RETURNING MATERIALS: 'bV1SSI_] PTace in book drop to remove this checkout from w your record. FINES w111 be charged if Book is returned after the date stamped be10w. ABERRATIONS IN STEROID METABOLISM IN THE RAT INDUCED BY TAMOXIFEN, AN ANTITUMOR AGENT AS ASSESSED BY METABOLIC PROFILING By John Joseph Leary, III A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1986 TI qv FOL p.‘ -A M ABSTRACT ABERRATIONS IN STEROID METABOLISM IN THE RAT INDUCED BY TAMOXIFEN, AN ANTITUMOR AGENT AS ASSESSED BY METABOLIC PROFILING By John Joseph Leary, III Tamoxifen (ICI 46,476) is a widely used and effective drug for the endocrine management of advanced breast cancer in postmenopausal women. While Tamoxifen therapy effectively causes tumor regression, researchers have shown that the drug triggers changes in the level of circulating steroids and peptide hormones. Because modulation of the endocrine system by this nonsteroidal antiestrogen may explicate some of the drug's adverse effects, it was considered important to assess the impact of this drug on overall steroid hormone secretion and metabolism in the normal, cycling female rat and the tumor—bearing female rat experimental models. The Sprague—Dawley rat is used as an animal model for human breast carcinoma because 7,12-dimethy1benz(a)anthra- cene-induced mammary tumors in this rat are morphologically similar to those occurring in the human 1*)- (D (I) (D (1') (I) John Joseph Leary, III being. In this study, 24—hour urine samples were collected from individual rats. The five types of experimental models were: l)normal, cycling female rats, 2)female rats treated with Tamoxifen, 3)ovariectomized rats, 4)tumor—bearing female rats, 5)tumor—bearing female rats treated with Tamoxifen. Steroids were isolated from the urine and methyloxime—trimethylsilyl derivatives were prepared. Urinary steroid profiles were generated with gas chromatography (GC) and gas chromatography—mass spectrometry (GC-MS). Steroid components were identified using mass spectral information obtained from an LKB—2091 gas chromatograph—mass spectrometer—data system (GC—MS—DS) fitted with a 15—meter DB—l megabore capillary column (0.53 mm i.d.). All samples were analyzed by the GCMET GC—only metabolic profiling routines developed at the MSU/NIH Mass Spectrometry Facility. This program performs reverse library searches on gas chromatographic data. The development of the GCMET routines is described. Significant changes were noted in the excretion of several corticosteroid metabolites during the estrous cycle of the normal, cycling female rat prior to and following Tamoxifen treatment. Tamoxifen, at the dose levels used in these studies, has the effect of physiologically "locking" the estrous cycle in an anestrus kl » A C TL Art “A; r; WW7; rt lbwu John Joseph Leary,III phase. A diminution in the excretion of corticosteroid metabolites is observed in the retention index region of 2900 to 3300. Changes were noted in the excretion of selected corticosteroid metabolites from tumor—bearing rats treated with Tamoxifen as compared with tumor—bearing rats receiving only saline injections. These steroids included trihydroxy-pregnan—one and isomers of tetrahydroxy—pregnan—one, which are quantitatively important metabolites found in rat urine. The alterations may be caused by cellular changes in the adrenal cortex brought on by Tamoxifen treatment. ACKNOWLEDGEMENTS I would like to take a moment to express my thanks to a few of the many individuals who made this dissertation possible. I would like to sincerely thank my advisor, Professor J.T. Watson, for his guidance and friendship during my graduate career. Many thanks also to Professor C.W. Welsch who provided the animal models for these experiments, and to Jim Vrbanac for his instruction in the art of steroid profiling. I am also indebted to the Watson group members and the staff of the MSU/NIH Mass Spectrometry for all of their help and their valued friendship. Financial support is acknowledged from the National Institutes of Health, and a BASF Wyandotte Summer Fellowship. I would also like to thank Mom, Dad, Mary, and Mark for their continued support and encouragement. Finally, I would like to thank my fiancée Tonya, for her heroic assistance in the preparation of this document. Her patience and love made this finale bearable. TABLE OF CONTENTS LIST OF TABLES ....................................... x LIST OF FIGURES .................................... xii I. Introduction ...................................... 1 A. Objectives ................................... 1 1. Refinement of metabolic profiling technology .............................. 1 2. Application of metabolic profiling to aberrations in steroid metabolism ....... 4 II. Metabolic Profiling .............................. 7 A. General concepts and historical review ....... 7 1. Introduction ............................ 7 2. Instrumentation for metabolic profiling .............................. 9 3. Metabolic profiling with CO and GC—MS ..................... 10 4. Urinary steroid profiles: normal and diseased states ............ 14 5. Automated metabolic profiling by GC-MS .............................. 19 a. Library search methods ............ 19 b. MSSMET: Profiling by GC—MS—DS ..... 20 vi B. Improvements in metabolic profiling technology ................................. 21 1. Choice of Internal Standard ............ 22 2. Profiling by GC—MS—DS .................. 29 3. Installation of the Megabore Capillary Column in the LKB 2091 GC—MS—DS ......................... 29 b. Development of a Library of Rat Steroid Metabolites .............. 39 3. GCMET: Metabolic Profiling by GC—DS....48 a. Development of GC-only metabolic profiling ........................ 48 b. A functional description of the GC—DS metabolic profiling routines ............... 58 i. GCMAN: the GCMET library manager ................... 58 ii. GCMET: analysis of gas chromatographic data files ................ 64 iii. GCSTAT: statistical manipulation of GCMET output files .............. 79 III. Tamoxifen ...................................... 83 A. Tamoxifen—Human Studies ..................... 83 1. Pharmacology of Tamoxifen .............. 83 2. Metabolism of Tamoxifen in humans ...... 84 3. Use of Tamoxifen in human breast cancer ......................... 88 4. Side effects of Tamoxifen treatment in breast cancer patients ............. 92 5. Steroid metabolism in breast cancer patients .............................. 96 Vii 6. Endocrine effects of Tamoxifen therapy ..................... 98 a. Normal, premenopausal women ....... 98 b. Postmenopausal women with breast cancer .......................... 101 B. Tamoxifen—Studies with Rats ................ 105 1. DMBA-induced rat mammary carcinoma....105 2. Treatment of DMBA—induced rat mammary carcinoma with Tamoxifen ............. 106 3. Pharmacology of Tamoxifen in the rat..107 4. Tamoxifen metabolism in the rat ....... 109 5. Endocrine effects of Tamoxifen therapy .................... 113 IV. Steroid Metabolism and Excretion in the Rat....118 A. Introduction ............................... 118 B. Urinary and fecal excretion of steroids....118 C. Excretion of steroids in the bile .......... 133 D. Steroid levels in plasma ................... 136 E. Steroid levels in the ovary ................ 138 F. Steroid production in the adrenal cortex...143 G. Steroid metabolism in the liver ............ 147 H. Steroid metabolism in DMBA—induced rat mammary carcinoma ......................... 150 viii V. Experimental Procedures ......................... 152 A. Introduction ............................... 152 B. Animal Model ............................... 152 1. Normal, cycling adult female rats ..... 153 2. Tamoxifen—treated, adult female rats..153 3. Ovariectomized, adult female rats.... 154 4. Tamoxifen—treated and control, tumor—bearing adult female rats ..... 154 C. Preparation and Analysis of Samples ........ 155 1. Sample preparation (method I) ......... 155 2. Sep—pak extraction ................... 158 3. Introduction .................... 158 b. Sep—Pak extraction efficiency....160 c. Sample contamination ............ 164 3. Enzymatic hydrolysis vs. solvolysis...166 4. Derivatization ....................... 172 VI. Experimental Results ........................... 175 A. Trends in excretion of urinary steroids....175 1. Normal cycling female rats ............ 175 2. Tamoxifen-treated normal adult female rats and ovariectomized adult female rats ..................... 185 3. Tamoxifen-treated and control adult female rats bearing palpable DMBA—induced mammary carcinomas ....... 193 B. Discussion ................................. 217 APPENDIX 1 ......................................... 221 LIST OF REFERENCES ................................. 230 ix TABLE 10 11 LIST OF TABLES PAGE Aberrations in urinary steroid excretion for various enzyme deficiencies .............. 16 Retention indices of internal standard candidates ................................... 27 Retention indices of steroid standards on packed and megabore capillary columns ........ 32 Retention indices of steroid metabolites found in the three LH—20 fractions of postpuberal pooled female rat urine. Retention indices were obtained with an LKB 2091 GC-MS fitted with a 6—foot packed column packed with 3% OV—101 .......... 43 Determination of the number of hydroxyl groups on rat urinary steroids. Samples analyzed on the LKB 2091 GC-MS fitted with a 15—m DB—1 megabore capillary column ........ 50 Result of analysis file (with compound names) ....................................... 72 "Total" result of analysis file .............. 73 Result of analysis file ...................... 74 Level of estradiol, testosterone and androsterone in rat plasma at various times during the estrous cycle .................... 138 Mean peak areas of fludrocortisone for triplicate injections of each blank sample on the Varian 2100 GC fitted with a 15—m DB—l megabore capillary column .............. 163 Excretion of corticosteroid metabolites for Tamoxifen—treated rat #1 .................... 194 J Lo 6 IAII) 12 13 14 15 16 17 18 19 Excretion of corticosteroid metabolites for Tamoxifen treated rat #6 .................... 194 Comparison of the mean excretion of corticosteroid metabolites for day 4 - 8 and 16 — 20 from tumor-bearing rat #6 (control) ................................... 199 Comparison of the mean excretion of corticosteroid metabolites for days 4 — 8 and 16 — 20 from tumor-bearing rat #8 ........................ 200 Excretion of trihydroxy—pregnan—one by tumor—bearing rats #6, 7, 8 and 9 ........... 204 Excretion of tetrahydroxy-pregnan—one (TMS), R.I. = 3317, by tumor—bearing rats #6, 7, 8 and 9 ....................................... 205 Excretion of tetrahydroxy—pregnan—one (TMS), R.I. =3127, by tumor-bearing rats #6, 7, 8 and 9 ....................................... 205 Excretion of tetrahydroxy—pregnen—one (MO-TMS), R.I. = 3146, by tumor—bearing rats #6, 7, 8 and 9 ......................... 210 Excretion of tetrahydroxy—pregnan-one (MO-TMS), R.I. = 3072, by tumor-bearing rats #6, 7, 8 and 9 ......................... 215 xi r.» ._ nun PlJ Fr fiJJ FIGURE 1 LIST OF FIGURES PAGE Internal standard candidates: A. fluoro— metholone, B. 1,4-pregnadiene—6°—methyl- 11a,17°,21—triol—20—dione, C. 5°-pregnane— 3a,11a,20a,21—tetrol, D. fludrocortisone, E. dexamethasone, F. cholesteryl butyrate ...... 25 Reconstructed total ion chromatograms, cortol (250 ng) and hydrocarbon standards C30 and C32 analyzed on the LKB 2091 GC-MS. A. 6—ft glass column packed with 3% 0V-101. B. 15—m DB—l megabore capillary column ................. 34 A. Gas chromatogram, 60-m DB—l narrow—bore (0.25 mm i.d.) capillary column. B. Gas chromatogram, 15—m DB-1 megabore (0.53 mm i.d.) capillary column. Both samples: Hydrocarbon standards C20 to C30. Other peaks in profile A are from a coinjected blank sample ................................... 35 A. Reconstructed total ion chromatogram, 6—ft glass column packed with OV—101. B. Reconstructed total ion chromatogram, 15—m DB-l megabore capillary column. Both samples: Hydrocarbon standards C20 to C30 analyzed on the LKB 2091 GC—MS. Scan speed: 4 sec/decade ................................... 37 A. Reconstructed total ion chromatogram, 15—m DB-1 megabore capillary. B. Gas chromatogram, 60—m DB—1 narrow—bore capillary column. Both samples: 24—hour urine from a tumor-bearing female rat, third day of Tamoxifen treatment. Stars indicate coinjected hydrocarbon standards, C20 to C36 .......................... 38 Reconstructed total ion chromatograms of three fractions of female rat urine. Samples were analyzed using the LKB 2091 GC-MS fitted with a 6—ft glass column packed with 3% OV-101 with a scan rate of 4 sec/decade ............... 41 xii 10 11 12 13 14 Abbreviated mass spectrum of a tetrahydroxy— pregnan-one (t0p) and its representation as a MSSMET library entry. The library entry contains the compound number (1830), the compound name, retention index (elutes at 74% of the time interval between straight—chain hydrocarbons C30 and C32), designate ion (DIN=562), confirming ions (CIN=562,472,652), and their relative intensities (999,510,250)...44 Reconstructed mass chromatograms of the designate ion (m/z 562), confirming ions (m/z 472 and 652), and the total ion current obtained from the GC/MS/DS analysis of a 24—hr urine sample obtained from tumor—bearing rat #9 on the third day of Tamoxifen treatment ..... 46 Shift in the retention index and molecular weight when trihydroxy—pregnan-one is derivatized with d9—BSA (profile B) rather than dO—BSA (profile A). The molecular ion for the compound appears at m/z 593 rather than m/z 566, and its retention index decreases from 3339 to 3314 .................... 49 GC-MS—DS analysis of a 24—hr rat urine sample collection. The lower profile is of the total ion current recorded during the analysis. The top profile is a reconstructed mass chromatogram of m/z 99, an ion characteristic of the coinjected hydrocarbon standards ........ 52 Two gas chromatograms analyses. The sample analyzed in the top profile is a 24—hr urine sample obtained from a normal female rat during the estrus phase of its estrous cycle, plus a series of coinjected hydrocarbon standards C20 through C36. The sample analyzed in the lower profile is a blank plus the coinjected hydrocarbon standards ...................................... 54 Organization of GCMET GC—only metabolic profiling routines ............................. 57 Flow diagram of the sample preparation procedure for the analysis of urinary steroids ...................................... 159 Two gas chromatographic profiles of aliquots of derivatized pooled female rat urine. The steroid conjugates in the sample displayed in the top profile were hydrolyzed with the enzyme preparation xiii C1]L.\t 15 16 17 18 19 20 21 22 23 24 obtained from Helix Pomatia. Enzymatic hydrolysis for the sample shown in the lower profile was achieved with beta-glucuronidase obtained from the Marine Mollusk. Coinjected hydrocarbon standards C22 through C34 are indicated with the even numbers. The internal standard fludrocortisone is indicated with an "F", and Sep-Pak contaminants are noted with "C"'s ......................................... 168 Relative concentrations of components eluting between hydrocarbons C30 and C32 during gas chromatographic analyses of 2 aliquots of derivatized pooled rat urine. Similar profiles were obtaine with both Helix Pomatia (*) and Marine Mollusk (0) enzyme preparations ........ 169 Gas chromatographic analyses of two aliquots of derivatized pooled female rat urine. The sample in the top profile underwent enzymatic hydrolysis and solvolysis. The sample in the bottom profile only underwent enzymatic hydrolysis. Coinjected hydrocarbon standards are indicated with even numbers ............... 171 Urine collection schedule for normal, cycling female rats ................................... 177 Gas chromatographic profiles obtained for three phases of the estrous cycle. A. proestrus, B. estrus, C. diestrus ............. 179 Excretion of trihydroxy-pregnan-one (R.I.=3015) by control rats #1(O), 3(X) and 4(*) .......................................... 181 Excretion of tetrahydroxy—pregnan—one (MO-TMS) (R.I.=3072) by control rats #1(O), 3(X) and 4(*) .......................................... 182 Excretion of tetrahydroxy—pregnen—one (R.I.=3146) by control rats #1(O), 3(X) and 4(*).... ...................................... 184 Excretion of tetrahydroxy-pregnan-one (TMS) (R.I.=3117) by control rats #3(X) and 4(*)....186 Excretion of tetrahydroxy-pregnan—one (TMS) (R.I.=3127) by control rats #3(X) and 4(*)....187 Urine collection schedule for Tamoxifen- treated rats .................................. 188 xiv 25 26 27 28 29 3O 31 32 33 34 35 36 37 38 Urine collection schedule for ovariectomized rats .......................................... 190 Gas chromatographic profiles obtained for a Tamoxifen—treated rat (top profile) and an ovariectomized rat (bottom profile) ........ 191 Gas chromatographic profiles obtained for a Tamoxifen-treated rat (top profile) and a normal,cycling female rat in the diestrus phase ......................................... 192 Urine collection schedule for tumor—bearing control and drug-treated rats ................. 195 Gas chromatographic profiles obtained from a tumor—bearing rat prior to and during Tamoxifen treatment ................ 197 Excretion of trihydroxy—pregnan—one (R.I.=3015) by tumor—bearing rats #6(X) (control) and 7(0) ............................ 202 Excretion of trihydroxy—pregnan—one (R.I.=3015) by tumor—bearing rats #7(0) and 8(X) ...................................... 203 Excretion of tetraydroxy-pregnan-one (TMS)(R.I.=3117) by tumor—bearing rats #6(*) (control) and 8(X) ...................... 206 Excretion of tetrahydroxy-pregnan-one (TMS)(R.I.=3117) by tumor—bearing rats #7(0) and 8(X) ................................ 207 Excretion of tetrahydroxy—pregnan-one (TMS)(R.I.=3127) by tumor—bearing rats #6(X) (control) and 7(0) ...................... 208 Excretion of tetrahydroxy-pregnan—one (TMS)(R.I.=3127) by tumor—bearing rats #7(0) and 8(X) ................................ 209 Excretion of tetrahydroxy-pregnen—one (R.I.=3146) by tumor—bearing rats #6(X) (control) and 7(0) ............................ 211 Excretion of tetrahydroxy—pregnen—one (R.I.=3146) by tumor-bearing rats #7(0) and 8(X) ................................ 212 Excretion of tetrahydroxy—pregnan-one (MO-TMS) (R.I.=3072) by tumor— bearing rats #6(X) (control) and 7(0) ......... 213 XV 39 4O Excretion of tetrahydroxy—pregnan—one (MO—TMS) (R.I.=3072) by tumor— bearing rats #7(0) and 8(X) ................... 214 Excretion of cholic acid (R.I.=3344) by tumor—bearing rat #9 ....................... 216 xvi I. Introduction A. Objectives The objectives of this research project were two—fold. First, to refine the analytical methodologies involved in the production of urinary steroid metabolic profiles. This included alterations in the sample preparation procedures as well as in the methods of analysis. The analytical techniques used in this project were gas chromatography — mass spectrometry (GC—MS) and gas chromatography with flame ionization detection (GC—FID). The second objective of this project was to apply the metabolic profiling methodology to the characterization of aberrations in steroid metabolism in the female, Sprague— Dawley rat induced by an antitumor agent, Tamoxifen. 1. Refinement of metabolic profiling technology A profile, according to Webster's Dictionary is a group of data representing quantitatively the extent to which an individual exerts traits as determined by tests and usually presented in the form of a graph. The metabolic profile would then be a two-dimensional representation of some aspects of one's physiological state. Most of the research in metabolic profiling over the past fifteen years has centered on using the gas chromatograph - mass spectrometer with an attached mini part SCPICI St _‘rr [1) minicomputer data system (GC-MS—DS) to obtain profiles of particular analytes extracted from physiological fluids. The use of the GC—MS—DS in metabolic profiling has been successful, particularly in the analysis of organic acids and steroids (1). Although packed column—GC—MS-DS could not chromatographically resolve certain overlapping peaks, such as 11a—hydroxyandrosterone and 17°— hydroxypregenanolone, examination of reconstructed mass chromatograms would confirm the presence or absence of these peaks. Bonded-phase, fused silica GC columns now provide the increased resolving power necessary to separate the majority of the components in a urinary steroid sample. These narrower peaks require a higher scan rate for the conventional magnetic sector mass spectrometer (0.2 — 0.5 seconds per scan) in order to acquire enough data points to characterize the chromatographic profile. However, as the scan rate increases, there is less time available to assess each peak in the mass spectrum. Fewer ions are detected and the accuracy of the ion current measurement decreases. One method to improve ion statistics, which is under investigation by the researchers at the MSU/NIH Mass Spectrometry Facility is the attachment of an integrating transient recorder (ITR) onto a GC—time—of—flight (TOF) mass spectrometer (2). The TOF instrument produces a complete mass spectrum from every pulse of ions leaving mass coan the resul 3.- (I'D mass spectrometer (2). The TOF instrument produces a complete mass spectrum from every pulse of ions leaving the ion source. The ITR sums the successive pulses, resulting in an increase in signal to noise so that the complete summed mass spectrum can be acquired more frequently than is possible by conventional data aquisition systems. This increase in the number of data points in the reconstructed mass chromatograms should improve the degree to which the chromatographic resolution can be represented. An alternative approach, and a major focus of this research project, would be to assume that analysis of urinary steroid profiles by GC/MS is not necessary for every sample (3). Fused silica capillary columns provide reproducible urinary steroid patterns. A protocol can be developed where all samples are analyzed by GC—FID; samples statistically aberrant or with identification ambiguities can be rerun using GC—MS-DS. This project involved the development of GC—only profiling for the characterization of rat urinary steroids. This process requires a dedicated computer system to acquire and store the gas chromatographic data. An integrated software system (GCMET) utilizing a reverse library search has been written at the MSU/NIH Mass Spectrometry Facility (1), and is complementary to Sweeley's mass spectral reverse library search (MSSMET) (15). Characteristic samples or 5am; GC-H NS—D samples with identification ambiguities are analyzed using GC-MS—DS. Both of the gas chromatographs (GC—FID and GC- MS-DS) are fitted with capillary columns with the same type of stationary phase so that each species to be analyzed will have the same retention behavior on both instruments. Straight—chain saturated hydrocarbons are coinjected with all samples analyzed by GC—FID and GC—MS— DS so that retention indices may be assigned to each chromatographic peak. The optimization of the chromatographic system designed for the analysis of urinary steroids will be described in chapter II. The development of a library of rat urinary steroid metabolites for GC-FID-DS and GC-MS—DS will also be described below in chapter II. 2. Application of metabolic profiling to aberrations in steroids metabolism. The development of a library of rat urinary steroid metabolites facilitated the analysis of GC profiles obtained from female Sprague-Dawley rats. Experimentation with these animals fulfilled the second objective of this project, to determine whether the antitumor agent Tamoxifen has an effect on steroid metabolism when administered to normal or tumor—bearing female rats. The drug is a nonsteroidal antiestrogen which acts not only at the breast tumor site, but also acts systemically to bind :etabc examin Broil] r; -4 L. 5%. ~“‘ FLA "o‘ t ‘ML - 7. a. 6“;er 9-1 ‘5 .,e p Eir‘ L u‘ E} to other tissues containing estrogen receptors. This systemic action as well as evidence that high doses of the drug causes adrenocortical cell necrosis in rats led to the query whether Tamoxifen—induced changes in the steroid metabolism might be responsible for side effects. To examine this question, the urinary steroid metabolic profiles of the following five groups of rats were compared: 1) normal, cycling female rats, 2) normal female rats treated with Tamoxifen, 3) ovariectomized rats, 4) female rats with chemically-induced breast carcinomas, and 5) tumor-bearing rats treated with Tamoxifen. This dissertation consists of seven chapters, the first of which contains these introductory remarks. Chapter II defines the concept of metabolic profiling, describes the analytical instrumentation employed, the applications of this technique, and the experiments conducted for this research project in order to refine the metabolic profiling technology. The third chapter contains a review of the literature describing the actions of the antitumor agent, Tamoxifen in man and in the rat. Chapter IV is a review of the literature describing the metabolism of steroids in various tissues in the rat and the excreted metabolites appearing in the urine, feces, and bile. Chapter V contains a description of the experimental procedures used to isolate and analyze urinary steroids and the improvements and modifications developed through this research project. Chapter VI explicates a particular application of metabolic profiling, i.e., the analysis of steroid metabolites from the urine of disease-free and tumor-bearing rats treated with Tamoxifen. Conclusions are formed and future directions are suggested in Chapter VII. II. Metabolic Profiling A. General concepts and historical review 1. Introduction The origins of metabolic profiling may be traced back to the work of Roger Williams in the 1940's (4). Using paper chromatography, he measured the concentrations of various urinary metabolites, such as amino acids, and presented his results graphically as a metabolic pattern. These vector diagrams utilized vector angles to represent various analytes and vector distances to represent the analyte concentrations. These vector diagrams included such analytes as amino acids, creatinine, sucrose, potassium chloride, sodium chloride, pH, and taste thresholds. Williams found that these metabolic patterns vary from individual to individual but remain quite constant from day to day for a single person. These metabolic patterns served as a type of fingerprint to help characterize an individual's physiological state. Horning and Horning (5,6) were the first investigators to use the term "metabolic profile" which they coined to describe "multicomponent gas chromatographic analyses for a group of metabolically or analytically related metabolites". They also used gas chromatography-mass spectrometry to identify the components of complex biological mixtures, and they 7 predicted the technique would be useful in distinguishing between normal and pathological states. The authors developed the necessary methodologies to analyze several different classes of metabolites including urinary steroids, urinary and blood sugars and sugar alcohols, urinary and blood acids, compounds of the Krebs cycle and related substances, and drug metabolites in blood and urine. The necessary derivatives were prepared prior to vapor phase analysis. The metabolic profile provides a two—dimensional cross section of a complex physiological state delineated by the sample source, chemical processing, and method of analysis (1). One must take note of the difference between metabolic profiling analyses and clinical screening analyses. Screening procedures are multiple analyte analyses, where a number of compounds are identified in a complex sample (e.g., urine or serum) and quantified These values are compared with a normal reference range of concentrations for each analyte. Each species is considered to be a single variable and may be used to help evaluate an individual's state of health. The analytes may be measured individually or simultaneously with an automated analyzer. In the metabolic profile, the relative relationships between the analytes are important, as well as their absolute values. The metabolic profile also provides an analysis of compounds which are chemically or analytically related. Since an initial sample separation procedure is generally used, this imposes limitations on the chemical types of analytes which can be determined in a particular profiling analysis. 2. Instrumentation for metabolic profiling Numerous analytical techniques have been utilized for metabolic profiling analyses, including planar chromatographic techniques, such as paper chromatography and thin-layer chromatography. For example, Armstrong et al (7). analyzed the phenolic acids in human urine using two dimensional paper chromatograms developed with reagents such as diazotized sulphanilic acids. Thin—layer chromatography offers less spot and band diffusion during chromatographic development, and densitometry can be applied to provide quantitative measurement of analytes. Shackleton and Mitchell (8) used silica gel plates for the thin—layer chromatographic measurement of 3a—hydroxy— delta—S—steroids. After color development with antimony trichloride, quantitative measurements were made with a reflective scanner. The identity of the steroids were confirmed with GC—MS. Column liquid chromatography has been used to gennerate metabolic profiles, but a classical anion— exchange chromatographic run proved to be quite tedious. Jolley and Freedmen (9) employed this system equipped with 10 colorimetric detection to measure the levels of carbohydrates in normal and diabetic urine and serum. Separation of monosaccharides and small oligosaccharides on a 1.5 m x 0.62 cm Dowex 1 column required 24 hours. On the other hand, high performance liquid chromatography (HPLC) has great potential as a metabolic profiling tool. HPLC separations can clean—up samples, and isolate and characterize analytes in the same analytical procedure, eliminating the extensive sample extraction and derivatization procedures necessary prior to GO analysis. The applicability of HPLC to metabolic profiling was demonstrated by Al Qureshi et al (10). who analyzed the amino acids in plasma and urine samples in less than one hour on reversed phase C18 columns. Although the column efficiency and speed of analysis is better in HPLC than classical column liquid chromatography, the technique is restricted by the sensitivity and dynamic range of its detection system. Exotic detection systems such as Fourier-transform infrared spectroscopy and mass—spectrometry are being combined with HPLC, which will undoubtedly improve its applicability to metabolic profiling. 3. Metabolic profiling with GC and GC-MS Gas chromatography (GC) has been the technique of choice for metabolic profiling since the experiments of Horning and Horning (5,6). The large dynamic range of the sample statior rate ( sample Versus Provid( 5872?} E“ additg, Chrgfia? 11 flame ionization detector (FID) is well suited to handle the wide variations in the concentrations of metabolites in physiological fluids which can encompass several orders of magnitude. An added benefit is that the FID response is relatively constant for similar chemical species in a sample. Experimental variables, such as choice of stationary phase, temperature programming rate, and flow rate can be optimized to produce the best separation of sample components. The two—dimensional output (intensity versus time) obtained from the strip chart recorder provides a convenient visual representation of all the sample components and their relative abundances. The addition of a mass spectrometer as a detector for the gas chromatograph occurred in the late 1950's and early 1960's. The mass spectrometer added a new dimension to the information obtainable from a GC. Instead of just chromatographic peak area versus time data, one obtains ion mass, ion intensity, and chromatographic time data. This abundance of information led to the logical addition of a minicomputer to form a gas—chromatograph-mass spectrometer-data system (GC—MS—DS). This instrumentation became the optimum analytical system for the metabolic profiler, and the profiling activity of the past fifteen years has centered on the GC—MS—DS. Usually, magnetic sector mass spectrometers are used to aquire mass spectral data for metabolic profiling. These instruments provide 12 accurate ion current measurements, a wide dynamic range for ion current measurement, and reproducible ion fragmentation patterns. Quadrupole mass spectrometers are also utilized, and their principle advantage is the ability to scan the mass axis at a faster rate, making it better suited to capture spectra from narrow peaks eluting from a gas chromatograph fitted with a high resolution capillary column. Hites and Biemann (11) repetitively scanned a range of masses with a mass spectrometer during a gas chromatographic run and stored the complete mass spectra on a computer. They then plotted ion intensity versus scan number for a particular m/z value to obtain a mass chromatogram. Plotting the total ion current (a summation of all intensities at all m/z values) versus scan number produces a total ion chromatogram which is analogous to the tracing resulting from the plot of a GC-FID signal versus time. Jellum et al. (12) used a GC—MS—DS system to perform multicomponent analyses of biological materials. Specifically, they sought to identify aberrations in the urinary excretion of organic acids which would be indicative of inborn errors of metabolism. The authors analyzed samples from 700 patients which led to the discovery of three inborn errors: methylmalonic aciduria, a-methylcrotonyl-COA carboxylase deficiency, and prrogl identi five decree: with 13 pyroglutamic aciduria. They utilized two methods of identifying mass spectra. The first method listed the five strongest peaks in the spectrum in order of decreasing intensity. The computer compares this list with its collection of authentic standard spectra. A "match percentage" is produced. For example, if all five peaks in the unknown spectrum and reference spectrum have equal m/z values and the same relative intensity 3 match of 100% is output. Less probable matches are output, such as 40% for two equal m/z values in the same relative intensities. The second method compares the five most intense peaks, but the emphasis is placed on their m/z values, irrespective of each ion's intensity. This method takes into account variation between different types of mass spectrometers, and variability in operational characteristics such as inlet and source temperatures, ionization efficiency, and whether the mass spectrum is recorded on the ascending or descending edge of the GC peak, all factors which may affect the intensity of the peaks in the mass spectrum. Jellum's technique involved semi—automated GC—MS—DS. Each unknown mass spectrum was visually inspected and the five most abundant peaks were arranged in order of decreasing intensity and this information was submitted to the computer via punched cards. The computer search required at least three minutes per unknown mass spectrum. 14 4. Urinary steroid profiles: normal and diseased states Steroid profiling by GC and GC-MS has been performed on numerous complex biological matrices including urine, plasma, saliva, cerebral spinal fluid, amniotic fluid, and various tissues (16). The general sample preparation scheme for these compounds includes extraction and isolation of steroids from the physiological matrix, separation of free steroids from steroid conjugates and derivatization. Sample purification is the rate—limiting step in the production of steroid profiles. Important differences in the steroid profiles of adults, newborns, children, and pregnant women have been observed (16,17). Evidence of clinical disorders such as Cushing's disease, Addison's disease, hyperaldosteronism, and deficiencies in steroid enzymatic pathways can be detected in steroid profiles by noting the elevation or diminution of particular steroid metabolites (16). Mental stress can cause a change in the pattern of excreted urinary steroid metabolites. Ludwig et a1. (18) obtained urine samples from doctoral candidates immediately after their defense, and then days later. Analysis of these samples showed a 10 - 20 fold increase of dehydroepiandrosterone (DHEA) excretion during that stressful time when compared with the normal, resting levels origir testos .4 abreflo .nyst TBS ( 15 state. The authors noted that the DHEA level rose starting the morning of the examination, reached the 20— fold normal level by noon, and then dropped to normal levels within a week. DHEA is a 17—ketosteroid of adrenal origin with about one—tenth the androgenic potency of testosterone. Phillipou et al. (19) summarized the aberrations in urinary steroid excretion for several inborn errors of corticoid biosynthesis, as listed in Table 1. Their metabolic profiling procedure was effective in discerning cases of adrenal hyperfunction resulting in high urinary levels of THE, THF, and allo—THF. These pathological states were caused by either adrenal hyperplasia or ectopic ACTH production. Cushing's syndrome, a generalized term for overproduction of adrenocorticosteroids, leads to excessive excretion of most adrenal steroid metabolites in the urine including THS (Leunissen and Thijssen, 1978) (20), and THF and THE (Luyten and Rutten, 1974) (21). Ta Enzyme 16 Table 1. Aberrations in urinary steroid excretion for several enzyme deficiencies. Enzyme Deficiency Urinary Steroids Elevated Suppressed 21-hydroxylase pregnanetriol THE, THF pregnanetriolone allo—THF 11a-hydroxylase THS THE, THF TH—DOC allo-THF 17a—hydroxylase pregnanediol androsterone pregnenediol etiocholanolone TH—DOC Patients with advanced cancer display abnormalities in cortisol metabolism. Werk et al. (22) noted that these patients excreted a larger ratio of 6—hydroxycortisol (6- OHF) to l7—hydroxycorticosteroids (17—OHCS). The increase in the excretion of 6—OHF was compensated for by a decrease in cortol and cortolone metabolites. The authors did not propose a mechanism for this effect, but used it as an index of abnormality. Pfaffenberger and Horning (23) found greatly elevated concentrations of 3b-hydroxy—17-ketosteroid metabolites in the urinary GC profile of a postmenopausal female with a DHEA—secreting adrenal tumor. Minowada et al. (24) suggested that the presence of excessive amounts of St— and 11b—hydroxysteroid metabolites in the urine, such as 17 3a,17a,21—trihydroxy—5b-pregnan—20—one (THS) and pregn—5- ene—3b,1la,203-triol, may be indicative of the presence of an adrenal carcinoma. This unique pattern of steroid metabolites reflects a decrease in the activity of the liver enzymes Sa—reductase and llb-hydroxy steroid dehydrogenase. The excessive secretion of these steroids are not observed in patients with adrenal adenomas, so the urinary metabolic profile may serve as a useful screening method for adrenal tumors. Pfaffenberger et al. (25) have found that a certain urinary steroid profile for premenopausal women could indicate an increased risk for the development of breast lesions. They noted that three—fourths of the premenopausal women with benign breast lesions in their study had unusual ratios of etiocholanolone to androsterone (Et/An) and tetrahydrocortisol to allotetrahydrocortisol (THF/a-THF) which were more characteristic of males rather than females. These patients had an average Et/An of 0.66 and THF/a-THF of 1.58, whereas only one third of the healthy control group displayed this pattern. Two thirds of the control group had the more normal pattern where Et/An averages 1.39 and THF/a-THF averages 2.98. These aberrations in steroid metabolism are due to changes in the activity of the liver enzymes reducing the 3-oxo—4—ene moiety of the steroidal A/B ring system, such as 5b—H and Sa-H oxidoreductases, which androst~ tetrahyé androger authors breast mastect' androsto .Jx‘ tenydrup 3 NOTE A mmrg mydI‘OX‘;( h‘COI’trjl 18 which convert testosterone to etiocholanolone and androsterone, and cortisol to tetrahydrocortisol and allo— tetrahydrocortisol. Another prognostic tool, suggested by Thomas et al. (26) is the measurement and comparison of urinary androgens with urinary corticosteroid metabolites. The authors compared the urinary steroid profiles of early breast cancer patients 48 hours prior to and 10 days post— mastectomy. Patients who excreted amounts of androsterone, (A), etiocholanolone (E) and dehydroepiandrosterone (DHEA) below the median values had a more rapid recurrence rate of cancer after mastectomy. A more significant test compares any of the three 17— hydroxycorticosteroid metabolites measured: a-cortolone, b—cortolone, and b—cortol, with any of four androgen metabolites: A, E, DHEA, and 11 —hydroxyandrosterone (11- OHA). When ratios were set up for the urinary concentrations of either A, E, DHEA, and 11—0HA, with the urinary concentrations of the 17-hydroxycorticosteroids, patients with values less than the median were more likely to have a earlier recurrence of the malignancy than patients with a ratio greater than the median. Comparing either 11-ketoetiocholanolone or 11—hydroxyetiocholanolone with the 17-hydroxycorticosteroids did not prove to be a statistically significant test. The predictive success of these androgen/corticosteroid ratios was greatest with 19 postmenopausal patients, and those with stage 2 disease (5 3 nodes). The authors mention that these prognostic indices should always be considered in the light of other crucial information, such as pathologic state and histiologic grade of the tumor. 5. Automated metabolic profiling with GC-MS—DS a. Library search methods Interpretation of the mass spectral information acquired during a GC run and stored on a minicomputer has historically been approached from two directions: the "forward library search" and the "reverse library search". The former technique involves either a manual or computerized comparison of each acquired unknown spectrum against each member of a library of reference spectra to find the best match. This data base search method tends to be time consuming, although the search may be expedited by pre-ordering the library file to list the most frequently occurring structural moieties or compound classes first. Forward library search algorithms are ill—suited for the analysis of overlapping or co-eluting GC peaks. When the unknown spectrum contains ions formed during the fragmentation of two different species, comparison with the library of reference spectra yields matches with poor confidence values. Sophisticated forward search routines have Overl store 0 I-n (I) 20 have been constructed which take into account peak overlapping and minor chromatographic peaks (13). The alternative method for automated interpretation of stored mass spectra utilizes a "reverse library search" suggested by Abramson in 1975 (14). In this procedure, all the unknown spectra are searched to obtain a match for a given reference library spectrum, rather than searching all the reference library spectra to find a match for each unknown spectrum. Abramson's "reverse search" involved searching an entire GC—MS run for matches to each library spectrum. b. MSSMET: profiling by GC—MS—DS The mass spectral data processing algorithm developed by Sweeley and coworkers (15), named MSSMET is a more selective technique utilizing hydrocarbon retention indices to delineate the time dimension. This off—line reverse library searching program allows for the qualitative and quantitative assessment of a given urine extract analyzed by GC—MS. To determine the presence or absence of a particular analyte (e.g., steroid) which is listed in the MSSMET library, the computer examines mass Spectral data acquired during a time "window" at the expected retention index for the compound. The presence of the compound is confirmed by detecting a peak in the mass chromatogram (profile of ion current at a specified 21 m/z value, the profile having the appearance of a gas chromatogram) corresponding to the designate ion (a characteristic peak in the mass spectrum of the compound). If this designate ion is found, the computer searches the mass spectrum for a series of other characteristic peaks ("confirming ions") to obtain a positive identification of the peak in the mass chromatogram. The program then integrates the areas of each gas chromatographic peak identified and calculates a correlation coefficient based upon the differences between the observed ratios of the designate and confirming ion intensities, and those listed for each MSSMET library entry. The calculated correlation coefficient must exceed a minimum value arbitrarily set by the operator for the compound to be considered "found". A "found-file" is created containing the retention time, retention index, integrated peak area, abundance of the designate ion, relative analyte concentration, and other identifying characteristics for each compound "found" to be in that particular sample analyzed by GC—MS—DS. A more thorough discussion concerning the development and use of a MSSMET library may be found in section 2b below. B. Improvements in Metabolic ProfilingfiTechnolpgy One of the major goals of this research project was to improve upon the existing procedures for the metabolic profiling analysis of urinary steroids. Although many of 22 these techniques have been evolving over the past twenty years, optimization of certain experimental portions was necessary, such as choice of quantitative internal standard, implementation of megabore capillary technology in the LKB 2091 GC-MS-DS, and development of GC—only metabolic profiling using GC-DS. The following three sections contain descriptions of experiments designed to optimize the chemical work-up and instrumental analysis of the steroids found in rat urine. Section 3 below describes the development of metabolic profiling using a gas chromatograph - data system. The second portion of that section contains a functional description of the GC metabolic profiling routines. This section can be used as reference documentation for the user of GCMET software. Other improvements in sample preparation and analysis are described below in the Experimental Procedures chapter. 1. Choice of Internal Standard The most widely used internal standard in the field of urinary steroid metabolic profiling is cholesteryl butyrate. /CM‘3 3\\ CH 5 CH-CH(CH ) 3 2 23 This steroid was chosen as an internal standard since its retention time was long in a steroid GC run (RI = 3417) and would not interfere with the identification of steroid metabolites through peak overlap. The compound is exogenous and would not interfere with the quantitation of steroid metabolites. There are several disadvantages to the use of cholesteryl butyrate as an internal standard for the analysis of urinary steroids. The major disadvantage with using cholesteryl butyrate is that the compound does not separate with the steroids during Sep—Pak extraction. The compound is retained on the octadecylsilyl cartridge as the steroids are eluted with methanol. It also is retained on Amberlyte XAD—2 columns. Therefore, this internal standard has to be added to the sample after extraction, conjugate hydrolysis, and group separation steps, with a subsequent loss of quantitative information. Another disadvantage is that cholesteryl butyrate does not have a chemical structure similar to the steroids normally characterized in urinary steroid profiles. The compound does not have alcohol or keto functions so it does not form a methyloxime—trimethylsilyl derivative. Many researchers have used cholesteryl butyrate nonetheless, including Horning and Horning (5,6), Maume et al. (123), Shackleton and Whitney (128), Setchell and Shackleton (129), and Vrbanac et al. (130). Pike et al. (131) added 24 cholesteryl butyrate prior to sample derivatization, but also added 5b-dihydro-epitestosterone directly to the urine sample as an additional internal standard. 5b- cholestan-Bb—ol had been used as a quantitative standard by Axelson (132). This compound is also added after extraction and prior to derivatization, when the TMS ether is formed. Axelson added the internal standard to a steroid mixture of known concentration, and also to another urine sample. After sample work-up and GC analysis, relative concentrations were calculated. An experiment was organized to evaluate a series of possible internal standard candidates. Rather than using members of the androstane or cholestane classes of steroids, various corticosteroids were chosen since the analytes of interest from the female rats were urinary metabolites of adrenocorticosteroids. The optimal candidate would not be endogenous in the urine of rats and would produce only one methyloxime—trimethylsilyl derivatized species. The standard should elute near the corticosteroid region of the chromatogram, but would not overlap with rat steroid peaks. The internal standard candidates, which are pictured in figure 1., included fluorometholone, fludrocortisone, and dexamethasone, which are synthetic fluorinated corticosteroids; and also 1,4—pregnadien-6a-methyl- Figure 1. 25 CH CH-CH(CH )” 5 3 2 .3\\ CH :5 Internal standard candidates: A. fluorometholone, B. 1,4—pregnadiene—6a—methy1—118,17a,21-triol- 3,20-dione, C. 5 —pregnane-38,118,208,21—tetrol, D. fludrocortisone, E. dexamethasone, F. cholesteryl butyrate. 26 11b,17a,21-triol—3,20—dione, and 5a-pregnane— 3b,11b,20b,21—tetrol. Each standard (10 mg) was weighed into silanized glass screw—top vials fitted with teflon cap liners, and dissolved in 10 ml of ethanol. A 20 ul aliquot was then removed and added to 20 ml of distilled water. This 20 ml volume was then treated as a normal urine sample and underwent Sep-pak extraction, enzymatic hydrolysis, and derivatization as described below. The derivatized standards were analyzed on a 60—meter DB—l capillary column (0.252 mm i.d., 0.1 um film thickness) in a HP5840 GC with a flame ionization detector. The column temperature program was from 180 to 300 at 1.25°/min. The retention indices of the internal standard candidates were calculated based on coinjected hydrocarbons (C20 through C36) and are listed in Table 2. 27 Table 2. Internal Standard Candidates Name Retention Index Fluorometholone 3147 and 3171 (1,4—pregnadien-9a—fluoro- 7b—methyl-11b,17a—diol 3,20-dione 1,4—Pregnadiene—6a-methyl— 3282 and 3307 11b,17a,21-triol—3,20-dione Sa-Pregnane—Bb,11b,20b,21-tetrol 3315 Fludrocortisone 3348 (11b,17a,21-trihydroxy- pregn—4—en—9a-fluoro— 3,20-dione Cholesteryl Butyrate 3417 Fluorometholone and 1,4—pregnadien-6a-methyl— 11b,17a,21—triol—3,20—dione each produced two peaks in the chromatogram, probably due to the formation of geometric isomers of the syn/anti type during methyloxime formation. Horning et al. (141) noted the possible formation of isomers when 3—keto-delta—4, 3—keto-5a-H, and 16—keto structures were derivatized. They noted that these isomers are produced in constant ratios (such as 55:45) but isomer formation is detrimental from a quantitative point of view. Thus, fluorometholone and 1,4—pregnadiene— 6a-methyl—11b,17a,21—triol—3,20—dione were rejected as 28 candidates. Dexamethasone was also rejected due to excessive chromatographic tailing. Fludrocortisone was chosen as the internal standard, rather than 5a—pregnane-3b,11b,20b,21—tetrol because the latter compound eluted in a rather crowded area of the chromatogram, and would be more difficult for the GCMET program to identify. Fludrocortisone elutes just after the corticosteroid region. Since fludrocortisone has a structure similar to those of the corticosteroids of interest, it can be added directly to the rat urine sample and can be extracted using the C18 Sep-pak cartridges. The extraction efficiency of fludrocortisone is examined below in chapter V. Fludrocortisone displays excellent extraction efficiency on the Sep—Pak cartridge as compared with cholesteryl butyrate, which cannot be eluted from the Sep—Pak cartridge using the steroid extraction procedure. Fludrocortisone acts as an efficient internal standard when added at the onset of the sample preparation procedure. Its presence aids in taking account of accidental sample losses occuring in the first extraction steps which would not be detected if cholesteryl butyrate was used as the internal standard. 29 2. Profiling by GC—MS—DS a. Installation of the megabore capillary column in the LKB 2091 GC—MS—DS The gas chromatograph attached to the LKB 2091 mass spectrometer is usually equipped with a coiled glass column (3m x 2mm i.d.) packed with 3% 0V—101 on Supelcoport (8O - 100 mesh). Although packed column chromatography does not provide adequate separation of all the steroidal components in human urine (3), most of these steroids can be identified and quantitated by examining selected ion chromatograms of the ions characteristic of a given steroid. Problems arise when a pair of poorly resolved steroids have identical mass spectra, such as the steroid epimers THF (3a,11b,17a,21-tetrahydroxy—5b- pregnan—ZO—one) and a—THF (3a,11b,17a,21-tetrahydroxy—53- pregnan-20-one). Packed columns also suffer from problems such as column bleed, poor column—to-column reproducibility, and the presence of active sites which necessitate periodic treatment of the column with a silylating agent, such as BSTFA (N,0- bis(trimethylsilyl)trifluoroacetamide. A narrow bore capillary column (0.32 mm i.d.) can be installed in the GC-MS—DS and will provide adequate resolution to separate the majority of components of a complex mixture. However, two factors must be taken into consideration when using capillary column gas 30 chromatography - repetitive scanning mass spectrometry. First, shorter scan cycle times are required with the narrow bore columns, such as 0.2 - 0.5 seconds/scan compared with 1 - 2 seconds/scan with packed column GC-MS. The greater scan rate is necessary to obtain an adequate number of data points to define the chromatographic profile. Sharp peaks are desirable for quantitative analysis of data obtained from GC—FID, but pose problems with GC—MS-DS if the appropriate scan rates can not be achieved with the instrument system available. The second factor to consider is that a complex mixture, such as urinary steroids, contains both major and minor components whose concentrations may differ by up to three orders of magnitude. Thus, if the lower limit of accurate quantitation for full mass range repetitive scanning GC—MS is 1.0 - 10 ng, in order to detect the minor components of a mixture it would be necessary to inject several hundred nanograms of the major components. Injecting such a concentrated amount of sample would result in the overloading of the narrow—bore capillary column. A compromise between packed columns and narrow-bore capillary columns can be found with the recently available megabore fused-silica capillary columns (0.53 mm i.d.). These very-wide-bore columns have a stationary phase consisting of polymer chains which are covalently bonded to each other in a cross-linking fashion and are also I} r 31 bonded to the silica surface. The stationary phase of the megabore column is far more stable than the stationary phase of the packed column, which is simply "coated" onto small particles of diatomaceous earth. The megabore columns display improved column—to-column reproducibility, compared with hand—packed GC columns. They also display improved high temperature stability (up to 325°C for the DB—l), and produce significantly less bleed than the packed column. In reference to the second factor concerning the narrow—bore capillary column mentioned above, the megabore capillary column has a sample loading capacity similar to a packed column and is more difficult to overload than the narrow-bore column. In reference to the first factor, the megabore column is capable of much greater resolution than a packed column, but is not as good as a narrow—bore column, which means that the peaks will not be as narrow and the mass spectral scan rate need not be so rapid. The LKB 2091 was refitted and replumbed to accept a 15—meter DB-1 fused—silica megabore capillary column. Steroid standards were analyzed on the megabore column and the results were compared with those obtained on a packed column. The retention times of all the standards were decreased. For example, at the same scan rate, the apex of the cortol peak eluted from the packed column during scan 251, and from the megabore column during scan 201. The man Inc.) c take on column, tempera the sta in the as liste 32 The manufacturer of the megabore column (J & W Scientific, Inc.) claims that an analysis on a 15-meter megabore will take one—third less time than on a 6-foot packed glass column, when both systems have identical flow and temperature conditions. The differing characteristics of the stationary phase of the two columns led to a decrease in the retention index of each of the steroid standards, as listed in Table 3. Table 3. Retention indices of steroid standards on packed and megabore capillary columns. Steroid Packed Column Megabore Column R.I. Total Ion R.I. Total Ion Intensity Intensipy 6b—Hydroxy 2753 1.5 x 106 2745 2.1 x 106 androst-4—ene— 3,17-dione Estriol 2916 3.4 x 105 2913 5.3 x 105 a—Cortolone 3080 1.6 x 105 3076 3.3 x 105 b-Cortol 3112 5.8 x 105 3103 9.3 x 105 Table 3. also lists the relative intensity of the total ion current for each of the standards. The intensity of each peak increased by an average of 63% on the m decre and Cs hydro: run, ( adeque peaks' 33 the megabore column. The level of background noise decreased by an average of 40% for estriol, a—cortolone, and cortol. The signal-to—noise (S/N) ratio for 6h— hydroxy—androst—4—en—3,17—dione, the first standard to be run, did not improve until the megabore column had been adequately baked out to remove contaminants and "ghost peaks". The improved chromatography available with the megabore column, compared with a packed column, is easily seen in Figure 2. The S/N ratio for cortol improves from 10:1 with the packed column (top profile "A") to 49:1 with the megabore column (bottom profile "B"). When changing from packed GC columns to megabore capillary GC columns the major advantage realized is the decrease in chemical noise, rather than any major increase in peak resolution. The megabore capillary column can not approach the resolving power of a narrow—bore capillary column. Figures 3 and 4 display four chromatograms of straight-chain hydrocarbon standards C20 through C30. The chromatograms in figure 3 are gas chromatograms displaying the response of the flame ionization detector. The top profile A was obtained using a 60—m DB—1 narrow-bore (.252 mm i.d.) capillary column. The other non—hydrocarbon peaks in the profile are from a coinjected sample blank. Notice, though, that the hydrocarbon peaks in the lower PrOfile B have been broadened, as this trace was obtained using a 15-m megabore capillary column (.53 mm i.d.). lul hUamCCuCh C>—.CFQZ q FigL Relative Intensity 30 Ix. Cortol 32 j "V ‘ l ‘ V r ‘ l V 1 V U l I j V I l V 1 V V l V’ V V V l V V V V ' Cortol Figure 2. Reconstructed total ion chromatograms, cortol (250 ng) and hydrocarbon standards C20 and C32 analyzed on the LKB 2091 GC-MS. A. 6-ft glass column packed with 3% 0V—101. B. 15—m DB—l megabore capillary column. Intensity Relative 20 ‘ji- gC;__ Time (min Time (min) Relative Intensity 2 20 0 22 Time I (min) ‘5} 20 22 I Time -9' 5 (min) Figure 3. 35 24 26 24 26 28 28 "\JLJLJ 1O 30 30 l 15 A. Gas chromatogram, 60-m DB-l narrow—bore (0.25 mm i.d.) capillary column. B. Gas chromatogram, 15-m DB-l megabore (0.53 mm i.d.) capillary column. Both samples: Hydrocarbon standards C20 to C30. Other peaks in profile "A" are from coinjected blank sample. 36 There is very little tailing observed and very little chemical noise present in this megabore GC—FID analysis. The chromatograms in figure 4 are reconstructed total ion chromatograms obtained on the LKB 2091 GC—MS-DS. In the top profile A, a six foot glass column packed with 3% 0V101 was used to achieve separation of the hydrocarbon standards. In the lower profile B, a 15 meter DB—l megabore capillary column was installed in the gas chromatograph. The lower profile shows a decrease in background noise and column bleed, a benefit of the bonded stationary phase of the megabore capillary column. For example, the signal—to—noise ratio for hydrocarbon 28 increases 2.3 to 8.5 when the megabore column is used. No dramatic decrease in peak width is observed when changing from the packed glass column to the megabore capillary column. This wide—bore column should be looked upon as a packed column alternative, since its chromatographic resolution is more similar to that achieved with a packed column rather than that achieved with a narrow—bore capillary column. An added benefit of using the DB-l megabore capillary column, rather than the 3% OV—101 packed column, is that the retention indices of the steroids analyzed by GC-MS will be quite similar to those obtained using GC—FID equipped with a 60 meter DB-l narrow—bore capillary column. Figure 5 displays two chromatograms of a urine 37 oumuoc\uom q ”woman :mom mz\vu MOON mug mzu co emuxaacmomu cuoNo muumvcmum conumuouv>x ”mmaaemm zoom .ce:aou sumaafinmu whopmwmoe Hume Emma .Emquum50unu ace ammo» vouunuumcouom .m .Ho~1>o Nm spa: woxuma canaou mmmam ammo .Emquum50ucu cOH ammo» wouuzuumcoumm .< .q musmflm amassz :mum am“ am" a: 03 am“ a: mg 93 a: 63 am am 2.. so am av am am a" a bp—bbe—lhlbybn—Ipbp—PDPD1—hbbb—blbbb—bnpn—nan—bPh-bebP—hbhb—bnib—D-Ph—bbph—pbnb—Phnh—n-bh_pnhD—nbp- r Om mm om «N mm Cu r .h .oemamu "use" I mmm mam mm" as" mm_ as" me so mm as p n P P — n P n P .— P h h b — P P b n — P b bl- _ p n P b — b n h P —r n h FBI—7|- rb E On mm N on mm ON I r _ e .«ssco_ ”use" m L. _ < Kitsuaiul aArieIaa LIN B l. 11.41, wt....l\# xuwmcvuch G>fiucmcx Relative Intensity I} on #- Figure 5. 38 F * * * k * * * i: *' F’ U ml K A: Reconstructed total ion chromatogram, IS-m DB—l megabore capillary column. ‘ B. Cas chromatogram. 60—m DB—l narrow-bore capillary column Both samples: 24-hour urine from a tumor-bearing female rat third day of Tamoxifen treatment. Stars indicate coinjected hydrocarbon standards, C26 to C3“. 39 sample from a tumor— bearin The top profile is a reconst obtained with a GC—MS fitted capillary column. The lower obtained with a 60-meter column. The intensity of peaks, designated with stars detected with a mass spectr ionization. Otherwise, t g rat treated with Tamoxifen. ructed total ion chromatogram with a lS-meter DB—l megabore profile is a gas chromatogram DB—l narrow-bore capillary the coinjected hydrocarbon is comparatively less when 9 ometer rather than by flame he two profiles are quite similar, although there is a decrease in resolution with the wider bore column. The congruence of the retention behavior of the steroids on both DB—1 phase columns aids in the creation of a steroid library for GCMET and MSSMET data analyses. b. Development of a library of rat steroid metabolites. A MSSMET library of human urinary steroid metabolites has been described by Vrbanac et al. (140) Unfortunately, this library is not applicable to the analysis of the steroids found in rat urine. As described above, these animals excrete the metabolites of corticosterone, rather than cortisol. In addition the rat produces steroid metabolites with a 5a—configuration, rather than 5b- configuration, which would change their relative retention indices. Thus, it was necessary to characterize the rat urinary steroids through manual interpretation of mass 40 spectra. In order to increase the concentrations of the steroid metabolites to be analyzed, a pool of female rat urine was collected and analyzed as described above. In addition, a group separation step was inserted after conjugate hydrolysis and prior to methoxine-trimethylsilyl derivative formation. The fractionation of free steroids involved adding the sample onto a Sephadex LH—20 column and eluting the steroids with a solvent composed of cyclohexane/ethanol (4:1) as described by Setchell and Shackleton (1973) (129). Three fractions were collected: 0 - 75 ml, 75 — 175 ml, and 175 — 275 ml. This group separation is used to simplify the chromatograms and obtain "clean" mass spectra. Fraction 1 contains androstane metabolites, pregnanediols and pregnantriols, and corticosterone metabolites. These corticosterone metabolites also appear in fraction 2 along with cortolones and cortols. Due to the overlap of peaks in each fraction, this procedure was not used to obtain any quantitative results. The total ion current mass chromatograms for each of the three fractions of pooled urine from postpubescent female rats are displayed in Figure 6. The first and second fractions contained the greatest amount of corticosteroid metabolites, which appear in the region of the chromatogram between hydrocarbons 30 and 34. Table 4 41 0-75 ml 30 .. 32 - 24 26 28 fifVV‘IVVVVIVVTVI‘VVVVI’VVVVIVYVV'VVVVIVIVV l V V V V I V1 V V l V V 75-175 ml n 28 1M 26 V V V V I V V V V l V V V V l V V V V I V V V V l V V V V l V V V V l V V V V l V V V V I V V V V 175-275Inl 1 - 28 24 26 3° 32 3‘ JL FVVVIVVYVIvvfiV'fivvvlfiVvi]V‘vvvlvaVIVVVV'V'"l"“l‘ 150 200 250 300 350 400 Scan Number l n Figure 6. Reconstructed total ion chromatograms of three fractions of female rat urine. Samples were analyzed using the LKB 2091 GC—MS fitted with a 6-ft glass column packed with 0V—101 with a scan rate of 4 sec/decade. 1ists 1 their r 3150 an prepUDE‘YI male rat pooled p :o thosc trihydrw fraction pregnan— tetrahvi 1 Each :ass spy entry am!- 42 lists the steroid metabolites found in each fraction, and their retention indices. Pooled urine samples from other types of rats were also analyzed by GC—MS. These sample sets included prepuberal female rats, as well as pre— and postpuberal male rats. The major steroid peaks identified in the pooled prepuberal female rat urine included (in addition to those listed in table 4): trihydroxy—pregnan—one, trihydroxy—pregnane-dione, and tetrahydroxy—pregnan-one in fraction 1; tetrahydroxy—pregnen—one and tetrahydroxy— pregnan—one in fraction 2, and trihydroxy-pregnen-one and tetrahydroxy-pregnen-one in fraction 3. Each compound identified by manual interpretation of mass spectra was added to a MSSMET library. Each library entry consists of a library number, compound name, retention index (TMI), designate ion (DIN), and set of confirming ions (CIN). Each confirming ion is paired with a relative intensity value, where the largest peak is assigned a value of 0.999 and the less intense peaks have smaller values. This information is obtained from a mass spectrum of each compound entered into the MSSMET library. A portion of the mass spectrum of tetrahydroxy-pregnan-one is displayed at the top of figure 7. The steroid displays a molecular ion at m/z 683. A characteristic loss of OCH3 from the methyloxime group yields a peak at m/z 652. 43 Table 4. Retention indices of steroid metabolites found in the three LH-20 fractions of post— puberal, pooled, female rat urine. indices were obtained with an LKB 2091 GC—MS Retention fitted with a 6-foot packed column packed with 3% 0V—101. Fraction 1 Steroid Dihydroxy—pregnen—one Pregnane—diol Dihydroxy—androstan—one Dihydroxy—pregnan—dione Trihydroxy-androstan—one Tetrahydroxy—pregnan-one Tetrahydroxy—pregnan-dione Tetrahydroxy-pregnan—dione Trihydroxy—pregnane-dione Trihydroxy-pregnane-dione Tetrahydroxy—pregnan—one Tetrahydroxy—pregnan—one Tetrahydroxy—pregnane-dione Tetrahydroxy—pregnan—one Retention Index 2753 2787 2898 2938 2969 3090 3324 3419 3009 3076 3114 3162 3318 3114 44 472 549 480 470 490 510 330 550 870 590 810 830 850 870 4: ENTRYTYPEzMSSMET NAM”: 1830 TETRAHYDROXY—PREGNAN-ONE TMh 3074 DIN: 562,1 CIN: 562,999, 472.510. 652.250 Figure 7. Abbreviated mass spectrum of a tetrahydroxy-pregnan-one (top) and its representation as a MSSMET library entry. The library entry contains the compound number (1830), the compound name, retention index (elutes at 74% of the time interval between straight—chain hydrocarbons C-30 and C—32), designate ion (DIN=562), confirming ions (CIN=562,472,652), and their relative intensities (999,510,250). 45 Other important peaks in the mass spectrum of tetrahydroxy-pregnan—one appear at m/z 472 and 562, which correspond to successive losses of trimethysilanol. The ions at m/z values 652, 562 and 472 are entered into the MSSMET library along with their relative intensities. When a computerized reverse library search is carried out with the MSSMET program, the computer searches a particular retention time "window" for the presence of a characteristic ion for a given library entry. If a GC peak for the designate ion is detected within the ' for the presence of confirming retention index "window' ions. For example, figure 8 displays a portion of the time axis of a GC—MS—DS analysis of a 24 hour urine sample from a tumor—bearing rat treated with Tamoxifen. Reconstructed mass chromatograms for the designate ion (m/z 562), confirming ions (m/z 472 and 652), and total ion current are displayed. The confirming ions must display peak maxima within one or two scans of the peak maximum for the designate ion for a positive identification. The MSSMET program calculates the peak areas for the designate and confirming ions and compares the relative intensities of these ions with the ratios of ion intensities listed in the MSSMET library. A correlation coefficient is calculated based on the similarity of relative intensities of the ion set for the observed peak and the library values. If the correlation «no u"\E < .nmmov Inoo» 46 .ucmEummuu cmwflxoeme mo awe vuflnu man so .o* you wcwummn Inossu Scum maqamw mean: palqm m «o mflmzamcm mn\mz\ou onu scum vocwmuno uconusu so“ Hmuou van .Ammo ccmmnq N\EV mac“ wcwauflwcoo .Amom N\EV cow oumcwflmmc can no mamuwoumaounu wmma umuoauumcoumm .m mpswflm omm 0mm 0mm own 0mm anew 9mm 0mm 0mm onm 0mm omm n. r 2:923 w. :0" .30. .vomaamw I800.“ H as v .353 IaOOu rI'I'Ir .OVonN lice" .0000? coeff: operal infort height "founc calcul listen calcul Of thc of the added and Lt: Into ,7 47 coefficient is greater than an arbitrary value set by the operator, the name of the compound and relavant information (e.g. retention index, retention time, peak height and peak area of the designate ion) is stored in a "found file". A relative concentration value is calculated for each "found" compound. The amount is listed relative to the amount of internal standard, and is calculated by dividing the peak area of the designate ion of the "found" compound by the area of the designate ion of the internal standard. The amount of internal standard added to the sample, the volume (ml) of urine analyzed, and the concentration (mg/ml) of creatinine are also taken into account when calculating the relative concentration of a "found" compound. Appendix I provides a listing of a MSSMET library of rat urinary steroids. An additional experiment was conducted to aid in the identification of rat urinary steroids. A deuterated silylating agent, N,0-bis(trimethyl—d9-silyl)acetamide (dg-BSA), (MSD Isotopes, Montreal, Canada) was used to derivatize an aliquot of pooled urine from tumor—bearing rats. Another aliquot of the pooled urine was derivatized with non—labelled BSA. Both aliquots received methoxyamine hydrochloride prior to silylation in order to prevent keto groups in ketosteroids and corticosteroids from undergoing enol ether formation during silylation. By comparing the mass spectra of analytes derivatized with either silyla‘ shift trimetl nolecui top prt not hat probahf index ( sample, with d5 15 inf} seen 1. indea, the iflt rather deuter; indeX analVQ‘ data s~ 48 either dg—BSA or do—BSA, one may determine the number of silylated hydroxyl sites on each molecule by measuring the shift in molecular weight. For example, the trimethylsilyl derivative of trihydroxy—pregnan—one has a molecular ion appearing at m/z 566, as displayed in the top profile, A, in figure 9. This particular steroid does not have a methyloxime moiety, as the keto group is probably in the hindered ll-position. It has a retention index of 3339. When another aliquot of the pooled urine sampled analyzed in profile "A" of figure 9 is derivatized with dg—BSA the molecular weight of trihydroxy-pregnan—one is increased by 27 to yield a molecular ion at m/z 593, as seen in the lower profile "B". The mass shift of 27 indicates the presence of 3 trimethysilanol groups due to the increase of 9 mass units when adding each —OSi(CD3)3 rather than -OSi(CH3)3 during trimethylsilylation. The deuterated trihydroxy—pregnan-one elutes at retention index 3314. Some other shifts in molecular weight for rat urinary steroids are listed in Table 5. 3. GCMET: profiling by GC—DS a. Development of GC—only metabolic profiling A system designed to be complementary to GC-MS—DS analysis, utilizing only gas chromatography and dedicated data system, has been developed at the Mass Spectrometry Facility at Michigan State University (1). Modern, bonded 49 j 100:— 0944. In]! see fi fi .1 AAMAA ""I'VV'I'V" "' ""If'—"T""I'VI'T‘V"1"I . «12:85 a 1003- 29000. / 7 32 34 +"'I"j'r'f"I'VTF""—T‘fi::fir'—f'—VITI'VIV"'I'A"—' 100:— 0912. un/z 593 A A/Lm V"' "fi '*"l""I"'VI'TV'I'"'I""1'fi'l"" :1003- 3050 «v: as .1 " 32 34 AA ""I""I""I""I"V'l"'+1""l""T""[V'f Figure 9. Shift in the retention index and molecular weight when trihydroxy-pregnan-one is derivatized with d -BSA (profile B) rather than dO-BSA (profile A). The molecular ion for the compound appears at m/z 593 rather than m/z 566, and its retention index decreases from 3339 to 3314. 50 Aflo+oov1msuzv Hmm Amsuzv owe AAo+ooV-Hm-zv owm Asmuzv mmo AAo+oov|~m|zv mmm Afimlzv wwo AAo+oov-N-Hm-zv cos AAo+ooV-Hm-zv owm Asmuzv wmo Axa+ooc-zv Ham A+zv cos Afio+ooV-xv ems Asm-zv «as A+zv mmm mCOH moam nqom Boom mofim mqmm .H.m (min) Figure 11. “Milan 54 24 26 28 30 34 36 p2 F CB 24 34 26 32 36 28 30 F CB 20 30 40 Two gas chromatographic analyses. The sample analyzed in the top profile is a 24—hour urine sample obtained from a normal female rat during the estrus phase of its estrous cycle, plus a series of coinjected hydrocarbon standards C 0 through C 6' The sample analyzed in the lower profi e is a blan plus the coinjected hydrocarbon standards. 55 only. If a hydrocarbon standard coelutes with an analyte peak, they may be chromatographically separated by varying either the column flow rate or temperature programming rate. Profiling by GC—DS has been undergoing evolutionary improvements over the past three years and has been a major focus of this research project. Initially, only manual interpretation of the tabular output from the gas chromatograph was possible. Coinjected hydrocarbon standards (C20 through C36) and an exogenous internal standard were visually identified and the retention indices and relative concentrations of the other peaks in the run were determined using the "RILEARY" FORTRAN program on the PDP 11/44 computer. Unfortunately, with this program it was necessary to manually supply the retention time and peak area of each constituent to be analyzed. This was rather tedious as each run contained over 50 peaks. The next improvement in GC-only profiling arose when an RS-232 link was installed between the 5840A GC and the PDP 11/44 computer. After each GC run, the raw chromatography data were transferred to the computer and stored. A program was written to calculate retention indices and relative concentrations after the user had manually entered the retention times and identity of each 56 hydrocarbon standard and the retention time and concentrations of the internal standard. The major advantage of the new GCMET software package is that the program utilizes a reverse library searching process and it is not necessary for the user to tell GCMET the retention times of the retention index standards or the quantitative standard. That information is contained in a library file created by the user for the particular set of analytes to be analyzed, using a constant set of experimental conditions. The hierarchy of the GCMET routines is displayed in Figure 12. After the raw data have been transferred to the computer, the program uses a two-pass search process to identify standards and analytes. First, the hydrocarbon standards are located and assigned retention indices. During the second pass, all the other GC peaks have their retention times converted to retention indices, based on the retention times of the hydrocarbon standards. The interactive GCMET software permits instant Operator evaluation of the quality of the identification assignments of peak identifications, as an override to those assigned by GCMET. The major advantages of GC—DS for metabolic profiling include the wide range of applications, and the relatively low cost of the necessary instrumentation. 57 .moCfl050p wCflHHmoua uaaonmuma :HCOIUU Bmzou we cofiumwficmwuo .NH mmswflm :35 3.2.52: 25.522. 2.825 .3 .38.. F 32.53. 225.25 TI] 5:32: .3 8:22»: - , - a: 38.: ca Ezra.» .2..me — 3: £3.22: “3 593: T :28: 22:: 33>»: - - , 33.23 F 5:. a: # ”manna—z 2:83 .35»: - as .5533... TE :53: _ 2:92.» 2:; 2.2358 fl 3: 222.525» :25. T1 58 The availability of a GC-MS-DS system, however, along with GC-DS results in a powerful metabolic profiling partnership. A sequential protocol can be developed where all samples are run, at relatively low cost on the GC—DS, then samples containing unknown species can be identified using a GC—MS-DS. When the GC—MS is fitted with a megabore capillary column, the retention characteristics of analytes on both instruments are quite congruent. b. A functional description of GC-DS profiling programs i. GCMAN: the GCMET library manager A library of compounds must be created for each specific application of GC-DS metabolic profiling. The library is specific for the types of compounds to be analyzed (e.g., steroids), as well as the experimental conditions used for the analysis (e.g., type of GC column, GC column length, carrier gas flow rate, split ratio, and temperature programming rate). To create a new library the operator logs on to the PDP 11/44 data system or boots up the IBM personal computer and then enters: GCMAN. The version number of GCMAN appears followed by a prompt: "GCMAN V1.6 GCMET library manager". If one wishes to edit an existing program the "OPEN library name" command is used, alternatively, the user can create a new library file by entering: 59 # CREATE LIBRARY FILENAME GCMAN asks for the name of the name translation file, which is a file which contains pairs of compound library numbers and compound names. All new library entries are added to the name translation file. If this file has not yet been created, the user must first enter: # CREATE NAME FILENAME When creating a new library GCMAN will also ask for comments which describe the new library file. Once the library file has been created the user adds compounds to the library and enters the variables requested: # ADD This will be compound number 136. Is this compound a retention index standard (NO)? Is this compound the default quantitative standard (NO)? Retention Index <0.00>: Delta Retention Index <0.00>: Area threshold <0): Normal Concentration <0.00>: High concentration <0.00>: Low concentration <0.00>: Resolution factor : Pattern factor weight <0.000>: 60 Pattern factor <(no linkages)>: Name for this compound (unknown 0): Each library entry is self—explanatory, except for the pattern factor weight and pattern factor, which are not being used at this time. In the future these variables may be used to improve the confidence of identifcation for metabolites which are excreted in a known pattern, (e.g., metabolite X is excreted in concentrations twice that of metabolite Y). Also, the concentration values given in ug per mg creatinine. If a mistake is made during an "add" session, typing control—Z will exit out of ADD and return the operator to the GCMAN monitor without changing the library. Alternatively, the user may finish the ADD session and then fix any mistakes with the EDIT command, by entering: # EDIT LIBRARY number Each attribute of the library entry may then be changed, but values will default to the existing attributes if no change is made. The EDIT command can also be used to alter the contents of the "name translation file", by entering: # EDIT NAME number 61 The compound name and compound retention index may be changed with this command. The DELETE command is used to remove entries from the library file and name translation file and name translation file. GCMAN confirms all deletions to protect against mistakes. The DELETE command formats are: # DELETE LIBRARY range # DELETE NAME range Deleting an entry from the name translation file automatically deletes the corresponding library entry. However, when the "DELETE LIBRARY number" command is entered, the name of the deleted compound is still retained in the name translation file. The LIST commmand displays the contents of a library or name translation file in various ways. A range of compounds in the library can be displayed (e.g., 1, 5, 3 — 16). If the range is not specified the whole library file will be listed. For example, # LIST 1, 5, 7 — 9 will list all the attributes of compounds # l, 5, 7, 8 and 9. # LIST BRIEF 5 62 will list the attributes of compound # 5 on a single line, rather than in a multiline format. The high, low, and normal concentrations are not listed, as well as the pattern factor information. # LIST WIDE 5 will list all the attributes of compound # 5. All the information will not fit on one line of an 80—column screen but will overlap to the next line. # LIST COMMMENTS # LIST NAME COMMENTS will produce a listing of descriptive comments entered into the header of a library file or a name translation file. All of the information produced by the LIST commands described above can be sent to other locations, rather than the user's terminal. For example, # LIST TO LPO: will send a complete library listing to the lineprinter, and # LIST TO TEST.LST will send the complete library listing to a data file called TEST.LST. 63 GCMAN libraries can be moved from one computer system to another by using the SCRIPT command. The SCRIPT command creates a script file of a range of compounds specified by the user. This script file is a list of commands that would have to be entered to recreate the library on another computer system, thereby eliminating manual typing. To make a script copy of a GCMAN library the user must OPEN the library and then enter: # SCRIPT CREATE A script copy is created and can be moved to another computer system (e.g., copy filename.SCR onto a floppy disk and insert into the disk drive of a different personal computer which can run the GCMET software package). To run GCMAN using a script file, the user types: GCMAN NUL, in the PDP 11/44 monitor, where the filename refers to the script file. When a GCMAN library editing session has been completed, the CLOSE command closes the currently open library file. Alternatively, typing EXIT will close the library file and leave the program. 64 ii. GCMET: analysis of gas chromatographic data files Once a gas chromatographic analysis has been performed, the GC integrator produces a report containing peak retention times and peak areas. A program called REFORMAT was written to convert raw gas chromatographic data obtained from any GC to one format amenable to GCMET analysis. The user enters: R REFORMAT, and is prompted for the raw data filename and the reformatted filename. Now the operator may perform a GCMET library search on the reformatted data: R $GCMET. The program will then print what version of GCMET is being used (e.g., GCMET V3.0a GC Metabolic Profiling). The GCMET program can be accessed at four different points during the analysis of a gas chromatographic data file. Initially, no data file is being analyzed. This state is designated as "No File". After the first pass analysis by the GCMET program, only the hydrocarbon standards are identified. This state is indicated as "Standards Found". If the user makes any changes in the assignment of standards, the state is changed to "Edit". Finally, when the other gas chromatographic peaks have been assigned retention indices and identified, the state is changed to "All Found". 65 To begin the analysis, the user enters the ANALYZE command along with the reformatted GC filename: # ANALYZE J1C3E3 This particular sample, J1C3E3, is from control rat # 3, first day of urine collection, estrus phase, analyzed at a temperature programming rate of 3°/minute on a 60 m DB—l narrow-bore capillary column. GCMET prompts the user for the library name, library number of the quantitative internal standard (e.g., fludrocortisone is # 9), amount of internal standard added to the sample (ug), and total amount of creatinine in the sample. Once this information has been entered the GCMET program makes its first pass through the data file, attempting to locate the coinjected hydrocarbon (retention index) standards and the internal quantitative standard. When the standards have been identified a graph appears at the terminal depicting the difference between the expected retention time of each standard (library value) and the experimental retention time as seen below. o.2oo+j E CALIBRATION vs. RETENTION rrne , o.150+ 7 : X 0. 100+ ' 0.050+ LO 9* u! 01* : It: . o.ooo+ —————————————————————— . ————————————————————————————————————— . 1 X x -0 . 050+ -1 2 g X g 8 1L100+ 66 The ordinate of this graph is a retention index scale and each standard appears at its expected retention index. The abscissa is a scale of positive and negative time values (minutes), indicating the deviation in retention times for the standards. This graph serves as a visual aid in quickly indentifying whether any standard assignments are questionable. It also serves as a quick diagnostic for the chromatographic system. For example, if the user is slow in hitting the start button on the integrator after sample injection, the resulting data will produce a rundelta versus retention time graph where the retention times of the standards will be early and yield negative rundelta values. Changes in column flow can also be detected by a nonlinear change in the rundelta pattern produced by the standards. If a retention index standard is not identified, the user will have to manually enter this standards with command: # SET STANDARD number TIME number GCMET usually will not locate a standard if its peak area is below the operator set threshold or if its retention time is outside of the "delta retention time" window. These two factors would lead to a low confidence level for the peak assignment. 67 It is very important for the operator to confirm the retention index standard assignments make by GCMET. As described above, the hydrocarbon standards are identified on the basis of retention time. If a standard elutes near another peak, misidentification is possible. Manual observation of the hydrocarbon standards in each run is useful, and if there still is doubt one can also analyze only the hydrocarbons and compare the two chromatograms. After GCMET identifies the retention index standards, verification is simple. First, the program will notify the user if a standard has not been found. Then the command LIST STANDARDS will give the status of each standard as well as its rundelta, confidence level, and peak number: # lis stan lib num status rundel conf peak num 1 Assigned -0.04 100. 4 2 Assigned —0.03 90. 12 3 Assigned 0.10 100. 18 4 Assigned 0.02 100. 25 S Assigned 0.06 100. 31 6 Assigned 0.10 100. 39 9 Assigned 0.12 100. 45 7 Assigned 0.17 80. 47 8 Assigned -0.06 100. 50 The assignment of each standard may be checked with the command "LIST nubmer". The program will ask the user what number of peaks on either side should also be displayed. 68 4 lis 6 Number of peaks on either side: 2 0 nun conf amount ri 37: 457.00 38: 1654.00 39: 6 2641.00 3200.00 40: 599.00 41: 429.00 If a standard has at 33.95 minutes rather Ari time 33.650 33.950 35.100 35.490 37.230 res unk unk unk unk unk manually assign the retention index standard: 0 set stan 6 time 33.95 0 lis 6 Number of peaks on either side: 2 # num conf amount ri ”ri time res 36: 235.00 33.450 unk 37: 457.00 33.650 unk 38: 6 1654.00 3200.00 33.950 unk 39: 2641.00 35.100 unk 40: 599.00 35.490 unk Typing LIST OPTIONS at any time during analysis will provide a listing of conf rundel 100. 0.10 been misassigned (e.g., # 6 eluted than 35.10 minutes) the user may conf rundel 100. -1.05 a GCMET the present state of the analysis and the valid commands available to thexuser: 0 lis op Current defaults: Library : Data file : Output name : Printing ROA files Quantitative Standard: Amount of Standard: Creatinine: Standard search confidence limit: Pattern factor upper limit: Pattern factor lower limit: Unassigned limit: Report area threshold : The current state is: Valid commands: ANALYZE GRAPH except RUNDELTA EXIT any SET command except FIND LIST QUIT COMPOUND sy:decrat.glb jlc3e3 jlc3e3 FLUDROCORTISONE 20.00 1.12 80. 80. 30. 30‘ no mg 2 z z x 12000. The standards have been edited. FINISH SCRATCH DELETE STANDARD 69 Once the standards have been correctly identified the user instructs GCMET to take a second pass through the GC data in order to assign retention indices and compare these peaks with the GCMET library. The command ANALYZE can be used to complete the peak searching, write the requested output files, close the present data file, and prompt the user for the name of the next GC data file to be analyzed. The FINISH command will also complete the peak searching, write the output files, close the present data file and keep the GCMET library open, but will not prompt the user for the next data file to be analyzed. FINISH leaves GCMET in a "No file open" state. Another command used to initiate the second pass of the GCMET program is FIND PEAKS. The program will complete the peak searching but then remain in the "all found" state without writing output files and closing the present data file. GCMET will indicate whether any major or minor peaks in the data file were not assigned corresponding library identities after the second pass. Peak identities can be checked with the LIST number command, or by using the LIST PEAKS command: 0 find peaks 10 small peaks were unassigned 0 lis 6 Number of peaks on either side: 2 fl num conf amount ri Ari time res conf rundel 37: 67 91. 8.54 3129.44 -0.44 33.650 unk 38: 69 79. 30.90 3144.04 2.46 33.950 unk 39: 6 100. 49.33 3200.00 35.100 unk 100. 0.10 40: 73 80. 11.19 3220.00 -1.00 35.490 unk 41: 76 88. 8.01 3309.23 -1.23 37.230 unk 70 fl lis peaks Start time <0.>: 20.0 End time (100.): 22.0 # num conf amount ri Ari time res conf rundel 15: 19 71. 15.43 2516.56 1.44 20.260 unk 16: 125 72. 74.87 2562.03 1.97 21.290 unk 17: 24 99. 9.56 2592.94 0.06 21.990 unk After the peaks have been identified the data file can be closed with the commands FINISH or ANALYZE. Prior to this point, though, the user must determine what type of data output is desired. GCMET can produce three types of output files. The ROA (result of analysis) file is the most simple, containing only library compound numbers and each corresponding confidence level and relative concentration value. This file can then be used with GCSTAT (GC statistical analysis routines). The NRA (named result of analysis) file is the most useful type of output available for immediate operator observation, providing information about the compound name, confidence level, relative concentration, retention index, delta retention index, and retention time. The final output file is the TRA (total result of analysis) file, which provides the name, confidence level, relative concentration, retention index, delta retention index, retention time, peak resolution factor (e.g., baseline—baseline separation between a peak and its neighbors), the peak area to peak height ratio, peak percent area of total area, and the 71 results of the first pass where the retention index standards are identified (e.g., confidence level and delta retention time). The three types of output files are displayed in table 6, 7 and 8. Any or all three of these commands may be entered during a GCMET analysis by entering the commands: # SET NRA # SET TRA # SET ROA Each command can be removed by entering: # SET NO NRA # SET NO TRA # SET NO ROA Chromatographic conditions may change over time. Slight changes in flow rate can effect the retention times of eluting peaks. The GCMET program allows the operator to take these changes into account and maintain correct peak identification abilities with the CALIBRATE command. After the first pass of the GCMET program and a rundelta versus retention time is displayed, the operator enters the CALIBRATE command. A calibration file is created and the offset is noted between the library retention time and 72 Table 6. Result of analysis file.(with compound names) * DBl:JIC3E3.NRA;2 ;File Name * Data taken from j1c3e3 ‘ ‘ D 29 4 1985 11 52 38 ;Date and Time I HP 5840A ' ;Gas Chromatograph U 0 . * Date of Analysis was Wed Aug 06 09:44:36 1986 P GCMET V3.0a ;GCMET Version F NRA ;Type of File L DECRAT.GLB;1 ;Library Name N RATRATE3 ;Name Translation File * name conf amount ri Ri time HYDROCARBON 28 100 20.75 2800.00 0.00 26.670 Unknown 2825 95 o 1.73 2824.58 0.42 27.200 Unknown 2907 97 3.55 2907.11 -0.11 28.980 tetrahydroxy-pregnene-dione 85 3.74 2988.25 -0.25 30.730 HYDROCARBON 30 100 19.56 3000.00 0.00 30.984 TRI-OH-PREGNAN-ONE 97 4.86 3014.52 0.48 31.280 TETRA-OH-PREGNAN-ONE (no) 93 2.19 3072.32 -0.32 32.460 Unknown 3121 96 1.63 3120.82 0.18 33.450 TETRAHYDROXY-PREGNENE-ONE 84 11.46 3145.31 1.19 33.950 HYDROCARBON 32 100 18.30 3200.00 0.00 35.067 Unknown 3327 96 2.34 3326.23 0.77 37.520 ri3339 84 136.75 3340.12 -0.62 37.790 FLUDROCORTISONE 90 6.62 3360.00 0.00 38.176 HYDROCARBON 34 100 21.11 3400.00 0.00 38.919 Unknown 3422 94 5.52 3425.04 -1.04 39.410 CONTAM 9O 89 18.96 3450.03 -2.03 39.900 HYDROCARBON 36 100 20.35 3600.00 0.00 42.841 73 Table 7. "Total" result of analysis file. * DBl:JIC3E3.TRA;3 ;File Name * ‘Data taken from jlc3e3 D 29 4 1985 11 52 38 ;Date and Time I HP 5840A ;Gaa Chromatograph U 0 * Date of Analysis was Wed Aug 06 09:44:34 1986 P GCMet V3.0a ;GCHET Version F TRA ;Type of File L DECRAT.GLB;1 ;Library Name N RATRATE3 ;Name Translation File *num conf ' amount ri Ri time res arht Zarea 4 100 20.75 2800.00 0.00 26.670 unk 0 3.591 100. 0.04 41 95 1.73 2824.58 0.42 27.200 unk 0 0.300 46 97 3.55 2907.11 -0.11 28.980 unk 0 0.614, 50 85 3.74 2988.25 -0.25 30.730 unk 0 0.647 5 100 19.56 3000.00 0.00 30.984 unk 0 3.385 100. 0.05 53 97 4.86 3014.52 0.48 31.280 unk 0 0.842 57 93 2.19 3072.32 -0.32 32.460 unk 0 0.379 135 96 1.63 3120.82 0.18 33.450 unk 0 0.282 69 84 11.46 3145.31 1.19 33.950 unk 0 1.983 6 100 18.30 3200.00 0.00 35.067 unk 0 3.167 100. 0.07 77 96 2.34 3326.23 0.77 37.520 unk 0 0.405 123 84 136.75 3340.12 -0.62 37.790 unk 0 23.669 9 90 6.62 3360.00 0.00 38.176 unk 0 1.146 100. 0.08 7 100 21.11 3400.00 0.00 38.919 unk 0 3.653 80. 0.08 87 94 5.52 3425.04 -1.04 39.410 unk 0 0.956 90 89 18.96 3450.03 -2.03 39.900 unk O 3.282 8 100 20.35 3600.00 0.00 42.841 unk 0 3.523 100. 0.09 74 Table 8. Result of analysis file. Zlfi'n'd *IZrdtd * e #- num 41 46 50 53 57 135 69 77 123 87 90 DBl:JlC3E3.ROA;5 ;File Name Data taken from jlc3e3 29 4 1985 11 52 38 ;Date and Time HP 5840A . ;Gas Chromatograph 0 Date of Analysis was Wed Aug 06 09:44:34 1986 GCMet V3.0a ;GCMET Version ROA ;Type of File DECRAT ;Library Name RATRATE3 ;Name Translation File conf amount 100 20.75 95 1.73 97 3.55 85 3.74 100 19.56 97 4.86 93 2.19 96 1.63 84 11.46 100 18.30 96 2.34 84 136.75 90 6.62 100 21.11 94 5.52 89 18.96 100 20.35 75 actual retention time of each standard and that value is added to the calibration offset in the GCMAN library. If the rundelta versus retention time graph is replotted, all rundelta values appear as zero. RUNDELTA vs. RETENTION TIME 1.000: 0.750% 0.500; 0.250; 0) 0003-"!E It a: —————— a; ———————————— 1 ------ x —————— x x -- 1 2 3 4 5 6 9 7 8 The calibration offset for each standard which are present in the GCMAN library can be graphically displayed by entering the command: # GRAPH CALIBRATION RUNDELTA vs. RETENTION TIME 0.200+ : ‘ * 0.150+ 7 , x o.1oo+ ' (0* m 0.0SO+ Lflfi manually assign a compound name to a particular peak. example If a peak is misidentified by GCMET, compound assigned to If compound number # 73 should be chromatographic peak the user may removes compound # SET COMPOUND 73 PEAK 41 as displayed 76 number below. 76 the 40th peak in the from # being 73 listed Assigning peak 41 assigned Compound 76 must be manually assigned or it a lis 77 Number of peaks on either side: 3 # num conf amount ri Ari time 39: 6 100. 49.33 3200.00 35.100 40: 73 80. 11.19 3220.00 —1.00 35.490 41: 76 88. 8.01 3309.23 —1.23 37.230 42: 77 86. 6.31 3324.10 2.90 37.520 43: 123 76. 368.72 3337.95 1.55 37.790 44: 82 84. 3.14 3354.36 —0.36 38.110 45: 9 90. 17.86 3360.00 38.220 8 set compound 73 peak 41 WARNING: This edit removes compound 76. You must SET this COMPOUND yourself, # lis 77 Number of peaks on either side: 3 # num conf amount ri Ari time 39: 6 100. 49.33 3200.00 35.100 40: 73 80. 11.19 3220.00 -1.00 35.490 41: 73 88. 8.01 3309.23 —1.23 37.230 42: 77 86. 6.31 3324.10 2.90 37.520 43: 123 76. 368.72 3337.95 1.55 37.790 44: 82 84. 3.14 3354.36 -0.36 38.110 45: 9 90. 17.86 3360.00 38.220 below has enter the command: the operator may For been gas chromatographic run. assigned to the 4lst gas as compound 73 to peak 41. will be lost. res unk unk unk unk unk unk unk otherwise it will be res unk unk unk unk unk unk unk cont 100.“ conf 100. 100. rundel 0.00 0.00 rundel 0.00 77 After all the peaks have been assigned either manually or automatically by the GCMET program the file can be closed using the FINISH command. Alternatively, the operator can choose to leave the program without saving any output files by entering the QUIT command or control- Z. The EXIT command also leaves the GCMET program, but closes and saves the output files. The EXIT command also closes the active file from the "standards found" state and automatically completes the peak finding process and the production of output files. The SCRATCH command can be used to abort analysis of an open CC data file but does not remove the user from the GCMET program. Other SET commands are available during GCMET analysis including: # SET AMOUNT STANDARD amount - assigns the amount of quantitative standard added. # SET CREATININE amount — assigns the total amount of creatinine found in the sample. # SET QUANTITATIVE library # — assigns a particular library entry to be the quantitative internal standard. # SET REPORT threshold — sets a cutoff peak area below which unassigned peaks are not reported. 78 # SET STANDARD library # AUTO - allows GCMET to assign the standard. # SET STANDARD library # TIME time — reassigns a standard based on retention time. # SET STANDARD library # VIRTUAL time — assign a virtual (imaginary) standard. # SET UNASSIGNED Z — assign a threshold for unknowns to be reported in the total result of analysis file. # SET WINDOW % — assign a cutoff value for standard assignments. An additional graphing command is available. GRAPH RI produces a plot of retention time versus retention index for the retention index standards. This command is useful when endogenous compounds are used as retention index standards since it shows the user the library compound number positioned elutes at a particular retention index. I 9r ri INDEX vs. RETENTION TIME 3600.* I . XX 8 3200.+ * 97 01:4“ 2800.+ * NI 9'? LJ * 2400.+ 2000.+ 1 1600.+ 1200.+ 800.+ 400.+ 0.4 - --------- 79 iv. GCSTAT: statistical manipulation of GCMET output files. This program has been designed to aid in the compilation of simple statistical information concerning the compounds identified concerning the compounds identified in GCMET analyses. GCSTAT calculates the mean, standard deviation, the relative error, and frequency of occurrence for each compound which appears in a reference set of GC runs selected by the user. For example, the user may wish to determine the mean excretion of a metabolite in the urine over a seven-day period. After GCMET analysis of th seven daily urine samples the user creates a reference set of those seven analyses. First, the user enters: # CREATE filename.set The GCSTAT program will prompt the user for the lowest possible confidence level above which compounds will be reported. The user must also enter the name translation file used in the GCMET analyses. Once a new reference set has been created the user Inust enter the RDA files to be included in the new set: # ADD filename.ROA 80 After all the desired ROA files have been added to the reference set a tabular output comparing individual ROA files with the reference set can be obtained with the COMPARE command. The table will include the mean, standard deviation, and the relative standard deviation of each compound in the reference set and the compound concentration for each ROA file entered into the set. The COMPARE command can be specified for a range of compounds (e.g., l, 2, 5 — 7) or all of the compounds in the reference set. The COMPARE TO clause allows the user to assign an output destination other than the user's terminal (e.g., LPO:). The OVERLAP command is used to compare two different sets of GCMET analyses. The user must first OPEN a previously created reference set and enter the command: # OVERLAP range The GCSTAT program asks the user to enter the name of the second set to be overlapped, and also for a threshold value for number of standard deviations away from the mean to extend for each compound. The program overlaps the mean :: (entered number x standard deviation) for a given compound in each set. As seen later in table 14, a percent overlap value is reported as is the mean and standard deviation for each compound in the two sets and which set has the larger mean. A "TO" clause can also be 81 used with OVERLAP to direct the results to a lineprinter or to a list file. Other GCSTAT commands include: # CLOSE — closes the reference set which is open. # COMMENTS — prints the reference set file header. # DELETE filename — removes an ROA file form the reference set. The filename is still saved in the list of ROA files that were ever in the reference set (NAMES ALL). # DELETE NAME filename — removes both the ROA file and record of the filename. This command must be used prior to the DELETE filename command. # NAMES ALL — lists all of the ROA filenames ever entered into the open reference set. # NAMES CURRENT — lists only the ROA filenames currently in the open reference set. # NAMES PAST — lists the ROA filenames which have been removed by the DELETE command. # EXIT, QUIT or CONTROL-Z — all allow the user to exit form the GCSTAT program. 82 # RI ON or OFF - either allows retention indices to be printed for each compound listed in the output reports generated by GCSTAT or disables that printing. # SORTBY output style — allows the user to generate output reports arranged by the alphabetical order of compound names (ALPHABET) by compound retention index (RI), or by compound library number (LIBRARY). III. TAMOXIFEN A. Tamoxifen — Human Studies 1. Pharmacology of Tamoxifen Tamoxifen (trans~1—(4—b—dimethyl aminoethoxyphenyl) but—l-ene) is a nonsteroidal antiestrogenic agent which has proven to be an effective, palliative treatment for certain, advanced breast cancers (27). The effects of Tamoxifen are species—dependent, exhibiting either estrogen antagonism, partial estrogen agonism, or both. Tamoxifen blocks the peripheral function of estrogen on tumor cells and other target tissues through competitive inhibition of estradiol binding to the estrogen receptor protein in cellular cytoplasm. The cellular proliferation of various breast tumors is induced by estrogen (28). Cytoplasmic estrogen binding proteins have been detected in these adenocarcinoma cells as well as in other estrogen target tissue (uterus, vagina, pituitary, and hypothalamus). These receptors bind l7b—estradiol, and the resulting hormone—receptor complex (true messenger unit) is transformed into protein which has a high affinity for the cell nucleus. The estrogen—receptor unit is translocated to the nucleus causing an increase in RNA synthesis. Cytoplasmic protein synthesis also increases leading to a 83 84 replenishment of the cytoplasmic estrogen receptors. About twenty—four hours after exposure to estradiol, DNA synthesis and cell division occurs. The generally accepted hypothesis for the mechanism of action of Tamoxifen involves the drug competing for and binding to the estrogen receptor protein in the tumor cytoplasm (29). The antiestrogen depletes the cytoplasm of free receptors. The Tamoxifen-receptor complex (false messenger unit) undergoes nuclear retention for a time period longer than that for the normal estrogen-receptor complex (30). Thus, Tamoxifen triggers an initial brief, estrogen—like response by binding to the estrogen receptor, followed by the antiestrogenic action. These cellular conditions severely impair the continued growth of the estrogen-dependent tumor. 2. Metabolism of Tamoxifen in humans Fromson et al. (32) investigated the metabolism of orally administered [ C]Tamoxifen in women with mammary carcinoma. They found that most of the radioactivity was slowly excreted in the feces, and only small amounts of 14C appeared in the urine. One patient excreted 9% of the dose in the urine and 26% in the feces over a eight day sampling period. A second patient voided 14% of the dose in the urine, and 51% of the radioactivity was recovered in the feces. 85 A serum half—life of eleven hours was noted after dosing with 20 mg Tamoxifen. The maximum serum level in one patient occurred at four hours, with approximately 0.10 ug Tamoxifen per ml of serum. After two weeks of Tamoxifen treatment a serum level of 0.013 ug was observed. The prolonged presence of Tamoxifen in the blood suggests the occurrence of enterohepatic circulation, especially for the hydroxylated metabolites. Some of these metabolites possess antiestrogenic activity, so biliary recirculation seems to prolong the duration of antiestrogenic behavior. Daniel et al measured the concentration of Tamoxifen and monohydroxytamoxifen in the plasma of patients with advanced breast cancer receiving twice daily doses of 20mg Tamoxifen. After sample preparation and derivatization, gas chromatography-mass spectrometry was used in the selected ion monitoring mode to detect tamoxifen (m/z 371.2249) and monohydroxytamoxifen (TMS) (m/z 459.2593). Daniel et al. (33) found that after an initial rise, the plasma concentration of Tamoxifen was maintained between 150 and 250 ng/ml plasma. Monohydroxytamoxifen's concentration was much lower, ranging from less than 1 ng per ml to 7.89 ng per ml. This level of monohydroxytamoxifen still exceeds, on average, the circulating level of estradiol. This metabolite has greater affinity for the estrogen receptor than Tamoxifen. 86 However, the concentration of Tamoxifen in the blood is 1,000 to 2,000 times that of estradiol and these proportions allow the drug to successfully compete with estradiol for the estrogen binding site. N-desmethyltamoxifen appears to be the quantitatively most important Tamoxifen metabolite. Its concentration was measured by Daniel et al. (34) in both plasma and primary breast cancer tumor tissue homogenates. This metabolite was found to be present at 1.5 times the concentration of the parent drug in plasma, and twice the concentration in tumor tissue. The mean plasma concentration of N—desmethyltamoxifen for fourteen patients whose length of treatment ranged from 15 to 180 days was 462 t 167 ng/ml. The mean concentration of Tamoxifen was 300 :t 162 pg/ml, and monohydroxytamoxifen was 6.7 :t 2.3 pg/ml as measured by high resolution GC-MS. The mean estradiol concentration was 0.15 t 0.04 pg/ml, as measured by radioimmunoassay. Another metabolite, N— desdimethyltamoxifen was reported to have a blood level of 40 ng/ml by Kemp et al. (35). Camaggi et al. (36) used an interesting technique to measure the concentration of Tamoxifen and its major metabolites. After sample extraction, they employed a high-performance liquid chromatographic analysis with post column on—line photocyclisation of the drug and 87 metabolites to the corresponding phenanthrenes. After a single 40mg dose of Tamoxifen, maximum plasma levels of Tamoxifen and desmethyltamoxifen were reached within two hours, to concentrations of approximately 55 and 40 ng/ml, respectively. High performance liquid chromatographic analysis with post—column fluorescence activation was used by Brown et al. (37) to determine the serum concentration of Tamoxifen and four metabolites: monohydroxytamoxifen, N— desmethyltamoxifen, metabolite E (trans—1(4— hydroxyphenyl)1,2—diphenylbut—1-ene) and metabolite Y (trans—1(4—hydroxyethoxyphenyl)1,2—diphenylbut—1—ene). Three different chromatographic systems were used. The major components: monohydroxytamoxifen, Tamoxifen, N— desmethyltamoxifen, were successfully eluted form a C18— reversed phase column using a mobile phase of absolute methanol with 0.04% diethylamine acetate. When a silica column was employed, with a methonal-water-triethylamine- acetic acid mobile phase, N-desmethyltamoxifen and Tamoxifen were eluted in an order opposite that of System I. A third system was needed to successfully separate metabolites E and Y. Again, a silica column was used, but hexane with 1.3% isopropyl alcohol was used as mobile phase. 88 3. Use of Tamoxifen in Human Breast Cancer Tamoxifen treatment has proven most successful with carcinoma of the breast in postmenopausal women. Patients whose tumors contain estrogen receptors are most likely to respond to Tamoxifen therapy. Only 15% of breast cancers in premenopausal women yield a positive estrogen receptor assay, compared with almost two—thirds of the postmenopausal. Patterson (38) summarized the results of forty studies examining the effectiveness of Tamoxifen in patients with advanced breast cancer. Sixty—two percent of patients who were ER positive underwent either complete remission, partial regression, or stabilization, whereas only twenty—seven percent of the ER negative patients responded favorably. There may be several explanations why receptor—poor patients benefit from Tamoxifen treatment. Histologically similar sections of the same tumor mass may show a wide range of ER content. Poulsen (39) reported a range as large as 0-300 fmol/mg protein. Tamoxifen may exert inhibitory effects on tumors lacking estrogen receptors by binding to the androgen receptor, for which the natural ligand is 5a—dihydrotestosterone, or to the "antiestrogen binding site" which has been found in certain human breast carcinoma cells. Tamoxifen is an inhibitor of prostaglandin synthetase, which may limit the growth of tumor cells and reduce metastasis by eliminating the 89 immunosuppressive effect of prostaglandin E produced by the cancer cell (40). The likelihood of a patient deriving a favorable response from Tamoxifen therapy increases with the age of that patient (41). For example, postmenopausal women under 60 years of age have a response rate of 31%. Those patients aged 61 to 70 have a response rate of 36%, and patients over 70 have a 45% response rate. Manni et al.(42) found that Tamoxifen-induced remissions in postmenopausal women with hormone responsive breast cancer were comparable to those obtained with surgical hypophysectomy (surgical ablation of the pituitary gland). Tamoxifen was administered orally to 113 patients with stage IV breast cancer (progressive metastatic mammary carcinoma). Fifty percent of the patients had objective remissions of over eighteen months. Manni et al. also noted that older age, positive estrogen receptor assay, and previous response to endocrine therapy predicted better response to Tamoxifen therapy. Ribeiro and Swindell (43) report the results of a seven year trial of adjuvant therapy for breast cancer with Tamoxifen. Following surgery, premenopausal women in the study received either 20mg Tamoxifen per day for one year or had an irradiation menopause. There was no statistically significant difference in the survival rate 90 for patients receiving either therapy. Similarly, the postmenopausal patients, who received either Tamoxifen treatment or no treatment (controls) after surgery, displayed no large statistically significant differences in response for either group. However, these postmenopausal patients did not receive estrogen receptor assays and were not divided into ER positive or ER negative groups. When both pre- and post— menopausal patient groups were combined and then divided by axillary (armpit) node status, there is noted a statistically significant survival benefit for those treated with Tamoxifen. Ribeiro and Swindell also noted that Tamoxifen—treated patients showed a significant delay in the first relapse of the disease, and also postmenopausal patients displayed a marked reduction in distant metastases . Viladiu et al. (44) reported the results of a randomized study where postmenopausal advanced breast cancer patients received either chemotherapy or chemotherapy plus hormonotherapy. The treatment regimens were: (A) cyclophosphamide, methotrexate, and 5- fluororacil (CMF) (chemotherapy), (B) CMF and Tamoxifen, and (C) CMF and medroxyprogesterone. Both (B) and (C) treatment groups displayed a greater proportion of positive responses (70.6% with Tamoxifen, 67.7% with medroxyprogesterone) than treatment group (A) (45% with 91 CMF alone). Combined hormonotherapy and chemotherapy may or may not lead to a longer duration of response than just chemotherapy treatment alone. The authors noted that survival rates for each group are dependent on the proportion of patients responding to the therapy, thus responders survived longer than nonresponders, regardless of which therapy they received. In general, between 25 and 60% of postmenopausal patients with advanced breast cancer showed improvement during Tamoxifen therapy, compared with an improvement rate 20 to 35% for patients receiving classical hormonal therapy (androgens and estrogens) (45). With patients that are more than 5 years postmenOpausal and have predominantly soft-tissue involvement, the response rate to Tamoxifen (41%) is similar to cytotoxic drugs (50%), with a smaller incidence of side effects for the patients receiving Tamoxifen. In summary, at present, Tamoxifen is the preferred therapeutic agent for the primary endocrine management of breast cancer in postmenopausal women. In the future, the drug may be used routinely as one component in a hormonotherapy/chemotherapy regimen, or as adjuvant therapy for the treatment of advanced or recurrent breast cancer. The drug exerts the majority of its antitumor effect through competition for the cytoplasmic estrogen 92 receptor. The fact that some patients lacking estrogen receptors benefit from Tamoxifen treatment provides clinical evidence that the drug may have a complex spectrum of physiological actions in addition to the antiestrogenic effects. 4. Side effects of Tamoxifen treatment in breast cancer patients Tamoxifen therapy is well tolerated by most postmenopausal breast cancer patients. The most frequent adverse reactions include hot flashes, nausea, and vomiting, which may occur in up to 25% of the users (40,46). These side effects, as well as pruritus vulvae and vaginal bleeding or discharge are due to the antiestrogenic effects of the drug. A smaller percentage of the patients exhibit classical estrogenic problems, such as fluid retention. Other minor adverse effects that have been reported include lassitude, fatigue, lightheadedness, dizziness, headache, mental confusion, depression, polydypsia, dryness of the skin and mucosa, rash, leg cramps, phlebitis, jaundice, anorexia, and edema. Tamoxifen has also caused thromboembolism, leukopenia, dermatomyositis, Peliosis Hepatitis, pulmonary embolism, and heart failure in very rare cases. Tamoxifen therapy has caused hypercalcemia and exacerbation of bone pain in breast cancer patients with 93 osseous metastases (47 - 55). The hormone-induced hypercalcemia appears within five to ten days after onset of Tamoxifen treatment causing neurological and gastrointestinal problems, which may be fatal unless controlled with supportive measures. These deleterious side effects probably arise from the biphasic nature of the drug. Initially, Tamoxifen acts as an estrogen agonist when binding to the estrogen receptor, causing accelerated tumor growth and increased bone resorption. After the drug—receptor complex has translocated to the cell nucleus, the tissue is rendered hormone unresponsive, since the cytosol estrogen receptors are not replaced. Frequent monitoring of serum calcium levels is recommended during the initial three weeks of Tamoxifen treatment. The appearance of mild hypercalcemia is not a reason to discontinue Tamoxifen therapy, since this side effect is indicative of a hormone—responsive tumor. Legha et al. (53) noted that ten patients out of a total of 470 patients with metastatic breast cancer developed hypercalcemia. All of these patients with hypercalcemia had bone metastases. Serum calcium levels jumped from a normal value of about 10mg/100ml up to highest levels of about 16mg/100ml. Although patients with osterolytic metastases are more susceptible to hypercalcemia during Tamoxifen therapy, the side effect only occurred in four percent of that patient group. The 94 serious hypercalcemia complications are usually shortlived and controlled with supportive measures and Tamoxifen dosage can usually be continued. Tamoxifen is an amphiphillic cationic compound, structurally similar to drugs such as chlorpromazine, thioridazine, and amiodarone. These drugs are characterized by a positively charged hydrophilic side chain and a hydrophobic moiety on the same molecule. These compounds have been shown to induce generalized lipidosis in several species of laboratory animals and in humans. They form tight but reversible bonds with polar lipids causing an accumulation of drug—polar lipid complexes is lysosomes and interfere with the intralysosomal catabolism of polar lipids. McKeown et al (56). reported that one patient receiving high doses of Tamoxifen (90mg, twice daily) developed a case of retinopathy. The patient's visual acuity decreased from right eye (RE) 20/30 and left eye (LE) 20/25 to RE 20/300 and LE 20/80 over a thirty—three month treatment period. She received 90 grams of Tamoxifen during this period. The patient developed bilateral intraretinal refractile opacities, cyctoid macular edema, and lesions of the level of retinal pigment epithelium. Tamoxifen's amphiphillic character causes lipidotic changes in the corneal epithelium which may lead to a lysosomal storage disorder. Pr' 19 da 811' I11: uh: ca' prev thr the COD? 95 Fortunately, low dosage schedules of Tamoxifen do not produce ocular changes. Beck and Mills (57) reported that 19 patients treated with Tamoxifen for periods of 3 months to 4 years at doses which did not exceed 20 mg twice daily. None of the patients displayed any retinopathy or superficial corneal opacities. Since eleven of the nineteen patients in this study obtained complete clinical regression of the metastatic breast cancer, Beck and Mills questioned the usefulness of high dose Tamoxifen therapy. Thrombosis is a side effect of Tamoxifen treatment which affects obese, postmenopausal women with breast cancer. Usually these patients had some degree of varicose veins. The condition is probably caused by the initial estrogen-like activity of the drug. Nevasaam et al. (58) described four cases of thrombotic changes in four patients receiving Tamoxifen. All four patients were successfully treated with anticoagulants. Only one patient could continue to receive Tamoxifen, though, together with an anticoagulant. Hendrick et al. (59) also warned against giving Tamoxifen to patients who have a predisposition to thrombosis. Again attributing the thromboembolisms to the paradoxical estrogen effects of the drug, the author compares the side effect to the changes in coagulation in women receiving oral contraceptives. These changes appear within ten days of 96 administration of the birth control drug, and persist for up to three weeks after cessation of treatment. 5. Steroid metabolism in breast cancer patients Menopause is the time in a woman's life when there is a permanent cessation of menstrual activity. This usually occurs between 35 and 58 years of age: 25% of all women reach menopause by age 47, 50% by age 50, 75% by age 52, and 95% by age 55. The aging ovaries show a diminishing response to the pituitary gonadotropins, partly due to the shrinking numbers of follicles. Estrogen secretion by the ovary decreases until the level of that hormone is inadequate to maintain the estrogen—dependent tissues such as the breasts and genital organs. Their tissues gradually atrOphy. The skin and bones begin to thin due to decreased protein anabolism. Other estrogen deficiency side effects such as hot flashes arise during this period. The incidence of atherosclerosis begins to rise, since estrogen helps to maintain low levels of cholesterol in the plasma. Bulbrook et al. (60) reported the plasma levels of estradiol, estrone and progesterone in premenopausal women to be about 100, 60 pg/ml, and 10 ng/ml, respectively. The postmenOpausal group of women displayed diminished average plasma levels of estradiol, estrone, and progesterone of about 9 pg/ml, 18 pg/ml, and 0.75 ng/ml. 97 There are interesting differences in steroid metabolism between normal, postmenopausal women and those with breast cancer. Kodama et al. (61) reported that patients with breast cancer excreted excessive amounts of tetrahydrocortisol relative to androsterone. Low urinary androsterone and etiocholanolone concentrations after mastectomy have been linked to early tumor recurrence (Thomas et al. 1982) (62). Thomas et al. (63) described a linkage between the ratio of androsterone to a—cortolone as a means of predicting possible recurrence. Measurement of urinary androgen metabolites alone may be used as indicators. Patients who excreted amounts of androsterone, etiocholanolone, and dehydroepiandrosterone below the median value had a greater tendency to develop tumors again. Obtaining an androgen/corticosteroid ratio proved the best prediction method. The probability of tumor recurrence decreases with a decrease in androsterone/a—cortolone ratio. MacIndoe and Woods (64) noted the presence of the steroid-metabolizing of the enzymes: 5a—reductase, 3a— hydroxy—steroid oxidoreductase, and 17b-hydroxysteroid oxidoreductase in the MCF-7 human breast cancer cells. Thus, androstenedione may be converted to testosterone, dihydrotestosterone, and androstanediol within the tumor cell. The presence of the aromatase enzyme‘ in the hormone-responsive MCF-7 cell line permits the formation 98 of estrogens (estradiol and estrone) from androgenic precursors (androstene-dione and testosterone). The role of these intracellularly produced steroids has yet to be ascertained. The enzymes may be involved with the hormonal regulation of the breast cancer cells. 6. Endocrine effects of Tamoxifen therapy Researchers have shown that antiestrogens trigger changes in the levels of circulating steroid (65) and peptide hormones (66,67,68). These alterations in the hormonal milieu possibly contribute to an indirect beneficial antitumor action of the drug; however, modulation of the endocrine system by Tamoxifen may also explicate some of the adverse effects of the drug. a. Normal, premenopausal women Groom and Griffiths examined the effect of daily Tamoxifen administration (20 mg/day) on the secretion of luteinizing hormone (LH) follicle-stimulating hormone (FSH), prolactin, estradiol and progesterone in the blood plasma of normal, premenopausal women. The treatment lasted for either five or ten days during the follicular phase of their menstrual cycle. Administration of Tamoxifen did not change the length of the menstrual cycle, or the amount of LH, FSH, and progesterone secretion. Prolactin levels were suppressed at midcycle but were otherwise normal through the rest of the cycle. m (1 PT" Hm thr. 99 There was a two— to eight-fold increase in plasma estradiol concentration during Tamoxifen treatment. Even though the antiestrogen's major mode of action involves competitive inhibition of estrogen binding at the breast cancer cell, the drug appears to have had a side effect of directly stimulating the ovaries of premenopausal women to produce estradiol. These findings were confirmed by Sherman et al. (69) They dosed normal, premenopausal women with Tamoxifen for two complete menstrual cycles. Estradiol levels had two maxima, at midcycle and during the luteal phase, and was present at twice the normal concentration. Average progesterone levels in this experiment (288.1:76.2 ng/ml) were found to be greatly increased over nontreatment cycle values (123.0126.9 ng/ml). The treated patients had normal but augmented follicular maturation resulting in the exaggerated secretion of estradiol and progesterone. This phenomenon may be the result of enhanced gonadotropin stimulation of a single maturing ovarian follicle or simultaneous maturation of multiple follicles. Tamoxifen may be exerting antiestrogenic influence on the hypothalamus and pituitary gland since there is a lack of gonadotropin suppression even in the presence of high levels of estrogen. The supranormal concentrations of estradiol and progesterone have an antiprogestational effect, leading to retarded endometrium development. 100 Tamoxifen was originally designed as a possible birth control drug. In another study, Tamoxifen administration to healthy female volunteers resulted in increased serum levels of estradiol, progesterone, and testosterone. Kokko et al. (70) speculated these increases were due to enhanced ovarian activity and superovulation. The level of 17b— hydroxy-steroid dehydrogenase activity did not change during Tamoxifen treatment, even though this enzyme's activity is promoted by increasing progesterone concentration. This enzyme catalyzes the reaction which converts testosterone to androst—4—ene—3,17—dione. Tamoxifen has been used to stimulate ovulation in anovular women. Tajima and Fukushima (71) reported daily hormone profiles for five patients with anovulatory cycles who were treated with Tamoxifen. It appears that the drug blocked estradiol receptors in the hypothalamus region, resulting in gradual increase in the secretion of follicular—stimulating hormone and luteinizing hormone. Unlike Kokko et al., Tajima and Fukushima believed that the initial site of action of Tamoxifen is on the hypothalamic—pituitary axis rather than the ovary. Since Tamoxifen treatment led to an increase in serum estradiol— level, reaching a maximum preovalutory level and then triggering a luteinizing hormone surge, the authors do 101 mention Tamoxifen's action may involve direct ovarian stimulation. b. Postmenopausal women with breast cancer Due to the diminshed ovarian activity in postmenopausal women, Tamoxifen treatment does not lead to increased ovarian secretion of estradiol and progesterone. However, changes in gonadotropin levels have been noted. Golder et al. (72) measured plasma concentrations of prolactin, follicle—stimulating hormone (FSH), luteinizing hormone (LH), and 17b—estradiol in patients with advanced breast cancer being treated with Tamoxifen. The patients received 20mg of the drug twice daily for a minimum of one month. The levels of 17b—estradiol and prolactin in the plasma did not significantly change during antiestrogen therapy. Plasma FSH levels were suppressed in the patients who responded to Tamoxifen treatment indicating some action on the hypothalamic—pituitary axis. After prolonged therapy (for over one month) the FSH levels returned to normal. After one week of treatment, the plasma LH concentrations were suppressed and remained low for the entire three month study period. Willis et al. (73) also found that by the second week of Tamoxifen treatment the levels of FSH and LH were significantly suppressed in 45 postmenopausal women with recurrent breast cancer. The authors mention that in -w PT. ef 39. Cl wh ef 102 premenopausal women Tamoxifen blocks the negative feedback effect of estradiol and stimulates continued gonadotropin secretion. In postmenopausal women, however, the circulating level of estradiol is low and the receptors which control gonadotropin secretion do not have many sites where the drug can bind and exert its antiestrogenic effects. Willis et al. also measured the basal levels of 17b— estradiol and androgens (testosterone and androstenediol) in the breast cancer patients and at 2, 6, and 12 weeks of Tamoxifen therapy. The levels of androgens remained within the normal range for postmenopausal patients. Patients who had remissions during Tamoxifen treatment displayed no significant change in l7b—estradiol levels, but the nonresponding patients had increasing estrogen levels up to 80.1 pg/ml during the twelfth week of treatment. The basal level was 52.6 pg/ml. One estrogen—like effect of Tamoxifen treatment is the stimulation of cortisol binding globulin (CBC) and sex hormone binding globulin (SHBG) production by the liver. Sakai et al. (74) calculated a mean rise in cortisol binding capacity for Tamoxifen treated breast cancer patients from 18.6 ug cortisol /100 ml plasma to 29.8 ug cortisol/100 ml plasma. The capacity for sex hormone binding globulin to bind dihydrotestosterone rises by 0.79 it 103 ug/lOO ml plasma. Tamoxifen may be stimulating CBC and SHBG production by binding to hepatic estrogen receptors. This effect diminishes in older patients because of the decreasing number of hepatic receptors. Levin et al. (75) studied the effect of Tamoxifen on the production, plasma concentration, and metabolism of cortisol in postmenopausal women with advanced breast cancer. After six weeks of Tamoxifen treatment the plasma cortisol concentration increased by 26% due to a significant increase in the level of CBG. They also noted that the cortisol production rate decreased by 35% and the resulting decrease in the fraction of the non—protein decrease in the fraction of the non—protein bound cortisol led to a decrease in the metabolic clearance rate (MCR) of plasma cortisol. These effects are similar to those observed during estrogen administration. During the six weeks of Tamoxifen treatment, Levin et al. found that the total excretion of cortisol (tetrahydrocortisone, tetrahydrocortisol, and allotetrahydrocortisol) decreased by 13%. Reduction of the llb-hydroxyl group increases and the ratio of excreted tetrahydrocortisol to excreted tetrahydrocortisone decreases by 28% for reasons that are not apparent. Interestingly, the effects of Tamoxifen on cortisol metabolism in receptor-positive and receptor— negative patients were not significantly different. This implies that the changes in cortisol metabolism were 104 produced systemically by the drug, rather than specifically at the tumor site. Wilking (76) examined the serum levels of several hormones (estrogens and adrenocortical steroids) in breast cancer patients and age-matched controls. The only significant difference observed was an increased dehydroepiandrosterone level for the breast cancer patients. Total estrone was significantly elevated in breast cancer patients after one week of Tamoxifen therapy, possibly due to changes in liver function induced by the drug, or the weak estrogenic effect of the Tamoxifen-receptor complex. A prompt increase in cortisol levels was also noted, probably due to an increase in the transcortin levels rather than direct stimulation of adrenocortical steroid biosynthesis. The knowledge that Tamoxifen can effect the endocrine system has been used to a therapeutic advantage, specifically in combination drug therapy for metastatic breast cancer. Megestrol acetate, an antitumor agent, induces a suppression of serum gonadotropin levels, leading to a decrease in the level of SHBG and estradiol. Megestrol acetate also increases plasma prolactin concentrations. When Tamoxifen is added to a megestrol acetate treatment program there is a cancellation of the hyper-response of prolactin to TRH stimulation (77). The 105 estradiol and cortisol levels were unaffected. Plasma gonadotropin concentration decreased and SHBG levels increased. B. Tamoxifen - Studies with rats 1. DMBA-induced rat mammary carcinoma The 7,12—dimethylbenz(a)anthracene (DMBA)—induced rat mammary carcinoma is one of the most widely used animal models of endocrine—dependent human breast cancer. These chemically-induced tumors are dependent upon pituitary and ovarian hormones for growth. This animal model closely resembles human breast cancer in its biologic properties, and in its responsiveness to chemotherapy or endocrine therapy. The rat mammary carcinomas contain both estradiol and progesterone receptors, and these receptors exhibit the same sedimentation characteristics as the steroid receptors in human mammary tumors. Huggins et al. (78) studied the incidence of mammary cancer in rats after a single feeding of DMBA. They administered various dosages of the polynuclear hydrocarbon to the rats, ranging from 1 mg to 50 mg. When rats were given the optimal dosage of 20 mg, they survived the ingestion of the carcinogen, and develOped palpable tumors by 42 :t 10 days. This dosage level produced mammary cancers in all 40 rats given DMBA. The estrous 0:" H} ad Te in af‘ 106 cycles of these rats were not upset by DMBA, but there was a temporary cessation of body growth. Huggins et al. determined the benefits of inducing mammary tumors with DMBA to be its extreme simplicity (compared to multiple feedings or grafting of tumor tissue) and the specificity of the appearance of tumors (in the mamma). 2. Treatment of DMBA-induced rat mammary carcinoma with Tamoxifen Welsch et al. (79) investigated the effect of Tamoxifen on the genesis of mammary carcinomas in female rats treated DMBA. The authors found that Tamoxifen not only delayed the appearance of mammary tumors in rats treated with Tamoxifen prior to DMBA administration, but also reduced the final yield of these neoplasms. When rats were injected daily with Tamoxifen (50 ug/lOO g body weight) for 33 days prior to and then after DMBA administration, the incidence of mammary tumors was reduced by 49%. Another set of rats received daily injections of Tamoxifen for 66 days, starting 33 days after carinogen treatment. The incidence of mammary carcinomas for this group decreased by 65%. The effectiveness of Tamoxifen treatment versus ovariectomy in DMBA-induced rat mammary carcinoma was compared by both Jordan (80) and Fiebig and Schmhl (81). In Jordan's protocol, rats either a) received 5mg of tf TE 107 Tamoxifen on days 1 and 2, b) received 5mg of Tamoxifen on days 1, 2, 30 and 31, c) were ovariectomized and received only peanut oil injections on days 1 and 2, or d) received only peanut oil injections on days 1 and 2. DMBA was administered to all rats on day 1. The rats treated with Tamoxifen on days 1 and 2 had < 10% of the number of tumors present in the control rats four months after DMBA injection. A second dose of Tamoxifen on days 30 and 31 prevented the development of tumors in all the rats in that group for up to 120 days. No tumors were present in the ovariectomized rats at 120 days. The control group (6 rats) had a total of 14 tumors at 120 days. Only one control rat did not develop mammary tumors. Fiebig and Schmhl found ovariectomy to be more successful than Tamoxifen treatment in suppressing DMBA-induced mammary cancer in rats. Tamoxifen treatment yielded one partial remission and seven no changes out of 13 rats. Ovariectomy led to five complete and seven partial remissions out of 14 rats. 3. Pharmacology of Tamoxifen in the rat The action of Tamoxifen in the rat with DMBA—induced mammary carcinoma is very similar to its action in the human with breast cancer. Jordan (82) suggested that Tamoxifen antagonizes the action of estrogen at the tumor level by competitively blocking the estrogen receptor 108 site. He noticed a linear correlation between positive response to Tamoxifen therapy and the levels of estrogen receptors in tumor biopsies: those tumors with the highest levels of estrogen receptors responded most favorably. Later, Jordan et al. (83)expanded the list of possible mechanism to explain Tamoxifen's actions. In addition to estrogen receptor blockade, the drug causes a decrease in the level of circulating gonadotrOpins (e.g., LH), reduces estradiol synthesis at the ovary, stimulates progesterone receptor synthesis, and inhibits prolactin release from the pituitary gland. It appears that Tamoxifen can inhibit carcinogenesis, as the drug will reduce estrogen—stimulated elevations in circulating prolactin and inhibit DNA synthesis in rat uteri. There is a Tamoxifen—dose related delay in the appearance of tumors in the DMBA-treated rat. The drug probably inhibits the tumor cell's growth and replication, rather than destroying all the tumor cells. When Tamoxifen treatment is halted, and the drug has been cleared from the rat, the tumor cells begin to flourish in the renewed cyclic hormonal environment. Macnab et al. (84) presented an interesting description of the idea that the behavior of Tamoxifen as a partial estrogen agonist is predicted by classical receptor theory. The authors used the immature rat uterus as a model to evaluate the interactions of Tamoxifen and 109 estrogen. They found that the effect of Tamoxifen is dependent on the level of estrogen in the system. At low levels of estradiol, the agonist effect of estrogen and antiestrogen are additive. At high levels of estradiol Tamoxifen acts as an estrogen antagonist. When Tamoxifen is considered to be a partial agonist, its behavior corresponds to classical receptor theory, where the partial agonist reduces the effect of the full agonist (estradiol), given relatively similar concentrations of partial agonist and full agonist acting on the same receptor. The authors pointed out that clinical applications of this classical receptor model may be difficult since measurement of total estrogen levels may not reflect the actual concentrations at specific receptor sites in different tissues. Another paradox is the fact that in older patients, circulating estrogen levels are lower, so Tamoxifen should act as an estrogen agonist. Indeed, the initial effect of Tamoxifen treatment in postmenopausal patient is estrogenic. However, as the antiestrogen-receptor complex internalizes in the nucleus of the tumor cell, this false messenger does not stimulate cell replication. 4. Tamoxifen metabolism in the rat The metabolism of Tamoxifen in the rat is very similar to that in humans. Fromson et al. (85) measured the level en 110 of Tamoxifen and metabolites in the urine, bile, and feces in the rat. Using [14C]Tamoxifen, they found that the drug was well absorbed, extensively metabolized, and slowly eliminated. The level of radioactivity in the urine was too low to permit identification of possible metabolites. A substantial amount of the radioactivity (between 22 and 53%) was found in the rat bile. The authors found that a large percentage of the metabolites in the bile were subsequently reabsorbed and remained in enterohepatic circulation until eliminated in the feces. In one metabolic pathway, Tamoxifen undergoes aromatic hydroxylation to form monohydroxytamoxifen (metabolite B). This compound is then hydroxylated at an adjacent position to form dihydroxytamoxifen (metabolite D). The next step involves partial methylation to form metabolite C. An alternative pathway leads to the deamination and aromatic hydroxylation of Tamoxifen followed by alcohol formation (metabolite F). All of these hydroxylated metabolites are conjugated as glucuronides and excreted in the feces, or are hydrolysed in the intestinal tract and undergo enterohepatic recirculation. 111 Jordan et al. (86) determined that monohydroxytamoxifen metabolite to have an even more potent antiestrogenic effect in the immature rat than Tamoxifen. The estrogenic and antiestrogenic activity of the drug and metabolite was determined by examining the change in immature rat uterine weight. Dosage with estrogen produces a full uterotropic response. Dosage with Tamoxifen produced a partial estrogenic and partial uterotropic effect. When Tamoxifen or monohydroxytamoxifen was given orally to rats after they had received an estrogen injection, an antiestrogenic effect was observed. Both of the antiestrogens caused a decrease in uterine wet weight gain when compared to estradiol—treated controls. Also, at a dosage level of 5 ug/day, monohydroxytamoxifen had a greater antiestrogenic effect than Tamoxifen. Jordan et al. concluded that some of Tamoxifen's antitumor activity is actually caused by the more active monohydroxy metabolite. The monohydroxylated metabolite of Tamoxifen has a high affinity for the estrogen receptor, and may help maintain the effect of Tamoxifen therapy. However, addition of another hydroxyl group ortho to the hydroxyl group in monohydroxytamoxifen causes a reduction in estrogen receptor affinity. Jordan et al. found this metabolite D to be less active as an antiestrogen than Tamoxifen. This metabolite, unlike Tamoxifen and 112 monohydroxytamoxifen was unable to stimulate an uterotrOpic response in the immature rat. Jordan and Allen (87) reported that even though monohydroxytamoxifen has a greater antiestrogenic potency than Tamoxifen, the monohydroxylated metabolite is a less potent antitumor agent. In this study, rats received DMBA by gavage, and four weeks later received treatments of Tamoxifen or monohydroxytamoxifen for various experimental periods. Unlike Tamoxifen, monohydroxytamoxifen did not produce a dose—related delay in the appearance of tumors in DMBA—treated rats, e.g. those rats treated with daily dosage of 0 - 2 ug monohydroxytamoxifen remained tumor free longer than the rats which received 50 ug of the drug. Monohydroxytamoxifen was more effective than Tamoxifen when given to rats between 60 and 90 days postcarcinogen treatment. Jordan and Allen speculated that this effect was due to the increased hydrophilicity of the monohydroxylated metabolite, and an increased affinity for the estrogen receptor. Tamoxifen appears to be the better antitumor agent, though, since it has a more prolonged biological activity, apparently a more important factor than antiestrogenic potency. Daniel et al. (88) attempted to correlate the tissue and plasma levels of Tamoxifen and two metabolites: monohydroxytamoxifen and N-desmethyltamoxifen, with the 113 clinical response in DMBA-induced rat mammary tumors. Drug and metabolite concentrations were measured using gas chromatography-high resolution mass spectrometry. The trimethylsilyl ether derivative of monohydroxytamoxifen was prepared as well as the heptafluorobutyrate of N- desmethyltamoxifen. The authors found that there is no correlation between the incidence of tumor regression and the plasma concentrations of Tamoxifen and N- desmethyltamoxifen. Rather, there was a correlation noted between the concentration of Tamoxifen and N— desmethyltamoxifen in cytosol fractions and the reduction of estrogen-receptor positive tumors. The tumors with the larger concentrations of estrogen receptors were the most likely to regress. 5. Endocrine effects of Tamoxifen therapy The hormonal changes induced by Tamoxifen therapy vary markedly from species to species. In the rat, as in the human, the drug is both a partial agonist and an estrogen antagonist. The drug is a pure agonist in both mice and guinea pigs and is an antagonist in frogs, goldfish, and chicks. Even the partial agonistic effect of Tamoxifen in the rat uterus is qualitatively different than the effect of estradiol benzoate, a pure agonist (89). The endometrial hypertrophy which results from Tamoxifen 114 treatment yields mitotic counts much lower than those observed with estrogen treatment. Patterson summarized some of the endocrine effects in the rat observed during Tamoxifen treatment. The drug will inhibit implantation of a fertilized ovum in the rat by both inhibiting the synthesis of estradiol by the ovary, and competitive inhibition of estradiol binding to its receptor. In the rat genital tract, the partial estrogenic behavior of Tamoxifen will produce partial cornification of vaginal smears at a rather high dosage levels (41 mg/kg per day), compared with 100% cornification stimulated by 4 ug/kg per day of dienoestrol. Patterson concluded that the endocrine effects are caused by Tamoxifen's actions directly at the organ site as well as indirectly on the hypothalamus and pituitary gland. The antitumor effect of Tamoxifen is probably not mediated through changes in plasma hormone levels. Nicholson and Golder (90) found little change in the levels of estradiol and prolactin in the plasma of Tamoxifen-treated rats bearing DMBA-induced mammary carcinomas. However, Jordan and Koerner (91) found that there was a slight difference between peripheral plasma levels of estradiol in the diestrus phase of untreated tumor bearing rats (105.2 :1 17.0 pg/ml) and the level in 115 the diestrus phase of Tamoxifen—treated (50 ug/day) tumor bearing rats (74.7 t 9.1 pg/ml). Estrogen—stimulated plasma prolactin levels were also significantly suppressed in the Tamoxifen-treated group. The neuroendocrine effects of a single dose of Tamoxifen given to an ovariectomized rat are complex. Bowman et al. (92) noted that the plasma prolactin level increased in a dose—related manner. The excretion of LH increased initially following Tamoxifen treatment but then dropped below the concentration found in control animals when measured 16 days post treatment. The LH secretion is probably increased when Tamoxifen occupies the estrogen receptors of the hypothalamus. FSH secretion was not significantly affected by the drug. Bowman noted that a single dose of Tamoxifen stimulated estrogen agonist and antagonist activity for up to 16 days. Welsh et al. (93) found that Tamoxifen inhibited the FSH-stimulated biosynthesis of progestins in cultured rat granulosa cells. The drug may inhibit 3b—hydroxysteroid dehydrogenase and 20a-hydroxysteroid dehydrogenase, as the production of progesterone and 20a-hydroxy—pregn—4-en—3— one, partially by suppression of pregnenolone biosynthesis. The antiestrogen may also be inhibiting the cholesterol side chain cleavage or other enzymes involved with cholesterol synthesis. 116 In contrast, Tamoxifen enhances FSH—stimulated aromatase activity. Welsh et al. added androst—4—ene- 3,17—dione to the rat granulosa cell culture and measured estrogen formation after eight hours. At 10 ng/ml FSH, Tamoxifen augmented FSH—stimulated estrogen production by 6.4-fold. A concentration of 10'7 molar Tamoxifen was required to enhance estrogen production. The drug may act independently of the hypothalamus and pituitary gland to stimulate ovarian estrogen production, and it appears that Tamoxifen has a direct estrogenic effect on aromatase induction. This unusual behavior of antiestrogens augmenting ovarian estrogen synthesis has important implications in the treatment of estrogen—dependent breast cancer . When Tamoxifen is administered to newborn female rats, abnormal reproductive development occurred. Chamness et al. (94) injected newborn Sprague-Dawley rats on days 1, 3, and 5 with 5 ug Tamoxifen. The drug caused the rats to stay in a constant state of diestrus. Many abnormalities were apparent in all drug-treated rats when autOpsied at four months of age. The uteri and ovaries were atrophied. The oviducts had acute and chronic inflammation, with abscesses and squamous metaplasia. The authors did not speculate whether Tamoxifen caused these effects directly or indirectly through the hypothalamus and pituitary gland. Their ultimate message was that Tamoxifen should d S 117 be used with caution in patients where the possibility of pregnancy exists. Tamoxifen has been shown to cause adrenocortical cell necrosis in female rats receiving large oral doses of the drug (100 — 130 mg/kg). Lullmann and Lullmann—Rauch (95) found degeneration of cells in the zona fasciculata and zona reticularis of rats after receiving the drug for a duration of between six and fourteen weeks. The authors speculated that this phenomenon was a specific endocrine effect of Tamoxifen therapy. The necrotic and degenerating cells were being replaced with abnormal vacuolated cells. The drug may be exhibiting a direct effect on adrenocortical steroid metabolism. The main topic of Lullmann and Lullmann-Rauch's article was that of the generalized lipidosis caused by Tamoxifen. The authors examined tissues of the liver, lung, lymph, adrenal gland, pituitary gland, retina, and autonomic ganglia and found lipidosis-like alterations in all of them. The drug has an amphiphillic cationic structure which can interfere with catabolism of polar lipids and leading to increased intralysosomal storage of these lipids. The cornea of these Tamoxifen—treated rats displayed lipidotic changes reminiscent of the retinopathy and corneal opacities observed in women receiving high doses of Tamoxifen (240 - 320 ng/day) (96). IV. Steroid Metabolism and Excretion in the Rat A. Introduction This section contains a review of the literature concerning the analysis of the products of the anabolic and catabolic processes for steroid hormones in the rat. A discussion of the steroid metabolites present in urine, feces, and bile will be included, as well as review of the secretion of steroids by ovarian, adrenal, and liver tissue. Although the rat is an excellent experimental model for medical research, there are many species differences between this mammal and primates. Concerning steroid metabolism, although both man and rat reduce the 4-ene—3- one A ring configuration, the rat produces 3b—hydroxy steroid metabolites and man usually produces 3a-hydroxy steroid metabolites. While the major corticosteroid in man is cortisol, the rat secretes corticosterone predominantly. The rat also has gut microbes which form microbial steroid metabolites during enterohepatic circulation . The sexual differences in rat steroid metabolism will also be examined. B. Urinary and fecal excretion of steroids Gustafsson (97) analyzed the excretion of monosulphurylated steroids in the urine of both germfree 118 119 and conventional rats. He identified the following steroids in the urine of germfree female rats: 5a— androstane—3a,l7b-diol, 3a,llb—dihydroxy-Sa—androstan-l7- one, 3a,7a—dihydroxy—Sa-androstan—l7-one, 5a-androstane- 3a,7a,17b-triol, 3a,16a-dihydroxy—Sa—pregnan-ZO-one, 3a,15a-dihydroxy—5a—pregnan—20—one, 3a(and 3b),l7a- dihydroxy—5a-pregnan-20-one, 5a—pregnane—3a,16a,20a-triol, 11b,21—dihydroxy-5a—pregnane—3,20—dione, 3a,11b,21— trihydroxy—Sa—pregnan-ZO—one, 3a,15a,21—trihydroxy-5a— pregnane—11,20—dione, and 3a,11b,15a,21-tetrahydroxy—5a— pregnan—ZO—one. The urine samples obtained from conventional female rats contained all the steroids listed above, with the exception of 3b,17a—dihydroxy—5a-pregnan— 20—one. These normal rats also excrete 33,19—dihydroxy- 5a—androstan—17—one, 3a—hydroxy—5a,l7a(and l7b)—pregnan— 20—one, 3a—hydroxy-5a—pregn-16—en—20—one, 3a,19—dihydroxy— 53,17a—pregnan-20—one, 3,l9—dihydroxy—5a,17a—pregnan—20- one, 3b,l5a—dihydroxy—Sa—pregnan—ZO-one, and several 15— hydroxylated C2104 steroids. By comparing the different steroid metabolites obtained from the urine of conventional and germfree female rats, Gustafsson surmised that 3a-hydroxy—5a,l7a-pregnan-20—one, 3a—hydroxy-5a— pregnan—ZO—one, and 3a-hydroxy—5a—pregn-16—en—20-one are intestinal microbial metabolites of 3a,16a-dihydroxy—Sa— pregnan—ZO—one produced during enterohepatic circulation. 120 In a striking display of contrasting steroid metabolism in male and female rats, no steroids were found in the monosulfate fraction of urine of either conventional or germfree male rats. Metabolites of pregnenolone and corticosterone are excreted unconjugated from the males, whereas female rats excrete both unconjugated and monosulphurylated metabolites. Neither male nor female rat urine contained diconjugated steroids. Eriksson and Gustafsson (98) studied the distribution and excretion of [4—14C] pregnenolone and [1,2—3H] corticosterone in the urine and feces in germfree and conventional male and female rats. They found that male rats excreted more unconjugated and polar steroids than the female rats. The conventional rats excreted more free steroids than the germfree rats, due to the intestinal sulphohydrolase activity. The ratio of free to conjugated steroids increased from about 0.33 in the germfree female rats to 1.0 in the conventional rats. Also, they noted an increased presence of steroid metabolites in the urine of conventional rats, compared to those in the fecal excretion. The urinary/fecal ratio of excreted corticosterone metabolites was 0.20 for germfree and 0.89 for conventional female rats. Eriksson (99) further studied the metabolism of [4— HC] pregnenolone and [4—14C] corticosterone in germfree and conventional female rats. Pregnenolone metabolites at at Dr eXg 121 produced included 3a—hydroxy-5a,l7a(and 17b) pregnan—ZO— one, 3b—hydroxy—5a,17a-pregnan-20—one, 3b,15b(and 15a) dihydroxy—5a—pregnan-20—one, and 3a,16a-dihydroxy—Sa— pregnan—ZO—one in the free fraction of feces from conventional rats. The adrenal gland and ovaries were involved with the production of pregnenolone metabolites observed, since the liver does not have a 3b-hydroxy- delta—S—steroid oxido—reductase for steroid hormones. Corticosterone metabolites included 3a(and 3b),20b- dihydroxy—Sa—pregnan—ll—one, produced by the action of llb—hydroxy steroid oxido-reductase on the mono— and disulfates of 3a,11b,21—trihydroxy—5a—pregnan-20-one. Most of the steroids identified by Eriksson had previously been found in the fecal monosulfate fraction of conventional female rats. These rats excrete more unconjugated metabolites than the germfree animals, due to the sulphohydrolase activity of the intestinal microflora. Other non—labelled or incompletely labelled fecal metabolites in conventional female rats included 3a,15a(and 15b)—dihydroxy-Sa—pregnan—ZO-one, 3b—hydroxy—5— androsten—l7-one, 5a—androstane—3b,17b—diol, 5a— androstane—3a,7a,17b—triol, 3a(and 3b),l9—dihydroxy—Sa— pregnan—ZO—one, and 5a—pregnane—3a(and 3b),20a—diol. In an earlier experiment, Eriksson et al. (100) examined the excretion of free pregnenolone and 122 pregnanediol isomers in the feces of germfree and conventional rats. They noticed that steroids with a 173 side chain configuration may be excreted with a 20-keto group. Those steroids are not reduced by 20b(or 20a)— hydroxy steroid oxido—reductase, which converts pregnanolones to pregnanediols. The feces of conventional rats contained 3a(and 3b)—hydroxy—5a-pregnan—20—one, and 5a—pregnane-3b(and 3a),20b-diol. Also, they noted that l6a—hydroxylated steroids can be metabolized to pregnanolone by the 16a—dehydroxylating and sulphohydrolase activity of the intestinal microflora. None of the aforementioned steroids were found in the free or monosulfate fecal fractions from germfree rats. Eriksson et al. proposed a scheme to describe the formation of 17a—pregnanolones in conventional rats from l6a—hydroxylated precursors through the action of bacterial microorganisms. 3a,16a-Dihydroxy—5a,l7b— pregnan—ZO—one is converted to 3a—hydroxy-Sa—pregn—l6-en— 20-one, and then to 3a-hydroxy—Sa,l7a—pregnan-20-one. The germfree rats did not excrete monohydroxy—monoketo pregnane derivatives, due to the absence of bacterial 16a— dehydroxylase activity. They excrete 3a,16a—dihydroxy—5a— pregnan-ZO—one. Bjorkhem et al. (101) analyzed the excretion of [4- HC]-pregnenolone metabolites in the urine and feces of Jlflldqlllf 123 developing male and female rats. They were interested in determining at what age differentiation of steroid metabolic processes occurs between male and female rats. They found that during the first, second, and third week of life no sexual differences were noted in the distribution of radioactive pregnenolone metabolites in the free steroid, monosulfate, and disulfate fractions. After that time, the female rats excreted an increasing amount of steroid sulfates. By the sixth or seventh week the female rats displayed a metabolite pattern identical to that in the normal adult. During the fourth through sixth week, the female rats excreted two isomers of androstane—triol—one, two isomers of pregnane—triol—dione, and one isomer of pregnane— tetrol-one as free steroids in the urine. Trace amounts of 3a,7a—dihydroxy-Sa—androstan—l7-one, 3a,19-dihydroxy— 5a-17-one, 5a—androstane—3a,7a,l7b—triol, 33,19—dihydroxy— 5a—pregnan-20—one, 3a,15a-dihydroxy—5a-pregnan—20—one, 3a,l6a—dihydroxy—Sa—pregnan-ZO-one, 3a,15a,20b-trihydroxy— 53,14b—pregnan—ll—one, and 3b,lSa,20b-trihydroxy—5a,14b— pregnan-ll—one were found in the monosulfate fraction of urine obtained in the fourth week from the female rats. By the sixth week they were also excreting 5a-androstane- 3a,153,17b—triol, 3a,17a—dihydroxy—Sa—pregnan-ZO—one, 5a— pregnane-3a,16a,20a—triol, and 3a,11b,15a—trihydroxy- 5a,14b—pregnan-20—one. Although present in the urine of 124 female rats four through seven weeks of age, 3a,16a- dihydroxy—5a-pregnan—20—one only appeared in the feces of female rats during the fourth week. 33—hydroxy—5a,17a- pregnan—ZO-one only appeared in the feces from rats six to seven weeks old. The only metabolites noted in the free fraction of female rat feces during the fourth week were 3b-hydroxy—5a,17a—pregnan—20—one, 3b,20b-dihydroxy—5a- pregnan—ll-one, and 3b,15a,20b-trihydroxy—Sa,14b-pregnan- 11—one. By seven weeks of age, the steroid metabolite pattern of the free steroid fraction of the female rat feces was quite similar to the adult pattern. Bjorkhem et al. also found no direct correlation between an increase in gonadal 3b—hydroxy—delta—S—steroid oxido—reductase activity and changes in steroid metabolism between male and female rats. This enzyme allows for the conversion of pregnenolone into progesterone. They found that the 3b—hydroxy-delta-5—steroid oxido—reductase activity in the adrenal gland increases between five and ten times between ages one to eight weeks, but there is not a significant increase in testicular enzyme activity during that period. In another experiment, Bjorkhem et al. (102) administered Za—cyano-4,4,17a-trimethyl-S—androsten—l7b— ol—3—one, a 3b-hydroxy—delta-5-steroid oxido—reductase inhibitor, to male and female, conventional and germfree 125 rats. They noted a 90% loss in ability of adrenal and gonadal tissues to oxidize 3b—hydroxy—delta—S—steroids. Most of the excreted metabolites were 3b—hydroxy-delta-5— steroids, analogues of the normally occurring saturated steroids. The corticosterone metabolites indentified in the urine and feces of cyanoketone treated germfree rats were 3a,11b,153,21—tetrahydroxy-5a—pregnan—20-one and 3a,15a,21—trihydroxy—Sa—pregnane—ll,20-dione, which are also found in the urine of untreated conventional and germfree female rats. Two unusual saturated 5b-C2102 steroids were identified: 3a—hydroxy—5b,l7a—pregnan-20— one and 3a—hydroxy—5b,17b—pregnan-20—one. Saturated metabolites of 3—oxo—delta—4—steroids were secreted by the gonads and adrenal indicating that the 3b-hydroxy—delta—5— steroid oxido—reductase was not totally inhibited by the cyanoketone. Male rats excrete metabolites more polar than those found in female rat urine and feces. Eriksson (103) identified the metabolites of [4-14C] pregnenolone and [4— 1%fl corticosterone in germfree and conventional male rats. Both in the urine and feces 5a-pregnane— 3a,11b,l6a,20b,21—pentol, 5a—pregnane—3a,llb,l6b,20a,21- pentol, 5a-pregnane-3a,11b,16b,20b,21—pentol, 5a—pregnane— 3b,1lb,16a,20a,21—pentol, 5a-pregnane-3b,11b,16a,20b,21— pentol, 5a—pregnane—3b,11b,16b,20a,21—pentol, 5a-pregnane— 3b,11b,l6b,20b,21—pentol, and Sa-pregnane— 126 3a,11b,16a,203,21-pentol constituted the major corticosterone metabolites obtained from the conventional males. They also excreted 3a(and 3b),20b—dihydroxy—5a— pregnan-ll-one and 5a—pregnane-3a(and 3b),1lb,20b,21- tetrol in the feces. This latter steroid was also the major metabolite found in the monosulfate fraction of germfree male rat feces, as well as 3a,7a—dihydroxy—Sa- androstan-17—one, 3a,11b—dihydroxy—5a—androstan—17—one, Ba,16a-dihydroxy—Sa—pregnan—ZO-one, 5a-androstane— 3a,16a,17b—triol, 5a—pregnane—3a,11b,20b,21—tetrol, and 5a—pregnane—3b,11b,20b,2l-tetrol. The disulfate fraction contained the aforementioned pregnane-tetrols, as well as 3a(and 3b), 11b,21—trihydroxy-Sa-pregnane-ZO—one. This study pointed out a major difference between steroid metabolism in male and female rats. All the radioactive metabolites of [4— 14C] pregnenolone and [4— 14C] corticosterone excreted by both the germfree and conventional rats are present in the free steroid fraction. The majority of the female rat steroid metabolites are excreted as monosulfates. The increased steroid conjugation may be attributed to increased sulphurylation activity noted in female rat liver preparations by Wengle (104). Another interesting sexual difference is that only female rats excrete sulfates of 15—hydroxylated C2105 steroids (such as 3,11,15,21— tetrahydroxy-pregnan—ZO—ones), and only male rats excrete 127 pregnane-3,11,16,20,21—pentols in urine and feces. One possible reason is that steroids containing a 20-oxo group (as in female rats) can undergo 21—or-l6—dehydroxylation, but steroids containing a 20,21-dihydroxy or 16a,20- dihydroxy structure (as in male rats) are resistant to this type of hydrolysis by gut microorganisms. Intestinal microorganisms have an important effect on the types of steroid metabolites present in the feces of rats. Eriksson and Gustafsson (105) examined the mono— and disulphurylated steroid metabolites in the feces of conventional and germfree rats. They found that the feces of germfree and conventional female rats contained only one common steroid, 5a—pregnane—3a,16a,20a—triol, which indicated the extensive metabolism conducted by the intestinal microflora. Normally, bacterial reduction dehydroxylates 16a—hydroxylated. C21 steroids, producing delta-16 intermediates, and 17a—pregnane metabolites such as 3a,15a(and 15b)-dihydroxy—5a,l7a—pregnan—20-one, and 2a,2a-dihydroxy—5a,17-pregnan—20-one. 3a(and/or 3b),l1b,21-Trihydroxy—Sa-pregnan-ZO-one, which appears in the disulfate fraction of the germfree female rat feces, may undergo 21—dehydroxylation by intestinal bacteria in the conventional animal to produce the corticosterone metabolite, 3b,11a—dihydroxy—5a-pregnan—20—one. Eriksson and Gustafsson also saw 3b,17a-dihydroxy—Sa—pregnan—ZO— one, Sa—pregnane—3a,l6a,20a—triol, 2a,3b,16a-trihydroxy— 20- C01 di. tr WE fE 0): 128 5a—pregnan—20—one, and 3a,15b,l6a—trihydroxy—5a—pregnan— 20—one in the germfree rats' monosulfate fraction. The conventional rat produced estriol, 3a,15a(and 15b)— dihydroxy-Sa,17b—pregnan-20—one, 5a—pregnane—3a,l6a,20— triol, 3a,17a—dihydroxy-Sa-pregnan-ZO—one. Microbial metabolites of 15a,21—hydroxylated steroids were found in the urine and feces of conventional male and female rats by Eriksson et al. (106) 33,15a—dihydroxy- 15b,l4b—pregnane—11,20—dione, 3b,11b,15a—trihydroxy— 5a,14b-pregnan—20—one, 3a,15a,20b—trihydroxy—5b,14b— pregnan—ll-one, 3b,15a,20b—trihydroxy—5a,14b-pregnan—11— one were present in both the urine and feces of female rats. In addition, 3b,153-dihydroxy—5a,14b—pregnane- 11,20-dione, 3a,11b,lSa—trihydroxy—le,l4b-pregnan—20-one, 5b,14b-pregnane—3a,11b,15a,20b-tetrol, and 5a,14b- pregnane-3b,11b,15a,20b—tetrol appeared in the feces of female rats. Male rat urine contained 3a,11b,15a- trihydroxy-Sb,14b—pregnan-20—one, 3b,11b,15a—trihydroxy- 53,14b—pregnan—20—one, and 3a,15a,20b—trihydroxy—5b,14b- pregnan—ll—one. Male rat feces contained those steroid metabolites as well as 33,15a—dihydroxy-5b,14b—pregnane— 11,20-dione, 3b,15a,20b—trihydroxy—5a,14b-pregnan—1l-one, and 5b,14b—pregnane-3a,11b,15a,20b-tetrol. Germfree female rats excrete C2105 steroids with oxygen substituents at the C—3, C-ll, C—15, C—20 and C—21 129 positions. These metabolites also appear in the urine of conventional female rats, but not in the feces. en 04 metabolites were present in the conventional female rats, where the intestinal flora remove the 21-hydroxy group through hydrolysis of the 21—sulfates from the 9fl,05 steroids. The On 04 are reabsorbed and partially resulphurylated in the liver, and the metabolites are excreted into the bile or urine. The feces of germfree rats contained 33,16a-dihydroxy- 5a—pregnan—20—one as the predominant 9fl.03 steroid. Gustafsson et al. (107) also noted the presence of the sulfates of 3b,l6a-dihydroxy-5a-pregnan—20—one, and 3b,16a—dihydroxy—pregn—S—en-ZO—one in the feces. They mentioned that the liver of human infants also produces l6a-hydroxy steroid metabolites, but the capacity for 16a— hydroxylation decreases with age. Gustafsson and Sjovall (108) examined the C19 steroids present in the feces of germfree rats. The two major metabolites were 3a,11b—dihydroxy-5a—androstan-l7—one, and 3a,7a—dihydroxy-5a—androstan—l7—one. The former steroid also appears in rat urine 5a-androstane—3a,7a,l7b—triol, 3a,19—dihydroxy—Sa-androstan-l7-one, 5a—androstane- 3a,15a,17b—triol, 3a,l5a-dihydroxy—Sa—androstan—17-one, and Sa-androstane—3a,17b—diol were also excreted. This last steroid was the only Cngg metabolite observed, 130 whereas conventional rats excrete dehydroepiandrosterone (3b-hydroxy—androst-S-en—l7—one). Gustafsson and Sjovall noted the predominance of a Sa configuration for these rat steroid metabolites (rather than 5b). Also, sulfate conjugation is the major form of eliminated steroids, and are mainly eliminated in the feces. Steroid glucuronides are mainly excreted in the urine. In a subsequent article, the authors examined the 15a- and 21-hydroxylated C21 steroids in germfree rat feces (109). The major metabolites were 3a,11b,15a,21— tetrahydroxy—Sa-pregnan—ZO—one, and 3a,15a,21-trihydroxy- Sa—pregnane—ll,20—dione. They also found the corticosterone metabolites 11b,21—dihydroxy—5a—pregnane— 3,20—dione, 3a,11b,2l—trihydroxy—Sa—pregnan—ZO—one, and 3a,21—dihydroxy-5a—pregnane-11,20-dione. 15a- Hydroxylation is an important step in the production of rat fecal steroid metabolites. In addition to the 15— hydroxy compounds mentioned above 3a,15a-dihydroxy—5a- pregnan—ZO-one was also observed. 15-Hydroxy steroids may also occur in conventional rats, as this position may escape metabolism by gut microorganisms. 3a,15a— Dihydroxy—5a—pregnan—20-one is a major fecal steroid from the conventional rat. Gustafsson (110) continued the study of steroid metabolite excretion from conventional rats and identified 131 nine C19 and six C21 fecal metabolites, including: 3b— hydroxy-androst-S-en—l7—one, 3a,11b—dihydroxy—Sa— androstan-17—one, 3a,7a—dihydroxy—Sa—androstan—l7—one, 5a- androstane— 3a,7a,17b-triol, 5a—androstane—3a,15a,17b- triol, 3a(and3b),19—dihydroxy—Sa—androstan—l7—one, 5a- androstane—3a(and 3b),17b—diol, 3a(and 3b), 163—dihydroxy- 5a—pregnan—20—one, 3a(and 3b),15a—dihydroxy—Sa—pregnan—ZO- one, and 3a(and 3b),19—dihydroxy—5a—pregnan—20-one. 3a,15a-Dihydroxy-5a—pregnan—20-one was the principal C2103 steroid excreted in the feces by the conventional animals. Only trace amounts of 3a,16a-dihydroxy-5a-pregnan-20-one were found in the feces of conventional rats even though it is the major C2103 steroid excreted into the feces by germfree rats. The 16a—hydroxyl group is probably eliminated by microbial action. Intestinal microorganisms also eliminate 21—hydroxyl groups, as 3a,15a,21- trihydroxy-Sa—pregnane—l1,20-dione and 3a,11b,15a,21— tetrahydroxy—5a-pregnan-20-one do not appear in the feces of conventional rats. Also, the intestinal flora's dehydrogenase activity leads to an increased excretion of steroids with a 3b,5a configuration. Germfree and conventional rats excrete different steroid metabolites due to absence or presence of intestinal microbes which have the ability to directly metabolize those steroids. The 21—dehydroxylation activity of intestinal microorgansims in vitro was further examined by Eriksson 132 et al. (111) They confirmed that steroids with 21— hydroxy—ZO—oxo structures can undergo microbial 21- dehydroxylation. Incubations of caecal microorganisms displayed the capacity to convert 3b,21-dihydroxy—5a- pregnan—ZO—one to pregnane-3,20-diols and pregnane- 3,20,21—triols. Incubation of 5a-pregnane-3b(and 20a),21,triol did not lead to pregnanediol formation. When the 20,21—diol structure is present microbial 21- dehydroxylation will not occur. Another in vitro study of the sulphatase and glucuronidase activity of caecal contents was performed by Eriksson and Gustafsson (112). The ratio between free and conjugated metabolites of administered [4—140] corticosterone was much higher in the conventional rats than in the germfree rats (13 times higher in female rats and four times higher in male rats). They noted the presence of sulphatase activity of microbial origin in the caecal contents of conventional rats. Sulfuric acid esters of the 3a,3b,17b and 21—hydroxyl groups were hydrolyzed by the caecal microflora. There was no sulphatase activity present in the caecal contents or intestinal mucosa from germfree rats. Glucuronidase activity was noted in both germfree and conventional rats, indicating that hydrolysis of glucuronides may be caused by mucosal cells that are sloughed off into the intestinal lumen as well as by intestinal microflora. 133 C. Excretion of steroids in the bile Baillie and Eriksson (113) investigated the mode of conjugation for the major biliary corticosterone metabolites in the female rats. The metabolites identified included 3a,11b,21-trihydroxy-Sa—pregnan—ZO-one (3a,5a—THB), 3a,llb,15a,21—tetrahydroxy-5a-pregnan—20-one (15—0H-3a,5a-THB), 3b,11b,21-trihydroxy-Sa—pregnan-ZO—one (3a,5b—THB), and 3b,11b,15,21—tetrahydroxy—Sa—pregnan—ZO— one (lSa-OH—3b,Sa—THB). Their experimental procedure included acetylation of the free hydroxyl group, followed by solvolysis of the sulfate group. Those free hydroxyl groups are then trimethylsilylated. The 15-OH—THB isomers were sulphated at the C-21 position, and the THB isomers were only conjugated at the C-3 position. Microsomal preparations from female rat liver have been shown to contain a steroid sulphate—specific le—hydroxylase which is only active in the presence of a 21-sulfate on the steroid (114). Steroid metabolites containing a sulfate conjugation at the C-3 position are poor substrates for the le—hydroxylase. The extent of 15b—hydroxylation in the metabolism of corticosterone is dependent on the activity of 21—sulphokinase with respect to those of the 3-sulphokinase and 3-keto—delta—4—steroid reductases. Begue et al. (115) measured the daily excretion of corticosteroid metabolites in the bile from male and female rats. They were interested in determining whether 134 a diurnal variation of steroidhormone excretion exists in these Sprague—Dawley rats. A cyclic excretion of corticosteroids was noted in both female and male rats. The minimum excretion occurs between 04:00 and 10:00, and maximum excretion occurs between 16:00 and 22:00. The common bile duct of rats were cannulated to allow for collection of bile. This operation disturbed the daily rhythm of corticosteroid excretion for at least two days. Begue et al. calculated mean values and standard deviations for the corticosteroid metabolites from female rats: total corticosteroids, 738 169 ug; monosulphurylated 33,11b,15a,21—tetrahydroxy—Sa—pregnan- 20—one, 300 16 ug, disulphurylated 3a,11b,21—trihydroxy— 5a—pregnan—20—one, 146 49 ug; disulphurylated 3a,11b,153,21—tetrahydroxy—5a—pregnan—20-one, 158 51 ug and disulphurylated 3b,11b,21—trihydroxy-5a—pregnan-20— one, 134 38 ug. When saline was injected to provoke a stress reaction, a maximum corticosteroid excretion is obtained between 10:00 and 16:00 hours. The normal excretion minimum still occurs between 4:00 and 10:00 hours, and a second excretion maximum occurs between 16:00 and 24:00 hours. During the first 24 hours after a saline injection the total corticosteroid excretion is 920 ug and then 730 ug during the second 24-hour period. A second stress reaction was induced with an intramuscular injection of 135 dimethylsulfoxide (DMSO). The total corticosteroid excretion during the first day after DMSO injection was much greater (859 ug) than the excretion on day two (490 ug), indicating a depletion of the adrenal steroid content during the first day. Begue et al. did not find diurnal variation in hydroxysteroid oxidoreductase or 15a-hydroxylase activity, as the ratio between different corticosteroid metabolites in bile remained relatively constant throughout the 24— hour period. The excretion of endogenous steroids and metabolites of [4-14C] pregnenolone in bile of female rats was examined by Cronholm et al. (116) The radioactive compounds recovered in the bile during the first 24 hour period after injection of [4-14C] pregnenolone included: 3a—hydroxy—5a-androstan-17—one, 3a-hydroxy—5a,l7b-pregnan— 20-one and 3a,16a—dihydroxy-5a—pregnan—20—one in the glucuronide fraction; 3a,7a-dihydroxy—Sa-androstan—17-one, 3a,l5a—dihydroxy—Sa-androstan—17-one, 3a,11b—dihydroxy—Sa— androstan-17—one, and 3a,16a—dihydroxy—Sa—pregnan—ZO—one in the monosulfate fraction; and 3a,7a—dihydroxy—5a— androstan-17—one, 3a,1lb—dihydroxy—Sa—androstan—l7—one, 5a—pregnane-3b,20b-diol, and 3a,21—dihydroxy-5a-pregnan— 20—one in the disulfate fraction. 136 Cronholm et al. found that steroid metabolism in the bile fistula rat is different from that in the intact animal. Specifically, when [4—14C] pregnenolone is administered to intact rats, labelled corticosterone metabolites are excreted in the urine and feces. In this study only a 21-hydroxylated corticosterone metabolite lacking an llb—hydroxyl group was found in the disulfate fraction of the bile. Labelled corticosterone metabolites such as 3a,11b,21—trihydroxy-Sa—pregnan—20—one, and 3a,11b,15a,21-tetrahydroxy-5a—pregnan-20-one, were not present. The labelled pregnenolone was probably poorly mixed into the adrenal steroid pools due to less enterohepatic circulation time in the bile fistula rat. There were fewer free steroids present in the bile than in the urine of conventional female rats. These steroids are probably metabolized by intestinal bacteria and are reabsorbed from the intestine to be excreted mainly in the urine, but to a smaller extent in the bile. The only intestinal bacterial steroidal metabolites identified in the bile were 3a—hydroxy—5a,l7a—pregnane-20-one, 3b,lla— dihydroxy—5a—pregnan—20-one, and 3b,20b—dihydroxy—5a- pregnan—ll—one. D. Steroid levels in plasma Butcher et al. (117) measured the plasma concentration of progesterone and l7b-estradiol at three hour intervals 137 throughout the four day estrous cycle of a group of female Sprague—Dawley rats. Using competitive protein binding, they found that plasma progesterone concentration was elevated from the ninth hour of metestrus until the ninth hour of diestrus. Progesterone has a second, larger peak, which occurs during proestrus. This compound's concentration rises in response to a surge in LH (luteinizing hormone) production and is indicative of a preovulatory phase. Estradiol plasma concentration rose late during metestrus and reached a maximum level at noon of proestrus. All five hormones (LH, FSH, prolactin, progesterone, and 17b-estradiol) measured by Butcher et al. reached a maximum on the day of proestrus. The plasma level of 17b—estradiol started to rise earlier (during metestrus) and reached a maximum before the other hormones. Dupon and Kim measured the levels of testosterone, androstenedione, and estradiol in rat peripheral plasma during the estrous cycle. Using radioimmunoassay, they found that starting from the afternoon of diestrus there was an increase in the plasma levels of estradiol, reaching a maximum level of 52.016.0 pg/ml during the 16th hour of proestrus. Testosterone and androstenedione exhibit similar patterns with values of 180.0136.0 and 138 210.0138.0 pg/ml, respectively. The levels of the three steroids on each day of the estrous cycle are given in Table 9, where E2 is estradiol, T is testosterone, and A is androsterone. Their findings suggest that androgenic steroids are also secreted in a cyclic fashion by the rat ovary. Table 9. Steroid levels in rat plasma during the estrous cycle. time(h) rats E2 T A estrus 10 4 1.6:1.4 83.0117.0 110.01270.0 metestrus 10 4 2.33:2.4 90.01: 9.5 93.1:19.0 diestrus 10 4 9.1:3.0 86.113.0 100.::19.0 16 4 27.18.7 140.:39.0 140.1 19.0 proestrus 1o 4 35.1.4.3 160.0:30.0 1402.4: 22.0 16 5 52.:6.0 180.0136.2 210.: 38.0 E. Steroid levels in the ovary Toorop and Gribling—Hegge (118) measured the levels of androgens and estrogens in the ovary of the 5—day cyclic rat. They did not report specific testosterone or 17b- estradiol concentrations because their testosterone 139 radioimmunoassay's antiserum cross—reacted with 5a— dihydrotestosterone and 5a—androstane-3a,l7b—diol, and their 17b—estradiol antiserum cross-reacted with estrone, estriol, and 17a-estradiol. They found ovulation was preceded by an increase and then a sharp decline in the level of ovarian estrogens and androgens. From metestrus until the 22nd hour of diestrus there is a gradual increase in the concentration of ovarian androgens and estrogens. These high concentrations of androgens and estrogens plateau until the 17th hour of proestrus, followed by a steep decline during the evening of proestrus. During the morning of proestrus, when both serum and ovarian levels of androgens and estrogens are high (15 and 8 pg/mg ovarian weight, respectively), the serum and ovarian concentrations of progesterone is low. Then as the level of androgens and estrogens drop, progesterone rises, indicating an inverse relationship between the ovarian production of progesterone and the androgens and estrogens. Szoltys (119) specifically examined rat ovarian follicles for the concentrations of estrogens and progestagens throughout the estrous cycle. Follicular estrogen levels were low during estrus and metestrus (less than 100 pg/mg follicular tissue). The concentration rose during diestrus and reached maximum levels during the 18th hour on the day of proestrus. During the evening of 140 proestrus the estrogen level decreases, and during the night before estrus and ovulation, the level in both plasma and follicles was very low (less than 100 pg/mg follicular tissue). Progestagen levels in the ovarian follicles could only be measured during the proestrus phase. Late on the day of proestrus a maximum level of 16 ng progestagens/follicle was measured. In the plasma, there are two progestagen peak levels, the first during metestrus or diestrus, and the second, higher one during proestrus. This second peak is of follicular origin. The source of the first peak may be the corpus lutea or interstitial tissue. Hashimoto et al. also found two peaks in the secretion of progestins, when measuring their levels in ovarian veous blood. These steroids were first extracted from the plasma, separated using thin layer chromatography, and quantitated by gas chromatography. One maximum level occurred on the evening of proestrus, about 4 ug/hour/ovary, and a second progesterone peak half as large occurred during early diestrus. Both maxima for the other measured steroid, 20a—hydroxy—pregn—4-ene—3—one, were of the same concentration, about 18 ug/hour/ovary. Tamoxifen therapy has been described as a type of chemical ovariectomy. Ovariectomy is said to benefit between 40 and 50% of women (usually premenopausal) with 141 breast cancer. Whereas ovariectomy removes the source of various hormones necessary for the continued growth of endocrine-dependent breast carcinoma, Tamoxifen blocks the effect of one of these hormones, estradiol, at the cellular level. Gustafsson and Pousette (120) analyzed the steroid excretion patterns in urine from ovariectomized and adrenalectomized rats using gas chromatography/mass spectrometry. By comparing the metabolites present after ovariectomy with those present after adrenalectomy the authors could distinguish the steroid of ovarian and adrenal origin. They found a series of steroid monosulfates in the urine of female rats which were of ovarian origin: 3a,7a—dihydroxy—Sa—androstan-l7—one, 3a,19—dihydroxy-Sa—androstan-l7—one, 5a—androstane- 3a,7a,17b-triol, 3a,llb—dihydroxy-Sa—androstan-l7-one, Sa— pregnane—3a,20b-diol, 33,19—dihydroxy—5a,l7a—pregnan-20- one, 3a,17a-dihydroxy-5a,17b—pregnan—20-one, 3a,153— dihydroxy—5a,17b—pregnan—20—one and 3a,16a-dihydroxy— 5a,17b-pregnan—20—one. Corticosterone metabolites, including 11b,21- dihydroxy—Sa—pregnan—B,20-dione, 3a,11b,21—trihydroxy-5a— pregnan—ZO—one, 3a,15a,21-trihydroxy-5a—pregnan-11,20- dione and 3a,11b,15a,21-tetrahydroxy—Sa—pregnan-ZO—one were of adrenal origin, and they did not appear in the 142 urine after adrenalectomy. The adrenal gland does not secrete the precursor of progesterone and C1903 metabolites present in rat urine, although progesterone metabolites from the adrenals may excreted in the bile. Gustafsson and Pousette also mention that enzyme hydrolysis with Helix Pomatia only liberates corticosterone metabolites. Begue et al. (121) found that postpuberal ovariectomy has no effect on the typical female pattern of corticosteroid metabolites found in the bile. When adult, ovariectomized rats are treated with testosterone propionate a male biliary corticosteroid metabolite pattern arises, indicating a suppression of 15a- hydroxylase activity by androgens. Corticosterone has shown to be efficiently 15a-hydroxylated in isolated female rat livers, where as male livers lack this enzyme activity. The female rats which were castrated after puberty excreted the following steroids into the bile: 3a,11b,15a,21-tetrahydroxy-5a-pregnan-20-one (monosulfate and disulfate), 33,11b,21—trihydroxy—5a—pregnan-20-one disulfate, and 3b,11b,21-trihydroxy-Sa-pregnan-ZO—one. Begue et al. concluded that sexual differences in corticosterone metabolism of rats are not dependent on continuous secretion of gonadal steroids and that the pattern of corticosterone metabolism is set prepuberally for the adult rat life. 143 F. Steroidgproduction in the adrenal cortex The adrenal cortex is the secretory source of three major types of steroids: glucocorticoids, mineralocorticoids, and androgenic steroids. These adrenal steroids are synthesized from cholesterol. The glucocorticoids, such. cortisol and corticosterone are produced in the zona fasciculata in a high energy process which requires a shuttling of the steroid back and forth between the microsomal endoplasmic reticulum and the mitochondria. These corticosteroids are important in the regulation of carbohydrate, water, bone, muscle, gastrointestinal, cardiovascular, central nervous system, and hematological metabolism. The physiological actions of glucocorticoids are beneficial when a man or animal is confronted with a stressful situation as they provide a source of energy, (through gluconeogenesis). They are also important anti-inflammatory agents. Mineralocorticoids, such as aldosterone are produced in the zona glomerulosa. The mineralocorticoids exert their main effect on salt and water metabolism. Aldosterone is involved in a negative feedback control loop regulating body fluid volume. This steroid stimulates active sodium reabsorption. The adrenal androgenic steroids such as dehydroepiandrosterone and androstenedione function 144 principally as anabolic agents. They are formed in the zona reticularis. Even though these adrenal l7- ketosteroids may be the source of about three fifths of all the androgens excreted in the urine, if a man or animal were castrated the adrenal androgens could not maintain their secondary sex characteristics. Prost and Maume (122) determined the major corticosteroids in pooled rat adrenals. After extraction of the steroids, methyloxime—trimethylsilyl derivatives were prepared and the samples were analyzed by gas chromatography and gas chromatography/mass spectrometry. They identified pregnenolone, progesterone, ll- deoxycorticosterone, and ll-dehydrocorticosterone as biosynthetic intermediates in the production of corticosterone. They noted the presence of cholesterol, 1l—dehydrocorticosterone, lS—hydroxy—ll-deoxy corticosterone (18-OH-DOC), 3b,11b,21—trihydroxy—5a— pregnan-3,20—dione (THBll), corticosterone, and b— sitosterol. Three phytosterols: b—sitosterol, campesterol, and stigmasterol were also identified, as was 5b—cholestan3b—ol, and 3b—hydroxy—5—cholesten—7—one, in the sterols. These compounds are components of rat food which become concentrated in the adrenals, and their biological effects are unknown. 145 Another in vitro study was conducted by Maume et al. (123), involving the characterization of adrenal corticosteroids by capillary gas chromatography and gas chromatography/mass spectrometry. The authors separated each sample into five zones with thin—layer chromatography prior to CC analysis. Zone 3 and b included QB 02 steroids such as progesterone and 20a—dihydroprogesterone. When 4,5a—epoxy—17b-hydroxy—3—oxo—2a-androstane carbonitrile, and inhibitor of 3b—hydroxy steroid dehydrogenase was added to the adrenal cell culture pregnenolone is the major product. The 203- dihydroprogesterone concentration is much smaller in the inhibited culture than in the non—inhibited culture. Zone c contained C2103 steroids such as ll-oxo—20a— dihydroprogesterone, deoxycorticosterone, and several isomers of X—hydroxy—ZOa—dihydroprogesterone. Zone d contained the more polar C2104 steroids, the major peaks being corticosterone and l8—hydroxy-deoxycorticosterone. McCredie et al. (124) investigated whether corticosterone and 18—hydroxy—deoxycorticosterone (18—OH- DOC) are synthesized in different cellular compartments of the adrenal cortex. They separated the subcellular adrenal fractions through centrifugation and analyzed the products of incubation with endogenous precursors or added progesterone and deoxycorticosterone (DOC). The authors found that 18—hydroxy—deoxycorticosterone is produced in 146 larger amounts than corticosterone in the intramitochondrial subcellular fraction, whereas a greater amount of corticosterone is produced in the extramitochondrial subcellular fraction. These findings indicate that there is subcellular compartmentation of steroidogenesis in the zona fasciculata/reticularis of the rat adrenal cortex. Their conclusions are in accordance with other findings that the major sites for 18 and 11b— hydroxylation are mitochondrial and 21-hydroxylation is largely microsomal. Another in vitro study using rat adrenal incubations was carried out by Gomez—Sanchez et al. (125). They used mass spectrometry and gas chromatography to identify the mineralocorticoid—like products of deoxycorticosterone metabolism in adrenal gland incubations. This tissue was able to convert deoxycorticosterone to 19- hydroxydeoxycorticosterone (19,21-dihydroxy-4—pregnene— 3,20-dione), 19—oxo-deoxycorticosterone (21—hydroxy—4- pregnene-3,19—20-trione), and l9-oic—deoxycorticosterone (19-oic—21—hydroxy—4—pregnen—3,20-dione). The production of 19-hydroxydeoxycorticosterone is probably the first reaction in the biosynthetic pathway leading to l9-nor- deoxycorticosterone, which is a mineralocorticoid with 1.5 - 5 times the potency of deoxycorticosterone. l9-nor— deoxycorticosterone is not produced within the adrenal cortex. However, Dale et al. (142) described the further 147 metabolism of 19-nor-deoxycorticosterone in adrenal incubates to l9—nor—corticosterone and 19—nor-18- hydroxydeoxycorticosterone. The 19-nor-18- hydroxydeoxycorticosterone demonstrated sodium-retaining activity similar to that of 18—hydroxydeoxycorticosterone. G. Steroid metabolism in the liver The liver is the major site of catabolic reactions which render steroid molecules inactive and increase their hydrophilicity. The major reactions are reductive in nature, such as reduction of the 4-5 double bond in C19 and C;n steroids, or reduction of the 20—ketone function to a 20a— or 20b-hydroxyl group. These hydroxyl functions can then serve as loci for the addition of either a glucuronide or sulfate conjugate, yielding metabolites which are more soluble and are more readily eliminated in the urine. Some of the differences in steroid excretion patterns observed between male and female rats can be attributed to the differences in metabolism of steroids in the livers of male and female rats. Eriksson and Gustafsson (143) studied the metabolism of corticosterone when added, i_ vitro, to isolated, perfused rat livers. They found the following 20—hydroxy steroids in the male livers: 5a— pregnane-3a,1lb,20b(and 20a),21—tetrol and 5a—pregnane- 3b,llb,20b,21—tetrol. The female livers produced 3a(and 148 3b),11b,15a,21-tetrahydroxy-Sa—pregnan-ZO-one and 3a(and 3b),11b,21-trihydroxy—Sa—pregnane—ZO—one, and only small amounts of Sa—pregnane—Bb,11b,20b,21-tetrol. 15b- Hydroxylase is only present in the female liver. The female livers produced more mono- and disulphurylated metabolites than the male livers. Bournot et al. (144) confirmed these results using both liver cells in culture and steroid extracts from liver organs. These findings coincide well with the sexual differences observed in urinary and fecal steroid excretion patterns for corticosterone metabolites. Eriksson (145) examined the substrate specificity of the sex—specific 15—hydroxylase present in the female rat liver. Sa—Dihydrocorticosterone and 3a,5a— tetrahydrocorticosterone, which are 5a—reduced steroids, proved to be the best substrate for le—hydroxylase in this isolated perfused liver. llb—Hydroxylated steroids were more effective substrates than the corresponding 11a- hydroxy, ll-dehydro, or 11-deoxysteroids. 3—Keto—delta—4— and 5b—reduced compounds were not as efficiently hydroxylated as the 5a—reduced steroids. This sulfate- specific hydroxylase system may be catalysed by a cytochrome P—450 (Gustafsson and Ingelman—Sundberg, 1975) (146). 149 Eriksson (145) also speculates that induction of 15b- hydroxylase in the developing rat is influenced by androgens and estrogens. For example, a castrated or estrogen treated adult male rat will not have 15— hydroxylating enzymes present in its liver. However, neonatal castration leads to a "female" pattern of corticosteroid excretion. When [4—14C] androstenedione and [4—14C] progesterone were added to isolated, perfused rat livers, other sexual differences in steroid metabolism are observed (147). The male livers produced 3,l6-dihydroxy-androstan-l7- one,androstane—3,16,17-triol, androstane—2,3,16,17-tetrol, 2,3-dihydroxy—pregnan—ZO—one,3,16—dihydroxypregnan—20-one, pregnane—2,3,20—triol, and 2,3,16-trihydroxypregnan—20- one. These 2—hydroxylated and 16—hydroxylated metabolites were found in the glucuronide fraction. The female livers converted androstenedione to predominantly 7-hydroxylated and 15-hydroxylated compounds: 3,7—dihydroxyandrostan—l7- one, androstane—3,7,l7-triol, and androstane—3,15,17- triol. These compounds were mainly found in the glucuronide fraction. Progesterone was mainly converted to isomers of 3,15-and 3,16-dihydroxy—Sa—pregnan—ZO-one, which were present as both sulfates and glucuronides. Again, the sexual differences in the urinary and fecal excretion of androstenedione and progesterone metabolites is a function of liver metabolism. 150 Desgres et al. (148) examined the sequence of the reduction and hydroxylation of progesterone in rat liver epithelial cell cultures. He found that the most probable order was 5a-reduction of progesterone followed by reduction of 5a—pregnane—3,20-dione at C—3 then 6a- hydroxylation, and finally 3a(or 3b)—reduction to produce a 3,6—dihydroxy—Sa—pregnan—ZO-one. H. Steroid metabolism in DMBA—induced rat mammary carcinoma The dimethylbenz(a)anthracene-induced rat tumor shares many characteristics with the human breast carcinoma. These rat tumors may be hormone responsive, perform steroid metabolism, and possess estrogen receptor proteins. Various researchers have studied what type of paraendocrine behavior these rat carcinomas exhibit. Miller et al. (126) studied the metabolism of dehydroepiandrosterone (DHEA) and testosterone by both the DMBA-induced rat mammary tumor and the human breast carcinomata. Both the rat and human tumor tissue transformed DHEA and testosterone to 5a— dihydrotestosterone, 5a—androstanediol and 16a-hydroxy testosterone. The production of these metabolites indicates the presence of 5a—reductase, l6a—hydroxylase and 3b—hydroxysteroid dehydrogenase. These results point out that although there are quantitative differences, 151 DMBA-induced rat carcinoma and human breast cancer can metabolize steroids in a similar fashion. The DMBA— induced mammary carcinoma is thus justified as an experimental animal model for human breast carcinoma. Eechaute et al. (127) found that DMBA—induced mammary tumors contained the following steroid metabolizing enzymes: Sa-reductase and 20a-hydroxysteroid dehydrogenase (highly active); 6a-hydroxylase and Ba— hydroxysteroid dehydrogenase (lower activity). No aromatization activity was detected. The author performed in vitro steroid metabolism studies on incubated tumor tissue. Sephadex LH-20 column chromatography, paper chromatography, and thin—layer chromatography were used to isolate the metabolites of radioactively—labelled steroids. The amount of each purified metabolite was assessed by liquid scintillation counting. The tumor tissue was able to convert testosterone to 5a-dihydrotestosterone, and to a smaller extent to 5a— androstanedione, Sa-androsterone, and Sa-androstanediol. Progesterone was found to be converted primarily to 203— hydroxy—4-pregnen—3—one, a progestational steroid more potent than progesterone. Other progesterone metabolites were reduced at either the Sa, or 20a position: 20a- hydroxy—Sa—pregnan—3—one, 5a-pregnan—3a,20a—diol, and Sa— pregnane—3,20—dione. V. Experimental Procedures A. Introduction The topics in this chapter cover the progressive development of biological and chemical methodologies necessary for the generation of rat urinary steroid metabolic profiles using capillary GC—FID and both packed and capillary column GC—MS—DS. The analysis of steroid profiles by CC or GC—MS-DS requires the following general procedures: (1) sample collection, (2) extraction of analytes or class of desired compounds from the sample, (3) derivatization to make the analytes amenable for gas phase analysis, (4) production of urinary steroid profiles using GC—DS and GC- MS—DS, and (5) manual or automated qualitative and quantitative analysis of the data. The experimental animal models are described in section B. The sample preparation procedure required for the extraction and derivatization of urinary steroids is described in section C. B. Animal model All the animals involved in the following experiments were female, Sprague—Dawley rats (Spartan Research Animals, Inc., Haslett, MI). The animals were housed individual, stainless steel, metabolism cages designed for 152 153 the collection of urine and feces. The animals were given a diet of Wayne Laboratory Meal (Allied Mills, Inc., Chicago, IL) and water ad libitum. The rats were kept in a temperature controlled (24 a: 1°C) and light-controlled (14 hr/day) room environment. Twenty—four hour urine samples were collected from each rat into silanized glass bottles. A mesh of cheesecloth over the top of each bottle prevented food and fecal particulates from falling into the urine sample. Each 24-hour urine sample was transferred into a silanized glass vial and stored at - 700C until sample extraction and derivatization. The following section describes the groups of rats used in this experiment. 1. Normal, cycling female rats One group of six 2—month old female rats was placed in metabolism cages. After one week of acclimatization, daily urine collection and vaginal swabs began and were continued for 18 days. 2. Tamoxifen-treated, normal adult female rats One group of six 2-month old rats was placed in individual metabolism cages. After one week of acclimatization, daily s.c. saline injections began. The saline (0.9% NaCl solution with gum arabic) treatment conditions the rat to expect daily injections and 154 handling. After 3 weeks all rats received daily injections of Tamoxifen (tamoxifen citrate ICI America, Inc.) for 18 days at a dose of 75 ug/lOOg body weight. The Tamoxifen solution was a 1:1 mixture by weight with gum arabic, added to distilled water. This dose level is effective in causing rat mammary carcinoma regression. Daily urine collections were taken from the 14th to the 18th day of Tamoxifen treatment. Daily vaginal swabs were also taken during this period. 3. Ovariectomized, adult female rats One group of six 28—day old female rats was bilaterally ovariectomized. At two months of age, each rat was placed in an individual metabolism cage. After four weeks, the rats began receiving daily saline injections. Daily urine collections were taken from the 14th to the 18th day of saline treatment. Daily vaginal swabs were also taken during this period. 4. Tamoxifen—treated and control adult, female rats bearing palpable DMBA—induced mammary carcinomas. Tumors were induced in two groups of six 50—day old female rats with singular injections of 7,12—dimethylbenz(a)antracene (DMBA) (Eastman Kodak Co., Rochester, NY) at a dose of 2 mg/lOOg body weight. The 155 carcinogen is mixed into a specially prepared emulsion (Upjohn Co., Kalamazoo, MI) prior to injection. This dose of DMBA induces multiple palpable mammary carcinomas in 100% of the treated rats within 2 months of carcinogen treatment. The vast majority of these tumors are carcinomas of ductal origin and are hormone (pituitary and ovarian) responsive. All animals injected with DMBA developed at least one and as many as six tumors. Six weeks after DMBA treatment, when all rats had at least one palpable mammary carcinoma, the rats were placed in individual metabolism cages. After one week of acclimatization all rats received daily saline injections, and urine collection began together with daily vaginal swabs. All rats received saline injections for ten days. During the next ten days, one group of 6 rats continued to receive saline injections and the second group of 6 rats received daily therapeutic doses of Tamoxifen. Palpable mammary carcinomas were measured with a vernier caliper before and during Tamoxifen treatment. C. Preparation and analysis of samples 1. Sample preparation (method I) Silanized glassware was used at all times. The Sep- pak C18 cartridges (Waters Associates, Inc., Milford, MA) are packed and shipped dry, so they must be conditioned prior to use. An organic solvent, methanol (3 ml) was 156 passed through the cartridge, followed by a rinse of doubly distilled water (5 ml). Fludrocortisone (20 ug) was added to each rat urine sample as an internal standard. A 500 ul aliquot of urine was removed for creatinine analysis. The urine sample was poured into a 20 ml silanized glass syringe, which had a conditioned Sep-pak cartridge fitted to its Leur tip. The plunger was inserted into the syringe and the sample was pumped through the cartridge at a rate of one drop per second. Doubly distilled water (5 ml) was passed through the cartridge at a rate of one drOp per second, as a was to remove interfering, hydrophilic species from the cartridge. The steroids and steroid conjugates were eluted from the cartridge with methanol (3 ml) into a 10 ml silanized test tube. After the samples were dried under a stream of nitrogen, sodium acetate buffer (4 ml, 0.5 M, pH 4.55) was added. b—Glucuronidase—aryl sulphatase from Helix pomatia (400 ul) (Behring Diagnostics, La Jolla, CA) was also added to each sample. The specific activity of this preparation was 5.8 I.U. per ml for b-glucuronidase and 3.7 I.U. per ml for aryl sulphatase. (One I.U. is defined as the amount of the enzyme which will liberate l u mole of phenolphthalein from phenolphthalein disulfate or phenolphthalein glucuronide per minute at 37° and pH 6.2 and 4.5, respectively). The samples were enzymatically 157 hydrolysed during a 48 hour incubation period in a 37° water bath. The Sep—paks were reconditioned by passing doubly distilled water ( 5 ml), methanol (3 ml), and doubly distilled water (5 ml) through the cartridges. After hydrolysis, each enzyme—buffer solution was passed through a reconditioned Sep—pak in order to separate the free steroids. The loaded cartridges were washed with doubly distilled water (5 ml), and the steroids were eluted with methanol (3 ml). The samples were dried under a stream of nitrogen in a 60 water bath. The methyloxime derivative was prepared by adding 100 ul of a solution of o— methoxyamine hydrochloride (100 ug/ul) (Supelco, Inc., Bellefonte, PA) in dry pyridine, and heating for one hour at 800C. Excess pyridine was removed under nitrogen in a 600C water bath, and the trimethylsilyl derivative is prepared by adding Sylon BTZ (150 ul) (Supelco, Inc., Bellefonte, PA) to each sample and heating for 24 hours at 800C. Sylon BTZ is a potent silylating agent prepared as a mixture of N,O—bis(trimethylsilyl)acetamide) (BSA), trimethylchlorosilane (TMCS), and trimethylsilylimidazole (TSIM) (3:2:3). A final sample clean—up step to remove polar compounds and excess silylating reagents was performed on the derivatized steroid fraction. Lipidex-SOOO (Packard 158 Instrument Co., Downers Grove, IL) in a solvent mixture of hexane—hexamethyldisilazane—pyridine—Z,2-dimethoxypropane (97:1:2:10 V/V) was packed in a silated pasteur pipet (5 30" x 7.0 mm o.d.). The solvent mixture (400 ul) was added to the derivatized sample, and the sample was transferred to the top of the Lipidex column. An additional 3.5 ml of solvent was passed through the column to elute the derivatized steroids. The solvent mixture was evaporated under nitrogen. Each sample was reconstituted with bis(trimethylsilyl)trifluoroacetamide (BSTFA) (25 ul) and pyridine (100 ul), and sealed in capillary tubes. 3 flow diagram of the sample preparation procedure is seen in Figure 13. 2. Sep-pak extraction a. Introduction Sep-Pak (Sample Enrichment and Purification) cartridges (Waters Associates, Inc., Milford, MA) were used for the extraction of free steroids and steroid conjugates from rat urine. These cartridges are small, self—contained, and packed with C” —substituted silica designed for reverse phase liquid chromatography. This high efficiency column is prepared using Waters' radial compression technology. The flexible walled cylinders are radially compressed to eliminate channels and void volumes 159 EXTRACTION ON C18 SEP—PAK v ENZYMATIC HYDROLYSIS v RE—EXTRACTION 0N C18 SEP-PAK V DERIVATIZATION V PURIFICATION ON LIPIDEX—SOOO Figure 13. Flow diagram of the sample preparation procedure for the analysis of urinary steroids. 160 in the packing structure, yielding a homogeneous column bed. To clean-up biological samples using Sep-pak C18 cartridges, the sample is loaded onto the column, and the components of the sample are retained on the column based on the affinity of those components for the active sites of the packing material. With these cartridges, hydrophilic (ionic) compounds are weakly retained and more non-polar, lipophilic compounds are strongly adsorbed. The cartridge is rinsed with distilled water to remove inorganic salts, amino acids, sugars, and hydrophilic proteins. The steroids are then eluted using a "stronger" solvent, e.g. methanol. b. Sep-Pak extraction efficiency Shackleton and Whitney (128) performed a thorough evaluation of Sep—Pak extraction of urinary steroids for gas chromatographic analysis. They compared this method with solvent partitioning and extraction with adsorbents such as the neutral polystyrene resin Amberlyte XAD-2. Using radiolabelled conjugated and unconjugated metabolites, the authors estimated the quantitative recovery of each method. They found that the Sep-Pak extraction and ethyl acetate extraction yielded approximately the same amount of steroids (about 99% recovery) but the Amberlyte XAD—2 extraction gave poor 161 recovery, especially for the less polar C19 steroids (e.g. etiocholanolone and androsterone). In another experiment, they evaluated the extraction efficiency for reference neutral steroids of various polarities. They found that there was not much difference between the extraction efficiency of ethyl acetate or Sep-pak, but the Sep—pak was more efficient as the steroid polarity increased. For example, 92% of the dehydroepiandrosterone was recovered by both methods, but 97% of the cortol was recovered with the Sep-Pak and 88% with the ethyl acetate extraction. Shackleton and Whitney concluded that steroid extraction with Sep—Pak is fast, economical, and as efficient as other extraction techniques. An experiment was conducted to determine the efficiency of the extraction of the internal standard, fludrocortisone, with the Sep—Pak cartridge. Three water blanks were prepared as described above. The enzymatic hydrolysis step was not included. Fludrocortisone (20ug) was added to each sample prior to Sep—Pak extraction. After extraction, each sample was dried and the methyloxime—trimethylsilyl derivative was prepared. Another group of three samples served as controls for this experiment. For each of these samples, fludrocortisone (20ug) was added directly to a test tube. Each sample was dried and then derivatized. In this way, the possible loss of fludrocortisone on the Sep—Pak cartridge for the 162 first set of samples could be determined by comparison with the second set which was not subjected to Sep-Pak extraction. Each derivatized sample was analyzed three times on a Varian 2100 GC fitted with a 15—meter DB-l megabore capillary column. The standard deviation for the gas chromatographic injections was 5.8% on average. The mean peak area for fludrocortisone contained in each sample is listed below. The overall relative standard deviation for each group of three samples was calculated from the means and standard deviations listed in the table below. 163 Table 10. Mean peak areas of fludrocortisone for triplicate injections of each blank sample on the Varian 2100 GC fitted with a 15-m DB-l megabore capillary column 20ug Fludrocortisone 20ug Fludrocortisone transferred by Sep-Pak transferred directly extraction Sample #1 2.3 x 10 Sample #4 2.7 x 10 #2 1.9 x 10 #5 2.8 x 10 #3 2.9 x 10 #6 2.3 x 10 mean = 2.4 x 10 mean = 2.6 x 10 standard standard 10 deviation = 2.6 x 10 II U1 'o X deviation For the three samples which received fludrocortisone prior to the Sep—Pak extraction, the relative standard deviation was calculated to be 21%. For the three samples where the fludrocortisone was added directly to the derivatization tube (no Sep-Pak extraction step), the relative standard deviation was calculated to be 10%. To test the null hypothesis that there is no significant difference between the mean peak areas for fludrocortisone in the two sample sets, a two—tailed t 164 test was applied. The pooled estimate of variance 32 was calculated to be 3.2 x 10 , with 4 degrees of freedom, and an observed value of t=0.44. The rejection region of t at the a=0.10 level of significance is —2.132 > t > 2.132. The observed value of t does not fall in the rejection region and it appears that there is not a significant difference between the means of these two sets. Thus, the fludrocortisone displays excellent extraction efficiency on the Sep—Pak cartridge. c. Sample contamination The possibility of introducing contaminant species into one's sample through Sep—pak extraction was not discussed by Shackleton and Whitney. To test this premise, an experiment was proposed to compare blank samples prepared using: 1) a normal Sep—pak cartridge, 2) a Sep-pak cartridge with its C18 packing material removed, and 3) the removed Sep-pak C18 packing material 'placed in a silated pasteur pipet, sandwiched between two silated glass wool plugs. Columns 1, 2, and 3 were conditioned by rinsing with methanol (3 ml) and doubly distilled water (5 ml). Doubly distilled water (20 ml) was passed through each column to mimic a urine sample. Each column was washed with doubly distilled water (5 ml), and methanol (3 ml) was passed through each column and collected. After drying under nitrogen, two internal standards, 20 ug each of cholesteryl butyrate and 3b,17b- 165 dihydroxy—17a—methyl—5b—androstane, were added. Cholesteryl butyrate was added as an internal standard because this experiment was conducted prior to the experiment which determined a more suitable internal standard for steroid analysis. The methyloxime— trimethylsilyl derivatives were prepared as is described below. Sample 1, obtained from the normal Sep-Pak produced two minor contaminant peaks, at retention indices 2287 and 2511. The earlier peak apparently is from the Sep—Pak column material as it appeared in the chromatogram of Sample 3, but not in Sample 2. Its early elution time would not interfere with identification and quantitation of steroid metabolites which elute after retention index 2400. The second contaminant peak, at retention index 2511 is introduced by the Sep—Pak cartridge, as it appears in Samples 1 and 2, but not in the sample obtained from the packing material. Since this peak occurs in a region of the chromatogram where androstanes would appear, it was important to identify this contaminant using GC-MS. Sample 2 was run on the HP5980 GC—MS with identical chromatographic conditions as those used for GC—FID. The contaminant yielded a base peak at m/z 149, characteristic of a plasticizer. The compound was added to the GCMET and MSSMET libraries, as the "Sep—Pak cartridge contaminant". 166 No other contaminant compounds were observed in CC traces of the three samples. 3. Enzymatic hydrolysis vs. solvolysis Several different enzyme preparations have been used to cleave steroid conjugates. Two of the most popular b— glucuronidases are: l) the digestive of the Roman snail (Helix pomatia) and 2) lypOphylized b—glucuronidase from the Marine Mollusk. It has been reported that the snail enzyme preparation has a higher sulphatase activity than the mollusk enzyme (133). Helix pomatia hydrolyses important 3b-hydroxy—5—ene and 3b-hydroxy-5b-steroid sulfates. However, this enzyme is not as efficient in cleaving 3a—hydroxy-5a—sulfates, as well as C21-steroids sulphated at position 20 and CH9—steroids sulphated at position 17. This could lead to an underestimation of certain steroid metabolites. Fortunately, these types of steroid conjugates are not major components of female rat urine (134). Also, 3b—hydroxy steroid dehydrogenase and 5-4-ene steroid isomerase may be present in this enzyme preparation, which would change 3b-hydroxy-5—ene steroids to 3—oxo-4—ene and effect metabolite quantitation. In order to compare the hydrolytic effectiveness of both Helix pomatia and Marine Mollusk b—glucuronidases, two aliquots of pooled female rat urine were prepared as described above, and each sample received either the b— 167 glucuronidase from Helix pomatia or Marine Mollusks. The 00 profiles of each derivatized pooled female rat urine sample are displayed in Figure 14. The traces are quite similar, with only slight deviation in the appearance of minor components with concentrations less than 1 ug (relative to the fludrocortisone internal standard (20 ug)), as displayed in Figure 15. Leunissen and Thijssen (135) promoted further cleavage of the steroid sulfate conjugates by including a solvolysis step in their procedure, after enzymatic hydrolysis (Method II). The hydrosylate is brought to pH 1 with concentrated HCl, NaCl (1.5 g) is added, and the liberated steroids and steroid sulfates were extracted into ethyl acetate (25 ml) by shaking for 60 minutes. The aqueous layer was removed and the remaining sulfates were cleaved by heating the ethyl acetate solution for 18 hours at 450. After solvolysis, an alkali wash step was performed, followed by the preparation of methoxime— trimethylsilyl derivatives. The authors measured the overall recovery of their sample preparation as 88 3%, using H—labelled cortisol. For the cortisol metabolites in human urine, the authors calculated a recovery of 95.0 318.3% specifically for the hydrolysis and solvolysis steps. Relative Intensity 22 24 Time—> (min) N N 24 Figure 14. 168 28 26 34 32 20 30 40 F 26 28 30 34 F 32 c Two gas chromatographic profiles of aliquots of derivatized pooled female rat urine. The steroid conjugates in the sample displayed in the top profile were hydrolyzed with the enzyme preparation obtained from Helix Pomatia. Enzymatic hydrolysis for the sample shown in the lower profile was acheived with beta—glucuronidase obtained from the Marine Mollusk. Coinjected hydrocarbon standards C 2 through C34 are indicated with the even numbers. T e internal standard fludrocortisone is indicated with an "F", and Sep-Pak contaminants are noted with "C" V S Relative Concentration (ug per mg cr) 10 169 ‘I‘ __ Q A... / .'. a ‘_ .___.a . . v / I l L ‘ J I I I I 3000 3025 3050 3075 3100 RETENTION INDEX Figure 15. Relative concentrations of components eluting between hydrocarbons C3 Oand C3 during gas chromatographic analyses of 2 aquuots of derivatized pooled female rat urine. Similar profiles are obtained with both Helix Pomatia (*) and Marine Mollusk (O) enzyme preparations. 170 The efficiency of the two sample extraction procedures mentioned above were compared to see which technique was better suited for the isolation of rat urinary steroids. Two aliquots of pooled female rat urine were prepared using either Method I, a variation of Shackleton's method, or Method II, Leunissen and Thjissen's hydrolysis plus solvolysis technique. The chromatographic results are displayed in Figure 16. It is important to note that the possible advantage of increasing the amount of cleaved steroid sulfates is more than offset by the disadvantage of sample loss through Method II's protracted procedure. The ratio of the peak area of fludrocortisone to the peak area of hydrocarbon 32 is 1.1 for Method I and 0.059 for Method II, indicating that about 90% of the internal standard is lost in the lengthier procedure. It was decided to perform all sample preparation with Method I, due to the small initial sample volumes (4 ~ 20 ml), the low concentration of steroids in each sample (ug range), and the possibility of sample loss using Method II. This choice is in agreement with the finding of Pike et al (131). They concluded that an acid hydrolysis step is probably not necessary, as the enzyme hydrolysis of radiolabelled steroid sulfate conjugates is more than 95% complete. 171 .memnssc co>o :ufl3. cosmoficcfl mum mconnmoouozc wouoomcwoo_ .mflmsaouoas oaumE»nco uco3umoc= saco maflmoua Eoqup any cw maasmm one .mfim>Ho>Hom can mammaouoxz oflumE>Nco uco3umoc= mammowa sou ecu :H maasmm one .mcflps one mameow eoaooq wouflum>fiumo mo muoscfiam ozu mo momsawcm ofinampwoumeouno moo .I 0!- - _____‘__lDJ' ' :90: E 1.711%} _ L w . L ”l.e r3 _ P. .0. ..gwfla A:.EV A- 05.9 5.. .6 I‘ ..' Kifguaiul GAIQBTGa ‘0 'oi mm vm No on mm mm em“ III ‘1: 172 4. Derivatization The extracted urine samples are subjected to a two— step derivatization procedure in order to increase the sample's volatility and to protect the hydroxyl and keto functionalities from degradation during vapor phase analysis. The first step involves conversion of the keto functions to methyloximes, as this prevents the keto groups from undergoing enolization during trimethylsilylation (136,137). The ll—ketone group does not react with methyloxime HCl under ordinary conditions due to the shielding effect of the methyl groups at positions 18 and 19. Sylon BTZ was chosen as the silylating reagent since it is one of the most potent derivatizing reagents available. The N,O—bis(trimethylsilyl) acetamide (BSA) is capable of silylating unhindered hydroxyl groups. For example, Chambaz and Horning (138) found BSA produced a 3a,20a,21-trimethylsilyl ether of cortol (Sa-pregnane- 3a,11b,17a,20a,21-pentol). The hydroxyls at the 11b and tert-l7a positions were not converted to silyl ethers. When trimethylchlorosilane (TMCS) is added to BSA and allowed to act on cortol, TMCS acts as a catalyst and the 3a,11b,20a,21—tetratrimethylsilyl ether is formed. Cortol is completely derivatized when trimethylsilylimidazole (TSIM) is added to the BSA and TMCS. Sylon BTZ, which 173 contains TSIM, BSA, and TMCS, is the reagent of choice when fully silylated derivatives are desired. Trimethylsilylimidazole (TSIM) is less volatile than BSA or TMCS and it is necessary to remove the excess amount of this derivatizing reagent after the silylating reaction. A chromatographic procedure utilizing Lipidex- 5000 was introduced by Axelson and Sjovall (139). Removal of this imidazole-containing reagent prolongs the life of the GC capillary columns used for steroid analysis. As described above the reaction mixture is passed through a pasteur pipet packed with Lipidex-5000. Ocassionally, the eluant form a Lipidex—5000 column is cloudy. In order to determine whether the "cloudiness" of the derivatized steroid solutions would lead to contaminant peaks int he steroid profiles, a blank sample of doubly distilled water (20 ml) was prepared using Method I. After passing the derivatized blank sample through a Lipidex—5000 column, the hazy eluant was split into two fractions. One fraction was dried under nitrogen and the resulting powdery blank was dissolved in pyridine and BSTFA. The second fraction was passed through a pasteur pipet fitted with a plug of glass wool and rinsing with the Lipidex solvent described above. The clear eluant was dried under nitrogen and dissolved in pyridine and BSTFA. Both samples were analyzed by GC and the 174 resulting blank chromatograms were identical, displaying coinjected only hydrocarbon peaks and Sep—Pak contaminant peaks. The samples cloudiness is probably due to passage of the Lipidex gel particles through the plug of glass wool at the base of the pasteur pipet. These particles do not interfere with the steroid profiles. After drying the resulting solution under nitrogen, the derivatized steroids are resuspended in a solution of pyridine and bis(trimethylsilyl)trifluoroacetamide (BSTFA). This silylating reagent is quite potent and keeps the steroids derivatized. BSTFA is volatile and will not interfere with the important peaks in the gas chromatogram. Also, this reagent produces less detector fouling and noise than BSA. GC detectors are fouled when silylating reagents are burned in the hydrogen flame, and silicon dioxide is formed, coating the detector. When BSTFA is consumed in the hydrogen flame hydrogen fluoride is produced. This compound reacts with silicon dioxide to produce completely volatile products. VI. Experimental Results This chapter involves the interpretation of the urinary steroid metabolic profiles obtained from the five animal models examined in this experiment: 1) normal cycling female rats, 2) normal female rats treated with Tamoxifen, 3) ovariectomized rats, 4) female rats with chemically—induced mammary carcinomas, and 5) tumor- bearing female rats treated with therapeutic dosages of Tamoxifen. Derivatized, 24—hour urine samples were analyzed on the Hewlett—Packard 5840A chromatograph, and chromatographic peak areas and times were stored on the PDP-11/44 data system. Data reduction was performed using the GCMET, the GC-only metabolic profiling procedure described above. The GCMET process eliminated countless hours of manual chromatographic data analysis and allowed for simple statistical comparisons of chromatographic data derived from the different animal models. A. Trends in the excretion of urinary steroids 1. Normal cycling female rats The adult female Sprague—Dawley rat has ovulatory estrous cycles of approximately five days duration time. The major phases of the estrous cycle are proestrus, estrus, and diestrus. During the proestrus phase there is a buildup of the uterine endometrium and development of 175 176 ovarian follicles. During estrus, ovulation occurs and there is a maintenance of conditions necessary for fertilization. Diestrus is a period of relative quiescence. Daily vaginal smears were obtained from each rat during the urine collection period and analyzed under a microscope. Vaginal cytology changes during the estrous cycle. Estrus begins when 25% of the cells are cornified and 75% are nucleated. The volume of urine excreted over a 24—hour period by an individual rat varied from about 3 ml to 20 ml, with the least amount of urine usually excreted during the estrus phase, and the greatest during diestrus. The estrous cycles of the six control rats were determined and the progression of phases are listed in the collection schedule displayed in Figure 17. Normally, the rat has one day each of proestrus and estrus followed by three days of diestrus. Only rats #3 and #4 displayed perfect cycling over the eighteen day collection period. Control rat #3 had only one complete estrous cycle. Rat #2 started with an irregular cycle (estrus, diestrus, proestrus) followed by 15 days of diestrus. Rat #5 remained in the diestrus phase except for one day of proestrus and three indistinct days: partially proestrus and estrus on one day, and partially estrus on two other days. Rat #6 had one day of estrus followed by one 177 .oUCflumaeCfi .a .msuummfio .Q .msuumm .m .mspumwoua..m "mm emumoflecfl mum mommna maoso msouumm .mumu.oamsow mafiaoso .Hmeuoc bow maseocom cowuomfiaoo mafia: .kx mtswfla OAII x003 m m m a a a m m a Q a m coauomaaou He 0mm Hopscoo memes. g mmwmo oucH w aflwmm mumm HH< 178 complete cycle. The remaining 12 collection days were predominantly diestrus, with three indistinct days. Qualitative examination of the gas chromatographic profiles obtained for the three different phases of the estrous cycle shows a change in the metabolite excretion pattern as seen in Figure 18. Focusing on the corticosteroid metabolite region (retention index: 2900 to 3300), visual comparison of the profiles would seem to indicate that excretion of steroid metabolites is at a maximum during diestrus, and at a minimum during estrus. However, calculation of the relative concentration of these analytes must take into account the volume of urine in each sample, as well as the creatinine concentration in the urine, and the peak area of fludrocortisone, the exogenous internal standard. Creatinine is a normal constituent of blood and urine. It is the end product of creatine metabolism and in humans about 0.02 gm/kg body weight is excreted per day by the kidney. Since the amount of creatinine excreted daily in the urine is relatively constant, this value is used to normalize urinary steroid relative concentrations. After analysis of the derivatized urine samples by GC- FID, the relative concentration for each analyte is calculated by the GCMET program using the following equation: 180 analyte (ug/mg cr) = (analyte P.A.)(amount of I.S.(ug) conc. (mg creatinine)(I.S.P.A.) P.A. — peak area; 1.8. — internal standard The excretion pattern of five corticosteroid metabolites are shown in Figures 19 through 23. A trihydroxy—pregnan—one (probably 33,11b,21—trihydroxy-5a- pregnan—ZO-one (97)) eluting at retention index 3015 displays a cyclic excretion behavior in Rat #3, where the relative concentration is greatest during estrus and is lower during diestrus (Figure 19). Rat #1 also displays a cyclic excretion behavior. Rat #4 has maximum excretion at estrus only in the first cycle. During the second cycle excretion of trihydroxy—pregnan—one decreases to a minimum at estrus. A methyloxime-trimethylsilyl derivative of tetrahydroxy-pregnan—one (retention index: 3072) displayed consistently cyclic excretion for control rats #1, 3, and 4 (Figure 20). The greatest urinary concentration was found during estrus (except for Rat #4 during the first cycle) and decreased excretion was observed for the diestrus phase. This steroid is probably 3a,11b,153,21—tetrahydroxy-Sa—pregnan—ZO—one, a metabolite of corticosterone detected by Gustafsson and Sjvall (109), Eriksson and Gustafsson (143), and Gustafsson (97). Another corticosterone metabolite, tetrahydroxy—pregnen— one elutes at a retention index of 3146. This 181 10.00 '1"- A m 1' 0 g aa>~— n. dr- 0 £3 a«)~— z o E -u- a 4.1” -1- o z 0 sub- (3 E zm1~— m .4» 030 r I I I I I l I ' I 1 o 1 2 3 4 5 0 7 0 9 10 COU££flONIfiW Figuna l9 Excretion of trihydroxy-pregnan-one (R.I.=3015) by control rats #1, 3 and 4 10.00 8.00 6.00 4.00 2.00 0.00 Figure 182 f E t E «p o. E ’ D I. " I " i -—1 n ' o P " I l -_. E 0 - o 0 o l 1 l l l l 1 1 l J I I I I I I I I I I 0 1 2 3 4 5 5 7 8 9 10 20 Excretion of tetrahydroxy-pregnan-one (HO-TMS) (R.I.=3072) by control rats #1, 3 and 4. 183 methyloxime—trimethylsilyl derivative has a molecular ion at m/z = 681 and displays characteristic ions including: M+—31 (m/z = 650), M+—90 (m/z = 591), M+—103 (m/z = 578), and M+—3l—90 (m/z + 560), M+—147 (m/z = 534), and M+—90—9O (m/z 501). Fragment ion peaks at m/z = 129 and 191 are also characteristic. The ion of mass 129 can be derived from a steroid with a 5—ene—3-TMS configuration in the "A" ring. The m/z = 191 ion is characteristic of polyhydroxylated steroids. The loss of 103 mass units from a steroid would correspond to the loss of —CH OTMS at the C—21 position. A possible structure for this steroid would be 3a,11b,15a,21-tetrahydroxy—pregn—5—en—20—one. For this metabolite, control Rat #1 excretes the compound in a cyclic fashion, but a cyclic pattern is less evident for rats #3 and #4 (Figure 21). Two other tetrahydroxy-pregnan—one isomers were detected in the urine of normal cycling female rats. These compounds eluted at retention indices 3117 and 3127. Mass spectral analysis of these species indicates that they were trimethylsilylated during derivatization, but did not produce methyloximes. This fact indicates that the keto group on each compound could be hindered by neighboring hydroxyl groups. These metabolites had a molecular ion peak at m/z = 654. Other characteristic ions included M+-15 (m/z = 639), M+—9O (m/z = 564), M+—15— 9O (m/z = 549), and M+-l3l (m/z = 523). This latter ion RELATIVE CONCENTRATION (00 PER M0 CR) 184 00.00 fir— 50.00 +- “L 40.00 -«- -II- 20.00 ~— ..I- D I I I l I I I I J I °°°° I I I I I I I I I_ I 0 I 2 3 4 5 o 7 a 9' In COLLECTION DAY Figure 21 Excretion of tetrahydroxy-pregnen-one (R.I.=3146) by control rats #1, 3 and4. 185 is indicative of the loss of C—20 and C—21, where C-21 has a trimethylsilyl group attached and a keto group is present at the C—20 position. Possible structures for these isomers could be 3a,11b,l6a,21—tetrahydroxy—5a- pregnan—ZO—one and 3b,11b,16a,21-tetrahydroxy—Sa—pregnan- 20—one. The isomer at retention index 3117 is excreted in a cyclic fashion by control Rats #3 and #4. Both animals excrete larger amounts of this compound during the estrus phase (Figure 22). The isomer eluting at retention index 3127 is excreted in a more irregular but vaguely cyclical manner by control Rats #3 and #4 (Figure 23). 2. Tamoxifen-treated, normal adult female rats and ovariectomized adult female rats. Daily 24-hour urine samples were collected from the six rats treated with Tamoxifen on the 14th through 18th day of drug treatment. Vaginal smears were obtained from each rat for each urine collection day. Therapeutic dosages of Tamoxifen disrupted the estrous cycle in each treated rat. The animal is locked into an "anestrus" or nonestrus state during drug treatement. Analysis of the vaginal cytology for the Tamoxifen-treated rats indicated that all of the rats were in anestrus for treatment days 14 through 18 (Figure 24). Treatment with Tamoxifen has been described as a type of "chemical ovariectomy" (40). Rather than surgically RELATIVE CONCENTRATION (UG PER M0 CR) 186 Imm)-~ &m>—~ P E D E amI—~ D E .4. P D D D P E ch~— D I U x P 2mI—— D I I I 03° Ev I I i i i i i i i o 1 2 3 4 5 5 7 a 9 10 COLLECTION DAY IHBUNBZZ Excretion of tetrahydroxy-pregnan-one (TMS) (R.I.=3117) by control rats #3 and 4. RELKHVE CONCENTRAUON (U0 PER M0 CR) 1 4.00 I 2.00 1 0.00 8.00 6.00 4.00 2.00 0.00 Figure 23 187 E P E P E D P I D D E l V P D D D l L L l J 1 l I I I I I I I I —I I 2 3 4 5 6 7 a 9 10 COLLECTION DAY Excretion of tetrahydroxy-pregnan-one (TMS) (R.I.=3127) by control rats #3 and 4. 188 .cm0 cofioomfiaou .Ommca mseummca. u < .mme umummcu wxoemh com mfisumcom cowuom—Foo mast: 4N Pzaflm m m H o AII x003 «moo IIIIIIIIIIIIIIIIIIII 6 «mos IIIIIIIIIIIIIIIIIIII m «mom IIIIIIIIIIIIIIIIIIII 4 fl««« IIIIIIIIIIIIIIIIIIII m «00¢ IIIIIIIIIIIIIIIIIIII N IIIIIIIIIIIIIIIIIIIIIII mumm mHmEmh . ”:3; < < < < w I.He emummqucmwflxoemH mews: g ocOEOmOuH cmwflxoeme mmwmo OucH » spasm mama HH< 189 removing the major source of estrogens through ovariectomy, Tamoxifen blocks the action of estrogen at the cellular level. For this experiment, daily 24—hour urine samples were collected from the six ovariectomized rats on the 14th through 18th day of saline treatment as described in Chapter V. Surgical removal of the ovaries causes a cessation of the estrous cycle. As with Tamoxifen treatment, the rat remains in an "anestrus" phase. However, the disrupting of the estrous cycle was not as complete for the ovariectomized rats as with the drug—treated animals. Ovariectomized Rat #3 went through estrus, diestrus, proestrus, and two more days of estrus during the urine collection period as displayed in Figure 25. Only Rats #1, 2, and 6 had five days of anestrus. Comparison of the urinary steroid metabolic GC-FID profiles for the Tamoxifen—treated and ovariectomized rats show that both types of treated rats display qualitatively more simplified profiles (Figure 26) as compared with the control rat urine profile. Figure 27 displays two gas chromatographic profiles. The top profile is from a Tamoxifen—treated rat and the lower profile is from a control rat in diestrus phase. Most of the peaks in the corticosteroid region (retention index 2900 to 3300) occur in both the drug-treated and control rat samples. The amount of metabolites excreted in the Tamoxifen—treated sample appears to be less than for the control sample. 190 .ommna uocmummocm .m .ommza monummcm .< .mumu commeouommum>o now maooonom comuooamoo meme: .mm muswmm m N H o AII x003 mmmmm IIIIIIIIIIIIIIIIIIII o mmmmm IIIIIIIIIIIIIIIIIIII m mmmmm IIIIIIIIIIIIIIIIIIII .0 mmmmm IIIIIIIIIIIIIIIIIIIII m mmmmm IIIIIIIIIIIIIIIIIIII N mumm mHmEmm .um:n< mmmmm IIIIIIIIIIIIIIIIIIII 2 83.58335 W ucmEumeB mcmflmm .\ mmwmo oucH w comuommaoo memo: Dump comma m mama Has 22 2.I 25 28 but Figure 26 191 3° 0110. III IMO 34 ICIL. Gas chromatographic profiles obtained for a Tamoxifen-treated rat (top profile) and an ovariectomized rat (bottom profile). 192 IIIII. JIM WI IIIIIIL «ILIWWIUIIMIIWIIUIIMAIMUIW 1 IL. IIIIIIIIII. ll I Figne 27 Gas chromatographic profiles obtained for a Tamoxifen-treated rat (top profile) and a normal,cycling female rat in the diestrus phase. I. 193 Indeed, calculation of the relative concentrations of the corticosteroid metabolites listed in the previous section indicated that the average amount of these compounds excreted was suppressed in the Tamoxifen—treated rats as seen in Tables 11 and 12. These tables were obtained with the GCSTAT program. The tetrahydroxy-pregnan—one isomer at retention index 3127 was not excreted at all by Tamoxifen—treated Rat #1. The large peak occurring to the left of the fludrocortisone peak in the drug—treated rat profile (Figure 27) was identified as cholic acid as described below. 3. Tamoxifen—treated and control adult, female rats bearing palpable DMBA—induced mammary carcinomas These two different animal models provided the most interesting information concerning the changes in steroid metabolism induced by Tamoxifen. In this experiment each rat served as its own control. Initially, both groups of tumor—bearing rats received saline injections for ten days as described in Chapter V. Over the following ten days, the control group (group 1) continued to receive saline injections and the other group Of six rats (group 2) received therapeutic dosages of Tamoxifen. The treatment schedule is outlined in Figure 28. Tamoxifen is very effective in causing the diminution in growth of DMBA— induced mammary tumors. At the onset of the saline 194 Table 11. Excretion of corticosteroid metabolites by Tamoxifen—treated rat #1. R.I. Name freq. %RSD Mean S.D. 3015 Trihydroxy- pregnan-one 5 20.0 2.0 i 0.4 3072 Tetrahydroxy— pregnan-one (MO-TMS) 5 20.0 3.8 i 0.7 3117 Tetrahydroxy- pregnan-one(TMS) 5 20.0 0.57 i 0.1 3127 Tetrahydroxy— pregnan—one(TMS) O 0.0 0.0 i 0.0 3146 Tetrahydroxy— pregnen-one (MO-TMS) 5 20.0 20.0 i 5.0 Table 12. Excretion of corticosteroid metabolites by Tamoxifen—treated rat #6. R.I. Name freq. %RSD Mean S.D. 3015 Trihydroxy— pregnan—one 5 20.0 6.0 i 1.0 3072 Tetrahydroxy— pregnan—one (MO—TMS) 5 20.0 4.1 i 1.0 3117 Tetrahydroxy— pregnan-one(TMS) 5 40.0 4.3 i 2.0 3127 Tetrahydroxy- pregnan—one(TMS) 4 70.0 2.1 i 2.0 3146 Tetrahydroxy— pregnen-one (MO—TMS) 5 30.0 23.0 i 7.0 195 .mumc cmummcu-o=cu 6cm Focpcoo newcmmn-coE:p toe upsumnom cowoumpfioo meat: mm 925mm o m w o m N o AII x003 omooozzaammmmmmmmmmmI I IIIIIIIIIIII NH zozaozamzzzzzzzzzzzz I IIIIIIIIIIIIIII II 3 amaoooomommamaanoamm IIIIIIIIIIIIIIIIIIIIIIIII IIIoH amommmmoommomaammaom IIIIIIIIIIIIIIIIIIIIIIIIIIII o aoaoomaoomooommaomoo IIIIIIIIIIIIIIIIIIIIIIIIIIII m ooaooooooammammmmaon IIIIIIIIIIIIIIIIIIIIIIIIIIII n L Dawsommce CONHxOEmH zammzzoomomammmmmoom IIIIIIIIIIIIIIIIIIIIIIIIIIII o azzaaazzzzzaammaaooa ............................ m max ammmmmooamoawmmmnaoo IIIIIIIIIIIIIIIIIIIIIIIIIIII m mamemm aaammaaoaaomaaoomgo ............................ m 9:88 mammmmmmmmmmmmmmooom IIIIIIIIIIIIIIIIIIIIIIIIIIII m 385:. mmmmmommammammmaammaV IIIIIII 9 IIIIIIIIIIIIIIIIIIIII 21* .328va Ocmeommfi. mamamm \ m camow 4 composes; V was awoa a» .55: 39562 at... be 9:5 a 3 88 33 H 96069 w/MOO wflo 09 wdpfi cowwu mm rvaEmi $063. 196 treatment period all of the rats had at least one and as many as six tumors. The mean tumor diameter for group 1 was 17 mm and the tumors carried by group 2 had a mean diameter of 16 mm. After ten days of saline treatment the mean tumor diameter was 18 mm for group 1 and 18 for group 2. The second group then received Tamoxifen injections (75ug/100g body weight) for ten days, and the mean tumor diameter decreased to 11 mm. The control group (group 1) which continued to receive only saline injections had a mean tumor diameter of 23 mm at the end of this final urine collection period. Thus, the average increase in tumor diameter during the first ten day urine collection period was 5.5% for group 1 and 13.0% for group 2. For the next ten days, group 1 continued to receive only saline injections and the average tumor size increased by 24%. Group 2, receiving daily injections of Tamoxifen experienced a decrease in average tumor diameter of 26%. Both sets of tumor—bearing rats displayed irregular estrous cycles, probably due to the influence of the tumor burden on the endocrine system. As with the normal, cycling female rats treated with Tamoxifen, the tumor- bearing rats treated with the drug exhibited a diminution in the excretion of corticosteroid metabolites. Figure 29 displays gas chromatograms of derivatized urine samples from tumor—bearing rat #9 obtained prior to Tamoxifen treatment (top profile) and after 4 days of Tamoxifen 197 on ..ZD fix. 3 {xx/liq .CfifibonvucmESmmcu cmwwxosmp ocecsu can Amouvou Lowcq not ocwcmmnuco533 m Eocm vasomuno mmpveoca owcamcaopmeoczo moo an an E on _.__.—-— On ma ___.-._——- mu on “a.” - _ __~. ___——___-......_.- om magma 4%; «jfifieél{$5 >2 on 32:32. «a 198 treatment (lower profile). Most of the metabolites in the region of the chromatogram between hydrocarbons 30 and 32 appear to be excreted in lower amounts during Tamoxifen treatment. Calculating the relative concentrations of these metabolites confirm that drug treatment suppresses the excretion of some species. For example, Table 13 compares the mean excretion of four corticosteroid metabolites from tumor—bearing Rat #6 (control) for urine collection days 4 through 8 and 16 through 20. This control animal received only saline injections over this 20 day period. When one overlaps the mean excretion (for collection days 4 through 8) :tthe standard deviation for trihydroxy-pregnan-one for days with the mean excretion the standard deviation for the same compound for collection days 16 through 20 there is a 100% overlap. This GCSTAT feature allows the chromatographer to conduct a gross examination of two sets of gas chromatographic data. For tumor-bearing rat #6 there is not a significant change in the excretion of trihydroxy—pregnan—one, or of the tetrahydroxy—pregnan—one isomers. Table 14 displays the result of using the OVERLAP routine to compare the excretion of the corticosteroid metabolites from tumor-bearing rat #8 during five days of the saline treatment period and five days of the Tamoxifen treatment period. Using the same technique of overlapping the mean concentration of these corticosteroid metabolites 199 Table 13. Comparison of the mean concentration of certain corticosteroid metabolites in 24-hour urine collections on days 4-8 and 16—20 from tumor-bearing rat #6 (control). Urine Urine Collection Collection Days 4-8 Days 16-20 (Saline (Saline Treated) Treated) R.I Name Mean S.D. % Overlap Mean S.D. 3015 Trihydroxy- Pregnan-one 4.3 1:3.0 100 3 1 311 0 3117 Tetrahydroxy— Pregnan—one(TMS) 6.5 3:3.0 70 4 6 3:3 O 3127 Tetrahydroxy— Pregnan—one(TMS) 8.7 i7.0 100 5 4 i4 0 3146 Tetrahydroxy— Pregnen—one (MO-TMS) 47.0 i20.0 100 37.0 1‘9 0 200 Table 14. Comparison of the mean concentration of certain corticosteroid metabolites in 24—hour urine collections on days 4-8 and 16—20 from tumor-bearing rat #8. Urine Urine Collection Collection Days 4—8 Days 16-20 (Saline (Tamoxifen Treated) Treated) R.I Name Mean S.D. % Overlap Mean S.D. 3015 Trihydroxy- Pregnan—one 4.8 i 0 6 0 1.1 i 1 0 3117 Tetrahydroxy— Pregnan—one(TMS) 6.3 i 3 0 0 0.0 i 0 O 3127 Tetrahydroxy— Pregnan—one(TMS) 7.9 i 3.0 O 1.8 i 0.6 3146 Tetrahydroxy— Pregnen-one (MO—TMS) 27.0 i 6.0 30 18.0 H- 0‘ .0 201 excreted during urine collection days 4-8 (saline treatment period) t the standard deviation with the mean concentration for urine collection days 16—20 (Tamoxifen treatment days 6—10) 2 the standard deviation one observes that there is no overlap for trihydroxy-pregnan—one and the two tetrahydroxy-pregnan—one isomers at retention indices 3117 and 3127. This provides a rough indication that the excretion of these metabolites changes when the animals receive Tamoxifen injections. Figure 30 shows the excretion of trihydroxy—pregnan— one for tumor—bearing rat #6 (control) and for tumor- bearing rat #7 (which received Tamoxifen injections from day 10 to day 20). Figure 31 compares the excretion of trihydroxy—pregnan—one from tumor—bearing rats #7 and #8. Both of these rats received the drug from day 10 to day 20. Group means i S.D. are shown below in Table 15. RELATIVE CONCENTRATION (UG PER M0 CR) 202 25.00 "v— T unh- P 20aI—~ .1... 15.00 P .J. E E D Inna—— 500—«- P E z/g/9\e .. E Start Tamoxifen D .- P .4 P/E U P/E 000 I I I .1 I I I I I I I I I_J I I I I I_J ' ”1'7 I I' I I I I I ITI I I I I I I I I*n o 1 2 3 4 s a 7 a 9 "311121314151617181920 COLLECTlON DAY Figne 3O Excretion of trihydroxy-pregnan-one (R.I.=3015) by tumor-bearing rats #6 (control) and 7. RELATIVE CONCENTRATION (00 PER M0 CR) 203 25.00 ‘17‘ " P 20.00 ~— woo-m P x 10.00 .._ 1r- D 4» D D D 5.00 "‘"‘ M fa ~~ Start D .. D D P E D Tamoxifen D x D P I I I I I I I I III I I I III I 03° I I I I I I I I I I I I I I I I o I 2 3 4 s a 7 a 9 10 II 12 13 I4 Is I: I7 Is In 20 COLLECTION DAY Figure 31 Excretion of trihydroxy-pregnan-one (R.I.=3015) by tumor-bearing rats #7 and 8. 204 Table 15. Excretion of trihydroxy-pregnan—one by tumor— bearing rats by tumor—bearing rats #6, #7, #8 and #9. Urine Urine Collection Collection Rat # Days 4 - 8 Days 16 - 20 (Saline (Tamoxifen Treated) Treated) Mean S.D. Mean S.D. 6 (control) 6.2 3.0 3.1 1.0 7 15.0 5.0 5.7 1.0a 8 4.8 0.6 1.1 1.0a 9 4.2 1.0 1.9 0.2a a Means are statistically different by Student's t-test (p < 0.01). Applying Student's t-test one finds that the excretion of trihydroxy—pregnan-one during days 4—8 and 16-20 are not statistically different (p < 0.1). On the other hand, the excretion level decreases significantly during Tamoxifen treatment for rats #7, 8 and 9 (p < 0.01). Both of the trimethylsilylated isomers of tetrahydroxy-pregnan-one (R.I. = 3117, 3127) are excreted in significantly lesser amounts by the drug—treated tumor— bearing rats (p < 0.05) as seen below in Tables 16 and 17. The daily excretion of these metabolites by rats #6, 7 and 8 is displayed in figures 32, 33, 34 and 35. The tetrahydroxy-pregnen—one isomer eluting at R.I. = 3146 only displays suppressed excretion (p < 0.05) during Table 16. Table 17. 205 Excretion of tetrahydroxy—pregnan—one (TMS), R.I. = 3117, by tumor-bearing rats #6, 7, 8 and 9. Urine Urine Collection Collection Rat # Days 4—8 Days 16-20 (Saline (Tamoxifen Treated) Treated) Mean S.D. Mean S.D. 6 (control) 6.5:: 3.0 4.6 i 3.0 7 9.5 i 4.0 2.7 i 1.0a 8 6.3 t 3.0 0.0 i 0.0a 9 2.8 i 1.0 0.64 i 1.03 aMeans are significantly different by Student's t—test (p < 0.05). Excretion of tetrahydroxy-pregnan-one (TMS), R.I. = 3127, by tumor-bearing rats #6, 7, 8 and 9. Urine Urine Collection Collection Rat # Days 4-8 Days 16—20 (Saline (Tamoxifen Treated) Treated) Mean S.D. Mean S.D. 6 (control) 9.9 16.0 5.4 i 4.0 7 15.0 i8.0 2.0 i 2.03 8 7.9 1“3.0 1.8 i 0.63 9 2.2 i 1.0 0.76 i 0.33 aMeans are significantly different by Student's t-test (p < 0.05). RELKHVE CONCENTRAHON (UG PER M0 CR) 206 15.00 —r‘ 10.00 .3- P/E 1&- " P/E 5.00 -I- D D 1" Start U “ Tamoxifen . D1) D1) D i . 4 I I I I J I I I I I I I I I14 03° I-I I I I I I I I I I r I I I I‘I‘I‘T‘T o I 2 3 4 5 a 7 s 9 uaIItzIJI4I5IaI7IsI920 DAY Figne 32 Excretion of tetramdroxy-pregnan-One (TMS)(R.I.=3117) by tumor-bearing rats #6 (control) and 8. RELATIVE CONCENTRATION (00 PER M0 CR) 207 10.00 ~7- 14.00 -~- 12.00 -"- 10.00 ~— 8.00 ~~ 6.00 ~— 4.00 —— 2.00 -- E Start D Tamoxifen D 000 I O .4 p u..— “ a... ‘ ..— U! .4. O Figne 33 Excretion (TMS)(R.I. #7 and 8. I I I I I I I I I 4F**‘T‘+‘T I I I I I I I I I a 9 "311121314151617181920 COLLECTION DAY ‘1 of tetrahydroxy-pregnan-one =3117) bytumor-bearing rats RELATIVE CONCENTRATION (00 PER MG CR) 208 zauI—— awn-«- IamI-— + . .m— PE 1000 / P/E fir- aaI—— U D D “ Start Tamoxifen D an ‘ D 000 I I l I I I I I-I I I I I I I I I I J I - I I I TTI I I 1 I I I TTI I I I I I,I'1 01234567aoIOIIIzIaI4IsIOI7Is‘1920 COUEEDONIMW Figne 34 Excretion of tetrahydroxy-pregnan-one (TMS)(R.I.=3127) by tumor-bearing rats #6 (control) and 7. RELATIVE CONCENTRATION (U0 PER M0 CR) 209 m.m .5?— 4)— 20.00 -~'- 15.00 ~— 10.00 -I-- 5.00 ~- D Start . d- . D . Tamox1fen D D 'QD ”I II . P I I I I I I J I l J I I I 000 1 l 1 l I l L ‘ T I j r I I j T FT T I r r 1 2 7 8 0 10 11 12 13 14 15 10 17 18 10 20 COLLECTION DAY Figne~35 Excretion of tetrahydroxy-pregnan-one (TMS)(R.I.=3127) by tumor-bearing rats #7 and 8. 210 the drug treatment period for tumor—bearing rat #7 as seen in figures 36 and 37. Tumor—bearing rats #8 shows a statistically significant change only at the p < 0.10 confidence level, as seen below in Table 18. Table 18. Excretion of tetrahydroxy-pregnen-one (MO-TMS), R.I. = 3146, by tumor—bearing rats #6, 7, 8 and 9. Urine Urine Collection Collection Rat # Days 4 - 8 Days 16 - 20 (Saline (Tamoxifen Treated Treated) Mean S.D. Mean S.D. 6 (control) 47.0 t 20.0 37.0 t 9.0 7 44.0 2 10.0 23.0 t 9.0a 8 27.0 t 6.0 18.0 t 6.0b 9 7.1 t 2.0 8.8 t 2.0 aMeans are significantly different by Student's t—test (p < 0.05). bMeans are significantly different by Student's t—test (p < 0.10). Finally, the methyloxime—trimethylsilyl derivative of tetrahydroxy—pregnan-one eluting at R.I. = 3072 displayed a significant change in excretion during drug treatment only for tumor-bearing rat #9, as seen below in Table 19. The excretion of this isomer by rats #6, 7 and 8 is shown in figures 38 and 39. RELATIVE CONCENTRATION (UG PER M6 CR) 211 unoa~~ 4.. anmI~— 60.00 ~4— __ P/E 40.00 —~ ME ‘I‘ 20.00 “I— Start -r Tamoxifen i _ I I I I I I I I I I I I .l Lil I I I I I °9° I I I I I I I I 1 I I I I*T I I I I I 1 o I 2 3 4 s s 7 a o uIIIIzIaI4I5IsI7IaIszo COLLECTION DAY Figne 36 Excretion of tetrahydroxy-pregnen-one (R.I.=3146) by (control) and 7. tumor-bearing rats #6 RELATIVE CONCENTRATION (UG PER M0 CR) 212 70.00 me P “I 00.00 —I- m- 50.00 ~9— .. P E E 40.00 ->— w.m --I._. 20.00 -—- 10,00 __ Start Tamoxifen -- i 000 1 1 1 l I 1 L l 1 l 1 1 I I I J I l [J ' IIIIIIITIIEIIIIIEIII 012 5 4 5 0 7 0 91011121314151017101920 COLLECTION DAY Figure 37. Excretion of tetrahydroxy—pregnen-one (R.I.=3146) by tumor—bearing rats #7 and 8. RELATIVE CONCENTRATION (00 PER M0 CR) 213 25m>~r JI. 20m:-— .5... l " P/E v " E IaaI—— IE amI—— " Start Tamoxifen D D 4" 4 000 L,l I I I I I I I .l I I I I I I14 I I I ° T ITI I I I I I I I I I I I I I T’I T I 0 I 2 3 4 5 5‘7 5 o "311121314151617181920 COLLECTION DAY Figne 38 Excretion of tetrahydroxy-pregnan-one (MO-TMS) (R.I.=3072) by tumor- bearing rats #6 (control) and 7. RELATIVE CONCENTRATION (UG PER M0 CR) 214 25.00—0— «I- 20.00-‘— :1:- ch- 15.00."- ‘I- Infilfi- .I- 5.00-- Start " Tamoxifen l" 4 000 I I I I I I I l I I I I I I I I I I 1_J ' IIIIIIIIITIIIIIIIIIO 012 3 4 5 0 7 8 0'101112131415161718‘1920 COLLECTION DAY Figue 39 Excretion of tetrahydroxy-pregnan—one (MO-TMS) (R.I.=3072) by tumor- bearing rats #7 and 8. 215 Table 19. Excretion of tetrahydroxy—pregnan—one (MO—TMS), R.I. = 3072, by tumor-bearing rats #6, 7, 8 and 9. Urine Urine Collection Collection Rat # Days 4 - 8 Days 16 — 20 (Saline (Tamoxifen Treated) Treated) Mean S.D. Mean S.D. 6 (control) 16.0 t 5.0 12.0 t 3.0 7 10.0:I:2.0 7.2:I: 4.0 8 11.0 t 2.0 14.0 t 5.0 9 3.3 1 1.0 4.8:h 0.7a aMeans significantly different by Student's t—test (p < 0.05). A compound eluting at retention index 3344 displayed an unusual excretion behavior from tumor-bearing rat #9. During the drug treatment period the urinary concentration of this compound rose from less than 20 ug per mg creatinine on day 11 to over 100 ug per mg creatinine on day 15. This peak is indicated with an asterisk in Figure 28. This excretion behavior is displayed in Figure 40. Mass spectral analysis of this compound indicated that the molecular ion is at m/z = 696. Other characteristic ions include M+-15 (m/z = 681), M+—90 (m/z = 606), M+—90—9O (m/z e 516), M+-90—90—9O (m/z = 426). An peak is present at m/z = 253, corresponding to an ion representing bare RELATIVE CONCENTRATION (UG PER M0 CR) 216 10000 'r- [I I: 75.00 ~— 50.00 —— “' START 25.00 __ D TAMOXIFEN D D m- D P D P E -- D I I I I I I I J I I I I I _I °'°° I I I I I I I I I I I I I I I Figure 40. Excretion of cholic acid (R.I.=3344) by tumor—bearing rat#9 217 steroid ring. This ion information suggests that this compound is cholic acid. This was confirmed by exact mass measurement using high resolution GC/MS (150). The unusual excretion pattern for cholic acid was Observed only for tumor-bearing rat #9. This metabolite did appear in the urine of other tumor-bearing rats, but there was no statistically significant difference between drug—treated and control urine samples. B. Discussion The significant change in excretion of corticosteroid metabolites observed with the tumor-bearing rats treated with Tamoxifen may be caused by cellular changes in the adrenal cortex noted previously by Lullman and Lullman— Rauch (95). These authors noted that cells in the zona fasciculata and zona reticularis degenerated after the rat had received large doses of Tamoxifen (100 - 130 mg/kg body weight). Subtle changes in the plasma concentration of cortisol in postmenopausal women with advanced breast cancer treated with Tamoxifen was noted by Levin et al. (75). Total urinary excretion of cortisol metabolites decreased by 13%. The ratio of excreted tetrahydrocortisol to tetrahydroxortisone decreased by 28%. These changes in steroid excretion are probably due to the systemic effect of Tamoxifen, rather than by the antiestrogenic action of 218 the drug on the tumor cells as the changes in cortisol metabolism occurred in estrogen receptor-negative patients (who do not respond to this therapy) as well. An extension of the metabolic profiling analysis discussed in this dissertation would be to analyze the urinary steroid profiles of human breast cancer patients before and after Tamoxifen treatment. Human urinary steroid metabolites are well characterized. Unlike urinary rat steroids, reference human steroids are commercially available. A MSSMET urinary steroid library is in place at the MSU Mass Spectrometry Facility. This project would be the logical extension of the present study. A major disadvantage inherent in the analysis of 24— hour urine collections from individual rats is the low urine volumes (3 — 30 ml) generated daily. Individual rat urine samples contain concentrations of estrogen and progesterone metabolites that are too low to be detected with repetitive scanning GC—MS. Many of the literature studies cited above used large volumes of pooled rat urine ( > 100ml) to obtain metabolite levels high enough to allow for steroid identification. In the present study, pooled urine samples were analyzed in order to identify urinary steroid for the MSSMET library. However, the individual rats in the five experimental models served as their own controls, so that changes in steroid excretion over time could be determined for each rat. If those 219 samples were pooled, the individual variations possibly would be averaged out. There are other disadvantages when analyzing rat urinary steroid profiles including: the lack of reference steroids, interfering nonsteroidal compounds eluting in the chromatographic region of interest, and enterohepatic circulation of steroids in the rat leading to microbial metabolism of steroid metabolites. A major benefit of this experiment was the ability to apply the GCMET GC—Only metabolic profiling software to the gas chromatographic data. As a new program, the "bugs" in the system had to be worked out. GCMET analysis of the rat urinary steroid profiles provided an excellecnt opportunity to examine the utility of this technique. Narrow-bore capillary GC columns provided very reproducible retention times and peak areas. One derivatized urine sample was injected five times. Components with relative concentrations greater than 1 ug per mg creatinine displayed relative standard deviations which were generally < 0.5%. Exceptions included peaks which occurred as shoulders on larger peaks, and weren't identified by the integrator. Components with concentrations less than 1 ug per mg creatinine displayed larger relative standard deviations again indicating that 220 more concentrated, pooled samples would yield better statistical information. In summary, Tamoxifen was shown to have an effect on the excretion of certain corticosteroids in the urine of tumor—bearing rats. These metabolites are quantitatively the most important urinary steroids and they were concentrated enough to be detected in 24—hour urine collections form individual rats. The GCMET GC—Only metabolic profiling routines facilitate the analysis of urinary steroid profiles, eliminating countless hours of manual calculation of retention indices and relative concentrations from integrator output. APPENDIX I APPENDIX I. A MSSMET library of urinary rat steroids *ENTRYTYPE:MSSMET NAM:*3000 TMA:38,15 DIN:71,1 CIN:71, 1000, 85, 625 *ENTRYTYPEzMSSMET NAM:*2000 TMR:.199 DIN:71,1 CIN:71,1000,85,625 *ENTRYTYPEzMSSMET NAM:*2200 TMR:.325 DIN:71,1 CIN:71,1000,85,625 *ENTRYTYPE:MSSMET NAM:*2400 TMR:.479 DIN:71,1 CIN:71,1000,85,625 *ENTRYTYPE:MSSMET NAM:*2600 TMR:.651 DIN:71,1 CIN:71,1000,85,625 *ENTRYTYPE:MSSMET NAM:*2800 TMR:.823 DIN:71,1 CIN:71,1000,85,625 *ENTRYTYPE:MSSMET NAM:*3000 TMR:1.00 DIN:71,1 CIN:71,1000,85,625 *ENTRYTYPEzMSSMET NAM:*3200 TMR:1.161 DIN:71,1 CIN:71,1000,85,625 *ENTRYTYPEzMSSMET 0PN:/TM 3 0PN:/TY 3 OPNz/LU 1 0PN:/RX 15 0PN:/RQ 16 0PN:/MH 1000 0PN:/CR o *ENTRYTYPE:MSSMET NAM:& FLUDROCORTISONE TMI:3360 DIN:379,1 CIN:379,999,623,200,533,300 221 222 *ENTRYTYPE:MSSMET OPN:/TY 4 OPNz/MH 50 0PN:/TH 50 OPN:/WM 192 OPN:/CF 40 OPNz/CQ 41 *ENTRYTYPEzMSSMET NAM:* 1290 PREGNENE-DIONE(PROCESTERONE) TMI:2355 DIN:357,1 CIN:267,467,339,324,357,999,372,771 *ENTRYTYPE:MSSMET NAM:* 1340 UNR31 TMI:2453 DIN:576,1 CIN:576,999,561,35,471,55 *ENTRYTYPE:MSSMET NAM:* 1350 ESTRATRIENE-DIOL-ONE TMI:2454 DIN:459,1 CIN:4S9,999,369,428,412,891 *ENTRYTYPEzMSSMET NAM:* 1360 UNR32 TMI:2469 DIN:574,1 CIN:499,624,574,999,589,698 *ENTRYTYPE:MSSMET NAM:* 1370 PREGNENE-OL-DIONE TMI:2468 DIN:387,1 CIN:297,62,371,31,387,999,402,412 *ENTRYTYPEzMSSMET NAM:* 1400 UNR34 TMI:2486 DIN:410,1 CIN:410,999,470,556,500,267 *ENTRYTYPE:MSSMET NAM:* 1410 UNR35 TMI:2521 DIN:546,1 CIN:456,42,531,30,546,999 *ENTRYTYPE:MSSMET NAM:* 1420 PREGNENE-DIOL-DIONE? TMI:2542 DIN:4S9,1 CIN:459,999,475,123,490,278 *ENTRYTYPE:MSSMET NAM:* 1440 UNR36 TMI:2580 DIN:470,1 CIN:470,999,439,44,380,50 *ENTRYTYPE:MSSMET NAM:* 1470 ESTRADIOL TMI:2588 223 DIN:416,1 CIN:326,65,401,206,416,999 *ENTRYTYPE:MSSMET NAM:* 1500 UNR40 TMI:2613 DIN:592,1 CIN:558,68,577,71,592,999 *ENTRYTYPE:MSSMET NAM:* 1520 UNR42 TMI:2639 DIN:622,1 CIN:442,227,607,185,622,999 *ENTRYTYPE:MSSMET NAM:* 1525 UNR43 TMI:2649 DIN:495,1 CIN:470,58,495,999,510,162 *ENTRYTYPE:MSSMET NAM:* 1570 UNR47 TMI:2730 DIN:420,1 CIN:369,95,405,228,420,999 *ENTRYTYPEzMSSMET NAM:* 1580 SALPHA-ANDROSTAN-l7ALPHA-METHYL-3BETA,17BETA-DIOL(IS) TMI:2750 DIN:4435,1 CIN:345,172,3600,34l,435,999,450,234 ;*ENTRYTYPE:MSSMET ; NAM:* 1595 DIHYDROXY-PREGNENE-0NE-3BETA, ; 16ALPHA-DIHYDROXY-5DELTA-P-20-ONE(FRU) ; TMI:2752 ; DIN:476 ; CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 1600 DIHYDROXY PREGNANE-ONE ; TMI:2759 ; DIN:507 ; CIN: *ENTRYTYPEzMSSMET NAM:* 1620 Salpha-PREGNANE-BBETA,20BETA-DIOL TMI:2787 DIN:470,1 CIN:470,999,455,208,439,22 *ENTRYTYPEzMSSMET NAM:* 1630 UNRSO TMI:2795 DIN:442,1 CIN:337,27,427,173,442,999 *ENTRYTYPE:MSSMET NAM:* 1650 PREGNANEDIOL TMI:2852 DIN:470,1 CIN:439,25,455,202 *ENTRYTYPE:MSSMET NAM:* 1660 UNR52 224 TMI:2850 DIN:469,1 CIN:274,3,289,4,454,181,469,999 ;*ENTRYTYPE:MSSMET ; NAM:* 1700 DIHYDROXY-SALPHA-ANDROSTANE-17-ONE (FRU) ; TMI:2898 ; DIN:450 ; CIN: *ENTRYTYPE:MSSMET NAM:* 1710 UNRSS TMI:2930 DIN:367,1 CIN:367,999,457,242 *ENTRYTYPE:MSSMET NAM:* 1720 ANDROSTADIENE-DIOL-ONE- (6ALPHA,17BETA-DIHYDROXY-ANDROSTADIENE-3-0NE) TMI:2934 DIN:446,1 CIN:341,203,356,214,431,307,446,999 *ENTRYTYPEzMSSMET NAM:* 1730 DIHYDROXY-PREGNANE-DIONE (FRU) TMI:2936 DIN:492 CIN: *ENTRYTYPE:MSSMET NAM:* 1740 TRIHYDROXY-ANDROSTANE-ONE (6BETA,16ALPHA,17BETA TRIHYDROXY-SALPHA-l-20-ONE) TMI:2969 DIN:523,1 CIN:433,250,523,999,538,500 *ENTRYTYPE:MSSMET NAM:* 1770 TRIHYDROXY-PREGNANE-ONE (e.g., THB) TMI:3009 DIN:566,1 CIN:461,581,476,949,551,602,566,999 *ENTRYTYPEzMSSMET NAM:* 1780 3ALPHA,lSALPHA,21-TRIHYDROXY-5ALPHA PREGNANE-11,20-DIONE TMI:3012 DIN:580,1 CIN:400,112,490,226,565,47,580,999 *ENTRYTYPE:MSSMET NAM:* 1790 TRIHYDROXY-PREGNANE-TRIONE TMI:3060 DIN:533,1 CIN:443,753,533,999,623,916 *ENTRYTYPEzMSSMET NAM:* 1800 PREGNENE-PENTOL TMI:3062 DIN:621,1 CIN:621,999,531,294,636,471,726,588 *ENTRYTYPE:MSSMET NAM:* 1810 TRIHYDROXY-PREGNANE-ONE TMI:3063 DIN:595,1 CIN:415,152,505,135,580,333,595,999 225 *ENTRYTYPE:MSSMET NAM:* 1820 PREGNENE-TRIOL-ONE TMI:3074 DIN:562,1 CIN:503,358,562,999,593,295 *ENTRYTYPE:MSSMET NAM:* 1830 TETRAHYDROXY-PREGNANE-0NE (FRU) TMI:3074 DIN:562,1 CIN:562,999,593,358,683,96 *ENTRYTYPEzMSSMET NAM:* 1831 TRIHYDROXY-PREGNADIENE-DIONE (FRU) TMI:3076 DIN:580 CIN: *ENTRYTYPEzMSSMET NAM:* 1840 TETRAHYDROXY-PREGNANE-ONE (FRU) TMI:3076 DIN:593,1 CIN:503,413,593,999,683,555 *ENTRYTYPE:MSSMET NAM:* 1850 UNR58 TMI:3062 DIN:428,1 CIN:398,412,413,297,428,999 *ENTRYTYPEzMSSMET NAM:* 1860 UNR59 TMI:3086 DIN:534,1 CIN:444,999,503,749,534,999 *ENTRYTYPE:MSSMET NAM:* 1870 PREGNADIENE-TRIOL-ONE TMI:3079 DIN:562,1 CIN:472,733,547,27,562,999 *ENTRYTYPE:MSSMET NAM:* 1880 TRIHYDROXY-PREGNANE-DIONE TMI:3100 DIN:578,1 CIN:578,999,594,344,519,299,609,304 *ENTRYTYPE:MSSMET NAM:* 1890 TETRAHYDROXY-PREGNENE-ONE (FRU) TMI:3103 DIN:472,1 CIN:472,999,562,961,652,723 *ENTRYTYPE:MSSMET NAM:* 1900 UNR6O TMI:3105 DIN:686,1 CIN:596,611,671,348,686,999 ;*ENTRYTYPE:MSSMET ; NAM:* 1901 TETRAHYDROXY-PREGNENE-0NE (FRU) ; TMI:3109 ; DIN: ; CIN:652 226 ;*ENTRYTYPE:MSSMET ; NAM:* 1902 TETRAHYDROXY-PREGNANE-ONE (FRU) ; TMI:3113 ; DIN: ; CIN:654 *ENTRYTYPE:MSSMET NAM:* 1910 PREGNENE-TRIOL-DIONE TMI:3133 DIN:578,1 CIN:398,91,488,165,563,9l,578,999 *ENTRYTYPE:MSSMET NAM:* 1920 UNR61 TMI:3119 DIN:486,1 CIN:486,999,561,40,576,218 *ENTRYTYPE:MSSMET NAM:* 1930 UNR62 TMI:3146 DIN:639,1 CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 1931 TETRAHYDROXY-PREGNANE-DIONE ; TMI:3146 ; DIN: ; CIN:580 *ENTRYTYPEzMSSMET NAM:* 1940 TETRAHYDROXY-PREGNADIENE-ONE TMI:3145 ’ DIN:650,1 CIN:470,517,560,738,650,999 *ENTRYTYPEzMSSMET NAM:* 1950 TRIHYDROXY-ANDROSTANE-ONE TMI:3131 DIN:655,1 CIN:475,404,550,152,565,459,655,999 *ENTRYTYPE:MSSMET NAM:* 1960 TETRAHYDROXY-PREGNANE-ONE TMI:3129 DIN:564,1 CIN:549,469,564,999,639,206,654,472 *ENTRYTYPE:MSSMET NAM:* 1970 TRIHYDROXY-PREGNENE-ONE TMI:3147 DIN:562 CIN: *ENTRYTYPE:MSSMET NAM:* 1980 TRIHYDROXY-ANDROSTENE-ONE TMI:3157 DIN:295,1 CIN:295,999,385,77,475,12,565,369 *ENTRYTYPE:MSSMET NAM:* 1990 TETRAHYDROXY-PREGNENE-ONE TMI:3149 DIN:591,1 CIN:591,999,650,999,666,218,681,785 227 *ENTRYTYPEzMSSMET NAM:* 2000 TETRAHYDROXY-PREGNANE-0NE TMI:3166 DIN: CIN:654 *ENTRYTYPE:MSSMET NAM:* 2001 TRIHYDROXY-ANDROSTANE-ONE TMI:3171 DIN:565,1 CIN:550, 522,565,999,640,696,655,429 *ENTRYTYPE:MSSMET NAM* :2040 DIHYDROXY-PREGNENE-ONE TMI:3224 DINN:476,1 CIN:371,87,386,185,461,105,476,999 *ENTRYTYPE:MSSMET NAM:* 2060 TRIHYDROXY-PREGNANE-ONE TMI:3229 DIN:551,1 CIN:461,369,476,578,551,999,566,414 *ENTRYTYPE:MSSMET NAM:* 2070 TRIHYDROXY-PREGNANE-ONE TMI:3233 DIN:476,1 CIN:386,852,476,999,566,221 *ENTRYTYPEzMSSMET NAM:* 2080 TETRAHYDROXY-PREGNENE-DIONE TMI:3233 DIN:695,1 CIN:515,295,605,225,695,999 *ENTRYTYPE:MSSMET NAM:* 2100 TETRAHYDROXY-ANDROSTENE-0NE TMI:3259 DIN:653,1 CIN:473,476,563,783,653,999 ;*ENTRYTYPE:MSSMET ; NAM:* 2101 PREGNANE-PENTOL ; TMI:3270 ; DIN: ; CIN:728, *ENTRYTYPEzMSSMET NAM:* 2110 TRIHYDROXY-PREGNANE-DIONE TMI:3293 DIN:490,1 CIN:490,999,565,297,580,466 *ENTRYTYPE:MSSMET NAM:* 2120 TETRAHYDROXY-ANDROSTENE-ONE TMI:3328 DIN:516,1 CIN:501,98,516,99,606,79,696,9 ;*ENTRYTYPE:MSSMET ; NAM:* 2121 PREGNANE-PENTOL ; TMI:3307 ; DIN: ; CIN:728 228 *ENTRYTYPEzMSSMET NAM:* 2150 TRIHYDROXY-ANDROSTENE-ONE TMI:3341 DIN:565,1 CIN:385,140,475,355,565,999 *ENTRYTYPE:MSSMET NAM:* 2160 TETRAHYDROXY-PREGNENE-ONE TMI:3332 DIN:578,1 CIN:563,351,578,999 *ENTRYTYPEzMSSMET NAM:* 2180 TETRAHYDROXY-PREGNENE-ONE TMI:3340 DIN:681,1 CIN:501,133,591,90,681,999 *ENTRYTYPEzMSSMET NAM:* 2200 TETRAHYDROXY-PREGNANE-ONE TMI:3334 DIN:682 CIN:592,130,607,187,682,999,697,25 ;*ENTRYTYPE:MSSMET ; NAM:* 2212 TETRAHYDROXY-PREGNANE ; TMI:3376 ; DIN: ; CIN:640 ;*ENTRYTYPE:MSSMET ; NAM:* 2213 TRIHYDROXY-PREGNENE-DIONE ; TMI:3376 ; DIN: ; CIN:607 *ENTRYTYPEzMSSMET NAM:* 2250 TRIHYDROXY-PREGNANE-DIONE TMI:3420 DIN:607,999 CIN:427,189,592,745,607,999,682,638 *ENTRYTYPE:MSSMET NAM:* 2260 CHOLESTERYL BUTYRATE TMI:3425 DIN:368,1 CIN:247,176,260,148,353,185,368,999 ;*ENTRYTYPE:MSSMET NAM:* 2310 4-ANDROSTEN-6BETA-OL-3,l7-DIONE TMI: DIN: CIN: *ENTRYTYPEzMSSMET NAM:* 2320 2ALPHA-HYDROXY-4DELTA-P-3,20,DIONE TMI: DIN: CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 2330 BBETA,118ETA,ZOBETA,21-TETRAHYDROXY-5ALPHA-P ; TMI: U. U. U. .0 .0 U. U. .0 .0 ° DIN: ° CIN: 229 ;*ENTRYTYPE:MSSMET ; NAM:* 2340 3BETA-HYDROXY-5BETA-P-ZO-ONE ; TMI: ; DIN: ; CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 2350 3BETA-HYDROXY-5DELTA-P-20-ONE ; TMI: ; DIN: ; CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 2360 3ALPHA,17ALPHA,DIHYDROXY-SALPHA-P-ZO-ONE ; TMI: ; DIN: ; CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 2370 16ALPHA,17BETA-DIHYDROXY-4DELTA-A-3-ONE ; TMI: ; DIN: ; CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 2380 3ALPHA,17ALPHA-DIHYDROXY-SALPHA-P-ZO-ONE ; TMI: ; DIN: ; CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 2390 5ALPHA-A-3ALPHA,17BETA,DIOL ; TMI: ; DIN: ; CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 2400 3ALPHA,llBETA,21-TRIHYDROXY—SALPHA—P-20-ONE ; TMI: ; DIN: ; CIN: ;*ENTRYTYPE:MSSMET ; NAM:* 2410 3BETA,11BETA-DIHYDROXY-SALPHA-A-17-ONE ; TMI: ; DIN: ; CIN: LIST OF REFERENCES 10. 11. 12. 13. 14. 15. 16. 17. LIST OF REFERENCES J.F. Holland, J.J. Leary and C.C. Sweeley, J. Chromatogr., 379 (1986) 3. J.F. Holland, C.C. Enke, J.T. Stults, J. Pinkston, B. Newcome, J. Allison and J.T. Watson, Anal. Chem., 55 (1983) 997A. C.H.L. Shackleton, J. Chromatogr., 379 (1986) 91. R.J. Williams, Biochemical Institute Studies IV. Individual Metabolic Patterns and Human Disease: An Exploratory Study Utilizing Predominantly Paper Chromatographic Methods, University Of Texas, Austin, TX. 1951, Ch. 1. E.C. Horning and M.G. Horning, Clin. Chem., 17 (1971) 802 — 809. E.C. Horning and M.G. Horning, J. Chromatogr. Sci., 9 (1971) 129 - 140. M.D. Armstrong, K.N.F. Shaw and P.E. Wall, J. Biol. Chem., 218 (1956) 293. C.H.L. Shackleton and F.L. Mitchell, Steroids, 10 (1967( 359. R.L. Jolley and M.L. Freeman, Clin. Chem., 14 (1968) 538. G.A. Qureshi, L. Fohlin and J. Bergstrm, J. Chromatogr., 297 (1984) 91. R.A. Hites and K. Biemann, Anal. Chem., 42 (1970) 885. E. Jellum, O. Stokke and L. Eldjarn, Clin. Chem., 18 (1972) 800. B.E. Blaisdell, Anal. Chem., 49 (1977) 180. F.P. Abramson, Anal. Chem., 47 (1975) 45. S.C. Gates, M.L. Smisko, C.L. Ashendel, N.D. Young, J.F. Holland and C.C. Sweeley, Anal. Chem., 50 (1978) 433. C.H.L. Shackleton and W. Chai, Endocrine Reviews, 6 (1985) 441. G. Spiteller, Spectra 10 (1985) 15. 230 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 231 H. Ludwig, M. Spiteller, H.J. Egger and G. Spiteller, Israel J. Chem., 16 (1977) 7. G. Phillipou, R.F. Seamark and L.W. Cox, Aust. N.Z. J. Med., 8 (1978) 63. W.J.J. Leunissen and J.H.H. Thijssen, J. Chromatogr., 146 (1978) 365. J.A. Luyten and G.A.F.M. Rutten, J. Chromatogr., 91 (1974) 393. E.E. Werk, J. MacGee and L.J. Sholiton, Metabolism, 13 (1964) 1425. C.D. Pfaffenberger and E.C. Horning, J. Chromatogr., 112 (1975) 581. S. Minowada, K. Kinoshita, M. Hara, K. Isurugi, T. Uchidawa and T. Niijima, Endocrinol. Japan, 32 (1985) 29. C.D. Pfaffenberger, L.R. Malinak and E.C. Horning, J. Chromatogr., 158 (1978) 313. B.S. Thomas, R.D. Bulbrook, M.J. Russell, J.L. Hayward, R.R. Millis, M.R.C. Path, Breast Cancer Research and Treatment, 4 (1984) 24. A. Manni, J.E. Trujillo, J.S. Marshall, J. Brodkey and O.H. Pearson, Cancer, 43 (1979) 444. M.P. Golder, M.E.A. Phillips, D.R. Fahmy, P.E. Preece, V. Jones, J.M. Henks and K. Griffiths, Europ. J. Cancer, 12 (1976) 719. R.C. Coombes, T.J. Powles, L.H. Rees, W.A. Ratcliffe, A.G. Nash, M. Henk, H.T. Fors, J.C. Gazet and A.M. Neville, Br. J. Cancer, 46 (1982) 30. E. Kokko, 0. Janne, A. Kauppila and R. Vihko, Am. J. Obstet. Gynecol., 143 (1982) 382. W. Eachaute, L. deThibaultdeBoesinghe and E. Lacroix, Cancer Res., 43 (1983) 4260. J.M. Fromson, S. Pearson and S. Bramah, Xenobiotica, 3 (1973) 711. C.P. Daniel, S.J. Gaskell, H. Bishop and R.I. Nicholson, J. Endocr., 83 (1979) 401. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 232 C.P. Daniel, S.J. Gaskell, H. Bishop, C. Campbell and R.I. Nicholson, Eur. J. Canc. Clin. Oncol., 17 (1981) 1183. J.V. Kemp, H.K. Adam, A.E. Wakeling and R. Slater, Biochemical Pharmacology, 32 (1983) 2045. C.M. Camaggi, E. Strocchi, N. Canova and F. Pannuti, J. Chromatogr., 275 (1983) 436. R.R. Brown, R. Bain and V.C. Jordan, J. Chromatogr., 272 (1983) 351. J.S. Patterson, B. Furr, A. Wakeling and L. Battersby, Breast Cancer Research and Treatment, 2 (1982) 363. H.S. Poulsen, Eur. J. Cancer, 17 (1981) 495. R. Nicholson and K. Griffiths, The biochemistry Of Tamoxifen action. In: "Advances in Sex Hormone Research", Vol. 14, J.A. Thomal, ed., pp. 119, (1980). J.S. Patterson, D.G. Edwards and L.A. Battersby, Japanese J. Cancer Clins., Suppl. (November) (1981) 157. A. Manni, J.E. Trujillo, J.S. Marshall, J. Brodkey and O.H. Pearson, Cancer, 43 (1979) 444. G. Riberio and R. Swindell, British J. Cancer Clin. Oncol., 21 (1985) 897. P. Viladiv, M.C. Alonso, A. Avella, M. Beltran, J. Borras, B. Ojeda and F.X. Bosch., Cancer, 56 (1985i0 2745. R.C. Heel, R.N. Brogden, T.M. Speight and G.S. Avery, Drugs, 16 (1978) 1. Y. Koseki, D. Zava, C.C. Chamness and W.L. McGuire, Endocrinology, 101 (1979) 1104. J.D. Veldhuis, Annals of Internal Medicine, 88 (1978) 574. D.J. Arnold, M.J. Markham and S. Hacker, J. Am. Med. Soc., 241 (1979) 2506. J.D. Veldhuis and R.J. Santen, J. Am. Med. Soc., 241 (1979) 2506. A.H.G. Paterson, Cancer Treatment Reports, 63 (1979) 2124. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 233 S.D. Lane, E.C. Besa and R.R. Joseph, Annals of Internal Medicine, 92 (1980) 572. T.X. O'Connell, Am. J. Surgery, 141 (1981) 277. 8.8. Legha, K. Powell, A.V. Buzdar and G. Blumenschein, Cancer, 47 (1981) 2803. B.L. Agrawal and L. Zelkowitz, Arch. Intern. Med., 141 (1981) 1240. A. Gotfredsen, C. Christiansen and T. Palshof, Cancer, 53 (1984) 853. G.A. McKeown, M. Swartz, J. Blom and J.M. Maggiano, Br. J. Ophthalmology. 65 (1981) 177. M. Beck and P.V. Mills, Cancer Treatment Reports, 63 (1979) 1833. K. Nevasaari, M. Heikkinen and P.J. Taskinen, Lancet, 2 (1978) 946. A. Hendrick and V.P. Subramanian, J. Am. Med. Assoc., 243 (1980) 514. R.D. Belbrook, M.C. Swain, D.Y. Wang, J.L. Hayward, S. Kumaoka, O. Tadatani, O. Abe and J. Utsunomiya, Europ. J. Cancer, 12 (1976) 725. M. Kodama, T. Kodama, S. Miura, M. Yoshida, H. Takagi, R. Totani and T. Ooki, J. Nat. Cancer Inst., 63 (1979) 599. B.S. Thomas, R.D. Bulbrook, J.L. Hayward and R.R. Millis, Eur. J. Cancer Clin. Oncol., 18 (1982) 447. B.S. Thomas, R.D. Bulbrook, M.J. Russell, J.L. Hayward and R.R. Millis, Breast Cancer Research and Treatment, 4 (1984) 27. J.H. MacIndoe and G.R. Woods, Endocrinology, 108 (1981) 1407. N. Wilking, K. Carlstrom, H. Skoldefors, N. Olofthene and A. Wallgren, Acta Chim. Scan., 148 (1982) 345. M.P. Golder, M.E.A. Phillips, D.R. Fahmy, P.E. Preece, V. Jones, J.M. Henks and K. Griffiths, Eur. J. Cancer, 12 (1976) 719. R.C. Coombes, T.J. Powles, L.H. Rees, W.A. Ratcliffe, A.G. Nash, M. Henk, H.T. Fors, J.C. Gazet and A.M. Neville, Br. J. Cancer, 46 (1982) 30. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 234 E. Kokk, O. Janne, A. Kauppila and R. Vihko, Am. J. Obstet. Gynecol., 143 (1983) 382. B.M. Sherman, F.K. Chapler, K. Crickard and D. Wycoff, J. Clin. Invest., 64 (1979) 398. E. Kokko, 0. Jnne, A. Kauppila and R. Vihko, Am. J. Obstet. Gynecol., 143 (1982) 382. C. Tajima and T. Fukushima, Fertility and Sterility, 40 (1983) 23. M.P. Golder, M.E.A. Phillips, D.R. Fahmy, P.E. Preece, V. Jones, J.M. Henk and K. Griffiths, Europ. J. Cancer, 12 (1976) 719. K.J. Willis, D.R. London, H.W.C. Ward, W.R. Butt, S.S. Lynch and B.T. Rudd, Br. Med. J., 1 (1977) 425. F. Sakai, F. Cheix, M. Clavel, J. Colon, M. Mayer, E. Pommatav and S. Saez, J. Endocr., 76 (1976) 219. J. Levin, M.J. Markham, E.S. Greenwald, D. Wollner, J. Kream, B. Zumoff and D.K. Fukushima, Cancer, 47 (1981) 1394. Wilking, K. Carlstrm, H. Skldefors, N.O. Theve and Wallgren, Acta Chir. Scand., 148 (1982) 345. >2 J. Alexieva—Figusch, M.A. Blankenstein, F.H. DeJong and S.W.J. Lamberts, EurOp. J. Cancer, 20 (1984) 1135. C. Huggins, L.C. Grand and F.P. Brillantes, Nature, 189 (1961) 204. C.W. Welsch, M. Goodrich—Smith, C.K. Brown, D. Mackies and D. Johnson, Oncology, 39 (1982) 88. V.C. Jordan, Europ. J. Cancer, 12 (1976) 419. H.H. Fiebig and D. Schmahl, Oncology, 34 (1977) 58. V.C. Jordan, Cancer Treatment Reports, 60 (1976) 1409. V.C. Jordan, K.E. Allen and C.J. Dix, Cancer Treatment Reports, 64 (1980) 745. M.W. MacNab, R.J. Tallarida and R. Joseph, Eur. J. Pharm., 103 (1984) 321. J.M. Fromson, S. Pearson and S. Bramah, Xenobiotica, 3 (1973) 693. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 235 V.C. Jordan, M.M. Collins, L. Rowsby and G. Prestwich, J. Endocr., 75 (1977) 305. V.C. Jordan and K.E. Allen, Europ. J. Cancer, 16 (1980) 239. C.P. Daniel, S.J. Gaskell and R.I. Nicholson, Eur. J. Cancer Clin. Oncol., 20 (1984) 137. J.S. Patterson, J. Endocr., 89 (1981) 67 P. R.I. Nicholson and M.P Golder, Europ. J. Cancer, 11 (1975) 571. V.C. Jordan and S. Koerner, J. Endocr., 68 (1976) 305. .P. Bowman, C.A. Jones, A. Leake and I.D. Morris, Br. Pharmac., 78 (1983) 617. LIU) T.H. Welsh, Jr., X. Jia, P.B.C. Jones, L. Zhuang and A.J.W. Hsueh, Endocrinology, 115 (1984) 1275. C.C. Chamness, G.A. Bannayan, L.A. Landry, Jr., P.J. Sheridan and W.L. McGuire, Biology of Reproduction, 21 (1979) 1087. H. Lullmann and R. Lullmann-Ranch, Toxicology and Applied Pharmacology. 61 (1981) 138. M.I. Kaiser—Kupfer and M.E. Lippman, Cancer Treatment Reports, 62 (1978) 315. J.A. Gustafsson, Eur. J. Biochem., 14 (1970) 560. H. Eriksson and J.—A. Gustafsson, Eur. J. Biochem., 15 (1970a) 132. H. Eriksson, Eur. J. Biochem., 16 (1970) 261. H. Eriksson, J.—A. Gustafsson and J. Sjovall, Eur. J. Biochem., 9 (1969) 286. I. Bjorkhem, H. Eriksson, J.—A. Gustafsson, K.—E. Karlmar and A. Stenber, Eur. J. Biochem., 27 (1972) 318. Bjorkhem, J.—A. Gustafsson and S.A. Gustafsson, Eur. Biochem., Eur. J. Biochem., 16 (1970) 557. LII—I H. Eriksson, Eur. J. Biochem., 18 (1971b) 86. B. Wengle, Acta Soc. Med. Upsal., 68 (1963) 154. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 236 H. Eriksson and J.—A. Gustafsson, Eur. J. Biochem., 16 (1970b) 252. H. Eriksson, J.—A. Gustafsson and J. Sjovall, Eur. J. Biochem., 19 (1971) 433. B.E. Gustafsson, J.—A. Gustafsson and J. Sjovall, Eur. J. Biochem., 4 (1968a) 568. J.—A. Gustafsson and J. Sjovall, Eur. J. Biochem., 6 (1968a) 227. J.-A. Gustafsson, Eur. J. Biochem., 6 (1968b) 236. J.—A. Gustafsson, Eur. J. Biochem., 6 (1968) 248. H. Eriksson, J.-A. Gustafsson and J. Sjovall, Eur. J. Biochem., 9 (1969) 550. H. Eriksson and J.—A. Gustafsson, Eur. J. Biochem., 13 (1970) 198. T.A. Baillie and H. Eriksson, FEBS Letters, 58 (1975) 170. J.—A. Gustafsson and M. Ingelman—Sundber, J. Biol. Chem., 250 (1975) 3451. R.—J. Begue, J.—A. Gustafsson and S.A. Gustafsson, J. Ster. Biochem., 4 (1973) 393. T. Cronholm, H. Eriksson and J.—A. Gustafsson, Europ. J. Biochem., 19 (1971) 424. R.L. Butcher, W.E. Collins and N.W. Fugo, Endocrinology, 94 (1974) 1704. A.I. Toorop and L. Gribling-Hegge, J. Endocr., 93 (1982) 25. M. Szoltys, J. Reprod. Fert., 63 (1981) 221. J.-A. Gustafsson and A. Pousette, Biochim. Biophys. Acta, 280 (1972) 182. R.—J. Begue, J.-A. Gustafsson and S.—A. Gustafsson, EurOp. J. Biochem., 40 (1973) 361. M. Prost and B.F. Maume, Hormonal Steroids in Biological Materials: A Study by Mass Fragmentography. In: "Mass Spectrometry in Biochemistry and Medicine", A. Frigerio and N. Castagnoli, ed. pp. 139, (1974). 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 237 B.F. Maume, C. MIllot, D. Mesnier, D. Patouraux, J. Doumas and E. Tomori, J. Chromatogr., 186 (1979) 581. E. McCredie, C. Goddard, C.P. Vinson and B.J. Whitehouse, J. Ster. Biochem., 14 (1981) 1269. C.E. Gomez—Sanchez, E.P. Gomez-Sanchez, C.H.L. Shackleton and L. Milewich, Endocrinology, 110 (1982) 384. W.R. Miller, A.P.M. Forrest and T. Hamilton, Steroids, 23 (1974) 379. W. Eechaute, L. deThibaultdeBuesinghe and E. Lacroix, Cancer Research, 43 (1983) 4260. C.H.L. Shackleton and J.O. Whitney, Clin. Chimica Acta, 107 (1980) 231. K.D.R. Setchell and C.H.L. Shackleton, Clin. Chimica Acta, 47 (1973) 381. J.J. Vrbanac, W.E. Braselton, Jr., J.F. Holland and C.C. Sweeley, J. Chromatogr., 239 (1982) 265. A.W. Pike, C. Moynihan, S. Kibler, P.G. Marsh and P.V. Fennessey, 306 (1984) 39. M. Axelson, B.-L. Sahlberg and J. Sjovall, J. Chromoatogr., 224 (1981) 355. C.H.L. Shackleton, J. Chromatogr., 379 (1986) 91. E. Vanluchene, W. Eechaute and D. Vandekerckhove, J. Ster. Biochem., 16 (1982) 701. W.J.J. Leunissen and J.H.H. Thijssen, J. Chromatogr., 146 (1978) 365. H.M. Fales and T. Luukkainen, Anal. Chem., 37 (1965) 955. J.P. Thenot and E.C. Horning, Anal. Letter, 5 (1972) 21. E.M. Chambaz and E.C. Horning, Anal. Letters, 1 (1967) 201. M. Axelson and J. Sjovall, J. Ster. Biochem., 8 (1977) 683. J.J. Vrbanac, J.D. Pinkston and C.C. Sweeley, Biomed. Mass Spectrom., 10 (1983) 155. 238 141. M.G. Horning, A.M. Moss and E.C. Horning, Anal. Biochem., 22 (1968) 284. 142. Dale , ibid. 211. 143. H. Eriksson and J.—A. Gustafsson, Europ. J. Biochem., 20 (1971) 231. 144. P. Bournot, M. Chessebeuf, G. Maume, A. Olsson, B.F. Maume and P. Padieu, Application of Mass Fragmentography to the Study Of Sex-Linked Metabolism of Testosterone and Coritcosterone in Liver Organ and in Liver Cells in Culture, In: "Mass Spectrometry in Biochemistry and Medicine", A. Frigerio and N. Castagnoli, ed., pp. 139, (1974). 145. H. Eriksson, Europ. J. Biochem., 64 (1976) 573. 146. J.-A. Gustafsson and M. Ingelman—Sundberg, J. Biol. Chem., 250 (1975) 3451. 147. H. Eriksson, J.-A. Gustafsson and A. Pousette, Europ. J. Biochem., 27 (1972) 327. 148. J. Desgries, L. Fay, D.—H. JO, M. Guiguet and P. Padieu, J. Ster. Biochem., 21 (1984) 391. 149. J.T. McClave and F.H. Dietrich, "Statistics", Dellen Publishing Co., San Francisco, CA, 1979, pp. 289-293. 150. C. McKewen, personal communication, 1986.