA and metabolites listed in Table 1.32Table 1. Concentration ranges, diagnostic ions and retention times of tBDMSderivatives of VPA and metabolites from standard reference samples.Compound Concentration Retention time Ionrange (mm) monitored(.ig/mL) m/zVPA 0.078 - 120.0 13.64 201(E)-2-ene VPA 0.074 - 19.076 16.58 199(Z)-2-ene VPA 0.004 - 0.924 14.64 199(E)-3-ene VPA 0.003 - 0.386 14.05 199(Z)-3-eneVPA 0.014-3.614 14.29 1994-EneVPA 0.008-2.000 13.88 199(E)-2,4-diene 0.012 - 3.044 17.78 197VPA(Z)-2,4-diene 0.004- 0.956 17.05 197VPA(E,E)-2,3’- 0.027 - 7.032 18.71 197diene VPA(E,Z)-2,3’- 0.004- 0.968 17.80 197diene VPA3-Keto VPA 0.016 - 4.000 21.79, 21 •99a 329b4-Keto VPA 0.008 - 2.000 20.03 2153-OH VPA 0.016 - 4.000 19.66, 19.93a 2174-OH VPA 0.016 - 4.000 12.21, 12.41 a 1 OOC5-OH VPA 0.008 - 2.000 22.14 331b2-PSA 0.008 - 2.000 22.18 331b2-PGA 0.008 - 2.000 22.50 345ba Isomersb Diderivative‘y-Iactone332.3 Instrumentation and equipment2.3.1 Gas chromatography with mass selective detectionA Model 5890 Series II gas chromatograph equipped with a split-splitlesscapillary inlet system, a Model 7673 autoinjector, a Vectra 486 computer with aModel 7360 controller and ChemStation software version A.02.01 (HewlettPackard Co., Avondale, PA, U.S.A.); a Model 5971A quadrupole mass selectivedetector (Hewlett-Packard Co., Avondale, PA, U.S.A.); a DB-1701 fused silicacapillary column (30m X 0.25mm i.d., 0.25 J.Lm film thickness; J and W Scientific,Rancho Cordova, CA, U.S.A.); a 2 mm borosilicate glass inlet liner (Hewlett-Packard Co., Avondale, PA, U.S.A.); silicone rubber septa (Thermogreen® LB-2)(Supelco, Bellafonte, CA, U.S.A.).2.3.2 Physiological monitoringA Beckman R-71 1 Dynograph Recorder (Sensormedics, Anaheim, CA,U.S.A.); strain-gauge transducers (Statham model P23Db, Gould Inc., Oxnard,CA, U.S.A.); Gould DTX disposable pressure transducers (Spectramed Inc.,Oxnard, CA, U.S.A.); cardiotachometers (Model 9857, Sensormedics, Anaheim,CA, U.S.A.); transit-time blood flow transducers (Transonic Systems Inc., Ithica,NY, U.S.A.); Apple lie computer (Apple Computers Inc., Cupertino, CA, U.S.A.);computer data acquisition system composed of Interactive Systems (DaisyElectronics, Newton Square, PA, U.S.A.), analog to digital converter and clockcard (Mountain Software, Scott’s Valley, CA, U.S.A.); roller pump for fetal urinecollection (DIAS, Ex 154, DIAS Inc., Kalamazoo, Ml, U.S.A.); IL 1306 pH/blood gasanalyzer (Allied Instrumentation Laboratory, Milan, Italy); Hemoximeter®34(Radiometer, Copenhagen, Denmark); YSI 2300 Stat Plus Analyzer (YSIBioanalytical Products, Yellow Springs, OH, U.S.A.).2.3.3 General experimentalA vortex-type mixer (VortexGenie®, Fisher Scientific Industries,Springfield, MA, U.S.A.); incubation oven (Isotemp® Model 350, Fisher ScientificIndustries, Springfield, MA, U.S.A.); IEC Model 2k centrifuge (Damon/IEC Division,Needham Hts., MA, U.S.A.); automatic Speed Vac® concentrator Model AS 290(Savant, Farmingdale, NY, U.S.A.); rotating-type mixer (Labquake® Model 415-110, Lab Industries, Berkeley, CA, U.S.A.); infusion pump (Harvard Model 944,Harvard Apparatus, Millis, MA, U.S.A.).2.4 Quantitation of VPA and metabolites using GC-MSThe quantitation of VPA and its 16 metabolites listed in Table 1 wasperformed using a GC-MS method developed in our laboratory (Yu et al., 1995).The procedures used are described below.2.4.1 GC-MS conditionsA DB-1 701 fused silica capillary column (30m X 0.25mm i.d., 0.25 m filmthickness) was used with helium as carrier gas and a column head pressure of 15p.s.i. Splitless injection was employed with an injector temperature of 250°C. Theinitial column oven temperature was 80 °C (held for 0.1 minutes) followed by anincrease to 100°C at a rate of 10°C/mm (0.1 mm hold time), a 2°C/mm increase to130°C (0.1 mm hold time) and a 30°C/mm increase to 260°C (8 mm hold time).The gas chromatograph interface temperature was held at 280°C. The detector35was operated in selected ion monitoring (SIM) mode with an ion sourcetemperature of 180°C, an emission current of 300 pA and an ionization energy of70 eV. The total run time required was 29.5 minutes.2.4.2 Preparation of standard samplesStandard samples were prepared by combining 200 i.iL of standard solution,100 jiL of internal standard and 200 pi of blank biofluid. The resulting solutionswere mixed, adjusted to pH 2 with 1 M HCI and further adjusted to a total volumeof 700 pL by the addition of distilled water. The solutions were then extracted twicewith ethyl acetate (1.5 mL) by gentle rotation for 30 mm. The combined organicphase was vortex-mixed with anhydrous Na2SO4, centrifuged (10 mm, l000xg,25°C), transferred to a 15mm culture tube and concentrated to approximately 50pL in a Savant concentrator. MTBSTFA (50i.LL) was added and the samples werederivatized at 60 °C for 1 h. A 1.tL aliquot of sample was injected into the GCMSD.In order to quantify total conjugated and unconjugated VPA and metabolitesin biofluids, 100 pL of control biofluid, 100 pL of internal standard solution, 100 j.tLof distilled water and 200 EIL of standard solution were combined, adjusted to pH12.6- 13 with 4 M NaOH, and incubated for 1 hat 60°C. After cooling, thesolutions were adjusted to pH 2 with 4 M HCI and distilled water was added to afinal volume of 1 mL. These samples were then extracted and derivatized asdescribed above.362.4.3 Determination of calibration curvesCalibration curves were obtained from a plot of the area ratio of the VPA ormetabolite peak to that of the internal standard versus the concentration of thecompound of interest. For the experiments described in the present work,[2H7]VPA was used as the internal standard for VPA and 4-OH VPA, 4-ene[H7]VPA was used for 4-ene VPA, (E)-2-ene[2H7]VPA was used to analyzeall of the remaining monoenes as well as the dienes, 2-MGA was used for 2-PSAand 2-PGA, and 3-keto[2H7]VPA, 4-keto[2H7]VPA, 3-OH[H7]VPA and 5-OH[2H7]VPA were used to quantitate the same corresponding unlabelledmetabolites. The retention times and m/z values for the diagnostic ions of thetBDMS-derivatized internal standards are listed in Table 2.2.4.4 Assessment of GC-MS assay in sheep biofluidsThe accuracy of the assay in serum and urine was determined by analysisof six samples of spiked blank biofluid at both a high and low concentration of eachcompound of interest. The experimentally determined concentrations were thencompared to the nominal (spiked) concentration of the sample. The intra-dayvariability of the analysis was determined by six-fold repeated analysis of spikedblank biofluid at both a high and low concentration of each compound of interest.The precision is reported as the coefficient of variation for each compound. Theinter-day variability was determined by an eight-fold repeated analysis of a set ofseven standard samples containing each of the compounds of interest, over aperiod of sixteen days.372.4.5 Analysis of experimental samplesSamples of the blotluid to be analyzed (200 pL for quantitation ofunconjugated compounds; 100 j.iL for hydrolysis and quantitation of conjugatedcompounds except in the case of maternal urine (10 pL)) were combined with1 OOjiL of internal standard and the appropriate amount of distilled water to give afinal volume of 700 jtL. The samples were then mixed, extracted and derivatizedas described above for the standard sample preparations. The samples wereanalyzed using the retention times listed in Table 1 and the m/z values of thediagnostic (M-57)+ ions of the tBDMS derivatized compounds for both unlabelledand [13C4]-labelled parent compound and respective metabolites. The m/z valuesfor the [13C4]-labelled compounds correspond to the values listed for theunlabelled compounds in Table 1 plus 4 m/z units.2.4.6 Protein binding assessmentThe protein binding of VPA and [13C4]VPA in maternal and fetal serumwere assessed using standard ultracentrifugation techniques (Centrifree®Micropartition System, Amicon Division, W.R. Grace & Co., Danvers, MA, U.S.A.).Following thawing, aliquots of serum (500 jiL) were placed in the sample chamberof the ultracentrifugation tubes and centrifuged at 3400 rpm (15 cm radius;approximately 1 000xg) for 30 minutes at room temperature (22 °C). The typicalrecovery from this procedure was 200-300 j.tL of ultrafiltrate. The ultrafiltrate wasthen analyzed as described above.38Table 2. Diagnostic ions and retention times of the derivatized internal standardcompounds.Compound Retention time m/z(mm)[2H7]VPA 13.44 208(Z)-2-ene[H7]VPA 14.43 206(E)-2-ene[2H7]VPA 16.36 2064-ene[2H7]VPA 13.70 2063-keto[2H7]VPA 21.77, 21.97 a 336b4-keto[2H7]VPA 19.99 2223-OH[2H7]VPA 19.61, 19.89 a 2245-OH[2H7]VPA 22.11 338b2-MGA 22.10 317ba Isomersb Diderivative392.5 Animal and surgical preparationThe procedures for these studies were approved by the University of BritishColumbia Animal Care Committee. Dorset and Suffolk breeds of sheep wereused. Pregnant ewes were brought into the animal unit at The Children’s VarietyResearch Centre at least one week prior to surgery. The ewes were kept ingroups of two or more in large adjacent pens. The animals received a standarddiet and were allowed free access to water.Pregnant animalsAseptic surgical techniques were employed throughout the surgicalprocedures. Food was withheld for at least 12 h prior to surgery. Surgery wasperformed at 115- 130 days gestation (term, 145 days) with the animals underhalothane (1-2%) and nitrous oxide (60%) in oxygen anesthesia following Lv.atropine administration (6 mg) to control salivation, induction of anesthesia with Lv.sodium pentothal (ig) and intubation. A midline abdominal incision was made inthe ewe and access to the fetus was gained through incision of the uterine wall inan area without placental cotyledons and major blood vessels. For thesimultaneous infusion experiments, sterile silicone rubber catheters (Dow Corning,Midland, Ml) filled with heparinized saline (12 U/mL) were implanted into the fetalfemoral artery (dimensions: 0.025 in l.D.; 0.047 in O.D.), carotid artery(dimensions: 0.025 in ID.; 0.047 in O.D.), lateral tarsal vein (dimensions: 0.025 inl.D.; 0.047 in O.D.), umbilical vein (dimensions: 0.020 in l.D.; 0.037 in O.D.),trachea (dimensions: 0.040 in l.D.; 0.085 in O.D.) and bladder (via suprapubicincision), as well as the maternal femoral artery (dimensions: 0.040 in l.D.; 0.085 inO.D.), femoral vein (dimensions: 0.040 in l.D.; 0.085 in O.D.) and amniotic cavity(dimensions: 0.040 in l.D.; 0.085 in O.D.) as described previously (Rurak eta!.,401988). The maternal and fetal bolus dosing studies only required catheterizationof the fetal femoral artery, lateral tarsal vein, trachea and amniotic cavity as well asthe maternal femoral artery and femoral vein. For the simultaneous infusionstudies, stainless-steel polytetrafluoroethylene (PTFE)-coated wire electrodeswere placed biparietally on the dura to record fetal electroencephalographicactivity and through the orbital ridge of the zygomatic bone of each eye to recordfetal electro-ocular activity. Also, a transit-time blood flow transducer (size 4S)was placed around the common umbilical artery to measure umbilical blood flow.The catheters and electrodes were tunneled subcutaneously and exited through asmall incision on the maternal flank and stored in a denim pouch when not in use.Amniotic fluid lost during surgery was replaced with sterile irrigation saline. Thematernal abdominal incision was then closed in layers. Ampicillin (500 mg) andgentamicin (80 mg) were administered intramuscularly to the ewe on the day ofsurgery and for the first four days following surgery while ampicillin (500 mg) andgentamicin (40 mg) were administered intravenously to the fetus at the time ofsurgery. Following surgery, ampicillin (500 mg) and gentamicin (40 mg) wereadministered into the amniotic cavity on a daily basis until delivery. Each catheterwas flushed daily with approximately 2 mL of heparinized (1 2U/mL) normal salinesolution.After surgery, the ewes were kept in holding pens with other sheep andallowed free access to food and water. Following a recovery period of at leastthree days, the sheep were transferred to a monitoring pen adjacent to and in fullview of the holding pen for experimentation purposes. For the simultaneousinfusion experiments, a maternal Foley® bladder catheter (16 fr) was inserted viathe urethra on the morning of the experiment for cumulative urine collection.41Nonpregnant animalsSimilar procedures as those described above were used for nonpregnantadult ewes. In these animals, catheters were implanted into a femoral artery andvein (dimensions: 0.040 in l.D.; 0.085 in O.D.). Also, a Foley® bladder catheter(16 fr) was inserted via the urethra on the morning of the experiment forcumulative urine collection.2.6 Physiological monitoringSix hours before, during and for six hours after the experimental period,maternal arterial, amniotic, fetal arterial and tracheal pressures were monitored ona polygraph recorder using strain-gauge manometers or disposable DTXtransducers. Fetal and maternal heart rates were measured from the arterial pulseby means of cardiotachometers. Fetal breathing activity was identified from thefrequency and amplitude of oscillations in the tracheal pressure recording, asdescribed previously (Rurak etal., 1988). All of these variables and the electrodeoutputs were continuously recorded on a polygraph recorder (Beckman R-71 1)and simultaneously converted to digital form using an analog to digital conversionfor computer analysis and storage (Kwan, 1989) The sampling rate for eachvariable was between 2.5 and 6 Hz depending on the number of variables beingrecorded. At one minute intervals, the variable measurements were averaged andthe fetal arterial pressure was corrected for amniotic pressure. The averagedvalues were saved on floppy disk every 30 minutes during monitoring. Blood p°2,pCO2, pH and whole blood base excess (BE) were monitored using an IL 1306pH/blood gas analyzer, while °2 saturation (02 sat.) and hemoglobin content (Hb)were measured using a Hemoximeter. Glucose and lactate concentrations werealso determined using a YSI 2300 Stat Plus Analyzer.42Fetal urine collection was performed using a computerized roller pumpassembly developed in our laboratory. The mechanism involved the triggering ofthe roller pump through a disposable transducer as the pressure in a gravity fedurine reservoir increased above a preset pressure value (typically 3 mm Hg). Acalibrated volume of urine would be pumped from the reservoir into either a sterilesample collection syringe (during experimentation) or back into the amniotic cavityvia the amniotic catheter (during control periods). The cumulative volume of urinepumped per minute was stored by the computer for future analysis.2.7 Experimental protocolsNonpregnant animal experimentsThree nonpregnant sheep were used in these experiments. Serial femoralarterial blood samples were collected for analysis at -5, 2, 6, 10, 15, 20, 30, 45, 60,90 mm., 2, 2.5,3, 3.5, 4, 6,8, 10, 12, 24, 48 and 72 hours following intravenousbolus administration of 1000 mg of VPA (50 % VPA, 50 % [13C4]VPA) over oneminute through a femoral venous catheter. Cumulative urine samples were alsoobtained for the following time periods: -30-0, 0-30, 30-60 mm., 1-2, 2-3, 3-4, 4-6,6-8, 8-10, 10-12, 12-24, 24-36, 36-48 and 48-72 hours. The blood samplescollected for drug analysis were transferred to empty blood collection tubes,allowed to stand at room temperature for 2 hours and then centrifuged at 3000 rpmfor 10 minutes. Following centrifugation, the supernatant was transferred to newborosilicate glass culture tubes with a polytetrafluoroethylene (PTFE) lined screwcap. The samples were stored at -20°C prior to analysis.43Maternal bolus administration experimentsFive pregnant sheep were used in these experiments. A 1000 mg dose ofVPA was prepared by the addition of a 5 % molar excess of base (using 4 NNaOH), followed by adjustment of the solution pH to 7.4 and final dilution to 10 mLwith 0.9 % saline for injection. The resulting solution was filtered through a 0.22.im nylon filter prior to administration. The 1000 mg dose of VPA was administeredvia the maternal femoral vein over one minute. The catheter was then flushed with10 mL of heparinized saline (12 U/mL). Blood from the maternal femoral artery (3.0mL) and fetal femoral artery (1.5 mL) as well as amniotic and tracheal fluids (3.0mL) were collected at the following times in relation to the time of drugadministration: -5,2, 6, 10, 15, 20, 30, 45, 60, 90, 120, 150, 180 and 210 minutes,4,5, 6,8, 10, 12, 24 and 48 hours. In addition, samples of maternal arterial (0.5mL) and fetal arterial (0.5 mL) blood were collected at -5, 2, 10, 30, 60 and 90minutes, 4, 8 and 24 hours for blood gas and acid-base analysis. The bloodsamples collected for drug analysis were transferred to empty blood collectiontubes, allowed to stand at room temperature for 2 hours and then centrifuged at3000 rpm for 10 minutes. Following centrifugation, the supernatant was transferredto new borosilicate glass culture tubes with PTFE-lined screw caps. The amnioticand tracheal fluid samples were transferred directly to new borosilicate glassculture tubes with PTFE-lined screw caps. All samples were stored at -20 °C untilanalysis.The volume of fetal blood collected during sampling was replaced, via thefetal tarsal vein, with an equal volume of maternal blood which had been collectedbefore beginning the experiment in 10 mL syringes containing a small amount ofheparin. The replacement blood was given following withdrawal of a total of 10 mLof fetal blood (i.e., after every fifth sample).44Fetal bolus administration experimentsFive pregnant sheep were used in these experiments. In a similar fashionto that described above, a 250 mg dose of VPA was prepared for the fetal dosingexperiments by the addition of a 5 % molar excess of base (using 4 N NaOH),followed by adjustment of the solution pH to 7.4 and final dilution to 3 mL with 0.9% saline for injection. Following filtration of the solution, the 250 mg dose of VPAwas administered via the fetal tarsal vein over one minute. The catheter as thenflushed with 3 mL of heparinized saline (12 U/mL). Samples were then collectedand handled as described above for the maternal dosing experiments.Simultaneous infusion experimentsFive pregnant sheep were used for these experiments. Loading doses of35.4 mg kg-1 maternal weight of VPA and 8.8 mg kg1 maternal weight of[13C4]VPA for simultaneous administration to the ewe and fetus, respectively,were prepared by the addition of a 5 % molar excess of base (using 4 N NaOH),followed by adjustment of the solution pH to 7.4 and final dilution to 10 mL (formaternal injection) or 3 mL (for fetal injection) with 0.9 % saline for injection. Theresulting solutions were filtered through a 0.22 im nylon filter prior toadministration. Adequate quantities of infusate for the six hour infusion of VPA tothe ewe at a rate of 0.1746 mg min1 kg1 maternal weight and of[13C4]VPA to thefetus at a rate of 0.0436 mg mirn1 kg1 maternal weight were prepared by theaddition of a 5 % molar excess of base (using 4 N NaOH), followed by adjustmentof the solution pH to 7.4 with HCI and final dilution to the appropriate volume with0.9 % saline for injection. The resulting solutions were filtered through a 0.22 jimnylon filter prior to administration. The respective loading doses wereadministered via the maternal femoral vein (VPA) or the fetal lateral tarsal vein([13C4]VPA) over one minute. The catheters were then flushed with 10 mL and 345mL of heparinized saline (12 U/mL), respectively. The infusions were then initiatedand continued for six hours at the rates listed above. Immediately before andduring the six hour infusion period, blood from the maternal femoral artery (3 mL),fetal femoral artery (3 mL), fetal carotid artery (2 mL) and umbilical vein (2 mL) aswell as amniotic and tracheal fluids (3.0 mL each) were collected at the followingtimes: -5, 5, 30 and 60 minutes, 2, 3, 4, 5 and 6 hours. In addition, samples offetal arterial (0.5 mL) and umbilical venous (0.5 mL) blood were collected at -5, 5,and 60 minutes as well as at 4 and 6 hours for blood gas and acid-base analysis.The blood samples collected for drug analysis were transferred to empty bloodcollection tubes, allowed to stand at room temperature for 2 hours and thencentrifuged at l000xg for 10 minutes. Following centrifugation, the supernatantwas transferred to new borosilicate glass culture tubes with a PTFE-lined screwcap. The amniotic and tracheal fluid samples were transferred directly to newborosilicate glass culture tubes with a PTFE-lined screw cap. Cumulative maternalurine samples were also obtained hourly beginning one hour before drugadministration and ending at the termination of the infusion. Followingmeasurement of total urine volume and urine pH at each collection interval, analiquot of urine was collected (3-4 mL) and the remainder discarded. The aliquotwas then transferred to a new borosilicate culture tube with a PTFE-lined screwcap for storage. Fetal urine was collected cumulatively as described above,beginning one hour before dose administration and proceeding hourly until the endof the infusion period. Each hour the volume of urine collected was measured andthen returned into the amniotic cavity via an amniotic catheter except for a 2-3 mLaliquot. Following pH measurement, the fetal urine aliquots were stored in newborosilicate culture tubes with PTFE-lined screw caps until analysis. All sampleswere stored at -20 °C until analysis.46The volume of fetal blood collected during sampling was replaced, via thefetal tarsal vein, with an equal volume of maternal blood which had been collectedbefore beginning the experiment in 10 mL syringes containing a small amount ofheparin. A volume of blood equivalent to that removed during a sampling periodwas returned following the collection of samples for that interval (typically 8-9 mL).Control infusion experiments (with E2181 and E2241) were performed inthe same fashion as that described above with the exception of a substitution ofsaline for all drug solutions used.2.8 Data Analysis2.8.1 Calculation of in utero fetal weightsThe fetal weight at the time of experimentation was estimated using theequation listed below (Gresham et aL, 1972) and a knowledge of weight at birthand the time interval between the experiment and birth.Log (fetal weight in utero) = Log (birth weight) - 0.153 * (ND) (eq.1)where ND is the number of days between experiment and birth.2.8.2 Pharmacokinetic analysis - i.v. bolus studiesSerum drug concentration versus time data from the bolus administrationstudies were analyzed using PCNONLIN (version 4.1) (Statistical Consultants Inc.,Lexington, KY), a nonlinear least squares data fitting program, in order todetermine the appropriate model for fitting and to obtain estimates ofpharmacokinetic parameters. Based on the effect of various weighting factors onthe weighted sum of squares residuals, Akaike Information Criterion (AIC) and47residual plots for the fit, data points for individual subjects were weighted using thereciprocal of drug concentration. The computer fitting of the data was used toobtain estimates of rate exponents and constants (e.g., A, a, B, ). The formulaeused in the estimation of further pharmacokinetic parameters were obtained fromGibaldi and Perrier, 1982, unless otherwise indicated.The distribution and terminal elimination half-lives,t1 and tii, werecalculated from the relationshipstl/2a = 0.693/a andt1 = O.693/3, respectively.The areas under the serum concentration-time curves (AUC00)were calculated asthe sum of the area under the curve (AUC0t-last) to the last time point sampled (Ct..last) as determined by trapezoidal approximation, and the estimated area to infinity(AUCt..iast°°) as determined by the relationship Ct..iast43 i.e.,AUC0°= AuC0t-Iast + Cti5t/13.The areas under the first moment curves (AUMC00)were calculated in a fashionsimilar with the portion from Ct..i5tto infinity determined by (tlast* Ct..iast/ + C.iast’132) i.e.,AUMC0°= AuMc0t-last + tlast* Ct..last/ + Ct..iast/132).Total body clearance (CLT) and apparent volume of distribution (Vdarea) weredetermined fromCLT = dose / AUC0°andVdarea = dose / (*AUC00),respectively.Vdss was calculated from48Vdss = (dose*AUMG0°/ (AUC00)2),while the mean residence time (MRT) was determined byMRT = (AUMC0°I AUC00).28.3 Calculation of placental and nonpiacental clearancesThe abbreviations used to describe the Szeto model are listed on thefollowing lines along with the appropriate parameter from an experiment involvingan infusion of VPA to the ewe and an infusion of a stable isotope labelled form ofVPA, [13C4]VPA, to the fetus:M - maternal compartmentF - fetal compartmentCLmf - maternal to fetal clearanceCLfm - fetal to maternal clearanceCLmo - maternal nonplacental clearanceCLf0 - fetal nonplacental clearanceCLm- total maternal clearanceCLf- total fetal clearanceJ - rate of infusion to the mother (e.g., VPA infusion rate)J’ - rate of infusion to the fetus (e.g.,[13C4]VPA infusion rate)Cfss - steady-state drug concentration in the fetus duringinfusion to the mother (e.g., fetal arterial [VPA])Cmss - steady-state drug concentration in the motherduring infusion to the mother (e.g., maternal arterial [VPA])C’f55 - steady-state drug concentration in the fetusduring infusion to the fetus (e.g., fetal arterial [[13G4]VPA])C’mss - steady-state drug concentration in the motherduring infusion to the fetus (e.g., maternal arterial [[13C4]VPA]).49The equations used to calculate placental, nonpiacental and total body clearancesbased on the model illustrated in Fig. 3 are as follows:CLm = J / [CmssCfss(C’mss I C’)] (eq.2)CLf = J’ I [C’fssC’mss(Cfss I Cmss)] (eq.3)CLfm = CLm (C’mss I C’fss) (eq.4)CLmf = CLf (Cf55 / Cmss) (eq.5)CLmo = CLm- CLmf (eq.6)CLfoCLfCLfm (eq.7)The total body clearance at steady-state can also be calculated using theinfusion rate and steady-state serum drug concentration as follows:CLf= J’ / C’fss (eq.8)CLm = J / Cmss (eq.9).The fetal placental clearance, CLfm, can be calculated using the Fickmethod if both fetal arterial (C’f55) and umbilical venous (C’uss) steady-state drugconcentrations as well as umbilical flow (Qum) are available during fetal druginfusion. The equation used to calculate CLfm in this fashion is as follows:CLfm = um ((C’f55 - C’uss) I C’f) (eq.1O)502.8.4 Renal clearances - simultaneous infusion studiesThe maternal and fetal renal clearances (CLr) of VPA were estimated fromthe data from the simultaneous infusion studies. The equation used to calculaterenal clearance is as follows:CLr(AXu/At)/Cmid (eq.8)where X is the amount of drug excreted in urine, t is time and Cmid is the serumconcentration of drug at the midpoint of the urine collection period.2.8.5 Statistical analysisPhysiological data was analyzed statistically using repeated measuresanalysis of variance (ANOVA) followed by planned paired comparisons (T-tests)with a significance level of p < 0.05 using Statview (version 4.0) (Abacus ConceptsInc., Berkeley, CA). Results from ANOVA will be reported as an F ratio with thenumber of degrees of freedom in the treatment condition (i.e., J - 1) and thenumber of degrees of freedom for the error term (Le., (J-1)(k-1) where k is thenumber of subjects) reported in brackets, followed by the level of significance (e.g.,p value). T-tests or ANOVA (with paired comparisons if necessary) were used forother statistical analysis when applicable.513. RESULTS3.1 Assessment of GC-MS assay in sheep biofluidsThe GC-MS assay used to quantitate VPA and sixteen of its metabolites forthe studies presented here was developed in our laboratory, and validated inhuman biofluids by Dr. DienChen Yu (Yu etal., 1995). Before utilizing the assayfor these experiments, the accuracy and precision of the assay protocols were alsoverified in sheep biofluids (Yu eta!., 1995). Figures 4a - 4d illustratechromatograms of VPA and its metabolites extracted from sheep serum. Theretention times and m/z values of the (M-57) diagnostic ions for VPA and itsmetabolites are listed in Table 1. The same information for the internal standardsis provided in Table 2. As seen in Figures 4a - 4d, the chromatographic conditionsprovided resolution of VPA and its metabolites including all of the isomericunsaturated compounds. The calibration curves used for the quantitation of VPAand its metabolites in sheep biofluids showed good linearity over the concentrationranges investigated as demonstrated by correlation coefficients of 0.995 or greaterin all instances. The accuracy and precision of the assay was assessed byanalysis of multiple urine and serum samples which had been spiked with knownamounts of VPA and metabolites. Tables 3 and 4 summarize the results of theassessment of the method in sheep urine and serum, respectively. Using a signal-to-noise ratio of 3:1 as an indicator, the limits of quantitation (LOQ) for themetabolites ranged between 3 and 20 ng/mL in all of the biofluids studied. Table 5describes the reliability of the assay by demonstrating stability in the slopes ofcalibration curves which were generated by repeated analysis of a set of standardsamples prepared in sheep serum over a period of sixteen days. Coefficients ofvariation for these slopes did not exceed 10 % except those of 4-OH VPA, 4-ketoVPA and 5-OH VPA which were 14 %. While the accuracy and reproducibility ofthe assay was not tested in amniotic and fetal tracheal fluids as described abovefor serum and urine, the percentage relative difference between the concentrationsof calibration standards, calculated using their peak area with the resultingregression equation and the nominal concentration, was routinely checked forirregularities during analysis of such samples.5280000800007537000070000A2ABB6000060000UUNN5000050000D0 AA40000N40000NCC630000200002000030000-E1000010000_______0.1011.50-12.5013.5013.5014.5015.50-‘---rnr---‘‘TIME(mm.)TIME(mm.)Figure4a.SIMchromatogramsof blanksheepserum(lower trace)andtBDMSdenvativesofVPAandVPAmetabolitesinspiked,controlsheepserum(uppertrace).ThecompoundsdetectedandtheiOflSmonitoredareasfollows:1—4-OHVPA-y-Iactones(m/z100);2—VPA(m/z201);3—4-eneVPA(m/z199);4—(E)-3-ene VPA(m/z199);5=(Z)-3-eneVPA(m/z199);6=(Z)-2-ene VPA(m/z199);7=(E)-2-ene VPA(m/z199);Continuedonnextpage.01 c)1400001270000111312000014A60000ABU100000U50000NND8000040000AANNC60000C30000EE4000020000200001000080.1frL..Ir,J•16.5017.5018.5019.5020.5021.50-.*?-çTh4--TIME(mm.)TIME(mm.)Figure4b.SIMchromatogramSofblanksheepserum(lower trace)andtBDMSderivativesofVPAandVPAmetabolitesinspiked,controlsheepserum(uppertrace).Thecompoundsdetectedandtheionsmonitoredareasfollows:8=(Z)-2,4-diefleVPA(mlz197);9=(E)-2,4-dieneVPA(m/z197);10=(E,Z)-2,3’-diene (m/z 197);11=(E,E)-2,3’-diefle(m/z197);123-OHVPA(mlz217);13=4-ketoVPA(m/z215);14=3-ketoVPA(m/z329);Continuedonnextpage.C),1100005000001516181790000A400000BUU N70000300000D AAN50000N C200000C21EE300001910000010000200__---J_l0.&.__irL-22.022.422.813.0014.0015.0016.00t--fr_f---_.J____UL.___.TIME(mm.)TIME(mm.)Figure4c.SIMchromatogramsofblanksheepserum(lower trace) andIBDMSdenvativesofVPAandVPAmetabolitesinspiked,controlsheepserum(uppertrace).Thecompoundsdetectedandtheionsmonitoredareasfollows:15=5-OHVPA(m/z331);16=2-PSA(m/z331);17=2-PGA(m/z345);18[2H7]VPA(m/z208); 19=4-ene[2H7]VPA(m/z206);20—(Z)-2-ene[2H7]VPA(m/z206); 21=(E)-2-ene[2H7]VPA(m/z206);Continuedonnextpage.CiiCii1600012000014000100000AA12000B U80000N10000D22D25A60000A8000N24c236000E40000/E400020000------200019.4019.8020.2020.6021.2021.6022.00;-TIME(mm.)TIME(mm.)Figure4d.SIMchromatogramsofblanksheepserum(lower trace)andtBDMSdenvatlvesofVPAandVPAmetabolitesinspiked,controlsheepserum(uppertrace).Thecompoundsdetectedandtheionsmonitoredareasfollows:223-OH[2H7]VPA(m/z224);23—4-keto[2H7]VPA(mlz222);243-keto[2H7]VPA(m/z336);25=2-MGA(m/z317);26=5-OH[2H7]VPA(m/z338).01 a,57Table 3. Accuracy and precision of assay with respect to the analysis of twonominal concentrations of VPA and metabolites extracted from sheep urine.aCompound Nominal Mean Coefficient of Relativeconcentration measured variation difference b(j.tg/mL) concentration (%) (%)(jig/rn L)VPA 20.000 19.320 0.72 3.40.156 0.160 6.4 2.6(E)-2-ene VPA 6.736 6.722 0.82 0.210.053 0.050 7.9 5.7(Z)-2-ene VPA 1.264 1.353 2.1 7.00.010 0.009 7.0 10(E)-3-ene VPA 0.386 0.393 1.9 1.80.003 0.004 3.8 3.3(Z)-3-ene VPA 3.614 3.422 2.3 5.30.028 0.030 6.8 7.14-ene VPA 4.000 4.101 1.9 2.50.031 0.029 2.0 6.4(E)-2,4-diene 3.044 3.189 3.1 4.7VPA 0.024 0.028 8.6 17(Z)-2,4-diene 0.956 1.011 3.4 5.8VPA 0.007 0.010 11 43(E,E)-2,3’- 7.032 6.723 3.2 4.3diene VPA 0.055 0.055 7.4 7.3(E,Z)-2,3- 0.968 1.026 2.3 6.0diene VPA 0.008 0.009 9.8 143-keto VPA 4.000 4.115 2.8 2.90.031 0.029 7.5 6.44-keto VPA 2.000 2.067 3.1 3.40.016 0.018 4.0 123-OH VPA 4.000 3.746 3.3 6.40.03 1 0.032 11 3.24-OH VPA 4.000 3.93 1 5.2 1.70.03 1 0.034 11 9.75-OH VPA 4.000 3.889 1.6 2.80.031 0.032 8.8 3.22-PSA 2.000 2.125 4.5 6.20.016 0.018 10 172-PGA 2.000 2.035 4.7 1.80.016 0.017 6.9 5.7a Based on analysis of six samples of spiked blank sheep urine.b Relative difference is defined as the difference between the measured meanand nominal concentration of a sample divided by the nominal concentration.58Table 4. Accuracy and precision of assay with respect to analysis of two nominalconcentrations of VPA and metabolites extracted from sheep serum.aCompound Nominal Mean Coefficient of Relativeconcentration measured variation difference b(pg/mL) concentration (%) (%)(pg/m L)VPA 20.000 19.435 2.5 2.80.156 0.160 3.7 2.6(E)-2-ene VPA 6.736 6.776 3.2 0.590.053 0.055 8.7 3.8(Z)-2-ene VPA 1.264 1.295 0.9 2.40.010 0.009 12 10(E)-3-ene VPA 0.386 0.396 3.3 2.60.003 0.004 5.2 33(Z)-3-ene VPA 3.614 3.214 6.8 1 10.028 0.030 8.6 7.14-ene VPA 4.000 4.020 2.7 0.50.031 0.029 5.5 6.4(E)-2,4-diene 3.044 3.168 2.0 4.1VPA 0.024 0.020 19 17(Z)-2,4-diene 0.956 1.016 2.2 6.3VPA 0.007 0.006 9.6 14(E,E)-2,3’- 7.032 6.838 1.5 2.8diene VPA 0.055 0.06 1 5.9 11(E,Z)-2,3’- 0.968 1.024 2.4 5.8diene VPA 0.008 0.008 2.4 03-keto VPA 4.000 4.110 2.1 2.80.031 0.030 3.1 3.24-ketoVPA 2.000 1.965 1.0 1.80.016 0.015 11 6.23-OH VPA 4.000 4.134 4.0 3.40.031 0.033 6.3 6.44-OH VPA 4.000 4.102 5.5 2.60.031 0.037 15 195-OH VPA 4.000 4.201 2.5 5.00.031 0.032 6.5 3.22-PSA 2.000 1.862 2.0 6.90.016 0.016 10 02-PGA 2.000 1.915 9.5 4.20.016 0.017 11 6.2a Based on analysis of seven samples of spiked blank sheep serum.b Relative difference is defined as the difference between the measured meanand nominal concentration of a sample divided by the nominal concentration.59Table 5. Assessment of inter-assay variability with respect to analysis of VPA andmetabolites extracted from serum.aCompound Mean slope Coefficient of_______________variation (%)VPA 5.104 2.03(E)-2-ene VPA 0.829 7.81(Z)-2-ene VPA 2.76 8.25(E)-3-ene VPA 0.897 7.92(Z)-3-ene VPA 0.891 8.324-ene VPA 1.023 7.67(E)-2,4-diene 6.244 6.72VPA(Z)-2,4-diene 8.634 7.74VPA(E,E)-2,3’- 1.511 7.04diene VPA(E,Z)-2,3’- 1.723 7.90diene VPA3-keto VPA 0.194 5.454-keto VPA 0.264 1 1.13-OH VPA 1.812 2.004-OH VPA 34.249 10.25-OH VPA 0.353 14.02-PSA 1.469 4.232-PGA 1.652 11.1a Determined by an eight-fold replicate analysis of a set of seven standardsamples prepared in serum.603.2 Nonpregnant animal bolus studiesStudies were conducted in three nonpregnant sheep to examine the effectof isotope labelling on VPA pharmacokinetics using the[13C4]VPA analogue. Theindividual weights for ewes #105, 617 and 1158 were 50.8, 59.9 and 53.5 kg,respectively, so that the mean weight of the animals in this study was 54.7 ± 4.7kg. Femoral arterial blood and urine were collected from all animals as previouslydescribed.3.2.1 Comparison of parent drug dispositionFigure 5 illustrates a representative disposition profile for VPA and[13C4]VPA in serum from a ewe (E1158) following the administration of a 1000 mgdose of 50% VPA / 50% [13C4]VPA (18.4 mg kg1 based on the mean animalweight). VPA reached maximal serum concentrations of approximately 90 .tgImLat the earliest sampling time following administration of the dose i.e, 2 mm. Thedisposition of VPA in serum was best described by a biexponential function of theform Aet + BeI. The terminal elimination half-lives,t11 for VPA and[13C4]VPA from serum were 3.5 ± 0.4 h and 3.2 ± 0.4 h, respectively, and werenot significantly different (p>0.05). The mean VPA / [13C4]VPA area under theserum concentration - time curve ratio (AUC,vpA/AUC,[13c]v )wasdetermined to be 1.03 ± 0.02. As denoted in Table 6, equivalent amounts of eachparent compound were recovered from urine in both conjugated and unconjugatedforms.-JEC,C0Ct’Ca)0C00C)D061100.0010.001.000.l0—a--— VPA.•... [13C4]VPA0.01I I I I I0 5 10 15 20 25Time (hours)Figure 5. Representative concentration versus time profiles for [13C4]VPA andunlabelled VPA in serum from a nonpregnant ewe following i.v. bolusadministration of 50% VPA /50% [13C4]VPA (1000 mg dose).62Table 6. Percentage (%) of dose excreted as parent drug in urine following i.v,bolus administration of a 1000 mg dose of 50% VPA / 50% [13C4]VPA to anonpregnant ewe (n=3).Unconjugated Conjugated TotalVPA 9.2 ± 0.7 37.8 ± 1.1 47.1 ± 0.3[13C4]VPA 8.5 ± 0.9 39.2 ± 1.2 47.7 ± 0.33.2.2 Comparison of metabolism of parent compoundsThe 11 metabolites of VPA detected in these animals following i.v. bolusdosing were the following: (E) and (Z)-2-ene VPA, (E) and (Z)-3-ene VPA, 4-eneVPA, 3-keto VPA, 4-keto VPA, 3-OH VPA, 4-OH VPA, 5-OH VPA and 2-PGA.The same metabolites were also detected in [13C]-labelled form suggestingsimilar metabolic mechanisms for the two compounds. As an example of thechromatographic results obtained, Figures 6a and 6b illustrate ion chromatogramsfor five metabolites of VPA detected in urine collected from a nonpregnant ewefollowing i.v. bolus administration of VPA (50 % VPA, 50 % [13C4]VPA). The ionpeaks for the13C4..analogues of these metabolites were virtually superimposableon the illustrated peaks in most instances.Tables 7 and 8 illustrate that not only were the same metabolites producedfrom the two parent compounds, but these products were produced to a similardegree. Table 7 demonstrates that the peak area ratio of unlabelled metabolite tolabelled metabolite in serum at each sampling time is repeatedly close to unity,thus suggesting similar dispositions for the compounds in serum. The recoveryratio of unlabelled metabolite to labelled metabolite from urine for each metaboliteillustrated in Table 8, also suggests similar dispositions for the labelled and63unlabelled compounds. Thus, [13C4]VPA appears to be bioequivalent to VPA inthe sheep model based on comparisons of AUCOO values, terminal elimination half-lives and recovery of metabolites.Table 7. Time-averaged peak area ratios from serum for the most prominentmetabolites detected following i.v. bolus administration of 50% VPA /50%[13C4]VPA (1000 mg dose).Metabolite Peak area ratio4-ene VPA / 0.97 ± 0.224-ene[13C4]VPA(E)-3-ene VPA / 1.02 ±0.18(E)-3-ene[13C4]VPA(E)-2-ene VPA/ 1.14±0.18(E)-2-ene[13C4]VPA3-ketoVPA/ 1.09±0.133-keto[l3C4]VPA4-OH VPA / 1.07 ± 0.314-OH[l3C4]VPA5-OH VPA/ 1.18±0.105-OH[l3C4]VPA2-PGA/ 1.29±0.172[13C4]PGA1200001100001000008000080000700006000050000200001000080000A 80000B 70000UN 60000N 40000CE 3000010000642019.80 20.000TIME (mm.)Figure 6a. SIM chromatograms of tBDMS derivatives of some representative VPAmetabolites collected in urine from a nonpregnant ewe following i.v. bolusadminstration. SIM chromatograms of tBDMS derivatives of some internalstandards are also shown. The compounds presented on this page are as follows:1 = 4-keto[2H7]VPA (m)z 222); 2 = 4-keto VPA (m/z 215); and 3 = 3-keto VPA(m/z 329). Continued on the next page.6512000110010000CAB7 4000070000,8 IN 3500050000.50000 \\ 25000\\ E30000 15000:j____ 10000TIME (mm.)Figure 6b. Continued from previous page. SIM chromatograms of tBDMSderivatives of some representative VPA metabolites collected in unne from anonpregnant ewe following Lv. bolus adminstration. SIM chromatograms oftBDMS derivatives of some internal standards are also shown. The compoundspresented on this page are as follows: 4 = 3-OH[2H7]VPA (m/z 224); 5 = 3-OHVPA (m/z 217); 6 = 5-OH[2H7]VPA (mlz 338); 7=5-OH VPA (m/z 331); 8 = 4-OHVPA—y-lactones (mlz 100).66Table 8. Ratio of amount of unlabelled VPA metabolite to amount of [13C4]VPAmetabolite recovered from urine for the most prominent metabolites following i.v.bolus administration of 50% VPA / 50% [13C4]VPA (1000 mg dose).Metabolite Metabolite ratio(E)-2-ene VPA / 0.90 ± 0.10(E)-2-ene[13C4]VPA3-keto VPA / 1.08 ± 0.023-keto[l3C4]VPA4-ketoVPA/ 1.16±0.314-keto[l3C4]VPA5-OH VPA/ 1.13±0.145-OH[l3C4]VPA2-PGA/ 1.13±0.042[13C4]PGA3.3 Maternal i.v. bolus studiesMaternal i.v. bolus studies were performed in five chronically instrumentedpregnant ewes in order to determine basic pharmacokinetic parameters andestimate fetal exposure to VPA. Table 9 lists experimental details including theidentification of the ewes on which maternal i.v. bolus experiments wereperformed, the maternal weight, the number of fetuses present and the fetalgestational age at the time of experiment. Maternal weight averaged 78.2 ± 6.7kg, while the mean gestational age at the time of experiment was 128.2 ± 2.7 days(term -‘145 days) . Twin fetuses were present in each case during this study.67Table 9. Information for experimental preparations used in the maternal i.v. bolusadministration studies.Ewe Maternal Number of Gestational Term Samplesnumber Weight fetuses age at expt. (days) collected*(kg) (days)54 82.0 2 127 137 MA, FA,AMN, TR1118 87.5 2 130 138 MA, FA,AMN, TR313 69.9 2 126 136 MA, FA,AMN1242 75.5 2 126 138 MA, FA,AMN,_TR2164 77.2 2 132 137 MA, FA,AMN, TRmean ± s.d. 78.2 ± 6.7 2 128.2 ± 2.7* MA = maternal femoral arterial bloodFA = fetal femoral arterial bloodAMN = amniotic fluidTR = fetal tracheal fluidFigure 7 illustrates a representative disposition profile for VPA in maternaland fetal serum as well as amniotic and fetal tracheal fluid, from a pregnant sheep(E54) following administration of a 1000 mg i.v. bolus dose to the ewe. VPAreached maximal concentrations of approximately 100 p.g/mL in the mother at theearliest sampling time following administration of the dose i.e., 2 mm. Thedisposition of VPA in maternal serum was best described by a biexponentialfunction of the form Aet + Be. The maternal pharmacokinetic parameterscalculated for individual animals from these data as well as mean values arepresented in Table 10. The mean terminal elimination half-life, for VPA inmaternal serum was 2.13 ± 0.49 h while the mean residence time (MRT) was 2.4868± 0.30 h. The weight normalized total body clearance (CLT) for the ewe was 0.083± 0.027 L/h/kg. The volume of distribution calculated (Vdarea) was 0.242 ± 0.036LIkg, while the nonparametric estimate of the same parameter (Vd) was 0.201 ±0.047 L/kg.VPA appeared very rapidly in fetal serum, with concentrations ranging from10-50 pg/mL at 2 mm following dose administration (see Figure 7). After thistime, drug levels declined in a fashion parallel to that observed in maternal serum.These serum levels resulted in significant fetal exposure to VPA as determined bythe ratio of the area under the serum concentration- time curves for fetal andmaternal serum (i.e., mean AUCQOFA/AUCOOMA = 0.410 ± 0.118). VPA alsoappeared rapidly in amniotic and fetal tracheal fluid however the concentrationsobserved in these fluids did not rise above 3 tg/mL at any time.Table 11 lists the eleven metabolites of VPA detected in maternal serumfollowing maternal drug administration. Of these, only the metabolites detected inthe highest concentrations in maternal serum were detected in fetal serum he., (E)2-ene VPA, 4-ene VPA, 3-keto VPA, 4-OH VPA and 5-OH VPA, and for thesemetabolites, maternal serum concentrations were higher than those in the fetus.The fetal blood gas, acid-base and metabolic parameters monitored (i.e.,P°2, pCO2 pH, BE, [HCO3i, [Hb], 02 sat., lactate and glucose levels), were allwithin the normal range for an ovine fetus prior to drug administration (see Table12). No physiologically significant changes were observed in these parametersover the course of the experiments. Furthermore, no significant changes werenoted in either fetal heart rate (F(71 ,213) = 1.182, p >0.05) or arterial pressure(F(71 ,21 3) = 0.690, p >0.05) following drug administration compared to preexperimental values (see Table 12). There was, however, a significant decreasein fetal breathing activity in the first hour following maternal drug administration(F(5,10)=11.848, p.<0.05) (see Table 13).691000100-J____E0)10011.01Figure 7. Representative VPA concentration versus time profiles for maternal andfetal serum as well as amniotic and fetal tracheal fluids following maternal (E54)Lv. administration of VPA (1000 mg dose).—0--- maternal serumfetal serumamniotic fluidfetal tracheal fluid0 2 4 6 8 10 12Time (hours)Table10.Pharmacokineticarametersintheeweestimatedfromthematernal i.v.bolusadministrationexperiments.Ewenumber54313111812422164mean±s.d.weight(kg)82.069.987.575.577.278.2±6.7A(ig/mL)77.3655.96100.5445.95132.0982.38±34.83B(ig/mL)37.2650.7742.9037.4057.4245.15±8.80cx(h-i)2.0649.94717.833.15811.638.926±6.481(h-i)0.34250.34360.23380.41650.34990.3373±0.0656t112(h)2.0232.0172.9641.6641.9812.130±0.490AUCOC145.3177.5218.3105.3179.3165.1±42.3(ug_h_mL-1) AUMCOO317.4460.4629.2230.8460.2419.6±(righ2mL-i)152.7MRT(h)2.1842.5942.8822.1922.5672.483±0.297Vd(Lkgi)0.18330.20910.15090.27570.18550.201±0.047Vdarea0.24510.23460.22390.30200.20650.242±(Lkg1)0.036CLT0.083950.080610.052350.12580.072250.083±(Lhikg-i)0.027AUCOOFA?0.3920.4020.2920.6070.3590.410±AUCOOMA0.118—.1071Table 11. The VPA metabolites detected in maternal serum following maternal i.v.bolus administration of VPA along with the range of peak concentrations (C max)typically observed in the five animals and the time following drug administration atwhich peak levels were observed (t max)VPA metabolite C max max(igImL) (h)(E)-2-ene VPA 0.1-0.12 2-3(Z)-2-ene VPA <0.01 0-1(E)-3-ene VPA 0.04-0.06 2-3(Z)-3-ene VPA 0.01 -0.02 2-34-ene VPA 0.02-0.05 1-23-ketoVPA 0.1-0.12 1-24-keto VPA 0.03-0.06 1-23-OH VPA 0.01 -0.02 1-24-OH VPA 0.2-0.4 2-35-OH VPA 0.08-0.1 1 1-22-PGA 0.03-0.05 2-372Table 12. Fetal blood gas, acid-base, metabolic and cardiovascular parameters(mean ± s.d.; n=5) prior to maternal bolus administration of VPA.Parameter Mean (± sd) controlvaluespH 7.350±0.028P°2 (mm Hg) 18.8 ±3.3pCO2 (mm Hg) 50.4 ± 4.4BE (meg/L) 1 1.4 ± 2.4[HC03] (meg/L) 26.6 ± 2.5[Hb] (g/dL) 2 10.25 ± 1.2102 sat. (%)3 43.4± 11.8[lactate] (mM) 1.04± 0.14[glucose] (mM) 0.71 ± 0.24heart rate (bpm) 159 ± 5arterial pressure (mm Hg) 52 ± 21. Whole blood base excess2. Haemoglobin concentration3. Oxygen saturation73Table 13. Fetal breathing movements monitored during maternal bolus doseadministration of VPA (1000 mg) experiments.Time Breathing activity(mean ± sd)(mm h1)control period* 33.45 ±1.150-1 h post dose 14.25 ± 4.75#1-2h 33.0±4.02-3 h 25.5 ± 5.53-4 h 37.5 ± 6.54-5h 36.0±2.0* Refers to the averaged pre-experimental breathing activity over 6hours for 3 animals.# Significantly different from the mean control value as determinedby paired comparisons carried out following a repeated measuresANOVA (F(5,1 0)=1 1.848, p<0.05).743.4 Fetal i.v. bolus studiesFetal i.v. bolus studies were performed in five chronically instrumentedpregnant ewes in order to determine basic fetal pharmacokinetic parameters forVPA. Table 14 lists experimental details including the identification of the ewes onwhich fetal i.v. bolus experiments were performed, the maternal weight, thenumber of fetus present, the fetal gestational age at the time of experiment, fetalweight at birth and fetal weight at the time of experiment. Maternal weightaveraged 76.8 ± 7.4 kg while the mean fetal weight at the time of the experimentwas 2.319 ± 0.344 kg. The mean gestational age at the time of experiment was129±5 days (term -145 days).Figure 8 illustrates a representative disposition profile for VPA in maternaland fetal serum as well as amniotic and fetal tracheal fluid from a pregnant sheep(E499) following administration of a 250 mg i.v. bolus dose to the fetus. VPAreached maximal concentrations of approximately 100 jig/mL in the fetus at theearliest sampling time following administration of the dose i.e., 2 mm. Thedisposition of VPA in fetal serum was best described by a biexponential function ofthe form Ae-at + Be. The fetal pharmacokinetic parameters calculated forindividual animals from these data as well as mean values are presented in Table15. The data from one of the five animals used in this portion of the study (El 242)was erratic and inconsistent with data from the other animals and was thus omittedfrom further pharmacokinetic analyses. The mean terminal elimination half-life,t12, for VPA in fetal serum was 3.37 ± 1.37 h while the mean residence time(MRT) was 4.52 ± 2.17 h. These values are not significantly different from thoseobserved in the ewe following maternal i.v. bolus dosing (p>0.05). The weightnormalized total body clearance (CLT) for the fetus was 0.529 ± 0.306 L/h/kg, avalue that was significantly higher than that observed in the ewe on a weight75normalized basis (p<0.05). The volume of distribution calculated (Vdarea) was 5.27± 2.84 L/kg, while the nonparametric estimate of the same parameter (Vd) was5.03 ± 3.32 L/kg.Table 14. Information regarding experimental preparations used in the fetal i.v.bolus administration studies.Ewe Maternal Gestational Term Fetal Fetal Number Samplesweight age at the (days) weight weight at of collected*(kg) time of at birth the time of fetusesexperiment (kg) experiment(days) (kg) #499 81.2 129 140 2.496 1.755 3 MA,FA,AMN, TR1136 71.7 123 136 3.538 2.238 1 MA, FA,AMN, TR1118 87.5 136 138 2.646 2.466 2 MA, FA,AMN313 69.9 130 136 3.22 2.606 2 MA, FA,AMN1242 73.5 130 138 3.353 2.529 2 MA, FA,AMN, TRmean 76.8± 129±5 138± 3.051± 2.319± 2± s.d. 7.4 2 0.455 0.344# Calculated using: log (experimental weight) = log (birth weight) - (0.0153)(# ofdays between experiment and birth) (Gresham et al., 1972).* MA = maternal femoral arterial bloodFA = fetal femoral arterial bloodAMN = amniotic fluidTR = fetal tracheal fluid761000-— Maternalserum100Fetal serum-JAmniotic fluid10 .. O— Fetal tracheal fluid.01—. • I • I • I • I • I0 2 4 6 8 10 12Time (hours)Figure 8. Representative VPA concentration versus time profiles for maternal(E499) and fetal serum as well as amniotic and fetal tracheal fluids following fetali.v. administration of VPA (250 mg dose).Table15.PharmacokineticparametersinthefetusestimatedfromfetalLv.bolusadministrationexperiments.Ewenumber313499111811361242mean±s.d.*fetalweight(kg)2.6061.7552.4662.2382.5292.266±0.372A(p.g/mL)72.38121.12127.959.5774.4895.24±34.31B(pgImL)22.2211.63015.2588.8025.19914.48±5.80cc(h1)7.7134.1521.8763.6542.8624.349±2.446I(h-i)0.14100.17680.40620.23540.04550.23990.1175t1126(h)4.9123.9201.7062.94415.2383.371±1.370AUCOO192.6106.1121.661.40140.1120.4±54.5(ug_h_mL1) AUMCOO1277.7561.4185.0285.31472.6577.3±493.3(.tg_h2_mL-1) MRT(h)6.6345.2911.5214.64710.514.523±2.166Vd(Lkg-i)3.3047.1041.2688.4547.4165.032±3.324Vdarea3.7147.5932.0527.72915.5075.272±2.841(Lkg-i)CLT0.20090.76490.33800.81 300.27900.5292±0.3057(L_h-ikg-i)AUCMA/0.4530.7290.6630.3830.8920.557±0.165AUCOOFATrachealfluidN/A0.052±0.066±0.022±0.042±0.046±0.022conc./fetalserum0.0780.0960.0160.031conc.([TR]/LEA])Amnioticfluid0.518±0.100±0.093±0.080±0.300±0.198±0.214conc./fetalserum0.2800.0670.0870.0410.254cone.([AM]/[FA])Mean±s.d.doesnotincludedatafromE1242(i.e.,n=4).—.178VPA appeared very rapidly in maternal serum as well as amniotic and fetaltracheal fluid following fetal dosing. In most instances, the maternal serum levelswere roughly equivalent to fetal serum levels by 2 h following dosing and remainedso for the remainder of the sampling period. Once again, the amniotic and fetaltracheal fluid concentrations remained low and did not increase extensively asdenoted in the time-averaged concentration ratios of VPA in these fluids relative tofetal serum (i.e., [TR]/ [FA] and [AM] / [FA]), which were 0.198 ± 0.214 and 0.046± 0.022 for amniotic and fetal tracheal fluids, respectively.The concentrations of metabolites detected in maternal and fetal serumfollowing fetal drug administration were substantially lower and more erratic thanthose detected following maternal administration, as would be expected followingadministration of a lower dose. The following metabolites were detected in bothmaternal and fetal serum collected during the first 6 hours following drugadministration although higher levels were typically observed in maternal serum:(E)-2-ene VPA, (E)-3-ene VPA, 3-keto VPA, 4-keto VPA, 4-OH VPA, 5-OH VPAand 2-P GA.The fetal blood gas, acid-base and metabolic parameters monitored (i.e.,P°2, pCO2 pH, BE, [HC03], [Hb], 02 sat., lactate and glucose levels), were allwithin the normal range for an ovine fetus prior to drug administration (see Table16). No physiologically significant changes were observed in these parametersover the course of the experiments. While there was a statistically significantincrease (F(7,28) = 2.500, p<0.05) in lactate concentration for approximately 90minutes following bolus administration, this increase, of approximately 0.22 mM, isprobably of limited physiological importance (see Figure 9). Once again, neitherfetal heart rate (F(71 ,284) = 0.774, p >0.05) nor arterial pressure (F(71 ,284) =0.869, p >0.05) were significantly affected by VPA administration (see Table 16).There was a significant decrease in fetal breathing activity lasting for79approximately two - three hours following fetal drug administration (F(8,1 6)=2.965,p<0.05) (see Table 17).Table 16. Status of fetal blood gas, acid-base, metabolic and cardiovascularparameters (mean ± s.d.; n=5) prior to fetal bolus administration of VPA.Parameter Mean (± Sd) controlvaluespH 7.33 1 ± 0.050P°2 (mm Hg) 19.8±3.1pCO2 (mm Hg) 48.0 ± 6.3BE (meg/L) 1 0.6 ± 2.9[HC03] (meg/L) 25.6 ± 2.6[Hb](g/dL)2 11.08± 2.85°2 sat. (%) 45.7 ± 6.3[lactate] (mM) 1.15± 0.28[glucose] (mM) 0.81 ± 0.17heart rate (bpm) 172 ± 6arterial pressure (mm Hg) 52 ± 31. Whole blood base excess2. Haemoglobin concentration3. Oxygen saturation80(‘:3Time (hours)Figure 9. Fetal femoral arterial lactate concentrations (mean ± Sd) monitoredduring fetal bolus dose administration (250 mg) experiments. The asterisk (*)indicates a statistically significant difference from control values (p<0.05).81Table 17. Fetal breathing movements monitored during fetal bolus doseadministration (250 mg) experiments.Time Breathing activity(mean ± sd) (mm h-i)control period* 28.2 ± 0.60-1 h post dose 7.7 ± 0.6#1-2h 19.3±152-3h 12.0±6.2#3-4 h 25.3 ± 14.64-5h 30.3±11.75-6h 19.3±5.16-7 h 30.3 ± 4.77-8h 31.7±11.5* Refers to the averaged pre-experimental breathing activity over 6hours for 3 animals.# Significantly different from the mean control value as determinedby paired comparisons carried out following a repeated measuresANOVA (F(8,16)=2.965, p<0.05).823.5 Simultaneous infusion studiesSimultaneous i.v. infusion studies were performed in five chronicallyinstrumented pregnant ewes in order to further assess maternal and fetal placentaland nonplacental drug clearances and drug biotransformation processes, as wellas the effects of VPA on the fetus. Table 18 lists experimental details including theidentification of the ewes on which simultaneous infusion experiments wereperformed, the maternal weight, the number of fetus present, the fetal gestationalage at the time of experiment, fetal age at term, fetal weight at birth and fetalweight at the time of experiment. Maternal weight averaged 74.2 ± 6.4 kg whilethe mean fetal weight at the time of the experiment was 2.170 ± 0.232 kg. Themean gestational age at the time of experiment was 127 ± 5 days (term ‘-145days).3.5.1 VPA and[13C4]VPA disposition in serumFigure 10 illustrates a representative disposition profile for VPA and[13C4]VPA in maternal and fetal serum from a pregnant sheep (E2181) followingsimultaneous administration of VPA and [13C4]VPA to the ewe and fetus,respectively. VPA and [13C4]VPA levels were at steady-state after 3 hours ofdrug administration in all cases (he., the slopes of the concentration vs. time databetween 3 and 6 hours were not statistically different from 0, p>0.05). The meansteady-state serum concentrations for VPA and [13C4]VPA were 69.42 ± 12.16and 18.63 ± 2.72 .tg/mL, respectively, in maternal femoral arterial serum, and38.96 ± 12.74 and 35.59 ± 10.41 p.g/mL respectively, in fetal femoral arterial serum(see Table 19). The ratios of VPA and[13C4]VPA concentrations in fetal serum tothat in maternal serum (i.e., Cfss / Cmss) were 0.55 ± 0.11 and 1.90 ± 0.38,83Table 18. Information regarding experimental preparations used in thesimultaneous infusion studies.Ewe Maternal Gestational Term Fetal Fetal Number Samplesweight age at the (days) weight weight at of collected*(kg) time of at birth the time of fetusesexpt. (days) (kg) expt.(kg) #313 69.9 123 136 3.220 2.036 2 MA, FA,AMN2101 81.6 135 141 2.927 2.369 1 MA, FA,AMN, TR,MUR, FUR2177 74.8 125 135 2.873 2.020 1 MA, FA, CAAMN, TR,MUR, FUR2181 65.8 128 134 2.418 1.957 2 MA, FA, UVAMN, TR,MUR, FUR2220 78.9 126 127 2.558 2.470 1 MA, FA,CA, UV,AMN, TR,MUR, FURmean 74.2± 127±5 135± 2.799± 2.170±±s.d. 6.4 5 0.317 0.232# Calculated using: log (experimental weight) = log (birth weight) - (0.01 53)(# ofdays between experiment and birth) (Gresham eta!., 1972).* MA = maternal femoral arterial bloodFA = fetal femoral arterial bloodCA = fetal carotid arterial bloodUV = umbilical venous bloodAMN = amniotic fluidTR = fetal tracheal fluidMUR = maternal urineFUR = fetal urine841000-VPA - maternal arteryVPA - fetal artery[13C]VPA - fetal artery-JE----- [‘3C]VPA - maternal arteryI> (N-.10— I I I I I0 1 2 3 4 5 6Time (hours)Figure 10. Representative VPA and[13C4]VPA serum concentration versus timeprofiles from ewe 2181 during simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus.85Table 19. Mean steady-state concentrations for VPA and [13C4]VPA in maternaland fetal femoral arterial serum obtained during the simultaneous infusion of VPAto the ewe and [13C4]VPA to the fetus.Ewe VPA - MA * VPA- FA * [13C4]VPA - FA* [13C4]VPA - MA *number (ig/mL) (pg/mL) (pgImL) (.tg/mL)E313 75.94 45.85 27.68 17.52E2101 58.92 28.07 39.09 17.70E2177 80.85 39.73 52.39 22.42E2181 77.52 56.03 29.37 20.15E2220 53.85 25.12 29.44 15.35mean ± sd 69.42 ± 12.16 38.96 ± 12.74 35.59 ± 10.41 18.63 ± 2.72* MA = maternal femoral arterial bloodFA = fetal femoral arterial bloodrespectively. Serum samples were also collected from the umbilical vein (UV) andfetal carotid artery (CA) when possible (see Table 18). Figure 11 illustrates arepresentative disposition profile for VPA and [13C4]VPA in the fetal femoral andcarotid arteries from E21 77 during the steady-state portion of the infusion. In bothanimals from which carotid samples were collected,[13C4]VPA serumconcentrations did not differ significantly between the two arteries during this timeperiod (mean [CA] = 47.2 jig mL1,mean [FA] = 41.6 jig mLl)(p>0.05). Althoughdifferences in unlabelled VPA concentrations between the two arteries were small,they did vary significantly from each other over the final three hours of the infusionas determined by paired comparisons (mean [CA] = 43.1 jig mL1, mean [FA] =34.7 jig mL1)(p<0.05). In a similar fashion, umbilical venous samples were86collected from two animals (see Table 18). Figure 12 illustrates a representativedisposition profile for VPA and [13C4]VPA in the fetal femoral artery and umbilicalvein from E21 81 during the simultaneous infusion. In both animals from whichumbilical venous samples were collected, neither VPA nor [13C4]VPA serumconcentrations differed significantly between the two vessels during this timeperiod (p>0.05).The protein binding of VPA and [13C4]VPA were measured in maternal andfetal arterial serum samples collected during the steady-state portion of theinfusion period as described earlier. The mean percentage of drug bound inmaternal serum was 49.7 ± 14.6% and 62.0 ± 9.1 %for VPA and[13C4]VPA,respectively. In fetal serum, the mean percentage of drug bound was 45.3 ± 8.1 %and 55.9 ± 7.4 % for VPA and[13C4]VPA, respectively. The mean steady-stateserum concentrations for unbound VPA and [13C4]VPA were 39.41 ± 6.11 and7.08 ± 1.03 pg/mL, respectively, in maternal femoral arterial serum, and 21.31 ±6.97 and 15.70 ± 4.59 pg/mL respectively, in fetal femoral arterial serum. Theratios of unbound VPA and [13C4]VPA concentrations in fetal serum to that inmaternal serum (i.e., Cfssu I Cmssu) were 0.60 ± 0.12 and 2.20 ± 0.44,respectively.871000-VPA-femoral arteryD[13C4]VPA-femoral artery----0---- VPA-carotid artery[13C4]VPA-carotid artery100-10— I I I I2 3 4 5 6Time (hours)Figure 11. Representative VPA and [13C4]VPA serum concentration versus timeprofiles in the fetal femoral and carotid arteries (E2177) during simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus.1000-JEC)10010Time (hours)VPA-femoral artery‘‘D[13C4]VPA-femoral artery----o---- VPA-umbilical vein[13C4]VPA-umbilical vein88Figure 12. Representative VPA and [13C4]VPA serum concentration versus timeprofiles for the umbilical vein and fetal femoral artery (E21 81) during simultaneousinfusion of VPA to the ewe and [13C4]VPA to the fetus.0 1 2 3 4 5 6893.5.2 Placental and nonplacental drug clearancesThe steady-state serum levels obtained in these studies were used tocalculate both maternal and fetal placental as well as nonplacental clearances asdescribed earlier (Section 2.9.3). Table 20 summarizes the clearance valuescalculated via the Szeto method from total (i.e., bound + unbound) serum druglevels from the simultaneous infusion experiments. The maternal total bodyclearance (CLm) calculated in this way was 3.6 ± 0.9 mL mirn1 kg1 which is higherthan, but comparable to, the CLm calculated using solely the infusion rate andsteady-state drug serum levels (2.6 ± 0.5 mL mirn1 kg-i). In a similar fashion, thefetal total body clearance (CLf) calculated via the Szeto method (63.0 ± 22.8 mLmm-1 kg-i) was larger than, but also comparable to, the value calculated from theinfusion rate and steady-state level (44.1 ± 9.4 mL mirn1 kg-i). The fetal ability toclear VPA across the placenta, represented by CLfm, was found to be entirelyresponsible for the fetal total body clearance of VPA (CLf) (t(8) = 1.616, p>0.05),while the maternal placental clearance accounted for approximately 30 % of thematernal total body clearance. Also, the fetal -to- maternal placental clearance,CLfm, was significantly larger than the drug clearance in the opposite direction,CLmf (t(8) = 5.934, p<0.05). Placental clearance values (CLfm) can also becalculated using the Fick method in animals in which both umbilical venous flowand steady-state drug concentrations are available, The mean CLfm calculated inthis fashion from the two animals with appropriate data, 20.5 ± 13.4 mL mirn1 kg-1,was substantially lower than the mean calculated via the Szeto method (67.3 ±22.5 mL mirn1 kg-i). Using the protein binding data reported above, the placentaland nonplacental clearances of free VPA in the ewe and fetus were alsocalculated. The results are reported in Table 21. The clearance values calculatedwith unbound drug data were roughly twice as large as the values calculated with90total drug data. The contribution made to total body clearance by placentalclearance was the same as that calculated using total drug concentrations i.e.,approximately 30 % of maternal total body clearance was accounted for byplacental clearance while 100 % of the fetal total body clearance was attributed toplacental clearance. Also, the fetal -to- maternal placental clearance, CLfm, wassignificantly larger than the drug clearance in the opposite direction, CLmf(p<0.05).3.5.3 VPA and[13C4]VPA disposition in fetal tracheal and amniotic fluidsBoth VPA and [13C4]VPA were detected in amniotic and fetal trachealfluids during the course of the infusions. Figure 13 illustrates a representativeconcentration- time profile for VPA and [13C4]VPA in amniotic and fetal trachealfluid from E21 77 during the simultaneous infusion period. The mean levelsobserved in amniotic fluid during the final three hours of the infusion period were9.6 ± 4.4 pg/mL and 8.3 ± 2.9 pg/mL for VPA and [13C4]VPA, respectively asreported in Table 22. The mean concentrations observed in fetal tracheal fluidduring the final three hours of the infusion were 4.2 ± 4.7 .tg/mL and 2.2 ± 1.3.tg/mL for VPA and [13C4]VPA, respectively. Although the levels observed in fetaltracheal fluid typically appeared to be lower than those observed in amniotic fluid,no statistically significant differences were observed (p>0.05). There was noevidence of the presence of conjugated VPA or [13C4JVPA in either amniotic orfetal tracheal fluids.91Table 20. Total body, placental and non-placental clearance parameterscalculated from serum steady-state total drug concentrations achieved during thesimultaneous infusion of VPA and [13C4]VPA to the ewe and fetus, respectively.The values in parentheses refer to the percentage of the total body clearance (Le.,CLm or CLf) accounted for by the reported value.Ewe Clearancenumber (mL mirn1)CLm CLf CLfm CLmf CLmo CLf0Ewe313 193 133 122 80 113 10(92%) (41 %) (59%) (8%)Ewe 2101 308 116 140 55 253 -23(100%) (18%) (82%)Ewe 2177 205 80 88 39 166 -9(100%) (19%) (81%)Ewe 2181 294 194 202 140 154 -8(100%) (48%) (52%)Ewe 2220 338 155 176 72 266 -22(100%) (21%) (79%)mean 268 135 146 77 190 -10(mLminl) (100%) (29%) (71%)mean ± sd 3.6 ± 63.0 ± 67.3± 37 ± 2.5 ± -4.3 ±(mL mirn1 0.9* 22.8 # 22.5 # 21 # 0.7 * 6.6 #kg -1)* per kg maternal weight# per kg fetal weightCLmf - maternal to fetal clearanceCLfm - fetal to maternal clearanceCLmo - maternal nonplacental clearanceCLf0 - fetal nonplacental clearanceCLm- total maternal clearanceCLf- total fetal clearance92Table 21. Total body, placental and non-placental clearance parameterscalculated from serum steady-state levels of unbound VPA and [13C4]VPAachieved during the simultaneous infusion of VPA and [13C4]VPA to the ewe andfetus, respectively. The values in parentheses refer to the percentage of the totalbody clearance (i.e., CLm or CLf) accounted for by the reported value.Ewe Clearancenumber (mL riinl)CLm CLf CLfm CLmf CLmo CLf0Ewe 313 438 268 247 182 257 21(92%) (41 %) (59%) (8%)Ewe 2101 699 235 283 125 573 -47(100%) (18%) (82%)Ewe 2177 464 160 177 88 376 -17(100%) (19%) (81%)Ewe 2181 666 392 408 318 349 -16(100%) (48%) (52%)Ewe 2220 766 313 357 163 603 -44(100%) (21%) (79%)mean 607 274 294 175 431 -21(mL minl) (100%) (29%) (71%)mean±sd 8.2± 128± 136± 82.8± 5.7± -8.8±(mL mirn1 1.9 * 46 # 45 # 47.7 # 1.6 * 1 1.9 #kg-1)* per kg maternal weight# per kg fetal weightCLmf - maternal to fetal clearanceCLfm - fetal to maternal clearanceCLmo - maternal nonplacental clearanceCLf0 - fetal nonplacental clearanceCLm- total maternal clearanceCLf- total fetal clearance93100-JE—SD)-10C0110.1VPA-amniotic fluid[13C4]VPA-amniotic fluidTime (hours)Figure 13. VPA and[13C4]VPA concentration versus time profiles in amniotic andfetal tracheal fluid from ewe 2177 during the simultaneous infusion of VPA to theewe and [13C4]VPA to the fetus.--e--- VPA-fetal tracheal fluid[13C4]VPA-fetal tracheal fluid0 1 2 3 4 5 694Table 22. Mean concentrations of VPA and[13C4]VPA in amniotic and fetaltracheal fluids during the final three hours of the simultaneous infusion of VPA tothe ewe and [13C4]VPA to the fetus.Ewe VPA in [13C4]VPA in VPA in fetal [13C4]VPA in fetalamniotic fluid amniotic fluid tracheal fluid tracheal fluid(pg/mL) (pg/mL) (igImL) (pgImL)E313 10.0 6.0 N/A* N/AE2101 5.9 9.0 1.5 1.4E2177 8.0 9.7 2.4 2.8E2181 16.9 12.0 1.8 0.9E2220 7.0 4.8 11.3 3.7Mean±sd 9.6±4.4 8.3±2.9 4.2±4.7 2.2±1.3*- not available953.5.4 VPA and[13C4JVPA in maternal and fetal urineThe mean rate of urine production during the infusion period of theseexperiments was 23.6 ± 11.4 mL h1 (0.666 ± 0.206 mL h1 kg-1) and 50.9 ± 19.7mL h1 (10.6 ± 4.4 mL h1 kg-i) for the fetus and ewe, respectively (see Table 23).Figure 1 4 illustrates the mean hourly fetal urine flow over the course of the infusionexperiments. As demonstrated in the figure, paired comparisons following anANOVA indicated that there was a significant increase in urine flow in the fourthand fifth hour of the infusion protocol relative to pre-infusion values(F(17,51)=4.531, p<0.05). This effect was not observed in saline infusion controlstudies. Table 24 reports the amounts of VPA and [13C4]VPA recovered fromfetal and maternal urine collected during the six hour simultaneous infusion period.Of the total amount of VPA and [13C4]VPA excreted, 18± 13 % and 10± 9 %,respectively, were excreted in maternal urine in unconjugated form, while 61 ± 19% and 76 ± 10 %, respectively, were excreted in fetal urine in unconjugated form.There was a significant difference in the percentage of drug excreted in conjugatedform between the ewe and fetus (F(3,9) = 19.682, p<0.05), with a higherpercentage of the limited amount of drug observed in fetal urine appearing inconjugated form. On the other hand, there was no significant difference foundbetween the percentage of conjugated VPA and [13C4]VPA excreted by the eweor by the fetus, respectively.The urine excretion data were also used in combination with serum druglevels to estimate maternal and fetal renal clearances of VPA and [13C4]VPA. Asshown in Table 25, the mean maternal values were 0.0917± 0.0639 and 0.521 ±0.071 mL min1 kg1 for unconjugated and total (unconjugated + conjugated) VPA,respectively. In contrast, the calculated fetal values of 0.0315 ± 0.0320 and960.0433 ± 0.0412 mL min1 kg1 were low and not statistically different from 0 (t(6)= 1.616, p>0.05).Table 23. Mean pH and volume of urine produced per hour by the ewe and fetusduring the six hour simultaneous infusion of VPA to the ewe and [13C4]VPA to thefetus.Ewe Rate of maternal Mean hourly Rate of fetal urine Mean hourlynumber urine production maternal urine production fetal urine(mL h-i) pH (mL h-i) * pH2101 78.8 7.63 17.4/20.3 6.682177 41.7 7.62 24.8/23.1 6.962181 33.4 7.50 13.2/13.5 6.722220 50.0 7.42 39.2 / 43.4 6.89mean ±sd 50.9± 19.7 7.54±0.10 23.6±11.4/ 6.81 ±0.1325.1 ±_12.9* Value based on volume collected per hour / value based on computer-collectedurine flow data.974030.=-J20a)D100begin infusionTime (hours)Figure 14. Fetal urine flow (mean ± Sd) for six hours before, during and six hoursfollowing the simultaneous infusion of VPA to the ewe and [13C4]VPA to the fetus.The asterisk (*) indicates a statistically significant difference from control values.*LO c)C—TC?C— C.J C) L()I I I I I0’- C%I C)coo) 0— C’J0)0—end infusion98Table 24. Summary of the cumulative amounts of VPA and [13C4]VPA recoveredfrom unhydrolyzed (unhydrol.) and hydrolyzed (hydroL) fetal and maternal urineduring the six hour simultaneous infusion period.Ewe Fetal urine (VPA) MAT urine (VPA)(mg) (mg)unhydrol. hydrol. unhydrol. hydrol.E2101 0.277 0.449 148.9 1147.3E2177 0.925 1.073 140.1 1369.5E2181 0.55 1.317 357 933.7E2220 2.485 4.613 70.58 621mean ± 1.059± 1.863± 179.14 ± 1017.9 ±s.d. 0.987 1.869 123.64 318.8Fetal urine([13C4]VPA) MAT urine([13C4]VPA)(mg) (mg)unhydrol. hydrol. unhydrol. hydrol.E2101 0.366 0.48 45.79 1765.3E2177 1.045 1.155 72.81 1659.8E2181 0.292 0.431 357.6 1507.9E2220 2.379 3.382 111.7 990.9mean± 1.020± 1.362± 146.98± 1481.0±s.d. 0.967 1.387 143.00 343.499Table 25. Maternal and fetal renal clearance values based on unconjugated andtotal (unconjugated + conjugated) VPA and [13C4]VPA levels in maternal and fetalurine, respectively, collected during the six hour simultaneous infusion period.Ewe number Maternal renal clearance Fetal renal clearance(mL ninl) (mL minl)u nconjugated total u nconjugated totalVPA VPA [13C4]VPA [13C4]VPAE2101 5.72 46.8 0.0239 0.0315E2177 4.28 43.7 0.0480 0.0531E2181 12.3 32.2 0.0260 0.0389E2220 4.15 34.4 0.194 0.281mean ± sd 6.61 ± 3.86 39.3 ± 7.1 0.0730 ± 0.101 ± 0.120mL min1 0.0814mean ± sd 0.0917 ± 0.521 ± 0.0315 ± 0.0433 ±mL min1 kg-1 0.0639 0.071 0.0320 0.04121003.5.5 Physiological effectsThe fetal blood gas, acid-base and metabolic parameters monitored (i.e.,P02, pCO2,pH, BE, [HC03], [Hb], °2 sat., lactate and glucose levels), were allwithin the normal range for an ovine fetus prior to drug administration (see Table26). Unlike the bolus administration studies in which no significant blood gas,acid-base or metabolic effects were observed, the initiation of the drug infusionscaused a short term but significant drop in fetal oxygen saturation (a meandifference of 13.5 %; F(5,20) = 4.830, p.<0.05; see Figure 15) and P°2 (a meandifference of 4 mm Hg; F(5,20) = 5.397, p.<0.05; see Figure 16) in theseexperiments. These changes were observed five minutes after the onset of drugadministration but had dissipated by sixty minutes post initiation. Furthermore,lactate levels increased significantly from control values (0.93 ± 0.16 to 1.76 ± 0.53mM; F(5,20) = 4.192, p.<0.05; see Figure 17) by five minutes into the infusionperiod and remained elevated throughout the infusion period. Lactateconcentrations returned to levels not different from control values within 24 hourspost infusion (p>0.05). Neither the oxygen related effects nor the increases inlactate were observed in control saline infusion experiments.In contrast to the bolus administration experiments, significant changeswere observed in both fetal heart rate and arterial pressure following initiation ofdrug delivery. There was a significant decrease (mean difference of 16.9 bpm;F(71 ,284) = 1.930, p<0.05) observed in fetal heart rate in the 15 minute periodimmediately following initiation of drug administration relative to control values.This decrease was temporary, however, with the fetus recovering within the firsthour of the drug infusion (see Figure 18). The change in heart rate wasaccompanied by a significant increase in fetal arterial pressure (mean difference of11.9 mm Hg in the first 15 mm; F(71 ,284) = 2.076, p<0.05) immediately following101infusion initiation. Following the initial increase, the arterial pressure declinedslowly but remained elevated relative to control values for approximately two hoursinto the infusion period, after which it returned to control values (see Figure 19).There were no significant changes observed in umbilical blood flow during thecourse of these experiments (F(71 ,1 42) = 0.553, p >0.05).Once again, administration of VPA produced a significant effect on fetalbreathing movements (F(12,36) = 5.837, p<0.05). As noted in Figure 20, there is asignificant decrease in breathing activity throughout the course of the infusionperiod relative to control values with the exception of the 3-4 h post-initiationperiod. The lack of a significant effect in the 3-4 h post-initiation period is due tothe large standard deviation in the data caused by one animal demonstrating amuch higher activity than the remaining three (26 mm h1 vs. 10, 7, and 9 mm h1).The breathing activity returned to control value levels within the first hour postinfusion. The simultaneous infusions did not produce a significant effect on eitherEOG (see Figure 21; F(1 9,38) = 1.456, p >0.05) or EC0G (see Figure 22; F(1 9,38)= 0.634, p >0.05) activity, although there is a trend toward lower EOG activity inthe first three hours of the infusion period relative to control values.102Table 26. Fetal blood gas, acid-base, metabolic and cardiovascular parameters(mean ± s.d.; n=5) prior to the simultaneous infusion of VPA and[13C4]VPA.Parameter Mean (± sd) controlvaluespH 7.344±0.016P°2 (mm Hg) 22.4± 3.1pCO2 (mm Hg) 49.5 ± 2.9BE (meg/L) 1 1.5± 1.0[HCO31 (meg/L) 26.5 ± 1.1[Hb](g/dL)2 9.96± 0.15°2 sat. (%) 53.3 ± 13.5[lactate] (mM) 0.93± 0.16[glucose] (mM) 0.72± 0.15heart rate (bpm) 169 ± 3arterial pressure (mm Hg) 46 ± 31. Whole blood base excess2. Haemoglobin concentration3. Oxygen saturationI806040200Time (hours)Figure 15. Fetal femoral arterial oxygen saturation (mean ± sd) monitored before,during and after the simultaneous infusion of VPA to the ewe and [13C4]VPA tothe fetus. The asterisk (*) indicates a statistically significant difference fromcontrol values (p%.1>200uJ100Time (hours)Figure 21. EOG activity (mean ± sd) for six hours before, during and six hoursfollowing the simultaneous infusion of VPA to the ewe and [13C4]VPA to the fetus.begin infusion end infusion11050 -40-30- J TIT:.Tr:T I 11 1.( 1k 18[jj :::: : :::: :::: :: ::::10- :::: ::::: : :::: :: :::: ::0-? ::•::•: :: T I•:•: :: •:• •:•: ::• •:•:It::•:•:•LO’t1 Q)C’J —o —C’J fl tO (C) t-(0 0)0’- C\\II I I I I 0) 0 —begin infusion end infusionTime (hours)Figure 22. ECoG activity (mean ± sd) for six hours before, during and six hoursfollowing the simultaneous infusion of VPA to the ewe and [13C4]VPA to the fetus.1113.5.6 VPA and[13C4]VPA metabolism during simultaneous infusionAll of the biofluids collected during these experiments were analyzed for thepresence of both unlabelled and labelled forms of 16 VPA metabolites. While VPAmetabolites were detected in maternal serum, fetal serum, maternal urine and tosome extent fetal urine, quantifiable levels of metabolites were not observed ineither amniotic or fetal tracheal fluids. Due to the termination of sampling at theend of the infusion period, metabolite recoveries could not be determined,however, the production of metabolites was monitored during the course of thesimultaneous infusions.3.5.6.1 VPA metabolites in maternal and fetal serumTwelve metabolites of VPA were detected in the serum samples collectedduring these experiments. The metabolites detected were the following: (E) and(Z)-2-ene VPA, (E) and (Z)-3-ene VPA, 4-ene VPA, 3-keto VPA, 4-keto VPA, 3-OHVPA, 4-OH VPA, 5-OH VPA, 2-PSA and 2-PGA. The time courses of the labelledand unlabelled forms of these metabolites in both maternal and fetal serum areillustrated in Figures 23-34. The metabolites detected were observed in bothunlabelled and labelled forms, with the exception of 3-OH VPA and 2-PSA, whichwere detected solely in their unlabelled form. Interestingly, the ratio of the level of13C4-labelled metabolite in fetal serum to that in maternal serum over the first twohours of the infusion period was less than one for each metabolite observed (seeTable 27). However, the ratios calculated for the last four hours of the study were,for the most part, higher than those observed for the same compound during thefirst two hours of the experiment. As in the bolus administration experimentsreported above, none of the diene metabolites of VPA were detected in maternal112or fetal serum. A portion of a representative chromatogram from an extractedsample of maternal serum is illustrated in Figure 35, in order to demonstrate theabsence of peaks due to diunsaturated metabolites of VPA in these biofluids.Table 27. The mean (± s.d.) ratio of metabolite concentration in fetal serumrelative to maternal serum in the first two hours and final four hours of thesimultaneous infusion of VPA to the ewe and [13C4]VPA to the fetus.metabolite 13C4-labelled metabolites unlabelled metabolites0-2hof 2-6hof 0-2hof 2-6hofinfusion period infusion period infusion period infusion period(E)-2-ene VPA 0.77 ± 0.53 0.87 ± 0.6 0.27±0.18 0.37 ± 0.03(Z)-2-ene VPA 0.84 ± 0.64 1.37 ± 0.06(E)-3-ene VPA 0.73 ± 0.30 1.95 ± 0.72 0.18 ± 0.17 1.05 ± 0.31(Z)-3-eneVPA 0.69±0.08 0.73±0.14 0.82±0.04 0.65±0.064-eneVPA 0.84±0.07 0.67±0.05 0.63±0.14 0.51 ±0.023-OH VPA 0.03 ± 0.07 0.27 ± 0.124-OH VPA 0.23 ± 0.20 0.48 ± 0.12 0.14 ± 0.14 0.55 ± 0.095-OH VPA 0.58 ± 0.29 1.99 ± 1.38 0.29 ±0.19 1.23 ± 0.723-keto VPA 0.70 ± 0.62 2.28 ± 0.60 0.21 ± 0.20 1.04 ± 0.254-keto VPA 0.20 ±0.14 0.28 ± 0.022-PSA 0.49 ± 0.29 1.00 ± 0.332-PGA 0.18 ±0.13 0.61 ± 0.17n=5n=5n=5n=5113(E)-2-ene VPA - maternal serum(E)-2-ene VPA - fetal serum(E)-2-ene [13C4]VPA - fetal serum----0---- (E)-2-ene [13C4]VPA - maternal serumTTI-JED)0: 0.10.010.00 1Time (hours)Figure 23. Mean (± sd) serum concentration versus time profiles for (E)-2-eneVPA and (E)-2-ene [13C4]VPA during the simultaneous infusion of VPA to the eweand [13C4]VPA to the fetus. The number of animals that contributed data to eachcurve is reported as an n-value beside the curve.0 1 2 3 4 5 6-JEzLC0I114(Z)-2-ene VPA - maternal serum(Z)-2-ene VPA - fetal serum(Z)-2-ene [13C4]VPA - maternal serumn=3n=4n=10.10.010.00 10.000 1Time (hours)Figure 24. Mean (± Sd) serum concentration versus time profiles for (Z)-2-eneVPA and (Z)-2-ene [13C4]VPA during the simultaneous infusion of VPA to the eweand [13C4]VPA to the fetus. The number of animals that contributed data to each0 1 2 3 4 5 6curve is reported as an n-value beside the curve.115-JE00.1ct 0.019)C?uJ(E)-3-ene VPA - maternal serum(E)-3-ene VPA - fetal serum1 (E)-3-ene[13C4]VPA - fetal serum(E)-3-ene[13C4]VPA - maternal serumn=3n=3n=3n=30.00 10 1 2 3 4 5 6Time (hours)Figure 25. Mean (± sd) serum concentration versus time profiles for (E)-3-eneVPA and (E)-3-ene [13C4]VPA during the simultaneous infusion of VPA to the eweand [13C4]VPA to the fetus. The number of animals that contributed data to eachcurve is reported as an n-value beside the curve.I I I1 2 3Time (hours)I I I4 5 6Figure 26. Mean (± Sd) serum concentration versus time profiles for (Z)-3-eneVPA and (Z)-3-ene [13C4]VPA during the simultaneous infusion of VPA to the eweand [13C4]VPA to the fetus. The number of animals that contributed data to eachcurve is reported as an n-value beside the curve.116p (Z)-3-ene VPA - maternal serum(Z)-3-ene VPA - fetal serum(Z)-3-ene[13C4]VPA - fetal serum(Z)-3-ene [13C4]VPA - maternal serumn=3n=3n=3n=3-JED)0I0.1-0.010.001•0-JED)00.011170.14-ene VPA - maternal serumffl 4-ene VPA - fetal serum““ 4-ene[13C4]VPA-fetal serum4-ene [13C4]VPA - maternal serumT n=5n=5n=5n=50 1 2 3 4 5 60.00 1Time (hours)Figure 27. Mean (± Sd) serum concentration versus time profiles for 4-ene VPAand 4-ene [13C4]VPA during the simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus. The number of animals that contributed data to eachcurve is reported as an n-value beside the curve.-JEC)C0118T3-OH VPA - maternal serum3-OH VPA - fetal serum10n =4n=10.10.01Figure 28. Mean (± sd) serum concentration versus time profile for 3-OH VPAduring the simultaneous infusion of VPA to the ewe and [13C4]VPA to the fetus.The number of animals that contributed data to each curve is reported as an nvalue beside the curve.0 1 2 3 4 5 6Time (hours)n=5n=5n=2n=1119p 4-OH VPA - maternal serum4-OH VPA - fetal serum4-OH [13C4]VPA - fetal serum4-OH[13C4]VPA - maternal serumT-JED)zLCo 0.10.010.00 1Time (hours)Figure 29. Mean (± sd) serum concentration versus time profiles for 4-OH VPAand 4-OH[13C4]VPA during the simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus. The number of animals that contributed data to each0 1 2 3 4 5 6curve is reported as an n-value beside the curve.n=5n=5n=4n=51205-OH VPA - maternal serum5-OH VPA - fetal serum5-OH[13C4]VPA - fetal serum5-OH [13C4]VPA - maternal serum-JED).2 0.10.010.001Time (hours)Figure 30. Mean (± Sd) serum concentration versus time profiles for 5-OH VPAand 5-OH[13C4]VPA during the simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus. The number of animals that contributed data to each0 1 2 3 4 5 6curve is reported as an n-value beside the curve.n=5n=5n=5n=51213-keto VPA - maternal serum— 3-ketoVPA-fetalserum3-keto[13C4]VPA - fetal serum0 3-keto [13C4]VPA - maternal serumTI .L-JED).2 0.10.010.001Figure 31. Mean (± Sd) serum concentration versus time profiles for 3-keto VPAand 3-keto [13C4]VPA during the simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus. The number of animals that contributed data to eachcurve is reported as an n-value beside the curve.0 1 2 3 4 5 6Time (hours)1224-keto VPA - maternal serum4-keto VPA - fetal serum4-keto[13C4]VPA - maternal serumT-JEn5.2 0.1n=5> 0.010a)-‘;j.0.00 1Time (hours)Figure 32. Mean (± Sd) serum concentration versus time profiles for 4-keto VPAand 4-keto [13C4]VPA during the simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus. The number of animals that contributed data to eachcurve is reported as an n-value beside the curve.0 1 2 3 4 5 6-JE0.1ci)C)00Co0Figure 33. Mean (± Sd) serum concentration versus time profile for 2-PSA duringthe simultaneous infusion of VPA to the ewe and [13C4]VPA to the fetus. Thenumber of animals that contributed data to each curve is reported as an n-valuebeside the curve.1232-PSA - maternal serum2-PSA- fetal serumn=5n=20.10.010.00 1 I I I0 1 2 3Time (hours)I I I4 5 6-J0)00.10C)0a1241p 2-PGA - maternal serum2-PGA - fetal serum2-[13C4]PGA - maternal serumn =4n=4n=30 1 2 3 4 5 60.01Time (hours)Figure 34. Mean (± Sd) serum concentration versus time profiles for 2-PGA and 2-[13C4]PGA during the simultaneous infusion of VPA to the ewe and [13C4]VPA tothe fetus. The number of animals contributing data to each curve is reported as ann-value beside the curve.12525000AB 20000UN 15000DA 10000NC 5000E0Time (mm.)Figure 35. SIM chromatograms of the tBDMS derivative of (E)-2-ene VPA (m/z =199; peak 1) and the mlz = 197 profile between 16.50 and 19 minutes, the timerange over which the diunsaturated metabolites of VPA are detected, from amaternal serum sample collected during the simultaneous infusion of VPA and[13C4]VPA to the ewe and fetus, respeCtively.116.50 17.00 17.50 18.00 18.501263.5.6.2 VPA metabolites in maternal and fetal urineBoth maternal and fetal urine were analyzed in order to assess theexcretion of VPA metabolites into each fluid. The majority of the metabolitesdetected in the serum samples were also detected in maternal urine samples. Thelevels of the five most prominent metabolites in maternal urine, 3-keto VPA, 4-ketoVPA, 3-OH VPA, 4-OH VPA, 5-OH VPA, are illustrated in Figure 36. Cumulativeamounts of these compounds typically ranged from 50 - 150 mg with largeramounts of unlabelled metabolites being observed than labelled metabolites inmost instances. The following metabolites were also detected in both unlabelledand labelled forms in maternal urine collected during the six hour infusion period:2-PSA, 2-PGA, (E) and (Z)-2-ene VPA and (E)-3-ene VPA. These compoundswere recovered in smaller and variable amounts that ranged from up to 10 mg for2-PSA and 2-PGA to less than 100 jig for (E)-3-ene VPA.In contrast, virtually no metabolites of VPA were detected in fetal urinecollected during this period. There were traces of the following unlabelledcompounds detected in fetal urine, however, the cumulative amounts recovereddid not exceed 100 jig in any instance: 3-keto VPA, 4-keto VPA, 3-OH VPA, 4-OHVPA, 5-OH VPA. This result is consistent with the lack of appreciable fetalnonplacental clearance of VPA reported above.127unhydrolyzed urine300-D hydrolyzed urine-Da)a)>0C-)a)a) Ho HH(U HH100TIT___ _<<< <<< <<< > > >> > > > > >o 0- I i- I;OjOQOoooO-C)-C) • cY) • C) , Y)d,C) C) Q 0 0- I I I) . cY) IC)MetaboliteFigure 36. Cumulative recovery of the most abundant metabolites of VPA inmaternal urine, collected during the simultaneous infusion of VPA to the ewe and[13C4]VPA to the fetus.1284. DISCUSSION4.1 Assessment of GC-MS assay in sheep biofluidsAs was observed with human samples, the chromatographic conditionsused in the current assay for ovine biological fluid samples provided resolution ofVPA and its metabolites including all of the isomeric unsaturated compounds(Figures 4a- 4d). For the first time, (E)- and (Z)-3-ene VPA were completelyresolved using tBDMS derivatives (Yu eta!., 1995). As can be seen in Tables 3and 4, the experimentally determined values for the spiked samples are in goodagreement with the calculated values in most instances. The GC-MS assay alsodemonstrated reasonable precision with coefficients of variation of < 10 % for VPAand most metabolites. The analyses which resulted in larger degrees of errorwere typically associated with low levels of metabolite that approached the limits ofdetection of the assay. The assay provided good sensitivity with limits ofquantitation (LOQ) for the metabolites ranging between 3 and 20 ng/mL in all ofthe biofluids studied. The sensitivity observed is probably due, in part, to theutilization of the intense (M-57) fragment of the tBDMS derivative which has beenfound to provide better detection limits with fatty acids than the less intense (M15) fragment of TMS derivatives (Phillipou, 1975; deJong etal., 1980; Woollard,1983; Darius and Meyer, 1994). The assay also makes use of stable isotopelabelled VPA and metabolites as internal standards because their structuralsimilarity to the compounds of interest produces chromatographic characteristicsakin to those of the compounds being examined (Roncucci et aL, 1975; Murphyand Sullivan, 1980). These findings suggest that the method is reliable andreproducible and hence, suitable for studies of VPA and its metabolites in pregnantsheep.129Although the accuracy and reproducibility of the assay in amniotic and fetaltracheal fluids was not verified in the same fashion as described for serum andurine, the chromatography and accuracy of calibration standards analyzed withexperimental samples strongly supports the notion that the results obtained fromthese fluids are meaningful. Nonetheless, the absence of metabolites of VPA insamples of these fluids collected during experimentation restricted the need forquantitation to VPA and[13C4]VPA using[2H7]VPA.4.2 Nonpregnant animal i.v. bolus studiesThe terminal elimination half-lives for VPA and [13C4]VPA in the sheep are3.5 ± 0.4 h and 3.2 ± 0.4 h, respectively. These values are shorter than half-livesobserved for VPA in adult humans which typically range from 8- 16 hours(Loscher, 1978; Bowdle etal., 1980; Gugler and von-Unruh, 1980). However, thisdiscrepancy is consistent with the differences noted in the half-lives of othercompounds such as metoclopramide and diphenhydramine, which also show morerapid elimination in sheep than humans (Yoo et al., 1986; Riggs eta!., 1988). Theobserved values are more consistent with the terminal elimination half-lives of VPAreported in other animal species such as the rhesus monkey (0.66 ± 0.5 h) (Lai etal., 1980) and the rat (0.98 ± 0.08 h) (Kobayashi eta!., 1990).It may be concluded from a comparison of the pharmacokinetic dataconcerning the dispositions of VPA and [13C4]VPA that they are equivalent interms of AUCQQ values andt12. However, in order to confirm the pharmacokineticbioequivalence of these two compounds, it is also necessary to verify that theirpathways of elimination are also satisfactorily equivalent. This issue wasaddressed by assessing the recovery of both unlabelled and labelled metabolitesfrom urine. Analysis of urine samples revealed that the recovery of labelled130metabolites is almost identical to that of unlabelled metabolites, thus furthersuggesting a similar disposition and fate for both VPA and [13C4]VPA (Table 8).The major metabolites observed in human serum are the p-oxidationproducts (E)-2-ene VPA and 3-keto VPA as well as (E,E)-2,3’-diene VPA and 3-ene VPA (Kassahun etal., 1990; Fisher etal., 1992a). The ewe appears toproduce all of these compounds with the exception of the (E,E)-2,3’-diene. In fact,none of the diunsaturated metabolites were found to be present in measurablequantities in the biofluids monitored. This result suggests that the sheep has areduced ability, compared to humans and other animal species such as the rat(Baillie and Rettenmeier, 1989), to biotransform VPA via these pathways. On theother hand, in a fashion similar to humans, the oxidation products 3-OH VPA, 4-OH VPA, 5-OH VPA, 3-keto VPA, 4-keto VPA and 2-PGA were all found insignificant amounts in the urine of the ewe. A more detailed discussion of thebiotransformation pathways utilized by sheep will follow in section 4.4.4.3 Maternal and fetal i.v. bolus studiesValproic acid is used to treat various epileptic disorders in pregnancy (Daviset aL, 1994). Therefore, a greater knowledge of its pharmacokinetics as well asthe in utero fetal exposure and fetal effects of VPA is of great importance. Thematernal and fetal Lv. bolus studies were conducted in order to obtain moredetailed information on the disposition and effects of VPA in the pregnant sheep.The appropriate drug doses for these studies were determined using informationgained on VPA disposition in sheep in the nonpregnant study. The maternal dosewas chosen to obtain serum levels comparable to those typically observed inhumans following VPA dosing i.e., 40 - 100 jig mL1. However, when choosing afetal dose it was necessary to consider the small size of the fetus relative to the131ewe and yet at the same time, respect the need to obtain measurable drug levelsacross the placenta in maternal serum. With this in mind, the fetal dose waschosen to be one-quarter of the maternal dose, a protocol similar to that used inprevious studies with other compounds in our laboratory. This dosing protocolresulted in detectable VPA concentrations in the ewe for approximately 12 hoursfollowing drug administration.4.3.1 PharmacokineticsThe disposition of VPA both in maternal serum following maternal dosingand in fetal serum following fetal drug administration was best described by abiexponential function. In each case, the data were tentatively fit to mono, bi andtriexponential functions, these fits being subsequently compared through themagnitude of their residual sum of squares and AIC criteria. Although increasingthe number of exponents in the equations typically improved the fitting of the data(Le., reduced the residual sums of squares), the standard errors of the estimate inthe tertiary phase of the triexponential equations were prohibitively large.Furthermore, the number of points placed in the tertiary phase during the datafitting was typically small, varying from two to four data points. Fitting the data witha biexponential equation usually resulted in the inclusion of at least nine datapoints in each phase. The goodness of fit of the data was also examined usingthree different weighting factors including no weighting, the reciprocal of drugconcentration and the reciprocal of the square of drug concentration. Based onthe effect of various weighting factors on the weighted residual sum of squares,AIC parameter and residual plots for the fit, data points for individual subjects wereweighted using the reciprocal of drug concentration. While the reciprocal of thesquare of drug concentration typically provided smaller weighted residual sum of132squares than either of the other two factors, the residual plots revealed thedevelopment of an unacceptable linear pattern in the weighted residual vs.weighted calculated Y-values plot. This is generally indicative of systemic errorand thus on this basis, data weighting with this factor was not used.As mentioned earlier, the data from El 242 following fetal bolus dosing ofVPA was erratic and inconsistent with the data from the other animals in the fetaldosing group. There were large fluctuations in fetal serum VPA concentrationsbetween two and six hours following dose administration, and significant levelspersisted in fetal serum up to 48 hours post drug administration. These factorsmade the data difficult to fit and the errors associated with the estimatedparameters were large. For these reasons, it was decided to omit the data fromEl 242 from further pharmacokinetic calculations.The terminal elimination half-life of VPA calculated for the ewe (and fetus) inthis study is shorter than that normally observed in both nonpregnant and pregnanthumans (Nau eta!., 1982; Nau, 1986a; Hendrickx eta!., 1988; Zaccara eta!.,1988; Davis eta!., 1994) (i.e., 9- 18 h) but once again, not unlike that observed inother animal species (Dickinson eta!., 1980; Nau eta!., 1982; Nau, 1986a;Kobayashi et a!., 1990) (0.3 - 6 h). The fetal elimination half-life was notsignificantly different than that observed in the ewe (p>O.05). This is not alwaysthe case in pregnant sheep, as illustrated by metoclopramide (MCP) kinetics,where the fetal elimination half-life is significantly longer than that observed in theewe (Riggs et a!., 1988). It has been suggested that this longer elimination half-lifemay be due to recirculation of the drug via fetal ingestion of amniotic and trachealfluids, where, unlike VPA, significant levels of MCP are present (Riggs eta!.,1988).VPA appeared rapidly in fetal serum following maternal dosing withconcentrations of approximately 10 -50 .tg/mL being observed two minutes133following drug administration. In a similar fashion, VPA was transferred tomaternal serum within two minutes of fetal drug administration. VPA was alsopresent in amniotic and tracheal fluids in measurable quantities at the earliestsampling time (2 mm. post-administration). The basic drugs, diphenhydramine(DPHM), MCP, ritodrine and labetalol, accumulate significantly in both amnioticand fetal tracheal fluid (Riggs et a!., 1987; Wright eta!., 1991; Yeleswaram et a!.,1992). This accumulation is marked in the case of MCP, where tracheal fluidconcentrations are up to 15-fold greater than those in fetal plasma (Riggs et a!.,1987). In contrast, VPA was not found to accumulate extensively in either fluid asindicated by mean time-averaged concentration ratios of tracheal fluid (0.046 ±0.022) and amniotic fluid (0.198 ± 0.214) to fetal serum. This is probably due todifferences in the physicochemical properties of the compounds, and is consistentwith observations in the adult that amphiphilic amine compounds are primarilytaken up bythe lung (Okumura eta!., 1978; Yoshida eta!., 1987; Yoshida eta!.,1989; Yoshida eta!., 1990). Although the exact mechanism involved in the uptakeof these drugs in the adult lung is uncertain, it has been suggested that bothpassive diffusion and carrier-mediated active transport processes across thepulmonary endothelial cells may be involved (Bend eta!., 1985). As will bediscussed in greater detail in section 4.4.2, VPA is known to cross otherphysiological membranes based on its detection in both brain tissue andcerebrospinal fluid (CSF) as well as in tears and saliva. VPA concentrationsobserved in CSF and tears correspond approximately to unbound drug levels inplasma, however, the low concentrations observed in saliva are usually erratic anddo not correlate with either plasma or CSF drug levels (Gugler and von-Unruh,1980; Monaco eta!., 1982; Rapaport eta!., 1983). The low levels of VPA detectedin tracheal and amniotic fluids may be indicative of an inability of the fetus to134eliminate VPA either into urine, a major component of amniotic fluid volume, orthrough excretion into lung fluid.The volume of distribution (0.242 ± 0.036 L kg-1, Table 10) of VPA in theewe following maternal dosing was small relative to values noted for othercompounds studied in our laboratory such as labetalol (3.02 ± 0.18 L kg-i)(Yeleswaram eta!., 1992), MCP (5.7 L kg-i) (Riggs eta!., 1988) and ritodrine (14.3± 3.5 L kg-i) (Wright, 1992). On the other hand, the apparent volume ofdistribution in humans following single-dose VPA administration ranges from 0.126- 0.175 L kg1 (Davis etal., 1994), values that are more comparable to the data forVPA in sheep. These findings suggest that VPA is primarily confined to thecirculation and extracellular fluids in both humans and sheep. In the currentstudies, the fetal volume of distribution of VPA (5.27 ± 2.84 L kg-1, Table 15) wassignificantly larger than that observed in the ewe on a weight normalized basis.This may be due in part to the rapid distribution of the drug to the ewe from thefetus via the placenta following fetal administration, since the total maternalvolume available for distribution and ability to clear VPA are much higher than inthe fetus. The fetal volumes of distribution observed for VPA, however, are notunusually large compared to values reported for other compounds following fetalbolus dosing such as ritodrine (8.7 ± 1.8 L kg-i) (Wright, 1992) and labetalol(19.01 ± 1.52 L kg-i) (Yeleswaram, 1992).The ability to collect serial samples of both maternal and fetal serum allowsus to obtain a reasonable estimate of the extent of fetal exposure to VPA followingmaternal dosing, via the ratio of the area under serum drug concentration versustime curves for the fetus and the mother (Mihaly and Morgan, 1984). Fetalexposure to VPA (i.e., placental transfer) has been reported in humans (Nau et a!.,1982; Hendrickx et a!., 1988) but these studies were based primarily on a singlepaired maternal venous and cord blood sample taken at the time of delivery. While135the fetal-to-maternal concentration ratio obtained in this fashion does provide anindication of the extent of fetal drug exposure, these estimates are highlydependent on the time of sampling relative to the time of drug administration(Anderson eta!., 1980b; Anderson eta!., 1980a).As indicated in Table 10, there is a considerable degree of fetal exposure tothis drug in the pregnant sheep (0.410 ± 0.118). The fetal/maternal AUC ratio forVPA in pregnant sheep was lower than the cord blood/maternal venousconcentration ratios reported in humans immediately following delivery, whichrange from 0.5 to 4.6 (Nau et a!., 1982; Hendrickx et a!., 1988). These differencescould be due to the time-dependency of the single point measurements in humansdiscussed above. However, the epitheliochorial sheep placenta differs from thehemochorial human placenta, having a lower permeability to hydrophilic, polarmaterials (Faber and Thornburg, 1983) which could play a role in the observeddifferences in fetal drug exposure. It is also important to consider the possibleinfluence of the extensive maternal and fetal hemodynamic changes that occurduring and immediately after delivery. These changes may affect the dispositionof a drug in the mother and/or fetus (Hamshaw-Thomas et aL, 1984), and hencemake the extrapolation of data obtained at delivery to the antepartum intrauterineperiod inappropriate.The estimated fetal exposure to VPA is intermediate between the valuesdetermined for the polar, basic compound ritodrine (with a ratio of approximately0.02) and other amine compounds studied previously, DPHM (ratio of “0.85) (Yooeta!., 1986) and MCP (ratio of 0.74) (Riggs eta!., 1988). Several factors play animportant role in the extent of drug exposure including physicochemical propertiesof the drug such as polarity, lipid solubility, degree of ionization and molecularweight as well as the maternal and fetal protein binding of the drugs (Rurak eta!.,1991). Since the protein binding of each of these compounds is similarto that of136VPA La, all between 30 and 50% for both the ewe and fetus (Rurak etal., 1991),it is unlikely that the differences in this factor play a determining role in the extentof fetal exposure. In contrast, the marked differences in lipid solubility between thecompounds correlate well with the differences in fetal exposure (Rurak etah,1991). For example, the chloroform/buffer partition coefficient of MCP is 1200times greater than that of ritodrine while its fetal to maternal AUC ratio isapproximately 37 times greater (Rurak et al., 1991). The large octanol/waterpartition coefficients (Craig, 1990) for MCP (417) and DPHM (1862) agree wellwith the significant fetal exposure observed with these compounds. In a similarfashion, one would expect a significant degree of VPA exposure based on itsreported octanol/water partition coefficient of 398 (Abbott and Acheampong, 1988)4.3.2 MetabolismThe metabolite profile in the ewe observed following maternal drugadministration was similar to that observed following VPA administration innonpregnant sheep. As noted in this earlier study, the diunsaturated metabolitesof VPA (i.e., 2,4-diene VPA and 2,3’-diene VPA) were not detected in the biofluidsanalyzed. However, the maximum concentrations of the other metabolitesobserved following bolus administration were not large for the most part (Table 11)and thus, it is possible that the diene metabolites were present at levels below thedetection limits of the assay. If this was the case, it was anticipated that perhapsthese metabolites would be detected in biofluids collected during the infusionstudies. The detection of 2-ene VPA, 3-OH VPA and 3-keto VPA suggest that thep-oxidation pathway for VPA biotransformation is active in sheep as are the VPAelimination pathways in the endoplasmic reticulum, as indicated by the detection of4-OH VPA, 5-OH VPA and 4-keto VPA. Although urine was not collected during137the bolus administration studies in pregnant sheep, data from the bolus studies innonpregnant sheep indicate that the glucuronidation pathway for VPA is alsoactive in sheep. Therefore, the primary metabolic pathways for VPA elimination inhumans are also active in sheep, although the absence of the diunsaturatedmetabolites suggests that the p-oxidation pathway may not play as prominent arole in the sheep as it does in humans. It has been proposed that some of thesemetabolites (e.g., 2-ene VPA and 2,3-diene VPA), which possess anticonvulsantactivity, may contribute to the effectiveness of VPA (Loscher, 1981 a; Loscher,1985), while others (e.g., 4-ene VPA and 2,4-diene VPA) may be responsible forthe rare but serious hepatotoxicity associated with VPA therapy (Kesterson etal.,1984; Kochen eta!., 1984; Nau and Loscher, 1984; Kassahun eta!., 1991;Kassahun eta!., 1994). The notable absence of the diene metabolites of VPA andthe significance of this absence will be further discussed in section 4.4.4.1,however, the absence of metabolites implicated in both VPA-associatedhepatotoxicity and metabolite-related antiepileptic effects suggests that the sheepmay not be well suited for the investigation of these issues.Most of the metabolites detected in maternal serum were also detected infetal serum albeit at much lower concentrations. The fact that higher levels ofmetabolites were observed in maternal serum following both maternal and fetaldosing may indicate that the biotransformation of VPA is occurring primarily in theewe. This issue was further investigated in the infusion studies and will bediscussed below.4.3.3 Physiological effectsThe bolus administration of VPA in these experiments had little effect onfetal blood gas, acid-base and metabolic status or on fetal cardiovascular138parameters (Le., heart rate and arterial pressure). These results are consistentwith previous studies in pregnant sheep in which administration of the GABAagonist muscimol failed to produce any significant changes in fetal cardiovascularor blood gas parameters (Johnston and Gluckman, 1983). The potentialinvolvement of GABA in the regulation of cardiovascular parameters such asarterial pressure and heart rate will be discussed further in section 4.4.3.Bolus administration of VPA to either the ewe or fetus did, however,produce a substantial depression in fetal breathing movements (FBM). This effectis consistent with reports on the effect of the GABA agonist, muscimol, on FBM infetal sheep and also with the effects observed with the GABA antagonist,picrotoxin in the same species (Johnston and Gluckman, 1983). GABA agonistshave also been shown to depress respiration in adult rats (Holzer and Hagmuller,1979), cats (Yamada etaL, 1981) and in premature, newborn rabbits (Hedner etal., 1980). In fetal sheep, breathing movements occur almost continuously fromearly gestation through to about 110 days (Bowes eta!., 1981). However,beginning at approximately 110 days of gestation, FBM are interrupted by apnoeicepisodes of increasing duration. Around 120 days gestation, these becomeclosely correlated with episodes of high voltage electrocortical activity (Bowes etaL, 1981). The relationship between the inhibition of breathing and this highvoltage activity is unclear although it has been suggested that central inhibition viathe inhibitory neurotransmitter GABA may be involved (Johnston and Gluckman,1983). The mechanism by which VPA increases GABA levels and hence affectsFBM remains uncertain. Increased GABA levels associated with VPAadministration are most likely associated with one or more of the following: 1.Activation of GAD (Fariello and Smith, 1989; Loscher, 1989); 2. Inhibition ofSSADH (Fariello and Smith, 1989; Zeise eta!., 1991); and/or3. Inhibition ofsynaptosomal GABA-T (Fariello and Smith, 1989; Loscher, 1993).139Regardless of the mechanism through which VPA increases GABA levels,regional brain studies in rats have conclusively demonstrated significant increasesin GABA levels following VPA administration (ladarola et aL, 1979; Loscher andVetter, 1985). These increases are particularly marked in midbrain regions, suchas the substantia nigra, that are thought to play a key role in seizure generationand propagation (ladarola et aL, 1979; Loscher and Vetter, 1985). These studiesindicated that the GABA increases induced by VPA in the substantia nigra occurpredominantly in the nerve terminals (ladarola and Gale, 1981; Loscher andVetter, 1985). Furthermore, it was noted that the time to onset of these effects onpresynaptic GABA levels was rapid Ia, significant increases were noted only fiveminutes after drug administration. It was also noted that the time course ofanticonvulsant activity correlated well with changes in GABA levels in the nerveterminals (Loscher and Vetter, 1985). The rapid onset of VPA effects on GABAlevels is consistent with the rapid changes noted in FBM in the bolus studies aswell as with the rapid changes in other physiological parameters noted in theinfusion studies. The relationship between GABA and breathing activity will bediscussed in greater detail in section 4.4.3.4.4 Simultaneous infusion studiesThe simultaneous infusion studies were designed to not only measurematernal and fetal clearances of VPA but to also provide more in depth informationconcerning the metabolic and renal elimination of this compound in both the eweand fetus. In addition, fetal physiological parameters were monitored in order toassess the effects of VPA on the fetus. The maternal dose was chosen in anattempt to attain maternal serum drug levels within the human therapeutic range.Fetal dosing was based on the maternal dose in the same fashion as that140described in section 4.3 for the bolus studies. The mean maternal steady-stateserum level of VPA was 69.42 ± 12.16 pg/mL while the fetal serum steady-statelevel of[13C4]VPA averaged 35.59 ± 10.41 pg/mL. The recommendedtherapeutic range for VPA is 40 - 100 pg/mL, although factors such as diurnalvariations in clearance, nonlinear protein binding and significant diurnalfluctuations of free fatty acids, which displace VPA from protein binding sites, allconfound the interpretation of the relationship between drug levels and clinicalefficacy (Davis et al., 1994). One should also consider the fact that maternalserum also contained a mean of 18.63 ± 2.72 pg/mL of [13C4]VPA and fetalserum levels of VPA were 38.96 ± 12.74 pg/mL on average. Thus, if thepharmacological actions of VPA are not affected by [13C4]-labelling, thecumulative concentration of VPA and [13C4JVPA should be considered to be theserum drug level relevant to observed physiological effects.The protein binding of VPA and [13C4]VPA in both maternal and fetalserum was lower than the approximately 85 to 95 % binding typically observed inthe monkey (Levy et al., 1977) and in humans with normal renal and hepaticfunction (Gugler and Mueller, 1978; Loscher, 1978). However, it is expected thatthe free fraction of drug will increase during pregnancy due to reduced maternalalbumin levels and increased maternal free fatty acid levels, which will competewith VPA for albumin binding sites. In fact, it has been reported that free VPAconcentrations can increase by up to 25 % in women at the time of deliveryrelative to nonpregnant women (Yerby eta!., 1992). It should also be noted thatprotein binding measurements were conducted in serum stored at - 20 °C prior touse. It is possible that the VPA binding to serum proteins is altered in storedserum because of factors such as continued lipolysis and production of free fattyacids (Haberer and Pollock, 1994). Although differences in binding between141maternal and fetal serum due to differences in albumin levels might have beenexpected, this was not found to be the case.Previous reports have noted differences in fetal upper and lower bodyconcentrations of glucose, lactate, oxygen partial pressure and oxygen saturation(Cohn et al., 1974; Charlton and Johengen, 1984; Tonn et aL, 1 995a). Thesedifferences are thought to occur via two reciprocal mechanisms (Tonn et aL,1 995b). One is the preferential shunting of oxygenated blood in the thoracicinferior vena cava (a mixture of umbilical venous and abdominal inferior vena cavalblood) through the foramen ovale into the left heart and hence primarily into theascending aorta. The other is the greater dilution of inferior vena caval blood withdeoxygenated superior vena caval blood in the right heart (which primarily suppliesthe descending aorta) compared to the dilution resulting from mixing of inferiorvena caval and pulmonary venous blood in the left heart. There is also thepossibility of incomplete mixing of ductus venosus (umbilical venous) blood withabdominal inferior vena caval blood in the thoracic inferior vena cava, andpreferential shunting of the former through the foramen ovale (Edelstone andRudolph, 1979). It has been suggested that similar effects would be observed withdrugs (Rudolph, 1985). However in studies involving fetal and maternaladministration of unlabelled and deuterated diphenhydramine (DPHM), Tonn eta!.(1 995b) found the opposite situation; drug concentrations were higher in the fetalfemoral artery than in the carotid artery, particularly with DPHM infusion via thefetal lateral tarsal vein into the inferior vena cava. Tonn et al. (1 995b) postulatedan explanation for this phenomenon by noting the previous observation of DPHMuptake by the fetal lung (Rurak eta!., 1991). Thus, in the left heart there would bemixing of drug-rich inferior vena caval blood with pulmonary venous blood that hada lower drug concentration, resulting in a lower DPHM level in the carotid artery.In contrast, they postulated that there would be minimal DPHM uptake by tissues142of the upper body of the fetus. Thus, drug levels in superior vena caval bloodwould be not much different from those in arterial blood, so that in the right heart,there would be minimal dilution of DPHM in inferior vena caval blood by mixingwith blood entering from the superior vena cava. This would lead to higher DPHMlevels in the femoral artery compared to the carotid artery. If this explanation iscorrect, then one would not expect to see a similar situation with drugs that are nottaken up by the fetal lung. As discussed earlier and below, there is minimalaccumulation of VPA in fetal lung fluid and hence likely limited uptake by the fetallung. It was thus of interest to find that in the simultaneous maternal and fetalinfusion experiments, the concentration of unlabelled VPA (i.e., that infused to theewe) was significantly higher in the fetal carotid artery than in the femoral artery,whereas this was not observed with labelled VPA infused via the fetal lateral tarsalvein. These results are opposite to those obtained by Tonn eta!. (1995b) forDPHM. Overall, the results obtained for the two drugs suggest that preferentialdistribution of maternally administered drug to the upper body of the fetus canoccur for compounds not taken up by the fetal lung. For xenobiotics that haveeffects on the fetal heart and/or brain (e.g., VPA), this preferential distribution tothe upper body could be of functional significance.Both VPA and[13C4]VPA were detected in amniotic and fetal trachealfluids during the course of the simultaneous infusions. Similar to the disposition ofVPA in the i.v. bolus studies reported above, VPA and[13C4]VPA did notaccumulate in these fluids to a significant degree, with the levels observed intracheal fluid being approximately an order of magnitude lower than theconcentrations observed in fetal serum. Comparisons and contrasts between theaccumulation of VPA and other compounds in amniotic and tracheal fluids havebeen previously discussed in section 4.3.1. As well, a further discussion of the low143levels of VPA observed in amniotic fluid in the infusion experiments will bepresented in section 4.4.2.4.4.1 Maternal and fetal clearance of VPA from serumThe serum steady-state levels of VPA and [13C4]VPA were applied to thetwo compartment, open model described by Szeto in order to calculate maternaland fetal placental and nonplacental clearances (Szeto et aL, 1 982b). Since onlyunbound drug can distribute easily between the maternal and fetal compartments,clearance estimates based on unbound drug concentrations will best estimateactual clearance values. Clearance calculations based on total drug levels on theother hand, will underestimate the true clearance rates. This factor is mostimportant when the extent of binding to maternal serum proteins is significantlydifferent from that observed for the fetus, which is not the case for VPA. If thereare differences in the extent of binding in the ewe and fetus, the relationshipbetween CLmf and CLfm will vary depending on whether total or unboundconcentrations are used in the calculations. The relative contribution of placentalclearance to total body clearance will, however, remain constant for both the eweand fetus. Thus, due to similar protein binding, calculations using total drug levelswill underestimate real clearance values. The relative proportions of CLmo / CLmand CLf0 I CLf will, however, remain constant, and the CLmf I CLfm ratiocalculated using total drug levels (0.51 ± 0.14) will closely approximate thatcalculated using unbound drug concentrations (0.58 ± 0.16).Maternal (CLm) and fetal (CLf) total body clearances calculated using totaldrug levels were 3.6 ± 0.9 mL mirn1 kg1 maternal weight and 63.0 ± 22.8 mL mirn1 fetal weight. Of these values, the maternal nonplacental clearance (CLmo)of 2.5 ± 0.7 mL min1 kg-1 maternal weight contributed to approximately 69 % of144CLm while the fetal nonplacental clearance (CLfo) of -4.3 ± 6.6 mL mm-1 kg-1 fetalweight apparently made no contribution to CLf. The CLm and CLf calculated usingunbound drug levels were 8.2 ± 1.9 mL mm-1 kg1 maternal weight and 128 ± 46mL mm-1 kg-1 fetal weight with the nonplacental contribution being the same asthat calculated using total drug levels.These results are interesting for several reasons. First of all, CLmo was lowcompared to other drugs that have been studied in sheep e.g., the reported CLmofor MCP and DPHM were 47±3 mL min1 kg-i and 44±3 mL mm-1 kg-1 maternalweight, respectively. Secondly, the total clearance value observed in sheep ishigher than the clearance observed in humans, which typically ranges from 0.4 -0.7 mL mm-1 kg1 (Levy and Shen, 1989). This result is consistent with sheep vs.human clearance comparisons for other drugs previously studied in our laboratorye.g., MCP (Riggs eta!., 1990) and DPHM (Yoo etaL, 1993). Thirdly, althoughhepatic blood flow is typically higher in sheep (- 0.5 - 3.0 L mimi) than in humans(‘ 0.5 - 1.5 L minl) (Altman and Dittner, 1972), VPA clearance is independent ofblood flow (Klotz and Antonin, 1977; Perucca eta!., 1978; Kaneko eta!., 1988) sothis would not be a contributing factor to observed differences in this case.However, since VPA is more highly protein bound in humans than in sheep andunbound drug is that which is available for elimination, one would expect to see ahigher rate of elimination from sheep unless the processes involved becamesaturated. Species differences in the renal ability to clear VPA will also beconsidered shortly.Perhaps the most notable result of these studies was the apparentcomplete absence of fetal nonplacental clearance (CLfo) of VPA, a finding that isunique in terms of the drugs studied thus far in this model (Yoo et al., 1993). TheCLf0 values calculated for compounds such as methadone, MCP and DPHM werecomparable to or larger than their respective CLmo values on a weight normalized145basis. In contrast, however, the reported nonplacental clearances of the acidiccompounds indomethacin (Anderson etal., 1980a) and acetylsalicylic acid (ASA)(Anderson et a!., 1 980b) are also small relative to values observed for other non-acidic compounds i.e., 4 ± 1 and 2 ± 0.2 mL mirn1 kg-1 fetal weight, forindomethacin and ASA, respectively. These results may be indicative of a lowfetal ability to effectively clear acidic compounds via nonplacental mechanismssuch as hepatic biotransformation and renal excretion. These potential routes offetal VPA elimination will be discussed below.The maternal (CLmf) and fetal (CLfm) transplacental clearances for VPAcalculated using total drug levels were 36.5 ± 20.8 and 67.3 ± 22.5 mL mirn1 kg1fetal weight, respectively, and 82.8 ± 47.7 and 136 ± 45 mL mirn1 kg1 fetal weightrespectively, based on unbound drug levels. These transplacental clearancevalues for VPA are lower than the transpiacental clearances reported for MCP(Riggs etaL, 1990), DPHM (Yoo etal., 1993) and methadone (Szeto eta!.,1 982a), but larger than those reported for acetaminophen (Wang et al., 1986) andmorphine (Szeto eta!., 1 982a). Similar to the estimates of fetal exposuredescribed earlier, it is most likely that these differences are due to differences inplacental permeability of these compounds in sheep. The fetal-to-maternaltransplacental clearance of VPA (CLfm) was found to be larger than the rate ofclearance in the opposite direction (CLmf). This finding is consistent with data forseveral other compounds that have been examined in sheep using the same twocompartment model, such as MCP, DPHM, morphine (Szeto eta!., 1982a) andmethadone (Szeto eta!., 1 982a). Yoo et a!., 1993 constructed a plot of the CLfm -CLmf difference versus CLfm for several compounds (acetaminophen, morphine,methadone, MCP and DPHM), with the clearance values based on free druglevels. Based on this plot, Yoo et a!. reported a linear relationship of y = 0.26x -8.4 between the magnitude of the difference between transplacental clearances (y)146and CLfm (x). For VPA, the magnitude of the difference in the calculatedclearance values (53.2 mL mirn1 kg-i fetal weight) was approximately twice that ofthe value determined from interpolation of the regression line (26.96 mL mirn1 kg1fetal weight) using 136 mL mirn1 kg-i fetal weight as the x-value. This suggeststhat the data from VPA agrees with this proposed relationship within the error limitsof the measured clearance values.As described earlier, the primary problems associated with applying theSzeto model (Szeto eta!., 1 982b) to calculating drug clearances in sheep are the“washout period” between maternal and fetal infusions if labelled drug isunavailable, and the need for measurable drug concentrations on the placentalside opposite the site of drug infusion. These concerns were not an issue in thepresent studies as a suitable stable isotope labelled analogue of VPA (along withan appropriate analytical assay) was available, and the transplacental transfer ofVPA was sufficient to easily provide measurable drug levels on either side of theplacenta. With umbilical venous serum and blood flow data available from two ofthe animals studied, the Fick principal could be applied to calculate CLfm valuesfor purposes of comparison to those calculated using the Szeto method. Theclearance values calculated in this manner, as denoted in section 3.5.2, werelower than, but not unreasonably different from, the values calculated using thetwo compartment, open model. Since the arterial-venous concentrationdifferences were small relative to the concentrations themselves, it is likely that theerror involved in calculating these estimates was significant and furthercompounded by the small sample size (i.e., n=2).1474.4.2 Maternal and fetal renal clearance of VPACollection of maternal and fetal urine and serum during the course of thesimultaneous infusions permitted the estimation of maternal and fetal renalclearances for VPA and [13C4]VPA, respectively. The maternal renal clearancesof unconjugated and total drug were 0.0917 ± 0.0639 and 0.521 ± 0.071 mL mirn1kg-1 maternal weight, respectively. The renal clearance of VPA accounts forapproximately 20 % of the total maternal nonplacental clearance (CLmo)observed. This clearance value is significantly less than the glomerular filtrationrate (GFR) in sheep which is reported to be approximately 2.4 mL mirn1 kgmaternal weight (Hill and Lumbers, 1988). These results are consistent withreports from humans in which the renal clearance of free drug (approximately 2 - 4mL mm-1) is small relative to the glomerular filtration rate (Levy and Shen, 1989).However, renal excretion appears to play a more prominent role in VPA eliminationin the sheep than the human, where its contribution is very minor. Although thesampling protocol employed in these studies did not include the collection of urineto the end of drug excretion i.e., the time when drug was no longer observable inurine, it appears that a higher percentage of a dose is excreted unchanged insheep since on average, roughly 2 % of the dose was recovered unchanged fromurine during the six hours of the infusion itself without post-infusion urinecollection. In comparison, approximately 1 to 3 % of a dose of VPA is recoveredunchanged from urine from adult humans following complete urine collection (Levyand Shen, 1989). These small rates of renal clearance of free VPA compared totypical glomerular filtration rates suggest that VPA is significantly reabsorbed in therenal tubule in both species. The low renal clearance of VPA relative to GFR alsosuggests that renal secretion of VPA is negligible in these species.148In contrast, there was essentially no renal excretion of VPA from the fetus.The excretion rates of 0.0315 ± 0.0320 and 0.0433 ± 0.0412 mL mm-1 kg-1 fetalweight for unconjugated and total drug are negligible and as mentioned previously,not significantly different from 0. These rates are significantly below the fetal GFRof approximately 1.03 mL min1 kg-1 fetal weight (Hill and Lumbers, 1988).Similarly small rates of fetal renal excretion have been reported for other acidiccompounds such as para-aminohippurate (PAH) (Elbourne etal., 1990),acetaminophen conjugates (Wang etal., 1986), morphmne-3-glucuronide (Olsen eta!., 1988) and most recently, diphenylmethoxyacetic acid (DPMA), a deaminatedmetabolite of DPHM (Tonn etaa’., 1 995b). It appears that while the mechanismsrequired for the renal secretion of organic cations (bases) are developed in thefetal Iamb, similar pathways for organic anions (acids) are not functional duringgestation (Elbourne etah, 1990). These pathways are, however, active in adultsheep. This disparity in fetal renal ability may not be the case in other species,however, as substantial excretion of PAH has been observed in fetal pigs (Alt etal., 1984). The absence of any significant renal excretion of VPA is consistent withthe lack of fetal nonpiacental clearance of VPA discussed above. Given the lategestational age of the fetal lambs studied in the present work, the onset ofeffective renal excretion of such compounds must occur after birth, however, thetime course of the onset and development of these processes remains unknown.Studies involving post-natal lambs of differing age may provide more informationconcerning the development of the renal excretory processes in sheep.The low fetal renal clearance of VPA may partially explain the relatively lowlevels of VPA and [13C4]VPA detected in amniotic fluid since fetal urine is themajor contributing component to the amniotic and allantoic fluid compartments.The appearance of drug in amniotic fluid via renal excretion from the fetal lambhas been observed for several compounds including meperidine, lidocaine and149ethanol (Morishima etaL, 1979; Szeto eta!., 1979; Clarke eta!., 1987). However,it has been demonstrated that accumulation of meperidine (Szeto eta!., 1979) andethanol (Clarke et al., 1987) still occurs in amniotic and allantoic fluids followingcomplete diversion of fetal urine. This suggests that other routes of entry such astransfer across the chorioallantoic membranes from the fetal and/or maternalcirculations may also play a role in drug transfer for some compounds (Rurak etal., 1991). It is thought that some acidic compounds that undergo little renalexcretion, such as morphine 3-glucuronide and acetaminophen glucuronide, mayalso cross the amniotic and allantoic membranes and enter these fluidcompartments (Wang eta!., 1986; Olsen eta!., 1988). However, the proteinbinding of these compounds tends to be very low e.g., plasma bound fractions of 3and 0% for acetaminophen glucuronide and morphine-3-glucuronide, respectively,and thus should not interfere with trans-membrane transport (Wang etal., 1986;Olsen etal., 1988). This may not be true for the more highly bound VPA since thedrug did not accumulate in amniotic fluid in the present studies. However, VPA isknown to cross other physiological membranes such as the blood brain barrier soit might be expected to traverse amniotic and allantoic membranes to some extent.VPA has been demonstrated to rapidly enter both cerebrospinal fluid (CSF)and brain tissue in mice (Loscher and Esenwein, 1978), rats (Hariton etal., 1984)and the rhesus monkey (Levy etaL, 1981). The successful treatment of statusepilepticus with single-dose oral or rectal administration of VPA suggests that rapidpenetration into the brain also occurs in humans (Vajda et al., 1978). The timecourse of VPA disposition in CSF has been reported to correlate well with bothtotal and free drug levels in plasma in both monkeys (Levy et a!., 1981) andhumans (Levy and Shen, 1989), indicating that there is relatively rapid equilibrationof VPA between brain and capillary blood (Levy and Shen, 1989). As mentionedearlier, the concentrations of VPA found in CSF have been found to correspond150fairly well with unbound drug levels in plasma (Rapaport et a!., 1983). This rapiduptake and equilibration of VPA has prompted work devoted to the investigation ofthe mechanisms of VPA transport across the blood-brain and blood-CSF barriers,particularly since it is unusual for a relatively strong carboxylic acid such as VPA,which is predominantly ionized at physiological pH (pKa = 4.56), to diffuse acrosslipoidal capillary membranes. Studies have suggested that VPA utilizes a carrier-mediated transport mechanism to cross these barriers (Frey and Loscher, 1978;Loscher and Frey, 1984). The exact nature of these mechanisms remains to bedetermined, although there is evidence, such as the observation of differing kineticrate constants for the influx and efflux of VPA from the rat brain (Cornfield et al.,1985), to suggest that separate transport systems may be involved in the influxand efflux of the drug.4.4.3 Physiological Effects4.4.3.1 Breathing activity and respirationIn a similar fashion to the effects noted following bolus dosing, VPA I[13C4]VPA administration in these studies produced a significant decrease in FBMover the course of the infusion period. A similar pattern of inhibition of fetalbreathing has been observed upon administration of both diazepam (Piercy eta!.,1977) and phenobarbitone (Boddy eta!., 1976). Since apnoeic episodesinterrupting FBM have been found to correlate with high voltage electrocorticalactivity (EC0G activity) after approximately 120 days gestation, the experimentsdetailed here incorporated the monitoring of ECoG activity in order to examine apotential effect of VPA on this parameter. In contrast to expectations, however,the simultaneous infusion of VPA and [13C4]VPA did not elicit a significant change151in EC0G activity. The correlation of fetal breathing activity with low voltage ECoGactivity further extends to a relationship between these conditions and rapid eyemovement (REM) sleep (Dawes eta!., 1972). For this reason EOG activity wasalso monitored in the present studies. In these experiments there was a trendtoward lower EOG activity during the first three hours of VPA I [13C4]VPAadministration but again, the effect was not significant. Since only three animalsproduced usable EOG traces, the significance of an effect on EOG may beestablished if data from additional animals are collected.There are several factors involved in the regulation of post natal ventilationincluding nonchemical contributions such as: stimulation of either pulmonaryreceptors known as stretch receptors or of irritant receptors which are locatedthroughout the airways, respiratory components of visceral reflexes andbaroreceptor stimulation, effects of afferents from proprioceptors, and effects ofafferents from higher centres such as the hypothalamus and the limbic system(Ganong, 1989). However, these contributions typically play a relatively minor rolein the regulation of respiration. The more dominant mediators of respiration arethe respiratory chemoreceptors i.e., the carotid and aortic bodies and collections ofcells in the medulla, that are sensitive to changes in arterial P02 and Hconcentration and which send impulses to the respiratory centre to attempt toadjust ventilation in order to compensate for these changes. The receptors in thecarotid and aortic bodies are thought to be the primary sources of stimulationcaused by changes in arterial P02 and H+ concentration and, while these sites arealso affected by carbon dioxide levels, modification of ventilation due to changes inis thought to be mediated mainly through medullary chemoreceptors.The acidity (i.e., H concentration) of brain interstitial fluid bathing certainregions of the ventral surface of the medulla has been reported to be a primarydeterminant of central ventilatory drive (Schlaefke, 1981; Kazemi and Johnson,1521986). It is thought that the effects of CO2 on respiration are mainly due to itsmovement into brain interstitial fluid, where it in turn increases the H+concentration and stimulates the receptors sensitive to H+ in this region of thebrain (Ganong, 1 989). It is, therefore, likely that amino acid neurotransmittersinvolved in H+ metabolism in the brain, via metabolism of compounds such asammonia and bicarbonate, can indirectly alter ventilatory drive. GABA andglutamate, both of which are derived from the amino acid glutamine, areassociated with CO2 fixation in the citric acid cycle in the brain and hence couldaffect respiration via this pathway.There have been numerous reports concerning the importance ofGABAergic systems in the control of respiration. As mentioned above, a brainregion frequently reported to be involved in influencing respiratory activity andwhere GABA plays a role, is the chemosensitive site known as the intermediatearea (or Schlaefke’s area) on the ventral surface of the medulla (Yamada et al.,1982).Administration of GABA or the GABA agonist, muscimol, to cats viaintraventricular injection or directly onto the ventral surface produces a depressionin respiration culminating in apnea (Yamada etal., 1982; Gatti eta!., 1987). Asimilar role for endogenous GABA was observed in studies in which GABAdegradation in the brain was prevented by inhibition of GABA-T resulting in adecrease in ventilation (Hedner et aL, 1984). Furthermore, administration of theGABA antagonist, bicuculline, had the opposite effect of GABA and muscimol onrespiratory activity, and reversed the depressant effects of both of thesecompounds (Yamada etaL, 1982; Gatti etal., 1987). It is, therefore, likely thatGABA is involved in ventilatory control both indirectly, via H homeostasis, anddirectly through excitation of inhibitory neurons in brain regions such asSchlaefke’s area in the cat. In the cat, activation of these GABAergic receptors153with projections to medullary respiratory motorneurons may promote a decrease intidal volume and inspiratory force (Yamada eta!., 1982; Kazemi and Hoop, 1991).However, in rats, the main effect of increased endogenous GABA levels appearsto be a decrease in respiratory frequency caused by a prolongation of expiratoryduration and this is accompanied by a less significant decrease in tidal volume(Hedner etaL, 1984) This suggests that there may be species variability in themode of respiratory depression caused by GABA and this may, in part, be due todifferences in the distribution of GABA receptors between different brain regions(Kneussl eta!., 1986). It has been suggested that increases in Clflux promotedby the interaction of GABA with the GABAA receptor may underlie this mechanismof central depression of ventilation despite the differences in the observed mode ofventilatory depression (Kazemi and Hoop, 1991).There is now considerable evidence to suggest that prostaglandin E2(PGE2)plays an important role in the regulation of fetal breathing movements(Thorburn, 1992). Although prostaglandin synthesizing enzymes are present andactive to some degree in the brainstem during late fetal life, the major source ofPGE2 in the fetal circulation is the placenta, thus suggesting that the placentaplays a significant role in the regulation of FBM (Fowden eta!., 1987; Thorburn,1992). Thorburn (1992) suggests that the plasma concentration of PGE2,superimposed upon the descending state-related inhibition of FBM during highvoltage electrocortical activity, may set the level of activity of the respiratoryneurones. He hypothesizes that an increase in fetal plasma PGE2 levels downregulates PGE2 receptors and may inhibit endogenous PGE2 synthesis in thebrainstem. The down-regulation of the PGE2 receptors increases the threshold ordecreases the sensitivity of brainstem neurones to PGE2, so that FBM can resumedespite the presence of high PGE2 levels. Thorburn suggests that by graduallyincreasing the threshold to PGE2,the fetus can adapt to the progressive increases154in plasma PGE2 levels and hence, maintain its characteristic episodic breathingpattern.Studies involving the denervation of the peripheral chemoreceptors haveindicated that these chemoreceptors may contribute to the control of FBM byexerting a tonic stimulation of the respiratory centre (Murai eta!., 1985), althoughthe same authors later concluded that the effects of PGE2 on FBM are notmediated by these receptors (Murai et a!., 1987). Thorburn’s (1992) interpretationof the data presented by Murai and colleagues (Murai etaL, 1985; Murai eta!.,1987) is that PGE2 maintains a tonic inhibitory influence on the peripheralchemoreceptors which, in turn, maintain a tonic stimulatory drive on the centralrespiratory neurones. These data and those reported above concerning GABAsuggest that both PGE2 and GABAergic mechanisms exert inhibitory influences onfetal breathing activity, although the actions of these compounds appear to beindependent.The importance of fetal breathing activity should not be underestimated. Ithas been determined that the intrapulmonary pressure exerted by lung fluid alongwith the periodic oscillations in pulmonary pressure caused by breathing in uteroare essential for normal lung growth and in the general promotion of thedevelopment of the respiratory muscles (Wigglesworth and Desai, 1979; Ligginsand Kitterman, 1981). It is conceivable that if VPA were to induce significantchanges in FBM over the long term, the development of the fetal respiratorysystem could be adversely affected. However, long term fetal VPA administrationis required to determine whether the drug can inhibit fetal breathing on a prolongedbasis.1554.4.3.2 Cardiovascular effectsUnlike the i.v. bolus dosing studies, in the simultaneous infusionexperiments there were transient decreases in fetal oxygen saturation and P°2These changes in blood gas parameters were accompanied by an increase in fetallactate levels that persisted throughout the course of the infusions. Thesetransient changes in the oxygen content suggest that there was a transientdecrease in cardiac output in the initial stages of the infusion studies. This notionis supported by the observed decrease in fetal heart rate immediately followinginitiation of drug administration. The initial bradycardia was also accompanied bya short-term increase in arterial pressure. The appearance of these effects in onlythe simultaneous infusion studies is probably attributable to the higher drugconcentrations present in the fetus following simultaneous dosing.Lactate is produced from pyruvate via a reversible reaction which iscatalyzed by lactate dehydrogenase. The reduction of pyruvate to lactate involvesthe oxidation of the reduced form of nicotinamide adenine dinucleotide (NAD+)and hence lactate concentrations will depend on not only pyruvate concentration,but also on pH and the cellular redox state i.e., the NADH/NAD ratio (Ganong,1989). Similar to fatty acids, pyruvate enters the citric acid cycle followingconversion to an acetyl CoA form. The rise in lactate levels observed in thepresent study may be a result of a shift in the lactate/pyruvate equilibrium causedby an increase in pyruvate levels which in turn was due to an increase incompetition for conversion to acetyl C0A. Following the infusion period, thelactate/pyruvate equilibrium returned to normal levels as the accumulated pyruvateentered the citric acid cycle and hence lactate levels returned to pre-infusionvalues, as noted in the 24 hour fetal blood sample.156The studies on the influences of GABA on respiratory control describedabove have typically also noted cardiovascular effects following administration ofGABA or the GABA agonist, muscimol. GABA and drugs with GABA agonisticproperties promote both a decrease in arterial pressure and a decrease in heartrate, primarily via action at the ventral surface of the medulla (Yamada etal., 1982;Gillis eta!., 1984), the same region thought to be the primary source of GABA’srespiratory activity. Administration of a GABA antagonist, bicuculline, in this regionreversed the effects of both GABA and muscimol on arterial pressure, respirationand heart rate in the rat (Keeler eta!., 1984), but produced little effect on heart ratein the cat model (Yamada eta!., 1982). GABA antagonists administered to the catwithout the previous administration of GABA or a GABA agonist caused anincrease in arterial pressure, primarily via action on the forebrain region but alsoon the intermediate area of the ventral surface of the medulla (Yamada et al.,1982; Gillis eta!., 1984). These latter data suggest that endogenous release ofGABA in these areas might control sympathetic outflow to the vasculature.Studies in the rat using the direct application of bicuculline to the intermediate areaof the ventral surface of the medulla further corroborates this conclusion (Keeler etal., 1984). Furthermore, studies revealing that GABA does not further reducearterial pressure in the presence of a peripherally administered a-receptorantagonist, phentolamine, while a direct acting vasodilator such as hydralazine iscapable of further dilatory action, also support the notion of a GABAergic inhibitoryinfluence over sympathetic outflow, as does the observed prevention of the effectsof bicuculline in the rat by blockade of ganglia in the peripheral sympatheticpathway with pentolinium (Keeler eta!., 1984).While the bradycardia observed in the present work is consistent with thesereports, the accompanying increase in arterial pressure is not. There have beensome species differences noted in the effects of GABA and GABAergic157compounds although the trends of effects have been consistent. The factorsinvolved in the observed increases in fetal arterial pressure remain unclear.4.4.3.3 Modelling of physiological effectsThe studies undertaken here have revealed a number of physiologicalchanges induced in the fetus through the administration of VPA. Despite theability to measure both the concentration of VPA present in serum and themagnitude of effects, it is difficult to establish a concentration-effect relationshipthrough the modelling of this data for various reasons. First of all, the nature of thesampling protocols used is such that there are relatively few concentration-effectdata pairs with which to work. For example, the most robust of the physiologicaleffects, the reduction in FBM following VPA administration, has been measured asa function of the number of minutes spent breathing per hour. Reducing the lengthof each sampling point Le., measuring the number of active minutes per half hour,is not a viable option in this instance because the normal cycle of fetal activity isapproximately 40 minutes requiring sampling periods be at least that long. Theother effects such as the transient increase in arterial pressure and decrease infetal heart rate are also difficult to model because of the limited number ofconcentration points collected in the infusion studies (8). Aside from theselimitations caused by the experimental protocols used, the majority of methods ofanalysis applied to kinetic data treat the measured effect as the terminal event of astochastic process of one or more steps. That is, it is assumed that the drug inquestion interacts with its site of action, a signal resulting from this action istransduced, propagated and the effect is observed (Veng-Pedersen and Gillespie,1988). However, this approach cannot be applied to many of the effects noted inthe present work as the modulation of pharmacologic responses by homeostatic158mechanisms may greatly complicate the interpretation of the data. One must alsoconsider the potential effects caused by the production of an active metabolite oran endogenous modulator in the system in question. For these reasons, it was feltthat the data collected were inadequate for use in statistically-sound modellingalgorithms. However, sampling protocols in future studies could be designed toprovide a larger amount of data which may provide some information via standardmodelling techniques.4.4.4 Metabolism4.4.4.1 VPA metabolites in maternal and fetal serumTwelve metabolites of VPA were detected in serum during the simultaneousinfusion experiments. Of these twelve, ten were also observed in 13C4-Iabelledform. The metabolites detected in these experiments corresponded to thoseobserved following i.v. bolus dosing of VPA. As expected, the levels ofmetabolites observed were typically higher in the infusion studies (see Figures 23-34) although these higher levels did not reveal the presence of any of the dienemetabolites of VPA as anticipated.As noted earlier, there is substantial evidence that VPA enters into the fattyacid type f-oxidation pathways in sheep as confirmed by the presence of the twomono-ene metabolites of VPA, 2-ene VPA and 3-ene VPA, as well as thesubsequent f3-oxidation products, 3-OH VPA and 3-keto VPA. However, theabsence of detectable diene metabolites along with lower than anticipated levels ofthe primary f3-oxidation products e.g., 2-ene VPA suggest that unlike the case inhumans, 13-oxidation is not a major pathway of VPA elimination in the sheep. The1592,3’-diene VPA isomers are 13-oxidation products of 3-ene VPA which is formedprimarily through an isomerization of 2-ene VPA (Bjorge and Baillie, 1991) andhence should be observed if this pathway were present in sheep in a fashionsimilar to that observed in humans. Furthermore, not only is 2,4-diene VPA adirect microsomal 13-oxidation product of 4-ene VPA (Rettenmeier et al., 1985;Rettenmejer etal., 1986) but it can also arise as a microsomal metabolite of 2-eneVPA (Kassahun and Baillie, 1993). Studies in rats have determined thatcytochrome P450 2B112 is responsible for both the conversion of VPA to 4-eneVPA and of 2-ene VPA to 2,4-diene VPA (Panesar, 1993). This may not be true insheep since 4-ene VPA was observed in the present studies while 2,4-diene VPAwas not. All of the other commonly observed biotransformation pathways for VPAappear to be active as evidenced by the presence of 5-OH VPA and 2-PGA (cooxidation), 4-OH VPA, 4-keto VPA and 2-PSA (0)-i oxidation), and ödehydrogenation (4-ene VPA).The fact that 13-oxidation appears to play a less significant role in VPAmetabolism in sheep compared to other species, such as rats and humans, isconsistent with reports of the lower activities of hepatic metabolizing enzymes forfatty acids in ruminants compared to other species (Bell, 1979). It has beenreported that the ruminant liver’s capacity for desaturation, esterification andoxidation, including 13-oxidation, of short and long chain fatty acids is lower thanthat observed in other species such as the rat and the chicken (Bell, 1979). Thesedifferences may be greater for branched chain fatty acids such as VPA. Thesedata would suggest that the sheep is not well suited to the study of thebiotransformation of compounds which utilize these pathways (e.g., 13-oxidation).In vitro microsomal studies that compare the rates of VPA metabolism in sheep tothose in other species, such as the rat, would be useful for further assessing thesuitability of the sheep for this line of investigation.160The maternal and fetal metabolite profiles illustrated in Figures 23-34 furtheraddress the issue of the apparent absence of fetal nonplacental clearance in thismodel. As mentioned earlier, the ratio of the levels of13C4labelled metabolite infetal serum-to-maternal serum over the first two hours of the infusion period wasless than one for each metabolite observed in these studies (see Table 27). Sincethe[13C4]VPA was infused to the fetus in these studies, one would expect to findhigher levels of the labelled metabolites in fetal serum, if they were in factproduced in fetal tissues. The fact that the opposite case was observed suggeststhat [13C4]VPA undergoes transfer to the ewe followed by metabolism and theresulting labelled metabolites are then transferred to the fetus down aconcentration gradient. These data further support the apparent lack of fetalnonplacental clearance of VPA. It is interesting to note that ratios calculated forthe last four hours of the study were, for the most part, higher than those observedfor the same compound during the first two hours of the experiment. In fact,metabolites such as 5-OH VPA and 3-keto VPA in both unlabelled and labelledform appear to accumulate on the fetal side of the placenta, indicating that eitherthese compounds are being trapped there and/or there is an active transport ofthese compounds across the placenta in the fetal direction. It appears unlikely thatthese metabolites could be trapped in fetal serum due to a pH effect, as maternaland fetal serum pH generally agree well with each other. However, thisphenomenon may be better understood by assessing the protein binding of thesemetabolites in maternal and fetal blood, since differences in binding may contributeto the apparent fetal accumulation of these compounds.An examination of the oxidative biotransformation of VPA in liver, lung,brain and adrenal homogenates from human conceptuses with gestational agesranging from 50 to 77 days revealed the production of 3-OH VPA, 4-OH VPA and5-OH VPA by these tissues following incubation with sodium valproate (Rettie et161a!., 1986). The reaction rates observed in the fetal adrenal homogenate wereapproximately four-fold higher than those observed in liver homogenates and tenfold higher than the rates observed in fetal brain and lung. Rettie et al. alsoassessed the embryotoxicity of VPA and the metabolites formed in human fetaltissue in cultured whole rat embryos. It was found that while VPA itself (0.8 mmolL-1) was highly embryotoxic, none of the metabolites produced in the human fetaltissue exhibited significant embryotoxicity at equimolar concentrations (Rettie eta!., 1986). This implies that the hydroxylation of VPA in the human fetal tissues isa detoxification process (Rettie et aL, 1986). It is possible that fetal lamb tissuesalso have the ability to metabolize VPA, but the overall capacity may be insufficientfor this to be detected with the in vivo methodology used in these studies. Anassessment of fetal VPA metabolism using in vitro techniques may provide moreinformation concerning the metabolizing capacity of fetal lamb tissues.Furthermore, an examination of the onset and development of effective VPAbiotransformation in sheep using both in vivo and in vitro technologies with postnatal lambs of differing ages may provide a greater understanding of theprocesses involved in VPA disposition in sheep.4.4.4.2 VPA metabolites in maternal and fetal urineThe metabolites detected in maternal urine over the course of the infusionperiod are those that one would expect to observe based on knowledge of theVPA biotransformation products typically observed in humans. Unfortunately,urine was not collected beyond the infusion period, thus preventing the calculationof metabolite recoveries for comparison purposes. The complete absence ofmetabolites of VPA in fetal urine during the infusion period further suggests a fetalinability to excrete acidic compounds, and also supports the notion of an overallinability of the fetus to eliminate VPA via nonplacental mechanisms at this stage ofgestation.1621635. SUMMARY AND CONCLUSIONS5.1 Quantitation of VPA and metabolites in sheep biofluidsA gas chromatographic / mass spectrometric method for VPA and 16 of itsmetabolites (Yu et aL, 1995) has been developed and applied to the study of thematernal and fetal pharmacokinetics and metabolism of VPA in the chronicallyinstrumented pregnant sheep model. The assay provides a convenient method forthe separation and quantitation of VPA and 16 metabolites in a singlechromatographic run using electron impact ionization and the intense (M-57)fragment of the tBDMS derivative of the compounds of interest. The features ofthe assay include the following:I) The assay provides resolution of VPA and its metabolites including all ofthe isomeric unsaturated compounds;ii) For the first time, (E)- and (Z)-3-ene VPA have been completely resolvedusing tBDMS derivatives;iii) The assay demonstrates excellent inter- and intra-assay precision withcoefficients of variation of < 10 % for VPA and most metabolites;iv) The assay displays good sensitivity with limits of quantitation for themetabolites ranging between 3 and 20 ng/mL; andv) The assay makes use of stable isotope[2H7]-labelled VPA andmetabolites as internal standards.5.2 Nonpregnant animal bolus studiesStudies were conducted in three nonpregnant sheep to examine the effectof isotope labelling on VPA pharmacokinetics using the [13C4]VPA analogue.164The terminal elimination half-lives for VPA and [13C4]VPA in the sheep are 3.5 ±0.4 h and 3.2 ± 0.4 h, respectively, while the mean VPA /[13C4]VPA area underthe serum concentration - time curve ratio (AUCOO,VPA/AUCOO,[13C]V )wasdetermined to be 1.03 ± 0.02. The 11 metabolites of VPA detected in theseanimals following i.v. bolus dosing were the following: (E) and (Z)-2-ene VPA, (E)and (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. Comparison of the serum levels and urine recoveries ofthe labelled and unlabelled compounds suggests similar dispositions for each pairof metabolite analogues examined. Thus, [13C4]VPA appears to bepharmacokinetically bioequivalent to VPA in the sheep model based oncomparisons of AUCOO values, terminal elimination half-lives and recovery ofmetabolites.5.3 Maternal i.v. bolus studiesMaternal i.v. bolus studies were performed in five chronically instrumentedpregnant ewes in order to determine basic pharmacokinetic parameters andestimate fetal exposure to VPA. In these studies, VPA reached maximalconcentrations of approximately 100 p.g/mL in the mother at the earliest samplingtime following administration of the 1000 mg dose i.e., 2 mm. The disposition ofVPA in maternal serum was best described by a biexponential function of the formAe-at + Be-E’ with at1 for VPA of 2.13 ± 0.49 h in maternal serum and a weightnormalized total body clearance (CLT) for the ewe of 0.083 ± 0.027 Lh-1kg1. VPAappeared very rapidly in fetal serum and the resulting serum levels producedsignificant fetal exposure to VPA with a mean AUCOOFA / AUCOOMA of 0.410 ±0.118. VPA also appeared rapidly in amniotic and fetal tracheal fluid however theconcentrations observed in these fluids did not rise above 3 pg/mL at any time.165The same eleven metabolites of VPA detected in serum following i.v. bolusdrug administration in nonpregnant sheep were also observed in maternal serumfollowing maternal dosing. Of these, only the metabolites detected in the highestconcentrations in maternal serum were detected in fetal serum i.e., (E)-2-ene VPA,4-ene VPA, 3-keto VPA, 4-OH VPA and 5-OH VPA, and for these metabolites,maternal serum concentrations were higher than those in the fetus.Maternal i.v. bolus administration of VPA (1000 mg) did not produce asignificant effect on fetal cardiovascular parameters such as heart rate and arterialpressure or on the fetal blood gas, acid-base and metabolic parameters monitoredduring these studies. There was, however, a significant reduction in fetal breathingactivity for approximately one hour following maternal drug administration.5.4 Fetal i.v. bolus studiesFetal i.v. bolus studies were performed in five chronically instrumentedpregnant ewes in order to determine basic fetal pharmacokinetic parameters forVPA. VPA reached maximal concentrations of approximately 100 p.g/mL in thefetus at the earliest sampling time following administration of the 250 mg dose i.e.,2 mm. The disposition of VPA in fetal serum was best described by abiexponential function of the form Aeat + Be. Thet1 for VPA in fetal serumwas 3.37 ± 1.37 h, which was not significantly different from that observed in theewe following maternal i.v. bolus dosing. The weight normalized total bodyclearance (CLT) for the fetus was 0.529 ± 0.306 Lh1 kg-1, a rate which wassignificantly higher than that observed in the ewe on a weight normalized basis.VPA appeared very rapidly in maternal serum as well as amniotic and fetaltracheal fluid following fetal dosing. While maternal serum concentrations of VPAapproximated and paralleled those in fetal serum beginning roughly two hours166following drug administration, amniotic and fetal tracheal fluid concentrationsremained low and did not accumulate extensively. This finding contrasts with theobserved accumulation of various basic, amine drugs such as metoclopramide,diphenhydramine and ritodrine in amniotic and tracheal fluids.The levels of metabolites of VPA observed following fetal dosing were lowerthan those observed in earlier experiments and hence fewer metabolites weredetected in these studies. The following metabolites were detected in bothmaternal and fetal serum collected following fetal drug administration althoughhigher levels were typically observed in maternal serum: (E)-2-ene VPA, (E)-3-eneVPA, 3-keto VPA, 4-keto VPA, 4-OH VPA, 5-OH VPA and 2-PGA.In a fashion similar to the maternal dosing studies, fetal i.v. bolusadministration of VPA (250 mg) did not produce a significant effect on fetalcardiovascular parameters such as heart rate and arterial pressure or on fetalblood gas, acid-base and metabolic parameters, except for a small but significantincrease in arterial lactate concentrations in the first hour following dosing. Onceagain, there was a significant reduction in fetal breathing activity following drugadministration. The duration of the observed effect was greater following fetaldosing, however, lasting approximately 2- 3 hours post drug administration.5.5 Simultaneous infusion studiesSimultaneous i.v. infusions to steady-state studies were performed in fivechronically instrumented pregnant ewes in order to further assess maternal andfetal placental and nonplacental drug clearances and drug biotransformationprocesses, as well as the effects of VPA on the fetus. Steady-state serumconcentrations for VPA and [13C4]VPA were achieved after approximately threehours of drug infusion, with steady-state concentration ratios of VPA and167[13C4]VPA in fetal serum to that in maternal serum (i.e., Cfss I Cmss) of 0.55 ±0.11 and 1.90 ± 0.38, respectively. The mean percentage of drug bound inmaternal serum was 49.7 ± 14.6% and 62.0 ± 9.1 % for VPA and[13C4]VPA,respectively. In fetal serum, the mean percentage of drug bound was 45.3 ± 8.1 %and 55.9 ± 7.4 % for VPA and[13C4]VPA, respectively. Both total and unboundserum steady-state levels of VPA and [13C4]VPA were applied to a twocompartment, open model in order to calculate maternal and fetal placental andnonplacental clearances. Maternal (CLm) and fetal (CLf) total body clearancescalculated using total drug levels were 3.6 ± 0.9 mL min1 kg-1 maternal weightand 63.0 ± 22.8 mL mirn1kg-1 fetal weight. Of these values, the maternalnonplacental clearance (CLmo) contributed approximately 69 % of CLm while thefetal nonplacental clearance (CLfo) apparently made no contribution to CLf. TheCLm and CLf calculated using unbound drug levels were 8.2 ± 1.9 mL mirn1 kg1maternal weight and 128 ± 46 mL min1 kg1 fetal weight with the nonpiacentalcontribution being the same as that calculated using total drug levels. Theseresults are indicative of a low fetal ability to effectively clear acidic compounds vianonplacental mechanisms such as hepatic biotransformation and renal excretion.This finding was supported by an examination of the urinary excretion ofdrug from the ewe and fetus. The maternal renal clearance of 0.521 ± 0.071 mLmm-1 kg-i maternal weight accounts for approximately 20 % of the total maternalnonpiacental clearance (CLmo). In contrast, there was essentially no renalexcretion of VPA from the fetus. The excretion rates of 0.0315 ± 0.0320 and0.0433 ± 0.0412 mL min1 kg-1 fetal weight for unconjugated and total drug arenegligible. Similarly, small rates of renal excretion from the fetus have beenreported for other acidic compounds such as para-aminohippuric acid (PAH),diphenylmethoxyacetic acid (DPMA) and indomethacin (IND).168Twelve metabolites of VPA were detected in serum during the simultaneousinfusion experiments. Of these twelve, ten were also observed in 13C4labelledform. The metabolites detected in these experiments corresponded to thoseobserved following i.v. bolus dosing of VPA. The biotransformation pathways mostcommonly observed in humans appear to be active in sheep as well, however, theabsence of detectable diene metabolites along with lower than anticipated levels ofthe primary p-oxidation products e.g., 2-ene VPA suggests that unlike the case inhumans, 3-oxidation is not a major pathway of VPA elimination in the sheep.Some metabolites of VPA such as 3-keto VPA and 5-OH VPA appear toaccumulate on the fetal side of the placenta. For each metabolite observed inthese studies, the ratio of the levels of3C4labelled metabolite in fetal serum-to-maternal serum over the first two hours of the infusion period was less than one.Since the [13C4]VPA was infused to the fetus in these studies, one would expectto find higher levels of the labelled metabolites in fetal serum, if they were in factproduced in fetal tissues. The fact that the opposite case is observed suggeststhat[13C4]VPA undergoes transfer to the ewe followed by metabolism, with theresulting labelled metabolites being transferred to the fetus down a concentrationgradient. This data further supports the apparent lack of fetal nonplacentalclearance of VPA. In light of these findings, it is apparent that the time course forthe development of the fetus/neonate’s mechanisms of VPA elimination warrantsfurther investigation.Unlike the bolus administration studies in which no significant blood gas,acid-base or metabolic effects were observed, the initiation of the drug infusionscaused a short term but significant drop in fetal oxygen saturation and P2• Thesechanges were observed five minutes after the onset of drug administration but haddissipated by sixty minutes post initiation. Furthermore, lactate levels increasedsignificantly from control values by five minutes into the infusion period and169remained elevated throughout the infusion period. Lactate concentrations returnedto levels not different from control values within 24 hours post-infusion. An initialand short-lived bradycardia was observed along with an increase in arterialpressure which lasted for approximately two hours after initiation of drugadministration. It is likely that these factors are all interconnected but the exactmechanisms involved in producing the observed effects remain unclear, althoughGABAergic mechanisms affecting sympathetic outflow to the vasculature throughareas such as the ventral surface of the medulla are probably involved.Once again, the administration of VPA produced a significant decrease inFBM over the course of the infusion period. However, the simultaneous infusionsdid not produce a significant effect on either EOG or EC0G activity, although thereis a trend toward lower EOG activity in the first three hours of the infusion periodrelative to control values. As mentioned above, both the simultaneous infusionand the bolus administration of VPA to either the ewe or fetus produced asubstantial depression in fetal breathing movements. This effect is consistent withreports on the effect of the GABA agonist, muscimol, on FBM in fetal sheep andalso with the effects observed with the GABA antagonist, picrotoxin in the samespecies. The mechanism by which VPA increases GABA levels and hence affectsFBM remains uncertain, however, the increased GABA levels associated with VPAtherapy are most often associated with one or more of the following: 1. Activationof GAD; 2. Inhibition of SSADH; and/or 3. Inhibition of synaptosomal GABA-T.In summary, the following conclusions can be made based on the studiespresented here:1701. The stable isotope labelled analogue of VPA, [13C4]VPA, isbioequivalent to VPA in sheep based on comparisons of AU COO values, terminalelimination half-lives and recovery of metabolites.2. VPA undergoes rapid and extensive placental transfer in the pregnantsheep. However, unlike basic compounds studied previously in this model, VPAdoes not accumulate extensively in either amniotic or fetal tracheal fluids despitethe rapid appearance of VPA in these fluids following drug administration.3. While the elimination half-life of VPA is similar in the ewe and fetus, thefetus displays a significantly larger total body clearance and volume of distributionof VPA than the ewe on a weight normalized basis.4. Although substantial transpiacental clearance of VPA was observed, therates of placental clearance of VPA are lower than those that have been observedfor diphenhydramine, metoclopramide and methadone. This difference is likelydue to a lower placental permeability for VPA in the sheep. In contrast, morphineand acetaminophen have been reported to have lower rates of transpiacentalclearance than those observed for VPA.5. Maternal nonplacental clearance accounts for approximately 69 % of thetotal maternal clearance of VPA. In contrast, the fetal nonplacental clearance ofVPA is negligible which indicates that the fetal lamb depends on placental transfermechanisms for the elimination of VPA. This finding is unique in terms of thedrugs studied in this model to date.6. VPA undergoes significant metabolism in the adult sheep, however, thefetal lamb appears to contribute little to VPA metabolism. Metabolite profile datasuggest that biotransformation of VPA occurs almost exclusively on the maternalside of the sheep placenta.1717. The VPA biotransformation pathways commonly observed in humansare also active in sheep, however, p-oxidation does not appear to be a majorelimination pathway for VPA in the sheep.8. The accumulation of some VPA metabolites in the fetus suggests eitherion trapping of metabolites in the fetus or, more likely, some form of activetransport across the placenta toward the fetus.9. The fetus exhibits a very limited ability to eliminate VPA via renalmechanisms. This low fetal renal clearance of VPA is consistent with theobserved negligible nonplacental clearance of VPA.10. The primary physiological effect observed in these studies was asignificant reduction in fetal breathing activity following VPA administration. VPA’sinfluence on GABAergic neurotransmission probably plays a role in the occurrenceof this action.1726. BIBLIOGRAPHYAbbott, F. and Acheampong, A. Quantitative structure-anticonvulsant activityrelationships of vaiproic acid, related carboxylic acids and tetrazoles.Neuropharmacology 1988, 27, 287-294.Abbott, F. S., Kassam, J., Acheampong, A., Ferguson, S., Panesar, S., Burton, R.,Farrell, K. and Orr, J. Capillary GC-MS of valproic acid metabolites in serum andurine using tert-butyldimethylsilyl derivatives. Journal of Chromatography 1986,375, 285.Acheampong, A., Abbott, F. and Burton, R. Identification of valproic acidmetabolites in human serum and urine using hexadeuterated valproic acid and gaschromatographic mass spectrometric analysis. Biomedical Mass Spectrometry1983, 10, 586-595.Acheampong, A. and Abbott, F. S. Synthesis and stereochemical determination ofdiunsaturated vaiproic acid analogs including its major diunsaturated metabolite.Journal of Lipid Research 1985, 26, 1002-1008.Alt, J. M., Colenbrander, B., Forsling, M. L. and Macdonald, A. A. Perinataldevelopment of tubular function in the pig. Quarterly Journal of ExperimentalPhysiology 1984, 69, 693-702.Biology Data Book; 2nd ed.; Altman, P. L. and Dittmer, D. S., Eds.; Federation ofAmerican Societies for Experimental Biology: Bethesda, 1972, pp 1703-1705.Anderson, D. F., Phernetton, T. M. and Rankin, J. H. G. The measurement ofplacental drug clearance in near-term sheep: Indomethacin. Journal ofPharmacology and Experimental Therapeutics 1 980a, 213, 100-104.Anderson, D. F., Phernetton, T. M. and Rankin, J. H. G. The placental transfer ofacetylsalicylic acid in near-term ewes. American Journal of Obstetrics andGynecology 1980b, 136, 814-818.Anderson, D. F., Phernetton, T. M. and Rankin, J. H. G. Prediction of fetal drugconcentrations. American Journal of Obstetrics and Gynecology 1 980c, 173, 735-738.Ardinger, H. H., Atkin, J. F., Blackston, R. D., Elsas, L. J., Clarren, S. K.,Livingstone, S., Flannery, D. B., Pellock, J. M., Harrod, M. J., Lammer, E. J.,Majewski, F., Schnizel, A., Toriello, H. V. and Hanson, J. W. Verification of thefetal valproate syndrome prototype. American Journal of Medical Genetics 1988,29, 171-185.Bailer, M., Hussein, Z., Raz, I., Abransky, 0., Herishanu, Y. and Pachys, F.Pharmacokinetics of valproic acid in volunteers after a single dose study.Biopharmaceutics and Drug Disposition 1985, 6, 33-42.173Baillie, T. A. and Rettenmeier, A. W. Vaiproate biotransformation. In AntiepilepticDrugs; R. Levy, R. Mattson, B. Meldrum, J. K. Penry and F. E. Dreifuss, Eds.;Raven Press, Ltd.: New York, 1989.Beeley, L. Clincal Obstetrics and Gynaecology 1986, 13, 197-214.Bell, A. W. Lipid metabolism in liver and selected tissues and in the whole body ofruminant animals. Progress in Lipid Research 1979, 18,117-164.Bend, J. R., Serabjit-Singh, C. J. and Philpot, R. M. The pulmonary uptake,accumulation and metabolism of xenobiotics. Annual Review of Pharmacology andToxicology 1985, 25, 97-1 25.Bjorge, S. M. and Baillie, T. A. Inhibiton of medium-chain fatty acid f-oxidation invitro by valproic acid and its unsaturated metabolite, 2-n-propyl-4-pentenoic acid.Biochemical and Biophysical Research Communications 1985, 132, 245-253.Bjorge, S. M. and Baillie, T. A. Studies on the f’-oxidation of valproic acid in ratliver mitochondrial preparations. Drug Metabolism and Disposition 1991, 19, 823-829.Boddy, K., Dawes, G. S., Fischer, R. L., Pinter, S. and Robinson, J. S. The effectsof phenobarbitone and pethidine of foetal breathing movements in sheep. BritishJournal of Pharmacology 1976, 57, 311-317.Bohan, T. P., Millington, D. S., Roe, C. R. and Yergey, A. L. Valproylcarnitine: Anovel metabolite of valproic acid. Ann. Neurol. 1984, 16, 394.Bossi, L. Fetal effects of anticonvulsants. In Antiepileptic drug therapy inpediatrics.; P. L. Morselli, Ed.; Raven Press: New York, 1983; pp 37-64.Bourgeois, B. F. D. Valproate clinical use. In Antiepileptic Drugs; R. Levy, R.Mattson, B. Meldrum, J. K. Penry and F. E. Dreifudd, Eds.; Raven Press, Ltd.:New York, 1989; pp 633-642.Bowdle, T. A., Patel, I. H., Levy, R. H. and Wilensky, A. J. Valproic acid dosageand plasma protein binding and clearance. Clinical Pharmacology andTherapeutics 1980, 28, 486 - 492.Bowes, G., Adamson, T. M., Ritchie, B. C., Dowling, M., Wilkinson, M. H. andMaloney, J. E. Development of patterns of respiratory activity in unaesthetizedfetal sheep in utero. Journal of Applied Physiology 1981, 50,693-700.Bray, R. E., Boe, R. W. and Johnston, W. L. Transfer of ampicillin into fetus andamniotic fluid from maternal plasma in late pregnancy. American Journal ofObstetrics and Gynecology 1966, 96, 938-942.Bruni, J. and Wilder, B. J. Valproic acid: Review of a new antiepileptic drug.Archives of Neurology 1979, 36, 393-398.174Chadwick, D. W. Valproate monotherapy in the management of generalized andpartial seizures. Epilepsia 1987, 28, Si 2-Si 7.Chapman, A., Keene, P. E., Meldrum, B. S., Simiand, J. and Vernieres, J. C.Mechanisms of anticonvulsant action of valproate. Progress in Neurobiology1982, 19, 315-359.Chapman, J. B. and Cutler, M. G. Sodium valproate: Effects on social behaviourand physical development in the mouse. Psychopharmacology 1984, 83, 390-396.Charlton, V. and Johengen, M. Nutrient and waste product concentrationdifferences in upper and lower body arteries of fetal sheep. Journal ofDevelopmental Physiology 1984, 6, 431-437.Ching, M. S., Jones, D. B., Morgan, D. J., Mihaly, G. W., Hardy, K. J. andSmaliwood, R. A. Fetal exposure to cimetidine following chronic administration topregnant sheep. Research Communications in Chemical Pathology andPharmacology 1985, 50, 139-i 42.Christianson, A. L., Chesler, N. and Kromberg, J. G. R. Fetal vaiproate syndrome:Clinical and neurodevelopmental features in two sibling pairs. DevelopmentalMedicine and Child Neurology 1994, 36, 357-369.Clarke, D. W., Patrick, J., Wlodek, M. E., Smith, G. N., Richardson, B. and Brien,J. F. The role of fetal urinary excretion in the transfer of ethanol into amniotic fluidafter maternal administration of ethanol to the near-term pregnant ewe. CanadianJournal of Physiology and Pharmacology 1987, 65, 1120-1124.Cloyd, J. C. and Kriel, R. L. Bioavailability of rectally administered vaiproic acidsyrup. Neurology 1981, 31, 1348-1352.Cohn, H. E., Sacks, E. J., Heyman, M. A. and Rudolph, A. M. Cardiovascularresponses to hypoxemia and acidemia in fetal lambs. American Journal ofObstetrics and Gynecology 1974, 120, 817-824.Cornfield, E. M., Diep, C. P. and Pardridge, W. M. Blood-brain barrier transport ofvalproic acid. Journal of Neurochemistry 1985, 44, 1541-1550.Craig, P. N. Comprehensive Medicinal Chemistry. The rational design, mechanisticstudy and therapeutic application of chemical compOunds.; Pergamon: New York,1990; Vol. 6, pp 237-965.Darius, J. and Meyer, F. P. Sensitive capillary gas chromatographic-massspectrometric method for the therapeutic drug monitoring of vaiproic acid andseven of its metabolites in human serum. Application of the assay for a group ofpediatric epileptics. Journal of Chromatography 1994, 656, 343-351.175Davis, R., Peters, D. H. and McTavish, D. Valproic acid: A reappraisal of itspharmacological properties and clinical efficacy in epilepsy. Drugs 1994, 47, 332-372.Dawes, G. S., Fox, H. E., Leduc, B. M., Liggins, G. C. and Richards, R. T.Respiratory movements and rapid eye movement sleep in the foetal lamb. Journalof Physiology (London) 1972, 220, 119-143.deJong, H. P. J. M., Elema, J. and Berg, B. v. d. Gas chromatography massspectrometry of t-butyldimethylsilyl derivatives of organic acids. Biomedical MassSpectrometry 1980, 7, 359.Devane, C. L. and Simpson, J. W. Pharmacokinetics of imipramine and its majormetabolites in pregnant rats and their fetuses following a single dose. DrugMetabolism and Disposition 1985, 33, 438-442.Dickinson, R. G., Hooper, W. D., Dunstan, P. R. and Eadie, M. J. Urinary excretionof vaiproate and some metabolites in chronically treated patients. Therapeuticsand Drug Monitoring 1989, 11, 127-133.Dickinson, R. G., Lawyer, C. H., Kaufman, S. N., Lynn, R. K., Gerber, N., Novy, M.J. and Cook, M. J. Materno-fetal pharmacokinetics and fetal distribution of valproicacid in a pregnant rhesus monkey. Pediatric Pharmacology 1980, 1, 71-83.DiLiberti, J. H., Farnadon, P. A., Dennis, N. R. and Curry, C. J. R. The fetalvalproate syndrome. American Journal of Medical Genetics 1984, 19, 473-481.Dreifuss, F. E. Vaiproate toxicity. In Antiepileptic Drugs; third ed.; R. Levy, R.Mattson, B. Meldrum, J. K. Penry and F. E. Dreifuss, Eds.; Raven Press, Ltd.: NewYork, 1989.Dreifuss, F. E., Santilli, N., Langer, D. H., Sweeney, K. P., Moline, K. A. andMenander, K. B. Valproic acid hepatic fatalities: A retrospective review. Neurology1987, 37, 379-385.Edelstone, D. I. and Rudolph, A. M. Preferential streaming of ductus venosusblood to the brain and heart in fetal lambs. American Journal of Physiology 1979,237, H724-H729.Ekwuru, M. 0. and Cunningham, J. R. Phaclofen increases GABA release fromvalproate treated rats. British Journal of Pharmacology 1990, 99, 251 P.Elbourne, I., Lumbers, E. R. and Hill, K. J. The secretion of organic acids andbases by the ovine fetal kidney. Experimental Physiology 1990, 75, 211-221.Faber, J. J. and Thornburg, K. L. Placental Physiology; Raven Press: New York,1983.176Fariello, R. and Smith, M. C. Vaiproate mechanisms of action. In AntiepilepticDrugs; R. Levy, R. Mattson, B. Meldrum, J. K. Penry and F. E. Dreifuss, Eds.;Raven Press, Ltd.: New York, 1989; pp 567-575.Fisher, E., Siemes, H., Pund, R., Wittfoht, W. and Nau, H. Vaiproate metabolites inserum and urine during antiepileptic therapy in children with infantile spasms:Abnormal metabo lite pattern associated with reversible hepatotoxicity. Epilepsia1992a, 33, 165-1 71.Fisher, E., Wittfoht, W. and Nau, H. Quantitative determination of valproic acid and14 metabolites in serum and urine by gas chromatography I mass spectrometry.Biomedical Chromatography 1 992b, 6, 24-29.Fowden, A. L., Harding, R., Ralph, M. M. and Thorburn, G. D. The nutritionalregulation of plasma prostaglandin E concentrations in the fetus and pregnant eweduring late gestation. Journal of Physiology 1987, 394, 1-12.Fowler, D. W., Eadie, M. J. and Dickinson, R. G. Transplacental transfer andbiotransformation studies of valproic acid and its glucuronide(s) in the perfusedhuman placenta. Journal of Pharmacology and Experimental Therapeutics 1989,249, 318-23.Franceschetti, S., Hamon, B. and Heineman, U. The action of valproate onspontaneous epileptiform activity in the absence of synaptic transmission and onevoked changes in [Ca+] and [K+] in the hippocampal slice. Brain Research 1986,386,1-11.Frey, H.-H. and Loscher, W. Distribution of valproate across the interface betweenblood and cerebrospinal fluid. Neuropharmacology 1978, 17, 637-642.Gaetani, E., Lauren, C. F. and Vitto, M. Ion trap detector-capillary gaschromatography of valproic acid and its mono-unsaturated metabolites in serumusing methyl ester derivatives. Journal of Pharmaceutical and Biomedical Analysis1992, 10, 193-1 97.Ganong, W. F. Review of Medical Physiology;l4th ed.; Appleton & Lange:Norwalk, 1989.Gatti, P. J., DaSilva, A. M. T. and Gillis, R. A. Cardiorespiratory effects producedby injecting drugs that affect GABA receptors into nuclei associated with theventral surface of the medulla. Neuropharmacology 1987, 26, 423-431.Gillis, R. A., Yamada, K. A., DiMicco, J. A., Williford, D. J., Segal, S. A., Hamosh,P. and Norman, W. P. Central gamma-aminobutyric acid involvement in bloodpressure control. Federation Proceedings 1984, 43, 32-8.Godin, Y., Heiner, L., Mark, J. and Mandel, P. Effects of di-n-propylacetate, ananticonvulsive compound, on GABA metabolism. Journal of Neurochemistry 1969,16, 869-873.177Gram, L., Larsson, 0. M., Johnsen, A. H. and Schousboe, A. Effects of valproate,vigabatrin and aminooxyacetic acid on release of endogenous and exogenousGABA from cultured neurons. Epilepsy Research 1988, 2, 87-95.Granneman, G. R., Wang, S. I., Machinist, J. M. and Kesterson, J. W. Aspects ofthe metabolism of vaiproic acid. Xenobiotica 1984, 14, 375-387.Gresham, E. L., Rankin, J. H. G., Makowski, E. L., Meschia, G. and Battaglia, F. C.An evaluation of fetal renal function in a chronic sheep preparation. Journal ofClinical Investigation 1972, 51, 149-156.Gugler, R. and Mueller, G. Plasma protein binding of valproic acid in healthysubjects and in patients with renal disease. British Journal of ClinicalPharmacology 1978, 5, 441 -446.Gugler, R., Schell, A., Eichelbaum, M., Froscher, W. and Schulz, H.-U. Dispositionof vaiproic acid in man. European Journal of Clinical Pharmacology 1977, 12,125-1 32.Gugler, R. and von-Unruh, G. E. Clinical pharmacokinetics of valproic acid. ClinicalPharmacokinetics 1980, 5, 67- 83.Haberer, L. J. and Pollock, G. M. Disposition and protein binding of valproic acid inthe developing rat. Drug Metabolism and Disposition 1994, 22, 113-119.Hamshaw-Thomas, A., Rogerson, N. and Reynolds, F. Transfer of bupivacaine,lignocaine and pethidine across rabbit placenta: Influence of maternal proteinbinding and fetal flow. Placenta 1984, 5, 61-70.Hariton, C., Ciesielski, L., Simler, S., Valli, M., Jadot, G., Gobaille, S., Mesdjian, E.and Mandel, P. Distribution of sodium valproate and GABA metabolism in CNS ofthe rat. Biopharmaceutics and Drug Disposition 1984, 5, 409-414.Hedner, J., Hedner, T., Bergman, B. and Lundberg, D. Respiratory depression byGABA-ergic drugs in the preterm rabbit. Journal of Developmental Physiology1980, 2, 401-407.Hedner, J., Hedner, T., Wessberg, P. and Jonason, J. An analysis of themechanism by which gamma-aminobutyric acid depresses ventilation in the rat.Journal of Applied Physiology: Respiratory, Environmental & Exercise Physiology1984, 56, 849-56.Hendrickx, A. G., Nau, H., Binkerd, P., Rowland, J. W., Rowland, J. R., Cukierski,M. J. and Cukierski, M. A. Vaiproic acid developmental toxicity andpharmacokinetics in the rhesus monkey: An interspecies comparison. Teratology1988, 38, 329-345.Hill, K. J. and Lumbers, E. R. Renal function in adult and fetal sheep. Journal ofDevelopmental Physiology 1988, 10, 149-159.178Holzer, P. and Hagmuller, K. Transient apnea after systemic injection of GABA inthe rat. Naunyn-Schmiedeberg’s Archives of Pharmacology 1979, 317, 3 15-320.ladarola, M. J. and Gale, K. Cellular compartments of GABA in brain and theirrelationship to anticonvulsant activity. Molecular Cell Biochemistry 1981, 39, 305-330.ladarola, M. J., Raines, A. and Gale, K. Differential effects of n-dipropylacetate andamino-oxyacetic acid on gamma-aminobutyric acid levels in discrete areas of ratbrain. Journal of Neurochemistry 1979, 33, 1119-1123.Jager-Roman, E., Deichl, A., Jakob, S., Hartmann, A.-M., Koch, S., Rating, D.,Steldinger, R., Nau, H. and Helge, H. Fetal growth, major malformations, andminor anomalies in infants born to women receiving valproic acid. Journal ofPediatrics 1986, 108, 997-1004.Janz, D. Antiepileptic drugs and pregnancy: Altered utilization patterns andteratogenesis. Epilepsia 1982, 23, S53-S63.Johnston, B. M. and Gluckman, P. D. GABA-mediated inhibition of breathing in thelate gestation sheep fetus. Journal of Developmental Physiology 1983, 5, 353-360.Kaneko, S., Otani, K., Fukushima, Y., Ogawa, Y., Nomura, Y., Ono, T., Nakane,Y., Teranishi, T. and M.Goto. Teratogenicity of antiepileptic drugs: Analysis ofpossible risk factors. Epilepsia 1988, 29, 459-467.Kasilo, 0., Romero, M., Bonati, M. and Togoni, G. Information on drug use inpregnancy from the viewpoint of a regional drug information centre. EuropeanJournal of Clinical Pharmacology 1988, 35, 447-453.Kassahun, K. and Baillie, T. A. Cytochrome P-450 mediated dehydrogenation of 2-n-propyl-2(E)-pentenoic acid, a pharmacologically-active metabolite of valproicacid, in rat liver microsomal preparations. Drug Metabolism and Disposition 1993,21, 242-248.Kassahun, K., Farrell, K. and Abbott, F. Identification and characterization of theglutathione and N-acetylcysteine conjugates of (E)-2-propyl-2,4-pentadienoic acid,a toxic metabolite of valproic acid, in rats and humans. Drug Metabolism andDisposition 1991, 19, 525-535.Kassahun, K., Farrell, K., Zheng, J. and Abbott, F. Metabolic profiling of valproicacid in patients using negative-ion chemical ionization gas chromatography-massspectrometry. Journal of Chromatography 1990, 527, 327-341.Kassahun, K., Hu, P., Grub, M. P., Davis, M. R., Jin, L. and Baillie, T. A. Metabolicactivation of unsaturated derivatives of vaiproic acid. Identification of novel179glutathione adducts formed through coenzyme A-dependent and -independentprocesses. Chemico-Biological Interactions 1994, 90, 253-275.Kazemi, H. and Hoop, B. Glutamic acid and gamma-aminobutyric acidneurotransmitters in central control of breathing. [Review]. Journal of AppliedPhysiology 1991, 70, 1-7.Kazemi, H. and Johnson, D. C. Regulation of cerebrospinal fluid acid-basebalance. Physiology Reviews 1986, 66, 953-1037.Keeler, J. R., Shults, C. W., Chase, T. N. and Helke, C. J. The ventral surface ofthe medulla in the rat: Pharmacologic and autoradiographic localization of GABAinduced cardiovascular effects. Brain Research 1984, 297, 217-224.Kelly, T. E. Teratogenicity of anticonvulsant drugs. I. Review of the literature.American Journal of Medical Genetics 1984, 19, 413-434.Kesterson, J. W., Granneman, G. R. and Machinist, J. M. The hepatotoxicity ofvalproic acid and its metabolites in rats. I. Toxicologic, biochemical andhistopathologic studies. Hepatology 1984, 4, 1143-1152.Klotz, U. and Antonin, K. H. Pharmacokinetics and bioavailability of sodiumvalproate. Clinical Pharmacology and Therapeutics 1977, 21, 736-743.Kneussl, M. P., Pappagianopoulos, P., Hoop, B. and Kazemi, H. Reversibledepression of ventilation and cardiovascular function by ventriculocisternalperfusion with gamma-aminobutyric acid in dogs. American Review of RespiratoryDisease 1986, 133, 1024-8.Knott, C. and Reynolds, F. Placental transfer of anticonvulsants in the rabbit: Theeffect of maternal protein binding and fetal flow. Placenta 1986, 7, 333-338.Kobayashi, S., Takai, K., Iga, T. and Hanano, M. Effect of pregnancy on thedisposition of valproate in rats. Journal of Pharmacoblo Dynamics 1990, 13, 533-42.Kobayashi, S., Takai, K., Iga, T. and Hanano, M. Pharmacokinetic analysis of thedisposition of valproate in pregnant rats. Drug Metabolism and Disposition 1991,19, 972-6.Koch, S., Gopfert-Geyer, I., Hauser, I., Hartmann, A., Jakob, S., Jager-Roman, E.,Nau, H., Rating, D. and Helge, H. Neonatal behavioral disturbances in infants ofepileptic women treated during pregnancy. In Prevention ofphysical, mental andcongenital defects, Part B: Epidemiology, early detection and therapy, andenvironmental factors, progress in dilnical and biological research.; Alan R. Liss,Inc.: New York, 1985; Vol. 163B; pp 453-461.Kochen, W. and Scheffner, H. On unsaturated metabolites of valproic acid (VPA)in serum of epileptic children. In Antiepileptic Therapy: Advances in Drug180Monitoring; S. I. Johannessen, P. L. Morselli, C. E. Pippenger, A. Richens, D.Schmidt and H. Meinardi, Eds.; Raven Press: New York, 1980; pp 1 11-120.Kochen, W., Sprunck, H. P., Tauscher, B. and Klemens, M. Five doublyunsaturated metabolites of valproic acid in urine and plasma of patients on valproicacid therapy. Journal of Clinical Chemistry and Clinical Biochemistry 1984, 22,309-317.Krauer, B. and Krauer, F. Drug kinetics in pregnancy. Clinical Pharmacokinetics1977,2, 167-1 81.Kwan, E. The effects of acute haemhorrhage and subsequent anaemia on thefetal lamb. M.Sc. Thesis, University of British Columbia, 1989.Laan, J. W. V. d., Boer, T. D. and Bruinvels, J. Di-n-propylacetate and GABAdegradation. Preferential inhibition of succinic semialdehyde dehydrogenase andindirect inhibition of GABA-transaminase. Journal of Neurochemistry 1979, 32,1769-1780.Lai, A. A., Levy, R. H. and Martis, L. Pharmacokinetic profile of valproic acid inrhesus monkeys following single bolus and constant rate intravenous infusion.Therapie 1980, 35, 221-232.Laishley, R. S., Carson, R. J. and Reynolds, F. Effect of adrenaline on thedistribution of bupivacaine in the rabbit fetus. British Journal of Anaesthiology1989, 63,721-725.Larsson, 0. M., Gram, L., Schousboe, I. and Schousboe, A. Differential effects ofgamma-vinyl GABA and valproate on GABA-transaminase from cultured neuronsand astrocytes. Neuropharmacology 1986, 25, 617-625.Lee, R. D., Kassahun, K. and Abbott, F. S. Stereoselective synthesis of thediunsaturated metabolites of valproic acid. Journal of Pharmaceutical Sciences1989, 78, 667-671.Levy, G. Pharmacokinetics of fetal and neonatal exposure to drugs. Obstetrics andGynecology 1981, 58, S9-S1 6.Levy, G. and Hayton, W. L. Pharmacokinetic aspects of placental drug transfer. InFetal Pharmacology; L. Boreus, Ed.; Raven Press: New York, 1973; pp 29-39.Levy, R. H. and Lai, A. A. Valproate: Absorption, distribution, and excretion. InAntiepileptic drugs; 2 ed.; D. M. Woodbury, J. K. Penry and C. E. Pippenger, Eds.;Raven Press: New York, 1982; pp 555-565.Levy, R. H., Lockard, J. S. and Ludwick, B. T. Nonlinear plasma protein bindingand CSF concentration of valproic acid in monkey. Epilepsia 1981, 22, 229-235.181Levy, R. H., Lockard, J. S., Patel, I. H. and Lai, A. A. Efficacy testing of vaiproicacid compared to ethosuximide in monkey model: I. Dosage regimen design in thepresence of diurnal oscillations. Epilepsia 1977, 18, 191-203.Levy, R. H. and Shen, D. D. Valproate absorption, distribution, and excretion. InAntiepileptic Drugs; 3 ed.; R. Levy, R. Mattson, B. Meidrum, J. K. Penry and F. E.Dreifudd, Eds.; Raven Press, Ltd.: New York, 1989; pp 583-599.Liggins, G. C. and Kitterman, J. A. Development of the fetal lung. In The Fetusand Independent Life; Pitman Ltd.: London, 1981; Vol. Ciba FoundationSymposium 86.Loiseau, P., Brachet, A. and Henry, P. Concentrations of dipropylacetate inplasma. Epilepsia 1975, 16, 609-615.Loscher, W. Serum protein binding and pharmacokinetics of valproate in man,dog, and mouse. Journal of Pharmacology and Experimental Therapeutics 1978,204, 255 -261.Loscher, W. Effects of inhibitors of GABA transaminase on the synthesis, binding,uptake, and metabolism of GABA. Journal of Neurochemistry 1980, 34, 1603-1608.Loscher, W. Anticonvulsant activity of metabolites of valproic acid. Arch. mt.Pharmacodyn. 1 981a, 249, 158-163.Loscher, W. Valproate induced changes in GABA metabolism at the subcellularlevel. Biochemical Pharmacology 1981b, 30, 1364-1 366.Loscher, W. Pharmacological evaluation of various metabolites and analogues ofvalproic acid. Anticonvulsant and toxic potencies in mice. Neuropharmacology1985, 24, 427-435.Loscher, W. Vaiproate enhances GABA turnover in the substantia nigra. BrainResearch 1989, 501, 198-203.Loscher, W. Effects of the antiepileptic drug valproate on metabolism and functionof inhibitory and excitatory amino acids in the brain. Neurochemical Research1993, 18, 485-502.Loscher, W. and Esenwein, H. Pharmacokinetics of sodium vaiproate in dog andmouse. Arzneim. Forsch. /Drug Research 1978,28, 782-787.Loscher, W. and Frey, H.-H. Kinetics of penetration of common antiepileptic drugsinto cerebrospinal fluid. Epilepsia 1984, 25, 346-352.Loscher, W. and Siemes, H. Valproic acid increases y-aminobutyric acid in CSF ofepileptic children. Lancet 1984, 2, 225.182Loscher, W. and Siemes, H. Cerebrospinal fluid y-aminobutyric acid levels inchildren with different types of epilepsy: effect of anticonvulsant treatment.Epilepsia 1985, 26, 314-319.Loscher, W. and Vetter, M. In vivo effects of aminooxyacetic acid and valproic acidon nerve terminal (synaptosomal) GABA levels in discrete brain area of the rat.Biochemical Pharmacology 1985, 34, 1747-1756.Lundberg, B., Nergardh, A. and Boreus, L. 0. Plasma concentrations of valproateduring maintenance therapy in epileptic children. Journal of Neurology 1982, 228,133-1 41.Macdonald, R. L. and Bergen, G. K. Valproic acid augments GABA-mediatedpostsynaptic inhibition in cultured mammalian neurons. Brain Research 1979,170, 558-562.Maitre, M., Ciesielski, L., Cash, C. and Mandel, P. Comparison of the structuralcharacteristics of the 4-aminobutyrate:2-oxoglutarate transaminases from rat andhuman brain, and of their affinities for certain inhibitors. Biochemical andBiophysicalActa 1978, 522, 385-399.Maitre, M., Ossola, L. and Mandel, P. In vitro studies into the effect of inhibition ofrat brain succinic semialdehyde dehydrogenase on GABA synthesis anddegradation. FEBS Letters 1976, 72, 53-57.McLean, M. J. and Macdonald, R. L. Sodium valproate, but not ethosuximide,produces use- and voltage-dependent limitation of high frequency repetitive firingof action potentials of mouse central neurons in cell culture. Journal ofPharmacology and Experimental Therapeutics 1986, 237, 1001-1011.Meinardi, H., Kleijn, E. v. d., Meijer, J. W. A. and Rees, H. v. Absorption anddistribution of antiepileptic drugs. Epilepsia 1975, 16, 353-365.Meschia, G., Battaglia, F. C. and Bruns, P. D. Theoretical and experimental studyof transplacental diffusion. Journal of Applie Physiology 1967, 22, 1171-1178.Meunier, G., Carrax, G., Meunier, Y., Eymard, P. and Aimart, M. Proprietespharmacodynamique de l’acide n-dipropylacetique. Therapie 1963, 18, 435-438.Mihaly, G. W. and Morgan, D. J. Placental drug transfer: Effects of gestational ageand species. Pharmacology and Therapeutics 1984, 23, 253-266.Millington, D. S., Bohan, T. P., Roe, C. R., Yergey, A. L. and Liberato, D. J.Vaiproylcarnitine: A novel drug metabolite identified by fast atom bombardmentand thermospray liquid chromatography-mass specrometry. Clinical Chimica Acta1985, 145, 69-76.183Monaco, F., Piredda, S., Mutani, R., Mastropaolo, C. and Tondi, M. The freefraction of vaiproic acid in tears, saliva, and cerebrospinal fluid. Epilepsia 1982,23, 23-26.Morishima, H. 0., Finster, M., Pedersen, H., Fukunaga, A., Ronfeld, R. A.,Vassallo, H. G. and Covino, B. G. Pharmacokinetics of lidocaine in fetal andneonatal lambs and adult sheep. Anesthesiology 1979, 50, 431 -436.Murai, D. T., Lee, C. H., Wallen, L. D., Clyman, R. L., Mauray, F., Clements, J. A.and Kitterman, J. A. Denervation of peripheral chemoreceptors decreasesbreathing movements in fetal sheep. Journal of Applied Physiology 1985, 59.Murai, D. T., Lee, C. H., Wallen, L. D., Clyman, R. L., Mauray, F. and Kitterman, J.A. Effects of prostaglandins on fetal breathing do not involve peripheralchemoreceptors. Journal of Applied Physiology 1987, 62, 271-277.Murphy, P. J. and Sullivan, H. R. Stable isotopes in pharmacokinetic studies.Annual Review of Pharmacology and Toxicology 1980, 20, 609-621.Nau, H. Species differences in pharmacokinetics and drug teratogenesis.Environmental Health Perspectives 1986a, 70, 113-129.Nau, H. Transfer of valproic acid and its main active unsaturated metabolite to thegestational tissue: Correlation with neural tube defect formation in the mouse.Teratology 1986b, 33, 21-27.Nau, H., Helge, H. and Luck, W. Vaiproic acid in the perinatal period: Decreasedmaternal serum protein binding results in fetal accumulation and neonataldisplacement of the drug and some metabolites. Journal of Pediatrics 1984, 104,627-634,Nau, H. and Krauer, B. Serum protein binding of valproic acid in fetus-mother pairsthroughout pregnancy: Correlation with oxytocin administration and albumin andfree fatty acid concentrations. Journal of Clinical Pharmacology 1986, 26, 215-221.Nau, H., Kuhnz, W., Egger, H.-J., Rating, D. and Helge, H. Anticonvulsants duringpregnancy and lactation: Transplacental, maternal and neonatal pharmacokinetics.Clinical Pharmacokinetics 1982, 7, 508-543.Nau, H. and Loscher, W. Valproic acid: Brain and plasma levels of the drug and itsmetabolites, anticonvulsant effects, and -aminobutyric acid (GABA) metabolism inthe mouse. Journal of Pharmacology and Experimental Therapeutics 1982, 220,654-659.Nau, H. and Loscher, W. Valproic acid and metabolites: Pharmacological andtoxicological studies. Epilepsia 1984, 25, S14-S22.184Nau, H., Wittfoht, W., Schafer, H., Jakobs, C., Rating, D. and Helge, H. Valproicacid and several metabolites: quantitative determinants in serum, urine, breastmilk and tissues by gas chromatography-mass spectrometry using selected ionmonitoring. Journal of Chromatography 1981, 226, 69.Okumura, K., Yoshida, H. and Hori, R. Tissue distribution and metabolism ofdrugs. Ill. Accumulation of drugs by the isolated perfused rat lung. Journal ofPharmacoblo Dynamics 1978, 1, 230-237.Olsen, G. D., Sommer, K. M., Wheeler, P. L., Boyea, S. R., Michaelson, S. P. andCheek, D. B. C. Accumulation and clearance of morphine-3-glucuronide in fetallambs. Journal of Pharmacology and Experimental Therapeutics 1988, 247, 576-584.Omtzigt, J. G., Nau, H., Los, F. J., Pijpers, L. and Lindhout, D. The disposition ofvaiproate and its metabolites in the late first trimester and early second trimester ofpregnancy in maternal serum, urine, and amniotic fluid: effect of dose, comedication, and the presence of spina bifida. European Journal of ClinicalPharmacology 1992, 43, 381-8.Panesar, S. Induction of VPA metabolism in rat liver microsomes bycarbamazepine and carbamazepine-1 0,11 -epoxide. Ph.D. Thesis, University ofBritish Columbia, 1993.Perucca, E., Gatti, G., Frigo, G. M. and Crema, A. Pharmacokinetics of vaiproicacid after oral and intravenous administration. British Journal of ClinicalPharmacology 1978, 5, 313-318.Phillipou, G. Subnanogram detection of t-butyldimethylsilyl fatty acid esters bymass fragmentography. Lipids 1975, 10, 714-716.Phillips, N. I. and Fowler, L. J. The effects of sodium valproate on y-aminobutyratemetabolism and behaviour in naive and ethanolamine-O-sulphate pre-treated ratsand mice. Biochemical Pharmacology 1982, 31, 2257-2261.Piercy, W. N., Day, M. A., Neims, A. H. and Williams, R. L. Alteration of ovine fetalrespiratory-like activity by diazepam, caffeine and doxapram. American Journal ofObstetrics and Gynecology 1977, 127, 43-49.Piper, J. M., Baum, P. H. C. and Kennedy, D. L. Prescription drug use before andduring pregnancy in a Medicaid population. American Journal of Obstetrics andGynecology 1987, 157, 148-156.Plasse, J. C., Revol, M., Chabert, G. and Ducerf, F. Neonatal pharmacokinetics ofvalproic acid. In Progress in Clinical Pharmacy, D. Schaaf and E. v. d. Kleijn,Eds.; Elsevier/North Holland Biomedical Press: Amsterdam, 1979; pp 247.185Rapaport, W. G., Mendelow, A. D., French, G., MacPherson, P. and Teasdale, E.Plasma protein-binding and CSF concentrations of valproic acid following acuteoral dosing. British Journal of Clinical Pharmacology 1983, 16, 365-369.Rettenmeier, A. W., Gordon, W. P., Barnes, H. and Baillie, T. A. Studies on themetabolic fate of valproic acid in the rat using stable isotope techniques.Xenobiotica 1987, 17,1147-1157.Rettenmeier, A. W., Gordon, W. P., Prickett, K. S., Levy, R. H. and Baillie, T. A.Biotransformation and pharmacokinetics in the rhesus monkey of 2-n-propyl-4-pentenoic acid, a toxic metabolite of vaiproic acid. Drug Metabolism andDisposition 1986, 14, 454-464.Rettenmeier, A. W., Howald, W. N., Levy, R. H., Witek, D. J., Gordon, W. P.,Porubek, D. J. and Baillie, T. A. Quantitative metabolic profiling of vaiproic acid inhumans using automated gas chromatographic/mass spectrometric techniques.Biomedical and Environmental Mass Spectrometry 1989, 18, 192-199.Rettenmeier, A. W., Prickett, K. S., Gordon, W. P., Bjorge, S. M., Chang, S. L.,Levy, R. H. and Baillie, T. A. Studies on the biotransformation in the perfused ratliver of 2-n-propyl-4-pentenoic acid, a metabolite of the antiepileptic drug valproicacid. Evidence for the formation of chemically reactive intermediates. DrugMetabolism and Disposition 1985, 13, 81-96.Rettie, A. E., Rettenmeier, A. W., Beyer, B. K., Baillie, T. A. and Juchau, M. R.Valproate hydroxylation by human fetal tissues and embryotoxicity of metabolites.Clinical Pharmacology and Therapeutics 1986, 40, 172-177.Rettie, A. E., Rettenmeier, A. W., Howald, W. N. and Baillie, T. A. Cytochrome P450-catalyzed formation of 4-ene VPA, a toxic metabolite of vaiproic acid. Science1987, 235, 890-893.Reynolds, F. and Knott, C. Pharmacokinetics in pregnancy and placental drugtransfer. Oxford Reviews of Reproductive Biology 1989, 11, 389-449.Riggs, K. W., Axelson, J. E., Gruber, N. C., McErlane, B. A., McMorland, G. H. andRurak, D. W. Metoclopramide pharmacokinetics in pregnant and nonpregnantsheep. Journal of Pharmaceutical Sciences 1988, 77,373-379.Riggs, K. W., Rurak, D. W., Taylor, S. M., McErlane, B. A., McMorland, G. H. andAxelson, J. E. Fetal and maternal placental and nonplacental clearances ofmetoclopramide in chronically instrumented pregnant sheep. Journal ofPharmaceutical Sciences 1990, 79, 1056-1061.Riggs, K. W., Rurak, D. W., Yoo, S. D., McErlane, B. A., McMorland, G. H. andAxelson, J. E. Drug accumulation in lung fluid of the fetal Iamb after fetal ormaternal administration. American Journal of Obstetrics and Gynecology 1987,157, 1286-1 291.186Roncucci, R., Simon, M.-J., Jacques, G. and Lambelin, G. Stable isotopes in drugmetabolism and pharmacokinetics. European Journal of Drug Metabolism andPharmacokinetics. 1975, 1, 9-20.Rosenfeld, W. E., Leppik, I. E., Gates, J. R. and Mireles, R. E. Vaiproic acidloading during intensive monitoring. Archives of Neurology 1987, 44, 709-710.Ross, S. M. and Craig, C. R. Studies on y-aminobutyric acid transport in cobaltexperimental epilepsy in the rat. Journal of Neurochemistry 1981, 36, 1006-1011.Rubin, P. C., Craig, G. S., Gavin, K. and Sumner, D. Prospective survey of use oftherapeutic drugs, alcohol and cigarettes during pregnancy. British Medical Journal1986, 292, 81-83.Rudolph, A. M. Animal models for study of drug exposure. In Prenatal DrugExposure: Kinetics and Dynamics; C. N. Chiang and C. C. Lee, Eds.; U.S.Department of Health and Human Services: Rockville, MD, 1985; Vol. NIDAResearch Monograph 60.Rurak, D. W., Wright, M. R. and Axelson, J. E. Drug disposition and effects in thefetus. Journal of Developmental Physiology 1991, 15, 33-44.Rurak, D. W., Yoo, S. D., Kwan, E., Taylor, S. M., Riggs, K. W. and Axelson, J. E.Effects of diphenhydramine in the fetal lamb after maternal or fetal administration.Journal of Pharmacology and Experimental Therapeutics 1988, 247, 271-278.Scheffner, D., Kong, S., Rauterberg-Ruland, I., Kochen, W., Hofman, W. J. andUnkelbach, S. Fatal liver failure in 16 children with valproate therapy. Epilepsia1988, 29, 530-542.Schlaefke, M. E. Central chemosensitivity: a respiratory drive. Review ofPhysiology and Biochemical Pharmacology 1981, 90, 171-244.Schoben, F., Kleijn, E. v. d. and Gabreels, F. J. M. Pharmacokinetics of di-npropylacetate in epileptic patients. European Journal of Clinical Pharmacology1975, 8, 97-1 05.Simler, S., Ciesielski, L., Klein, M., Gobaille, S. and Mandel, P. On themechanisms of action of the anticonvulsant, sodium di-n-propylacetate. C.R.Societe de Biologie (Paris) 1981, 175, 114-119.Simpson, J. L., Mills, J. L., Morey, A., Metzger, B. E., Knopp, R. H., Holmes, L. B.,Jovanovic-Peterson, L., Aarons, J. H., Brown, Z., Allen, M. V. and Elias, S. Drugingestion during pregnancy: infrequent exposure in a contemporary United Statessample. American Journal of Perinatology 1989, 6, 244-251.Singh, K., Abbott, F. S. and Orr, J. M. Capillary GCMS assay of 2-n-propyl-4-pentenoic acid (4-ene VPA) in rat plasma and urine. Research Communications inChemical Pathology and Pharmacology 1987, 56, 211-223.187Slater, G. E. and Johnston, D. Sodium valproate increases potassiumconductance in aplysia neurons. Epilepsia 1978, 19, 379-384.Snead, 0. C. y-Hydroxybutyrate model of generalized absence seizures: furthercharacterization and comparison with other absence models. Epllepsia 1988, 29,361-377.Snead, 0. C. and Miles, M. V. Treatment of status epilepticus in children withrectal sodium valproate. Journal of Pediatrics 1985, 106, 323-325.Stacey, T. E., Weedon, A. P., Haworth, C., Ward, R. H. T. and Boyd, R. D. H.Fetomaternal transfer of glucose analogs by sheep placenta. American Journal ofPhysiology 1978,234, E32-E37.Szeto, H. H. Pharmacokinetics in the ovine maternal-fetal unit. Annual Reviews ofPharmacology and Toxicology 1982, 22, 221-243.Szeto, H. H., Kaiko, R. F., Clapp, J. F., Larrow, R. W., Mann, L. I. and Inturrisi, C.E. Urinary excretion of meperidine in the fetal lamb. Journal of Pharmacology andExperimental Therapeutics 1979, 209, 244-248.Szeto, H. H., Umans, J. G. and McFarland, J. A comparison of morphine andmethadone disposition in the maternal-fetal unit. American Journal of Obstetricsand Gynecology 1 982a, 143, 700-706.Szeto, H. H., Umans, J. G. and Rubinow, S. I. The contribution of transpiacentalclearances and fetal clearance to drug disposition in the ovine maternal-fetal unit.Drug Metabolism and Disposition 1982b, 10, 382-386.Taberner, P. V., Charington, C. B. and Unwin, J. W. Effects of GAD and GABA-Tinhibitors on GABA metabolism in vivo. Brain Research Bulletin 1980, 5, 621 -625.Thorburn, G. D. The placenta, PGE2 and parturition. Early Human Development1992, 29, 63-73.Thurston, J. H., Carroll, J. E., Hauhart, R. E. and Schiro, J. A. A single therapeuticdose of vaiproate affects liver carbohydrate, fat, adenylate, amino acid, coenzymeA, and carnitine metabolism in infant mice: Possible clinical significance. LifeSciences 1985, 36, 1643-1651.Thurston, J. H., Carroll, J. E., Norris, B. J., Hauhart, R. E. and Schiro, J. A. Acutein vivo and in vitro inhibition of palmitic acid and pyruvate oxidation by valproateand valproylcoenzyme A in livers of infant mice. Ann. Neurol. 1983, 14, 384-385.Tonn, G. R., Kwan, E., Hall, C., Rurak, D. W. and Axelson, J. E. Hepatic first-passmetabolism of diphenhydramine in fetal lambs following umbilical venousadministrations: A comparison to adult sheep. Journal of Pharmacology andExperimental Therapeutics 1 995a, Submitted, TBA.188Tonn, G. R., Kwan, E., Hall, C., Rurak, D. W. and Axelson, J. E. Pharmacokineticsand pharmacodynamics of diphenhydramine in late gestational fetal lambsfollowing simultaneous infusion of stable isotope labeled and unlabeleddiphenhydramine to fetal lambs and maternal sheep. Journal of Pharmacology andExperimental Therapeutics 1 995b, Submitted, TBA.Vajda, F. J. E., Donnan, G. A., Phillips, J. and Bladin, P. F. Human brain, plasma,and cerebrospinal fluid concentration of sodium valproate after 72 hours oftherapy. Neurology 1981, 31,486-487.Vajda, F. J. E., Mihaly, G. W., Miles, J. L., Donnan, G. A. and Bladin, P. F. Rectaladministration of sodium valproate in status epilepticus. Neurology 1978, 28, 897-899.Vayer, P., Cash, C. D. and Maitre, M. Is the anticonvulsant mechanism ofvalproate linked toits interaction with the cerebral y-hydroxybutyrate system?Trends in Pharmaceutical Sciences 1988, 9, 127-129.Veng-Pedersen, P. and Gillespie, W. R. A system approach to pharmacodynamicsI: Theoretical framework. Journal of Pharmaceutical Sciences 1988, 77, 39-47.Vorhees, C. V. Fetal hydantoin syndrome in rats: Dose-effect relationships ofprenatal phenytoin on postnatal development and behavior. Teratology 1987, 35,Vorhees, C. V., Minck, D. R. and Berry, H. K. Anticonvulsants and braindevelopment. Progress in Brain Research 1988, 73, 229-244.Waddell, W. J. and Marlowe, C. Transfer of drugs across the placenta.Pharmacology and Therapeutics 1981, 14, 375-390.Wang, L. H., Rudolph, A. M. and Benet, L. Z. Pharmacokinetic studies of thedisposition of acetaminophen in the sheep maternal-placental-fetal unit. Journal ofPharmacology and Experimental Therapeutics 1986, 238, 198-205.Whittle, S. R. and Turner, A. J. Effects of the anticonvulsant sodium valproate on‘-aminobutyrate and aldehyde metabolism in ox brain. Journal of Neurochemistry1978, 31, 1453-1 459.Whittle, S. R. and Turner, S. J. Effects of anticonvulsants on the formation of yhydroxybutyrate from y-aminobutyrate in rat brain. Journal of Neurochemistry1982, 38, 848-851.Wigglesworth, J. S. and Desai, R. Effects on lung growth of cervical cord section inthe rabbit fetus. Early Human Development 1979, 3, 51-65.Wilder, B. J. Treatment considerations in anticonvulsant monotherapy. Epilepsia1987, 28, S1-S7.189Woollard, P. M. Selective silylation using tert-butyldimethylsilyl reagents: Their usein the quantification of fatty acids. Biomedical Mass Spectrometry 1983, 10, 143.Wright, M. R. Ritodrine: Analysis, comparative maternal and fetalpharmacokinetics and pharmacodynamics in sheep. Ph.D. Thesis, University ofBritish Columbia, 1992.Wright, M. R., Rurak, D. W., Weyde, M. P. V. D., Taylor, S. M. and Axelson, J. E.Clearance and disposition of ritodrine in the fluid compartments of the fetal lambduring and after constant rate fetal intravenous infusion. Journal of Pharmacologyand Experimental Therapeutics 1991, 258, 897-902.Yamada, K. A., Hamoshi, P. and Gillis, R. A. Respiratory depression produced byactivation of GABA receptors in the hindbrain of the cat. Journal of AppliedPhysiology 1981, 51, 1278-1 286.Yamada, K. A., Norman, W. P., Hamosh, P. and Gillis, R. A. Medullary ventralsurface GABA receptors affect respiratory and cardiovascular function. BrainResearch 1982, 248, 71-78.Yeleswaram, K. Analysis, pharmacokinetics, metabolism and pharmacodynamicsof labetalol in pregnant and nonpregnant sheep. Ph.D. Thesis, University of BritishColumbia, 1992.Yeleswaram, K., Rurak, D. W., Hall, C., Wright, M. R. and Axelson, J. E.Pharmacokinetics and pharmacodynamics of labetalol in the pregnant sheep.Journal of Pharmacology and Experimental Therapeutics 1992, 262, 683-691.Yerby, M. S., Friel, P. N. and McCormick, K. Antiepileptic drug disposition duringpregnancy. Neurology 1992, 42, 12-16.Yoo, S. D., Axelson, J. E., Taylor, S. M. and Rurak, D. W. Placental transfer ofdiphenhydramine in chronically instrumented pregnant sheep. Journal ofPharmaceutical Sciences 1986, 75, 685-687.Yoo, S. D., Rurak, D. W., Taylor, S. M. and Axelson, J. E. Transplacental andnonplacental clearances of diphenhydramine in the chronically instrumentedpregnant sheep. Journal of Pharmaceutical Sciences 1993, 82, 145-149.Yoshida, H., Kamiya, A., Okumura, K. and Hori, R. Contribution of monoamineoxidase (MAO) to the binding of tertiary basic drugs in lungmitochondria. Pharmaceutical Research 1989, 6, 877-882.Yoshida, H., Okumura, K. and Hori, R. Subcellular distribution of basic drugsaccumulated in the isolated perfused lung. Pharmaceutical Research 1987, 4, 50-53.190Yoshida, H., Okumura, K. and Hori, R. Contribution of monoamine oxidase (MAO)to the binding of tertiary basic drugs in isolated rat lung. Pharmaceutical Research1990, 7,398-401.Yu, D., Gordon, J. D., Zheng, J., Panesar, S. K., Riggs, K. W., Rurak, D. W. andAbbott, F. S. Determination of valproic acid (VPA) and its metabolites using gaschromatography with mass selective detection: Application to serum and urinesamples from sheep. Journal of Chromatography: Biomedical Applications 1995,in press, TBA.Zaccara, G., Messori, A. and Moroni, F. Clinical pharmacokinetics of valproic acid -1988. Clinical Pharmacokinetics 1988, 15, 367-389.Zeise, M. L., Lasparow, S. and Zieglansberger, W. Valproate suppresses Nmethyl-D-aspartate-evoked, transient depolarizations in the rat neocortex in vitro.Brain Research 1991, 544, 345-348.Zheng, J. J. Metabolic and pharmacokinetic studies of valproic acid using stableisotope techniques. M.Sc. Thesis, University of British Columbia, 1993.Zimmer, R., Teelken, A. W., Gundurewa, M., Ruther, E. and Cramer, H. Effect ofsodium vaiproate on CSF GABA, cAMP, cGMP and homovanillic acid levels inmen. Brain Research Bulletin 1980, 5, 585-588.Zimmerman, H. J. and Ishak, K. G. Valproate-induced hepatic injury: Analyses of23 fatal cases. Hepatology 1982, 2, 591 -597."@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "1995-11"@en ; edm:isShownAt "10.14288/1.0088329"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Pharmaceutical Sciences"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Pharmacokinetics, metabolism, placental transfer and fetal effects of Valproic acid in pregnant sheep"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/7367"@en .