Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. THE EFFECT OF ETHANOL ON CORTISOL METABOLISM IN MAN A thesis presented in fulfilment of the requirements for the degree of Master of Science in Biochemistry at Massey University PANOORA CARLYON EVANS 1979 ~i':C~ I . 8f ii AESTRACT Methods were developed for the estimation of human plasma cortisol by radioimmunoassay and urinary 68-hydroxycortisol (680HF) by colorimetry after separation by thin layer chromatography (TLC). In addition profiles of urinary neutral steroids were obtained by gas chromatographic separation of methoxime-trimethylsilyl derivatives from urine extracts on a glass capillary column. This approach was found to be rrore sensitive and reproducible than profile studies based on TLC separation and colorimetric estimation. Pilot studies of the plasma cortisol levels of normal subjects showed a consistent rise in cortisol during alcohol loading under the conditions of the observations, but in hospital patients admitted with acute alcohol intoxication, variability in the experimental conditions masked any consistent changes. Large variations in method reproducibility as well as subject differences affected results from the measurement of 680HF and chloroform extractable 17-hydroxycorticosteroids in one normal and four alcoholic subjects, rendering apparent initial differences insignificant. The results suggest, out do not derronstrate, that alcohol ingestion may divert normal cortisol metabolism into a pathway leading to the production of 680HF. Urinary steroid profiles obtained from two normal subjects, one normal subject under conditions of alcohol load and one alcoholic subject suggest that any effects of alcohol on cortisol metabolism are subtle and would require study of a large number of cases to define them. This work has served to delineate the faults and potential of various approaches to the study of cortisol metabolism and the_possible effects of alcohol thereon. It would seem that their application in carefully designed and well controlled experiments to a larger number of subjects is necessary to obtain the information desired. ACKNOWLEDGEMENTS The co-operation of Dr Louis Bieder and the staff of the Detoxification Unit at Palrnerston North Hospital in providing urine samples from intoxicated patients is gratefully acknowledged. My thanks is also due to Professor R. D. Batt and the members of the Alcohol Research Group at Massey University, in particular Mr K. G. Couchman for assistance with gas chromatography and mass spectrometry and my supervisor Dr R. M. Greenway for his advice and encouragement throughout this work. I am rrost grateful to Mrs M. R. Singleton for typing this manuscript. iii GENERAL A.C.T.H. BTZ CBE TSIM TMS GC GLC TLC PC RIA Kg Si gel 17-KS 17-0HCS EtOAc EtOH MeOH CH 2 c1 2 , DCM C-19 C-21 STEROIDS Abbreviation An Et 11-HAn 11-HEt 11-KAn 11-KEt DHEA Pd Pt Atr ABBREVIATIONS adrenocorticotrophic hormone Blue Tetrazolium (chloride) cholesterol n-butyl ether trimethylsilyl imidazole trimethylsilyl gas chromatography gas liquid chromatography thin layer chromatography paper chromatography radioi:rcmiunoassay Kieselghur silica gel iv steroid with keto group at 17-carbon position steroid with hydroxyl group at 17-carbon position ethyl acetate ethanol methanol dichloromethane steroid with no side chain at carbon 17 steroid with 2 carbon side chain at carbon 17 Trivial nrure androsterone etiocholanolone 11-hydroxyandrosterone 11-hydroxyetio- cholanolone 11-ketoandrosterone 11-ketoetiocholanolone dehydroepiandrosterone pregnanediol pregnanetriol androstenetriol Systematic name 5a-androstan-3a-ol-17-one I 5$-androstan-3a-ol-17-one 5a-androstan-3a,l1B-diol-l7-one sB-androstan-3a,11B-diol-17-one 5a-androstan-3a-ol-11,17-dione 5B-androstan-3a-ol-ll,17-dione 5-androsten-3aol-17-one 5B-pregnan-3a,20a-diol 5B-pregnan-3a,l7a,20a-triol 5-androsten-3B,l6a,l7B-triol E F THE THF a-THF aco Seo a.Cor 8Cor 680HE 680HF cortisone cortisol tetrahydrocortisone tetrahydrocortisol allo-tetrahydrocortisol acortolone 8cortolone acortol 8cortol 68hydroxycortisone 68hydroxycortisol V 4-pregnen-17a,21-diol-3,ll,20-trione 4-pregnen-118,17a,21-triol-3,20-dione 5f3-pregnan-3a,17a,21-triol-ll,20-dione 58-pregnan-3a.,118,17a,21-tetrol-20-one Sa.-pregnan-3a,118,17a,21-tetrol-20-one 58-pregnan-3a,17a,20a,21-tetrol-ll-one 58-pregnan-3a,17a.,208,21-tetrol-ll-one S8-pregnan-3a.,118,17a,20a,21-pentol 58-pregnan-3a,118,17a,208,21-pentol 5-pregnen-68,17a.,21-triol-3,ll,20- trione 5-pregnen-68,118,17a,21-tetrol-3,20- dione Abstract Acknowledgements Abbreviations TABLE OF CONTENTS CHAPTER 1 GENERAL INTRODUCTION The Effects of Alcohol on Human Endocrine Function Assessment of the Literature Hypothalamic-Pituitary-Gonadal Function Secretion of Catecholamines Hypothalarnic-Pituitary-Adrenocortical Axis A Review of the Literature on the Effects of Alcohol pn Cortisol Release and Metabolism The Aim of this Project CHAPTER 2 THE EFFECTS OF ETHANOL ON PLASMA CORTISOL IN THE HUMAN INTRODUCTION MATERIALS AND METHODS MATERIALS METHODS A Radioimmunoassay for Human Plasma Cortisol The standard curve Plasma samples The radioimmunoassay Background and total counts Correction -for free labelled cortisol Correction for non-specific binding by plasma Correction for procedural losses Calculation of cortisol concentration in plasma Antiserum dilution and incubation time Extraction of standards (a) Effect of dichloromethane residues (b) Effect of extraction of plasma (c) Effect of plasma proteins on assay Validation of Radioinununoassay for Plasma Cortisol Binding of plasma with no added antibody Coefficients of variation (a) Intra-assay variation (b) Inter-assay variation Recovery of unlabelled steroid from plasma Sens i ti vi ty Specificity of anticortisol-3-BSA serum Page ii iii iv 1 1 2 2 2 3 5 6 7 8 8 8 9 9 10 10 10 10 11 11 12 12 13 13 15 15 16 17 17 EXPERIMENTAL SUBJECTS Normals Alcoholics RESULTS AND DISCUSSION Normal Subjects Alcoholic Subjects Conclusion CHAPTER 3 EFFECT OF ETHANOL ON URINARY STEROID METABOLITE PROFILES INTRODUCTION Disturbance of Steroid Metabolic Pathways by Alcohol Separation Techniques in Steroid Analysis MATERIALS AND METHODS MATERIALS METHODS Collection of Urine Samples Standard Hydrolytic Procedure Solvent Extraction Thin Layer Chromatography Partition Sys terns Adsorption Systems Detection of Steroids on TLC Plates Detection o f the -4ene -3one Group Spray Reagents Extraction of ~hin Layer Material Scintillation Counting Quantitative Estimation of Steroids Experiment "A" Conclusions Experiment "B" Conclusions RESULTS AND DISCUSSION Page 18 19 20 20 24 25 26 28 28 29 29 30 30 32 32 32 33 34 34 35 37 37 39 CHAPTER 4 APPLICATION OF CAPILLARY COLUMN GAS LIQUID CHROMATOGRAPHY TO URINARY STEROID PROFILING INTRODUCTION A Review o f Recent Advances in Capillary ColUI1U1 Gas Chromatography of Steroids Preparation of Samples for GLC (a) Hydrolysis (b) Extraction and purification (c) Derivative formation Identification and Estimation of Steroids by GLC MATERIALS METHODS Gas Chromatography MATERIALS AND METHODS ( a) Gas chromatography (b) Inlet spl itter (c) Detector (d) Gas flows and regulation (e) Integrator ( f) Recorder (g) Operating conditions Steroid Standards Derivatisation of pure steroids Retention times Separation of pure steroid mixtures Reproducibility Standard Curves and Quantitation Human Urinary Steroid Profiles Urine Collection Hydrolysis Extraction Derivatisation RESULTS AND DISCUSSION URINARY STEROID PROFILES Normal Female Normal Male Alcohol Loading Experiment Alcoholic Female Conclusion Page 40 42 42 43 43 45 47 47 48 48 49 49 49 49 50 50 50 52 55 55 56 56 57 58 58 58 64 67 CHAPTER 5 6BETA-HYDROXYCORTISOL EXCRETION DURING ALCOHOL LOADING INTRODUCTION Microsomal Oxidation of Alcohol, Drugs and Steroids Alcoho l Metabolism Drug and Steroid Metabolism Cortisol Me t abolism: The Role of 68-Hydroxycortisol 6$-Hydroxycortisol as an Indicator of Hepatic Microsomal Oxidizing Capacity The Aim of this Study MATERIALS AND METHODS MATERIALS METHODS Collection of Samples Determination of 17-Hydroxycorticosteroids Method I (hydrolytic method) The Porter-Silber Reaction Method II (column method) Comparison of hydrolytic and column methods Extraction and Purification of 68-Hydroxycortisol from Urine Extraction Purification Conjugated 6$-Hydroxycortisol Method I Method II Specific Activities RESULTS AND DISCUSSION Normal Level of 6$-Hydroxycortisol Alcoholic Subjects Subject (1) Subject ( 2) Further subjects Conclusions CHAPTER 6 CONCLUSION BIBLIOGRAPHY Page 68 68 69 69 73 73 75 76 76 77 78 79 81 81 83 83 84 84 86 86 89 89 90 91 94 2i 2ii 2iii 2iv 2v 2vi 2vii 2viii 2ix 3i 4i 4ii 4iii 4iv 4v 4vi 4vii 4viii 4ix 4x 4xi LIST OF TABLES Comparison of Serial Dilutions of Plasma under Different Analytical Conditions Coefficients of Intra-assay Variation Between Assay Variation: Comparison of Points on Standard Curves over Four Assays Comparison of Experimental Conditions for Estimating Recovery of Unlabelled Cortisol from Plasma Specificity of Antiserum: Cross Reaction with Other Steroids Effect of Acute Doses of Alcohol on the Cortisol Page 14 15 16 17 18 Levels of Normal Subjects 20 Cortisol and Blood Alcohol Levels in Alcoholic Subjects 21-22 Correlation of Time of Day Variance with Plasma Cortisol and Blood Alcohol 23 Correlation of Time of Day with Blood Alcohol Level 24 Rf Values for Standard Steroids in Three TLC Systems 31 Retention Times of Steroid Standards 51 Reproducibility of Steroid Peak Areas over Three Days 53 Linear Dilution of Standard Steroid Derivatives 54 Protocol for Alcohol Loading Experiment 59 Alcohol Loading Experiment: Protocol and Fluid Balance for Day II, Alcohol Load 60 Alcohol Loading Experiment: Ratios of Steroid Peak Heights/CBE Alcohol Loading Experiment: Ratios of Steroid Peak Areas/CBE Alcohol Loading Experiment: Ratios of Steroid Pairs, Based on Peak Height/CBE Ratios Alcohol Loading Experiment: Ratios of Steroid Pairs, Based on Peak Area/CBE Ratios Summary of Alcohol Loading Experiment Alcoholic Female Experiment: Ratios of Steroid Peak Areas/CBE 61 61 62 62 64 65 4xii 4xiii Si Alcoholic Female Experiment: Ratios of Steroid to Internal Standard for Selected Steroid Pairs Alcoholic Female Experiment: Steroid Excretion Factors Reported as Being Associated with Increases, Page 66 66 or Relative Increases in Excretion of 6S-Hydroxycortisol· 71-72 Sii Reproducibility of Standard Curve for Porter-Silber Reaction Siii Siv 5v Svi Svii Sviii Six Comparison of Two Methods for Determining Urinary 17-Hydroxycorticosteroids Recovery of Unlabelled 6S-Hydroxycortisol Intra-assay Difference in 6S-Hydroxycortisol, Assay of Seven Urine Samples Specific Activities of 3 H-6~-Hydroxycortisol following Thin Layer and Paper Chromatographic Separations Excretion Levels of 6S-Hydroxycortisol and 17-Hydroxycorticosteroids by a Normal Subject Levels of Urinary 6S-Hydroxycortisol Obtained from Alcoholic Subject (1) 6S-Hydroxycortisol Levels in Alcoholic Subjects LIST OF FIGURES 78 80 82 83 85 86 87 89 Between Pages li 2i 2ii 2iii 2iv 2v 2vi Regulation of the Hypothalamic-Pituitary­ Adrenal Cortex_System Antiserum Binding Curves Typical Standard Curve for Cortisol Radioimmunoassay Effect of Solvent on Standard Curve Extracted and Non-Extracted Standard Curves Comparison of Standard Curves Parallelism of Standard Curve and Plasma Dilution Curve 2 and 3 9 and 10 10 and 11 11 and 12 12 and 13 12 and 13 14 and 15 2vii 3i 3ii 3iii The Effect of an Acute Dose of Ethanol on the Cortisol Levels of Normal Subjects Ethanol and Steroid Metabolism Related via NAD+/NADH Treatment of Urine Sample Prior to Fractionation of Steroids by TLC Schematic Representation of Experiment "A", Fractionation of Steroids by TLC 3iv TLC Experiment "A/•: Initial Separation of Between pages 19 and 20 25 and 26 29 and 30 35 and 36 Extracts in TLC System "W" 35 and 36 3v TLC Experiment "A": Separation of Eluate I(a) [Cortisone and C-19, 17 ketosteroids] in TLC System "X" 3vi 3vii 3viii 3ix 3x 3xi 4i TLC Experiment "A": Separation of Eluates Iae, Iba, Ibb and Ibc in TLC System "Y" (1.5 hr) TLC Experiment "A": Separation of Eluate I (b) in TLC System "W" (1. 25 hr) Schematic Representation of Experiment "B" Fractionation of Steroids by TLC TLC Experiment "B": Initial Separation of CH 2 cl 2 Extracts in TLC System "W" TLC Experiment "B": Separation of Eluates Uic, Uid, Sic and Sid in TLC System "Y'' TLC Experiment "B": Separation of EtOAc Extracts in TLC System "W" ( 3 hr) Connection of S.C.O.T. Column to Gas Chromatograph 4ii Connection of Glass-lined Stainless Steel Tubing to S.C.O.T. Column, ,using S.G.E. "Zero Dead Volume" Unions 4iii S.G.E. Inlet Splitter System: GISS-4A 4iv Apparatus for Drying Down Steroid Derivatives 4v GLC Profile of Synthetic Steroid Derivative Mixture 4vi GLC Steroid Profile from Normal Female 36 and 37 36 and 37 36 and 37 37 and 38 38 and 39 38 and 39 38 and 39 47 and 48 48 and 49 48 and 49 49 and 50 51 and 52 58 and 59 4vii 4viii 4ix 4x 4xi 4xii 4xiii Si Sii Ca) GLC Steroid Profile from 24 hr Urine of Normal Male Cb) Standard Steroid Mixture Cc) Derivatised Urinary Extract from Normal Male, Supplemented with Standard Steroids Alcohol Load Experiment: Urinary Steroid Profiles car Day I, sample (a) Cb) Day I, sample (b)" (cf Day I, sample ( c) Alcohol Load Experiment: Urinary Steroid Profiles (a) Day II, sample ( a) Cb) Day II, sample (b) ( c)_ Day II, sample Cc). Alcohol Load Experiment: Urinary Steroid Profiles ( a) Day III, sample ( a)_ (h)_ Day III, sample (b) cc>: Day I"II, sample Cc) Alcoholic Female Experiment: Urinary Steroid Profile Day I (high blood alcohol) Alcoholic Female Experiment: Urinary Steroid Profile Day II (control) Alcoholic Female Experiment: Urinary Steroid Profile Day III (control) Porter-Silber Reaction for 17-Hydroxycortico­ steroids: Typical Standard Curve Recovery of 3H-68-Hydroxycortisol after TLC and Paper Chromatography Between pages 58 and 59 58 and 59 58 and 59 61 and 62 61 and 62 61 and 62 61 and 62 61 and 62 61 and 62 61 and 62 61 and 62 61 and 62 64 and 65 64 and 65 64 and 65 77 and 78 81 and 82 1 CHAPIBRI GENERAL INTRODUCTION The Effects of Alcohol on Human Endocrine Function From reviews of this topic, such as that of Wright (1978), it may be concluded that the effects of ethyl alcohol on hormone secretion and metabolism are not large and dramatic, with the exception of its direct inhibition of the neurosecretion of the neurohypophyseal horrnones­ vasopressin and oxytocin. However, as shown from alcohol consumption figures, the average adult in any western society has ethanol present in his bloodstream for several hours per day throughout life, and the tissues of many are never ethanol-free. Under these conditions, even minor disturbances of endocrine balance may become clinically significant and worthy of investigation. The concentration of hormone to which a receptor tissue responds may be influenced by ethanol if this either (a) affects the rate of secretion of the hormone from the source tissue, or (b) toodulates its rate of catabolism or metabolic activation, particularly if this occurs in the liver where ethanol is actively oxidized to acetaldehyde and acetate. Examples of both types of interaction have been reported. Assessment of the Literature In spite of the considerable volume of literature on the endocrine consequences of alcohol ingestion, its interpretation is complicated by a \ number of problems. Comparison of the changes involving acute administration of alcohol with those due to prolonged intake, such as occur in chronic alcoholism, and comparison of the response of habitual drinkers with the respcnse of alcohol-naive subjects, has led to confusion and apparently conflicting results. In addition there has frequently been a failure to distinguish the endocrinological and metabolic effects of alcohol per se from those secondary to tissue (particularly liver) damage and a tendency to regard chronic alcoholics as a homogeneous group regardless of differences in drinking patterns, the type and quantity of liquor consumed, the history, nutritional status and the time interva+ between drinking and endocrine or 2 metabolic studies, all of which may profoundly affect the results obtained. Finally there is some difficulty in assessing data obtained before the introduction of nodern hormone assays such as specific radioimmunoassays and of correlating results obtained from animal and human studies. Hypothalamic-Pituitary-Gonadal Function Hepatic cirrhosis in men is commonly associated with both hypo­ gonadism and feminization. The similarities between the endocrine features of alcoholic and non-alcoholic cirrhosis initially suggested that it was the liver disease itself which was responsible for these changes. Recent evidence, however, suggests a possible direct effect of alcohol on testicular function. The changes described so far indicate that gonadal dysfunction may occur in the absence of overt liver disease but that in alcoholic cirrhosis, the cumulative effects of alcohol and hepatic • dysfunction may produce irore marked endocrine features. The subject is covered in some detail in the reviews by Adlercreutz (1974), van Thiel and Lester (1976) and Green (1977). Secretion of Catecholamines There is evidence from both human and animal studies that alcohol stimulates adrenal rnedullary secretion. Moderate doses of alcohol have been shown to produce a rise in both the plasma and urine catecholamines of normal human subjects (Perman, 1958; Anton, 1965), while similar effects have been observed in subjects with prolonged histories of drinking (Ogata et al, 1971). Hypothalarnic-Pituitary-Adrenocortical Axis The effects of alcohol on endocrine function have been studied :roost extensively in relation to the hypothalamic-pituitary-adrenal (H.P.A.) axis and work in this field has been reviewed by Schenker (1970), Marks and Chakraborty (19731 and Wright (1978}. The system and its regulation is shown schematically in Fig. li. Figure 1i Regulation of the Hypothalamic-Pituitary-Adrenal Cortex System Hypothalamus Neurosecretor, Cel l s Neura l Impulses "' V CRH > Adeno­ hypophysis ACTH > Adrenal Cortex ) CORTISOL Centra l METABOLIC EFFECTS Ne r vous System Short loops may exist between hypothalamus-pituitary and pituitary­ adrenal, imposing further regulation. Reproduced from Marks and Chakraborty (1973) . 3 Cortisol is the major adrenocortical steroid hor100ne found in human blood. The circulating levels of cortisol have been shown to rise rapidly in response to trauma e.g. injury, surgery, burns etc. (as reviewed by Alberti and Johnston, 1977). Cortisol is the major anti­ anabolic horrrone: its ability to inhibit protein synthesis is thought to be responsible for its unique anti-allergic and anti-inflanunatory effects. The secretion of cortisol is under the direct control of adrenocorticotrophic horrrone (A.C.T.H.) produced by the adrenohypophysis (anterior pituitary) in response to the neuroendocrine releasing factor C.R.H. (corticotrophin releasing horroone). The release of C.R.H. is, in turn, determined by the action of external stimuli on the central nervous system as well as a circadian "clock". A Review of the Literature on the Effects of Alcohol on Cortisol Release and Metabolism Although H.P.A. function appears to be definitely disturbed in chronic alcoholics (Stokes, 1971) the literature reports are often contradictory. This appears to be due, in part, to the absence of suitable techniques for measuring the horroones involved, as well as the multiplicity of possible physiological, psychological and sociological contributions. In man the effects appear to be dose related: while rooderate to large doses may activate adrenocortical activity through higher regulatory centres (rather than by direct action on the adrenal or pituitary), lower doses are less predictable and it has been postul~ted that they may even decrease the activity of a previously aroused H.P.A. system via a sedative effect on the central nervous system. Kissin et al. (1959) suggested that some of the observed abnormalities in the adrenocortical function of alcoholic subjects may be related to impaired liver function. A further investigation (Kissin et al., 1960) demonstrated increased urinary 17-0HCS and decreased plasma levels, accompanied by a marked diuresis, within two hours of a single dose of ethanol (1 g~ body weight) to alcoholic subjects. The similar effects of a water load seemed to indicate that the adrenocortical depletion may have been due to increased renal clearance, but a simultaneous water and 4 ethanol load produced a rise in plasma 17-0HCS with no appreciable change in urinary levels, suggesting an active stimulatory effect of ethanol on the adrenal cortex. Perman (1961) however, failed to show a significant change in urinary 17-0HCS two to three hours after a 1 g/Kg dose of ethanol to non-alcoholic subjects; there were no corresponding plasma steroid measurements. Margraf et al. (1967) found no significant difference in cortisol secretion rate or total excretion of 17-0HCS in alcoholic subjects as compared with non-alcoholic controls, although the distribution of the individual component steroids appeared to differ significantly from normal. In addition, 24 hour 17-ketosteroid excretion, response to A.C.T.H. and rate of metabolism of exogenous cortisol appeared to be lowered in alcoholics, while plasma corticosterone and its urinary metabolites were increased above normal levels, suggesting that alcohol affected steroid metabolism rather than adrenocortical function. In reviewing the literature Schenker (1970) suggests that chronic alcoholics show a plasma 17-0HCS level significantly higher than that of partially rehabilitated alcoholics, which in turn, is higher than that of non-alcoholics. A marked rise in an alcoholic's plasma 17-0HCS is often associated with gastro-intestinal disturbance or withdrawal. After 12 hours without alcohol acutely withdrawn, chronic alcoholics showed a 9 am plasma cortisol significantly higher than normal, which fell following the inges tion of small anounts of alcohol (Merry and Marks, 1972). This compares with a distinct rise in plasma cortisol following infusion (Jenkins and Connolly, 1968) or ingestion (Merry and Marks, 1969; Bellet et al., 1970) of ethanol to/by normal subjects. These findings suggest that withdrawal represents a state of considerable stress to the alcoholic the symptoms of which may be relieved by alcohol . Alcohol ingestion by non-alcoholics, however, raises plasma cortisol levels probably by increasing pituitary-adrenocortical activity, since no such effects were noted in non-alcoholic patients with clinical adrenal insufficiency (Bellet et al, 19701. The Aim of this Project The goal of the present research was to elucidate some of the effects of ethanol consumption on adrenal corticosteroid release and metabolism in an endeavour to clear up some of the apparent inconsistencies in the literature. Initially, attempts were made to cover effects on both the plasma level of cortisol and its conversion to metabolites and to study both normal and alcoholic subjects. Both approaches required establishment of modern methods of analysis, which occupied most of the time available for this project. 5 CHAPTER 2 THE EFFECT OF ETHANOL ON PLASMA C'ORI'ISOL LEVELS IN THE HUMAN INTRODUCTION 6 As outlined in Chapter 1 there is some controversy as to whether alcohol increases or decreases plasma levels of ACTH and cortisol, the predominant effect depending on the conditions and subjects of study. In an attempt to elucidate these effects, a radioimmunoassay (RIA) for plasma cortisol was established and validated to allow measurement of plasma cortisol in a variety of drinking situations. A number of RIA methods for the estimation of cortisol are outlined in the literature including those of Abraham et al (1972); Ruder et al. (1972); Loriaux et al. (1973); West et al. (1973); Farmer and Pierce (1974); Foster and Dunn (1974). In this work the method of Ruder et al. was adapted for use. Pilot studies were made on five normal subjects during acute ethanol loading experiments under laboratory conditions, and on a larger heterogeneous group of subjects who were admitted to Palmerston North General Hospital suffering from alcohol intoxication. One effect which confounds a study of plasma cortisol levels is the considerable diurnal variation present under normal conditions. Normal humans exhibit a peak of 120 ng/ml at about 7 am, with a subsequent decline to about 20 ng/ml in late evening (Kreiger, 1975). A large amount of data and covariance analysis methods are necessary to separate any effect of ethanol from the diurnal effect on plasma cortisol. Furthermore, some evidence also exists which suggests that ethanol metabolism in man (Reinberg et al., 1974; Sturtevant et al., 1975; 1~76) and in mice (Goldstein and Kakhana, 1977) is also subject to circadian variations. MATERIALS AND METHODS MATERIALS All chemicals, unless otherwise specified, were reagent grade or better, and supplied by May and Baker Ltd, British Drug Houses Ltd, or Sigma Chemical Co. Ltd. 7 Cortisol was obtained from Mann Research Laboratories, New York, U.S.A. A stock standard solution, containing 10 ng/ml, in ethanol,was stored at 4°c. A working standard of 1 ng/ml was prepared by freshly diluting the stock with ethanol. (1,2,6,7 (n)-3H)Cortisol was supplied in benzene:ethanol (9:1), by The Radiochemical Centre Ltd, Amersham, U.K. at a specific activity of 95 Ci/rranol (258 mCi/mg) and a radioactive concentration of 1 mCi/ml. 0.1 ml of this solution was diluted 1 in 250 (Et0H:H2o, 1:24) to give a stock solution of approximately 4 µCi/ml, which was stored at 4°C. The working solution used in the RIA, was prepared by diluting the stock 1 in 10 with assay buffer, immediately before use. A 1 in 100 dilution of the stock in ethanol was stored at 4°c for use as a radioactive "spike" in estimating procedural losses during the extraction of plasma. Anti-cortisol-3-BSA-serum, raised in rabbit, was kindly donated by Dr R. S. Fairclough, Ruakura Agricultural Research Centre, Hamilton, N. Z. .• This was stored frozen as 1 ml aliquots in sealed glass ampoules at a 1 in 200 dilution in assay buffer. These were further diluted 1 in 10 with assay buffer i:rranediately prior to use in the RIA. The assay buffer was 0.01 M phosphate buffered saline, pH 7.3, containing ~.1% gelatin. To prepare this, 20 ml 0.5 M NaH2Po4 was added to 100 ml water, the pH adjusted to 7.3 by addition of 0.5 M Na2HP04 , 0.35 g thiomerosal and 28.6 g sodium chloride added before making up to 3.5 1. After checking the pH, 1 g gelatin was dissolved per litre. Polyethylene glycol (PEG), 16.2% (w/v) solution of Carbowax 600 (Union Carbide or BDH) was added to assay tubes in such volumes as to give a final concentration of 12.5% in each tube. A 3\ solution of bovine gamma globulin (BGG) (Cohn Fraction II, Sigma Chemical Co.) in water was prepared daily. The scintillation fluid was 9 g PPO (2,5-diphenyloxazole, BDH) 8 and 300 mg POPOP (1,4 bis(2-(5 phenyloxazolyl)benzene) ,Sigma) dissolved in a mixture of 2 litres toluene and 1 litre of Triton X-100 (Rhon and Haas). Dichlorornethane (DCM) and ethanol were redistilled twice before use, the latter following overnight treatment with m-phenylarninediamine. For a plasma blank , a supply of wether plasma obtained from the Department of Physiology and Anatomy, Massey University was stripped of endogenous steroids (Torrey, 1971) by heating the plasma at 45°c for 2 min. with an equal volume of charcoal suspension (4 mg/ ml in phosphate buffer, pH 7.8). It was then centrifuged and filtered twice through Whatman No.42 paper . The 50% stripped-plasma solution was stored frozen in small aliquots. METHODS A Radioimrnunoassay for Human Plasma Cortisol The plasma extraction process was adapted from that described by Ruder et al. (1972 ). The standard curve Best recoveries were obtained when standards extracted from 0.1 ml charcoal stripped wether plasma were, employed. A series of standard tubes containing 0, 0.2 , 0.5, 1.0, 2.0, 5.0, 10.0 and 20 . 0 ng of unlabelled cortisol, following evaporation of the solvent,was equilibrated overnight at 4°c, each with a 0.1 ml aliquot of stripped wether plasma. The standards were extracted and assayed by the method described below for plasma samples. Plasma samples 0.1 ml aliquots of "unknown" plasma samples were extracted into 4.0 ml DCM ("Zipette" automatic dispenser) by shaking, horizontally, for 10 min. (equipoise-type shaker, Analyte Pty, Aust.). After standing briefly, to separate the phases, 0.4 ml aliquots (Gilsen automatic pipette, 0.2-1.0 ml) of the organic phase were removed to assay tubes and evaporated to dryness in a water bath at 40°c, under a vigorous stream of air. 9 The radioirnmunoassay This was performed in batches of 24 tubes, which is the capacity of the centrifuge used in the assay. To each tube of evaporated extract or extract plus standard was added: 1. 0.1 ml antiserum (AS} diluted 1:2,000 with assay buffer immediately before use. 2. 0.1 ml 3 H-cortisol ( 3HF}, diluted to approximately 0.4 µCi/rnl with assay buffer before use. 3. 0,1 ml 3% bovine gamma globulin (BGG} made up freshly in distilled water . The contents of each tube were mixed, briefly (vortex mixer),and allowed to stand covered for 1 hour at room temperature, followed by 2 hours at 4°c. These incubation conditions were shown to give the Sc}me results as an overnight incubation at 4° (Fig. 2i}. Free 3 HF was separated from that bound by the antiserum by the addition of 1.0 ml 16.2% PEG to precipitate protein. Tubes were allowed to stand 10 min. at 4°, then centrifuged at 2-4° for 10 min. at 7,700 g. The supernatant was aspirated off and discarded, and the precipitate redissolved in 1,0 ml distilled water with gentle warming. The solutions were decanted into scintillation vials, the tubes washed with two 5 ml aliquots of scintillation fluid which were added to the vials. Scintillation counting was performed in a Packard Tricarb liquid scintillation counter with automatic external standardization. Background and total counts Since the final counting solution was partially quenched and somewhat turbid, due to the presence of protein, the total added counts (TCl and background counts (BG) were determined as part of the assay. Duplicate BG and quadruplicate TC tubes were set up containing 0.1 ml assay buffer, 0,1 ml diluted AS, and 0.1 ml BGG, and treated in the same way as the rest of the assay tubes, except that the TC precipitates were redissolved in 0.9 ml water (instead of 1.0 ml) and washed into scintillation vials containing 0.1 ml of the dilute 3H cortisol used in the assay. 70 60 50 3 H aJ U) ·H .µ 40 s::: ,:.; >, .0 'O § 0 al .--l 30 0 U) ·H .µ H 0 u I ::c: C") di' 20 10 '-.:: ~ 500 Figure 2 i Antiserum Bi nding Curves ' 1000 ' ' 1500 No un l abelled cortisol added ' ' \ \ \ 1.0 ng unlabelled cortisol added ~ ~ ~ 2000 2500 Antise rum dilution ------ Represents incubation for 18 hours at 4°c. \ Represents incubation for 1 hour at room temperat ure followed by 2 hours at 4°c. 3000 10 Correction for free labelled cortisol The counts due to free 3H cortisol, which had been incompletely separated from the bound hormone, were assumed to be uniform throughout a run and were determined by including two or more blank tubes containing 0.1 ml assay buffer, 0.1 ml 3H cortisol and 0.1 ml BGG. The resulting counts were subtracted from the gross counts in all other assay tubes, excepting BG and TC. Correction for non-specific binding by plasma (NSBl All plasmas were found to bind the labelled cortisol to a small extent in the absence of antiserum, therefore "antiserum-blank" tubes were set up containing extracted male plasma with 0.1 ml assay buffer, 0.1 ml 3H cortisol and 0.1 ml BGG. The final counts, less those due to free 3 H cortisol, indicated the degree to which cortisol was bound by the plasma. Correction for procedural losses Extracted plasma samples and standards were corrected for procedural losses by the addition of 0.1 ml 3H cortisol (approximately 0.4 Ci/ml) to all plasma containing tubes. Following extraction a fourth 0.4 ml aliquot of solvent phase was transferred from extraction tubes to scintillation vials, evaporated and counted together with vials containing 0.1 ml 3 H cortisol to assess total added counts. The recovery from the extraction procedure was determined as a percentage of total counts added (always better than 90%), and the final assay results corrected accordingly. It was also necessary in the final calculation to correct the total counts and those of background and tree 3H cortisol \ for the counts added in the recovery estimation. calculation of cortisol concentration in plasma The following formula was used to calculate the percentage 3 H cortisol bound by antibody in each tube: 3 x-(a+b+c) % H cortisol bound=--'-----'-­ y 100 z 100 9 ~ a, rn ·rl .µ ~ >, .0 "O § 0 l'.Xl ...-l 0 rn ·rl .µ ~ 0 u I ::i:: I"') di' 50 40 30 20 10 0 Figure 2ii Typical Standard Curve for Cortisol Radioimrnunoassay 0.1 0.2 0.4 0.6 0.8 1.0 2.0 3.0 ng Unlabelled Cortisol (log scale) Standards have 0.1 ml stripped wether plasma added, are extracted into 4.0 ml dichloromethane and 0.4 ' ml aliquots taken for assay. where a = cpm BG b = cpm free 3H cortisol C = cpm NSB X = cpm sample or standard y = cpm TC and z = % recovery from extraction The percentage bound was plotted against ng unlabelled cortisol per tube for each standard and the curve of best fit drawn through the points (semi-log paper). 11 The cortisol levels of the unknown plasma samples, whose binding percentages had been determined as above, were then found by direct interpolation of the standard curve, bearing in mind the one-tenth dilution of plasma in the assay, i.e. only one-tenth of the 0.1 ml of plasma initially extracted was assayed. A typical standard curve is shown in Fig. 2ii. Antiserum dilution and incubation time h 3 . lb db Te percentage of H cortiso oun y the antiserum was estimated at 6 dilutions of the AS in the absence of unlabelled cortisol and in the presence of 1.0 ng unlabelled cortisol. Results are shown in Fig. 2i. There appeared little difference between incubating the assay overnight at 4° and 1 hour at room temperature followed by 2 hours at 4°. At a dilution of 1 in 2,000 the AS was shown to bind 47-53% of 3 H cortisol in the absence of cortisol and 12-15% in the presence of 1.0 ng cortisol. Extraction of standards Before it was finally decided to add plasma and extract standards - prior to assay, a number of standard curves were produced under varying conditions: (a) Effect of dichlorornethane residues: In an experiment designed to test the possibility that the assay might be affected by residues of solvents used to extract plasma samples, two identical series of unlabelled cortisol standards, in ethanol, were set up. 0.4 ml DCM was added to each tube of one curve; and all tubes were evaporated to dryness and subjected to identical RIA's. The curves so obtained are shown in Fig. 2iii. 50 40 § H QJ UJ •rl .µ ;il 30 >, .Q 'O § 0 Ill M 20 -0 UJ •rl .µ H 0 u I :r: I") di' 10 -- Figure 2iii Effect of Solvent on Standard Curve ---- ...... -- -- 0.02 0.05 0.10 0. 20 0.50 ng Unlabelled Cortisol (log scale) ---- Represents standard curve with no added solvent. ~presents standard curve with 0.4 ml dichloromethane added. 1.0 2.0 5.0 12 There appeared to be little difference in the linear regions of the two curves; however for cortisol concentrations less than 0.1 ng the slope of the curve to which DCM had been added was less and the percentage of 3 H corti so 1 bound ·in the absence of unlabelled cortisol, reduced by approximately 5%. (b) Effect of extraction of plasma: Standards containing 0, 0.2, 0.5, 1.0, 2.0, 5.0, 10.0 and 20.0 ng unlabelled cortisol were equilibrated with 0.1 ml wether plasma which had been stripped of endogenous steroids as previously described. The evaporated standards and plasma were equilibrated overnight at 4°C. Each was extracted into 4.0 ml DCM and three 0.4 ml aliquots rem:>ved from each. Losses during the extraction procedure were de~ermined by addition of 3 H. cortisol,as described for the routine assay. Tpe response curve so obtained was compared with that of a · set of standards to which no plasma had been added and which were assayed simultaneously (Fig. 2iv). The curves were similar and at their points of widest variance showed a difference of about 5% binding. They intersected at 0.04 ng cortisol and it appeared that omission of the extraction step led to over-estimation of cortisol concentration at low values and under­ estimation at high values. A sample of wether plasma extracted and assayed ­ with the two standard curves showed an endogenous cortisol concentration of 21.5-28.5 ng/ml when calculated using the standard curve extracted from plasma, and 14.5-23.0 ng/ml by interpolation from the non-extracted curve. At Ong cortisol the difference in binding between the two curves was around 2%, corresponding to l.~ ng/ml cortisol on the non­ extracted curve, which would not account for the discrepancy between the two methods in estimating the wether plasma sample. A non-specific binding of approximately 50 cpm by the plasma similarly fails to account for the difference. (cL Effect of plasma proteins on assay: Two sets of standards, one without plasma and one equilibrated with 0.1 ml stripped plasma were extracted and compared with one to which plasma was added but not extracted: 9 1-l Cl) Ul ·.-i .µ ~ >, .Q re, § 0 ~ ,..., 0 Ul •.-i .µ 1-l 8 I :r: M dP so 40 30 20 10 0.02 ..._ Figure 2iv Extracted and Non-extracted Standard Curves -... ........ ....................... ' ' . ' ' ' "' ~, . " 0.05 0.10 0.20 a.so ng Unlabelled Cortisol (log scale) Represents non-extracted standards. - - - - - Represents standards extracted following addition of plasma. '' ' ' 1.0 ', ' ' 2.0 50 40 !3 1-l Cl) Ill ·n 30 .µ ~ - ,:( :>, .Q 'O § 0 P'.l .-I 20 0 Ill ·n .µ 1-l 8 I :x: M dP 0 -------------- ....::::::::--.... Figure 2v Comparison of Standard Curves 0 .10 0.20 ng Unlabelled Cortisol (log scale) ----- Represents standards with no added plasma. ~ 0.50 - - - - Represents standards with plasma added, but not extracted. - Represents standards with plasma added and extracted. .. ~ "~ ~ ~~- 1.0 2.0 (i) contained O, 0.1, 0.2, 0.5, 1.0 and 2.0 ng unlabelled cortisol. (ii) contained O, 1.0, 2.0, 5.0, 10.0 and 20.0 ng cortisol. Each standard was equilibrated with plasma, extracted and one-tenth aliquots taken for assay as usual. 13 (iii) contained O, 0.1, 0.2, 0.5, 1.0 and 2.0 ng cortisol to each of which was added 0.1 ml of stripped wether plasma which had been diluted 1 in 10 with assay buffer to give a concentration identical with that in (ii) above. 0.1 ml assay buffer was added to all tubes in (i) and (ii), and the RIA performed as usual. The three curves so obtained are shown in Fig. 2v. The extracted and non-extracted plasma standard curves differed by up to 1% in amounts of 3H cortisol bound, well within experimental error. Plasma and non-plasma curves showed insignificant differences (±1%) between cortisol concentrations of 0.1-2.0 ng, 2% at 0.1 ng and 5% at 0 ng. Since 0.2-2.0 ng was considered close to the effective useful range of the curve, it was ~at that stage, considered valid to use a standard curve without added plasma and not to extract (merely dilute 1 in 10 with buffer) unknown plasma samples. RIA's were carried out in a total voltune of 0.4 ml (rather than 0.3) and the volume of added PEG increased to maintain a final concentration of 12.5%. Later trials, however, showed that the m:Jst reliable "recovery" results were obtained when plasma was added to the standard curve and all plasma and plasma-standard samples extracted as outlined for the routine method: see "recovery of unlabelled steroid" in the following section. Validation of Radioimrnunoassay for Plasma Cortisol The validation of this RIA is in accordance with the methods specified by the Journal of Endocrinology (J. Endocr. 63, 1-4, 1974). Binding by plasma with no added antibody Approximately 2% (i.e. 300 out of 15,000)of the total counts added to the plasma in the absence of antiserum appear to be bound by the plasma non-specifically. 14 Parallelism of standard curve and dilutions of plasma A dilution curve which was both linear and parallel to the standard curve was obtained only when the "pool" plasma was diluted with stripped wether plasma, then extracted into DCM, and when the standard curve was equilibrated with stripped-plasma and similarly treated (Table 2i). Even then, this linearity was shown to extend over a relatively small part of the standard curve i.e. approx. 20-100 ng/ml of plasma. The effect of dilution of a plasma pool containing 95,3 ng/rnl of cortisol by 2, 4 and 8 is shown in Table 2i. The eight-fold dilution lies on that part of the curve which is no longer linear and approaches the lower sensitivity limits. The standard and plasma curves are shown in Fig. 2vi. Table 2i Comparison of Serial Dilutions of Plasma under Different Analytical Conditions Experiment l 2 3 Undil. 180 112 95 3/4 115 2/3 107 1/2 84 48 46 1/4 29 20 1/8 17 5.5 Concentrations are in ng/ml plasma, and are the means of three estimates. All experiments used different plasma samples. All dilutions were obtained by adding stripped wether plasma to the undiluted plasma pools before taking aliquots for assay . Experiment 1: No plasma added to standards; neither plasma samples nor standards extracted. 2: No plasma added to standards; plasma samples extracted, standards not extracted. 3: Plasma added to standards; plasma samples and standards extracted. 3 H Q) C/l 40 ·M 30 .µ ~ >, ..Q 'd § 0 m 20 .-i 0 C/l .,-j .µ H 0 u :r: f") dP 10 0 Figure 2vi Parallelism of Standard Curve and Plasma Dilution Curve dilution) dilution) 0.10 0.20 0.40 0.60 ng Unlabelled Cortisol (log scale) Represents standard curve. - - - - Represents plasma dilution curve. dilution) 1.0 2.0 Coefficients of variation (a) Intra-assay variation: Two plasma pools containing high and low levels of cortisol were included in a number of assays, between three and nine estimations being performed in each run. Results, shown in Table 2ii, indicate intra-assay coefficients of variation ranging from 4.6\ to 23.6\, with only one coefficient above 20\. Table 2ii Coefficients of Intra-assay Variation A. LOW CORI'ISOL PLASMA POOL: Assay n Mean (ng) Standard c.v. Deviation (ng) · 1 9 0.269 0.026 9.8% 2 6 0.262 0.019 7.4% 3 6 0.262 0.020 7.8% 4 6 0.255 0.042 16.4% *5 5 0.167 0.039 23 . 6% B. HIGH CORTISOL PLASMA POOL (all pools different) Assay n Mean (ng) Standard c.v. Deviation (ng) 1 8 1.76 0 . 21 12. 0% 2 6 0.80 0.37 4.6% 3 3 1.12 0.12 10.7% *4 7 0.96 0.09 9.2% "n" is the number of separate estimates in each assay. C.V., the coefficient of intra-assay variation is calculated as the standard deviation divided by the mean. * these were assays in which both the standards and plasma samples were extracted prior to assay. 15 (b) Inter-assay variation: Coefficients of variation were compared over five successive assays of the low cortisol plasma pool. The mean of the intra-assay means was 0.243±0.43 (standard deviation), indicating a coefficient of inter-assay variation of 17.6% (where the coefficient of 16 variation between assays is calculated as the standard deviation of the means of each pool divided by the mean of the means over several assays). Table 2iii shows the comparison of variations between given points on the standard curve over four assays, where all coefficients are shown to be less than 5%, well within accep table limits. Table 2iii Between Assay Variation: Comparison of Points on Standard Curves over Four Assays ng Unlabelled Mean% 3H Cortisol Bound Inter-assay C.V.* Cortisol Added ± Standard Deviati on 0 54 .0±1. 3 2.4% 0.05 45.8±2.1 4.6 0.10 40. 3±1.2 2.9 0.20 33.9±0.6 1.8 0 . 40 26 . 3±0.7 2.5 0.60 21.3±0.4 1.9 1.00 15.5±0.4 2.7 2.00 9.6±0.3 2.7 * C.V., the coefficient of inter-assay variation is calculated as the standard deviation of the means of each pool divided by the mean of the means over four assays Recovery of unlabelled steroid from plasma Varying amounts of unlabelled cortisol were added to aliquots of the low cortisol, pooled plasma and the recovery estimated by RIA. Recoveries closest to 100% were obtained when plasma samples were extracted prior to assay and the concentrations calculated by interpolation of a standard curve which had been extracted from stripped plasma . In particular, non-extracted plasmas tended to give recove ries well in excess of 100%. A tendency to over-estimate the concentrations of plasma samples fortified with cortisol when results were based on non-plasma standard curves, was also shown. A number of recovery experiments, employing different assay conditions are shown in Table 2iv. Table 2iv Comparison of Experimental Conditions for Estimating Recovery of of Unlabelled Cortisol from Plasma ng Cortisol Added Experiment 1 2 3 4 5 0.1 126 122 100 95 85 0.2 110 113 110 101 0.4 118 112 0.5 110 93 95 0.6 0.7 155 129 17 1.0 98 114 All results are% recovery of cortisol added to the low cortisol plasma pool (approx. 26 ng/ml), and are the means of two to three estimates. Experiment 1 and 2: No plasma added to standards; plasma samples Sensitivity not extracted. 3 and 4: No plasma added to standards; plasma samples extracted. 5: Plasma added to standards; plasma samples extracted. The lowest concentration of cortisol which could be estimated in plasma, by this assay, was around 10 ng/ml; obtained by assaying one-tenth of an extract of 0.1 ml plasma. In routine assays, the lower sensitivity limit was taken as 20 ng/ml, and estimates lying below this level were reassayed using a larger volume of extract. Specificity of anticortisol-3-BSA serum The percent cross reaction of the AS with other steroids is shown in Table 2v. At the time of testing, samples of corticosterone and 11-deoxycortisol, which may have cross-reacted significantly, were not available. Of those steroids tested, only cortisone showed significant (9.6%) cross reaction. This was not, however, considered a problem for the purposes of this assay, since the normal plasma concentration of cortisone has been estimated at around 10% that of cortisol. Table 2v Specificity of Antiserum: Cross Reaction with Other Steroids Steroid Cortisone Deoxy corticosterone Tetrahydrocortisol 17-hydroxyprogesterone Progesterone B-Methazone Aldosterone Testosterone Androstenedione Estradiol-17B % Cross-reaction defined as (a/b) .100 % Cross-reaction 9.6 0.9 < 0.5 < 0 . 1 < 0.1 < 0.1 < 0.1 < 0.1 < 0.1 < 0.1 where a is the mass of unlabelled cortisol required to displace 50% of the bound 3H cortisol and bis the mass of cross reacting steroid required to displace 50% of the bound 3 H cortisol. EXPERIMENTAL SUBJECTS Normals 18 Plasma cortisol was measured in five normal subjects (including one female) who were part of a group studied to determine the effects of alcohol loading on a variety of blood metabolites (Couchman, 1974). These subjects were given an oral alcohol load of 0.5 g/kg body weight (as diluted vodka) at approximately 9.30 am in fasting condition. Blood alcohol levels were measured by K. G. Couchman, using gas chromatography of headspace gas equilibrated over a perchlorate extract of blood. 19 Alcoholics Sixty-seven blood samples were obtained from Palmerston North General Hospital from eleven patients (two were studied twice) admitted in acute alcohol intoxication. Blood alcohol levels were measured by the hospital laboratory using a gas chromatographic method. These samples were a heterogeneous group taken in varying numbers and at various times from the subjects whose intake of alcohol was not known. All of the patients would be classed as heavy drinkers or chronic alcoholics. 250 225 200 175 150. ,-j s ....... °' s:: ,-j 0 Ul 125 ·rl +J H 0 u CU s Ul CU ,-j p.. 100 75 so 25 0 30 Figure 2vii The Effect of an Acute Dose of Ethanol on the Cortisol Levels of Normal Subjects. 60 90 Time (min.) 120 150 180 D.W. R.G. J.T. D.S. 20 RESULTS AND DISCUSSION Normal Subjects The data from the five normal volunteers is given in Table 2vi and graphed in Fig. 2vii. In all subjects the highest cortisol level was found before alcohol was taken and fell throughout the study, except for the last sample in three subjects where a rise may have been associated with hunger. Since all of these studies were made in the period 8.30 am - 12.30 pm where the diurnal cycle of cortisol is still in a sharply declining phase, it is highly likely that much of this consistent decrease is due to normal diurnal cycling. The possibility remains that alcohol modifies the shape of this decline, but further studies with non-loaded controls would be necessary to demonstrate this. This pilot study underlines the need to take diurnal cycling of plasma cortisol into account when studying the effects of alcohol. Table 2vi Effect of Acute Doses of Ethanol on the Cortisol Levels of Normal Subjects Time: 0 15 min 30 min 60 min 90 min 120 min 150 min 180 J.T. 77 62 52 40 41 36 27 min 29 G.T. 133 115 92 68 65 41 42 116 R.G. 165 130 121 92 74 66 72 58 D.S. 145 122 85 61 41 37 27 26 D.W. * 260 208 174 142 90 77 63 78 * Female subject. Alcoholic Subjects The basic data from the 67 clinical samples is listed in Table 2vii. When all samples were combined the mean blood alcohol was 167±91 mg/100 ml (± standard deviation) and plasma cortisol was 163±80 µg/ml. 21 Table 2vii Cortisol and Blood Alcohol Levels in Alcoholic Subjects Patient No.l Time 1900 2050 2200 2350 2400 0050 0100 0150 0200 0250 Cortisol 108 190 160 184 164 142 108 94 78 92 BAC 460 368 290 271 170 179 212 216 184 216 Patient No.2(a)* Time 2150 2300 0100 0300 Cortisol 70 26 86 68 BAC 327 331 184 133 Patient No.2(b)* Time 2250 0050 0250 0550 Cortisol 105 172 86 98 BAC 299 258 161 138 Patient No.3 Time 1925 2175 2350 0550 Co:r;tisol 65 160 157 237 BAC 221 189 169 40 Patient No.4 Time 1650 1700 1800 1850 1900 1950 2000 2050 2100 2150 2200 2250 Cortisol 320 345 309 251 379 334 284 245 222 150 159 132 BAC 230 248 234 212 184 147 152 156 133 133 92 101 Patient No.S(a)* Time 1600 1750 1900 2050 2150 Cortisol 166 150 127 148 101 BAC 277 234 161 132 74 Patient No.5(b)* Time 2200 0050 0250 0450 0650 Cortisol 141 229 97 138 101 BAC 398 122 103 83 59 Patient No.6 Time 2350 0100 0300 Cortisol 70 49 214 BAC 151 64 54 Patient No.7 Time 1750 2050 0500 Cortisol 114 102 192 BAC 210 175 260 Continued ... Table 2vii continued Patient No.8 Time 0200 0400 0650 Cortisol 120 178 145 BAC 144 84 92 Patient No.9 Time 1350 1600 1775 1950 2150 2350 Cortisol 246 217 205 133 172 138 BAC 210 148 144 82 58 23 Patient No.10 Time 1550 1800 1900 2225 Cortisol 310 235 168 196 BAC 154 94 68 44 Patient No.11 Time 2100 0100 0300 0500 Cortisol 41 83 74 288 BAC 188 122 57 19 Cortisol levels estimated in ng/rnl on 67 sampl es derived from 11 patients over 13 admissions as shown. Blood alcohol (BAC) levels estimated by hospital laboratory in mg/100 ml of blood. * Patients nos 2 and 5 admitted twice. 22 A scatter diagram of plasma cortisol versus blood alcohol concentration for all samples showed little correlation between these variables. The coefficient of linear regression was -0.05. 23 In an attempt to reroove the effect of the diurnal rhythm on the variance of the plasma cortisol data a multiple regression analysis was carried out using the BAR 3 statistical programme on an IBM 1620 computer. In order to simulate the diurnal rhythm, the data on sample collection time on a 24 hour clock (T) was introduced into the computation as a polynomial combining 1st, 2nd, 3rd and 4th powers of T. As shown in Table 2viii, at no stage did correlation for the time-of-day variance lead to a significant correlation between plasma cortisol and blood alcohol, althoµgh the analysis suggested that any correlation was negative, i.e. high blood alcohol tended to be associated with low circulating cortisol levels. As seen from Table 2ix, the blood alcohol level was significantly correlated with time of day in the 1st and 3rd order polynomial runs, which may be explained by a tendency for heavy drinkers to consume rrore alcohol at certain times of day or for hospitals to admit intoxicated patients more comIOOnly at certain times. Table 2viii Correlation of Time-of-Day Variance with Plasma Cortisol and Blood Alcohol Power of Tin Regression Coefficient Significance time-of-day of BAC on cortisol Students polynomial (n) t p 0 -0.055 -0.39 NS l -0.112 -0.80 NS 2 -0.153 -0.96 NS 3 -0.187 -1.21 NS 4 -0.235 -1.48 NS Table 2ix Correlation of Time-of-Day with Blood Alcohol Level Power of Tin Regression coefficient Significance time-of-day of BAC on Tn polynomial (n) t 1 0.027 2.06 2 0.072 0.88 3 9Xl0- 0 -2.23 4 0.000 1. 30 Notes: T: Time of sample collection in hours (24 hour clock) BAC: Blood alcohol concentration NS: Not significant Conclusio n p <0.05 N.S. <0 .05 N.S. 24 It may be concluded that many factors operate in heavy drinkers to affect their cortisol levels but the present survey does not allow us to say whethe r alcohol in the blood is one of these. A more controlled experiment designed to minimise other factors known to affect cortisol release, including all types of mental and physical stress would be necessary to test the effect of alcohol per se. CliAPTER 3 EFFECT OF ETHANOL ON URINARY STEROID METABOLITE PROFILES INTRODUCTION Disturbance of Steroid Metabolic Pathways by Alcohol 25 There is fragmentary evidence in the literature to support the hypothesis that the usual pathways of steroid metabolism may be altered following the ingestion of significant quantities of alcohol. One example of this competition is reviewed in detail in Chapter 5. Isselbacher (1977) suggests that many clinical consequences of alcohol metabolism may be directly attributed to the generation of NADH and an increase in the NADH/NAD+ ratio in hepatic cells, such a change could influence the reductive metabolism of steroids (Chronholrn et al, 1971) (see Fig. 3i). Chronholrn and Sjovall (1970) in fact showed a six- to ten-fold increase in the ratio of 17S-hydroxy steroids to 17 keto steroids following ingestion of sufficient alcohol to produce a level of 30-60 mg of ethanol/100 ml of blood. Furthermore, use of deuterium-labelled ethanol showed a rapid incorporation of the label into the D-ring of dehydroepiandrosterone-3-sulphate, suggesting NADH coupling of the metabolic pathways. The relative sizes of the hydroxy- and keto-steroid pools correlated directly with the level of ethanol suggesting that their ratio may be determined by the NADH/NAD+ ratio in the liver cell cytoplasm. The further influence of NADH and NADPH on the biosynthesis and metabolism of steroids is indicated (Chronholrn et al., 1974) by the incorporation of deuterium from labelled ethanol into steroids, cholesterol and bile acids during NADH and NADPH dependent reductions. In this section of the work it was intended to examine the immediate and long-term effects of simultaneous alcohol metabolism on the patterns of steroid excretion (particularly cortisol metabolites) with a view to showing possible differences generated in the relative levels of "oxidized" (keto-) to "reduced" (hydroxy-) pairs of metabolites by an increased cellular level of NADH .and NADPH. ETHANOL METABOLISM CH 3 CIIO (acetaldehy de) Figure 3i Ethanol and Steroid Metabolism Related via NAD+/NADH + NAD NADH + H+ Metabolite ' (reduced) STEROID METABOLISM Metabolite · (oxidised) 26 Separation Techniques in Steroid Analysis Until recently the isolation, identification and quantitation of biological metabolites such as steroids were carried out by milligram­ level procedures based on such conventional methods as solvent extraction, chromatographic purification and crystallization using classical organic-chemical techniques. A major disadvantage of most of these was the requirement for comparatively large arrounts of tissue, urine, blood etc. Although many of the electronic and optical methods developed since 1940 for quantitative analysis (e.g. spectrophotometric and fluorometric techniques) require only microgram quantities, they are relatively non-specific. In the last two decades a large literature has appeared concerning the development of gas phase analytical procedures. Enjoying current P9pularity is the GC-MS-COM combination, involving gas chromatographic (GC} separation and estimation, mass spectroscopic (MS) identification and structural elucidation, with the results being analysed and collated by computer (COM). These methods are directly appli cable to the estimation and study not only of single compounds, but of multi-component systems such as urinary profiles within the microgram and sub-microgram range. Though elegant techniques, such as gas liquid chromatography (GLC) and high pressure liquid chromatography (HPLC) may now be unparalleled for biological steroid separations, the methodology is complicated and the capital cost of equipment high compared with the longer-established techniques of column, thin layer and paper chromatography. For many applications these latter methods are quite adequate, easier to use and far less costly. ' All established methods in current use (particularly for cortico- steroids) are reviewed fully elsewhere (Heftmann, 1973; Makin, 1975; Sandberg and Slaunwhite, 1975). Methods for the separation of corticosteroids by liquid column chromatography are well established and supports such as florisil (Eik-Nes et al, 1953}, alumina (Loras and Migeon, 1966), celite (Larsen, 1968) and, roore recently, Sephadex LH-20 (Setchell and Shackleton, 1973) have been used. Paper and thin layer chromatography have also found wide application for the separation and purification of corticosteroids. 27 (Paper: Bush, 1961; 1968; Hall et al., 1971; Daniilescu-Goldinberg and Giroud, 1974; Cawood et al., 1976. Thin layer: Butruk and Vaedtke, 1968; Duthie et al., 1969; Scandrett and Ros, 1976). Tables of systems suitable for both paper and thin layer chromatography of adrenal steroids are given by Smith and Hall Cl974). The TLC Irethods used in this chapter were based largely on those previously developed in this laboratory by Langford (1968). 28 MATERIALS AND METHODS MATERIALS Unless otherwise specified, all chemicals were reagent grade or better and supplied by May and Baker Ltd, British Drug Houses Ltd, or Sigma Chemical Co. Ltd. All solvents were redistilled before use. Kieselghur and Silica Gel GF-254 were supplied by E. Merck, Darmstadt, Germany. 8-Glucuronidase was supplied by Sigma Chemical company, St Louis, Mo., U.S.A. as a crude extract from Helix porra.tia, containing approximately 100,000 Fishman units of glucuronidase activity per ml. (1 Fishman unit defined as hydrolysing 1.0 µg phenolphthalein glucuronide per hour ~t pH 5.0 and 37°c}; and 3690 µMolar units of arylsulphatase activity per ml, using nitrocatechol sulphate as substrate, pH 5.0; 37°c. Cortisol, cortisone, THF, THE, allo-THF, a and 8-cortol, a and 8-cortolone, 11-hydroxyetiocholanolone, 11-ketoandrosterone and dehydroepiandrosterone were supplie d by Mann Research Laboratories New York, U.S.A. 11-ketoetiocholanolone and 11-hydroxyandrosterone were from Sigma Chemical CO.; and 68-hydroxycortisol was from Steraloids Inc., Pawling, New York. METHODS Blue Tetrazolium was supplied by Mann Research Labs. All water used was deionised and/or glass distilled. Collection of Prine Samples This ·section of the work was performed on aliquots of a conplete 24 hour urine collection from a normal female subject. The urine was collected in a plastic container, without preservative and stored at 4°C .throughout collection. The volume was made up to 1 litre with water. 100 ml was processed immediately and the rest stored frozen as 100 ml aliquots in polythene containers. 29 Standard Hydrolytic Procedure 100 ml aliquots of urine were adjusted to pH 5.0±0.2 with a few drops of glacial acetic acid, covered and incubated for 32 hours in a gently shaking water bath at 37°c. The pH of the hydrolysate was checked after 24 hours and adjusted if necessary. The activity of e-glucuronidase in the medium was also checked qualitatively after 24 and 32 hours incubation, as follows: 0.1 ml aliquots of hydrolysate were added to a test tube containing 0.25 ml of 0.002 M phenolphthalein glucuronide and 0.25 ml acetate buffer pH 5.0, and incubated at 37°c for 30 mins. The solution was adjusted to a pH~B by dropwise addition of 1 M NaOH. The presence of a pink, phenolphthalein colour in the alkaline solution indicated the continuing activity of the enzyme in the hydrolysis medium. A "blank" tube, containing 0.25 ml substrate and 0.35 ml buffer was routinely included. The hydrolysate was stored overnight at 4°c if extraction did not take place immediately. Solvent Extraction (adapted from Thrasher et al., 1969) The urine hydrolysate (100 ml) was transferred to a 1 litre separatory funnel and extracted once with two volumes of dichloromethane by gentle swirling for 10 minutes. Complete separation of the phases was effected by centrifugation. The organic phase was washed with one-tenth volume of 1 M NaOH, containing 15\ Na 2 so 4 (w/v), until a clear wash was obtained (2-3 times), then ~ashed once with one tenth volume l M acetic acid (containing 15% Na 2 so 4 } and finally with one hundredth volume of aqueous 15% Na2so4 . The extract was filtered through anhydrous Na2so4 into a 1 litre round bottoxred flask, and evaporated to dryness (Buchi "Rotavapor"). residue is hereafter referred to as Extract I. Approximately 15 g anhydrous Na2so4 were added to the aqueous phase remaining after the first extraction. This was extracted twice with two volumes of ethyl acetate, the combined extracts washed and dried as above and similarly evaporated to dryness. This residue is hereafter referred to as Extract II. This Figure 3ii Treatment of Urine Sample Prior to Fractionation of Steroids by TLC wash dry EXTRACT URINE l hydrolysis HYDROLYSATE EXTRACT l [Na0H/Na2So 4 ; 1/10 vol, 3 times] 1 l [H0Ac/Na2So 4 ; 1/10 vol, once] l l [H20/Na 2 so 4 ; 1/100 vol, once] l l [Filter through Na2so4 ] 1 l Evaporate to dryness l wash dry EXTRACT I EXTRACT II [ 30 Both extracts were transferred to conical test tubes : Extract I using aliquots of CH2c12 :Me~H (9:1) and Extract II using EtOAc:MeOH (1 : 1). Each was evaporated to dryness under a stream of dry nitrogen and redissolved in 1 . 0 ml of the appropriate solvent (Fig. 3ii). Thin Layer Chromatography Thin layers of 0.25 mm thickness were prepared by spreading a slurry of 30 g of aosorbant (silica gel or Kieselghur) in 65 ml water on glass plates 20 x 20 cm and 20 x 10 cm using a Desaga TLC spreader (Heidelburg, Germany). The p lates were left 10-15 min. to set and activated by heating at 110°c for 1-2 hours. The plates were developed in glass tanks (22 x 8 x 23 cm) lined with Whatman 3 MM filter paper and closed with ground glass covers. Before use the solvent level was adjusted to approximately 1 cm, following equilibration of the tank with the solvent for several hours. Continuous-elution chromatograms were developed in a "Shandon" continuous e lution TLC tank fitted with a lid and solvent reservoir. Partition Systems: The stationary phase was applied after the method of Langford (1968) by pre-running Kieselghur plates in solutions of ethylene glycol in acetone or methanol. The p lates were dried in a stream of cool air, leaving a uniform impregnation of the thin laye r and used immediately. System "W": Kg//GlycolfCH 2 c12 (Butruk and Vaedtke, 1968) Kieselghur plates were pre-run in 10-15% solutions of ethylene glycol in acetone, the steroids and extracts applied at the origin and developed in dichloromethane saturated with ethylene glycol. The Rf values for the steroid standards used in this study are given in Table 3i. The system, without over-running, is effective in separating cortisone, cortisol (and THE) from . some of their polar metabolites. Over-running for 1.5 hours completely separates cortisol and THE while the polar metabolites may separate to some extent. After 3 hours over-running the polar metabolites including the a- and ~-cortols and cortolones, are separated completely. 31 Table 3i Rf Values for Standard Steroids in Three TLC Systems Steroid System "W" System "Y" System "Z" Cortisol 0.57 0.37 0.65 Cortisone 0.91 a.so 0 .75 THF 0.20 0.21 0. 55 THE 0.46 0 . 70 a-THF 0.23 0.57 a Cortol 0.06 0.07 0.34 ff Cortol 0.07 0.07 0.34 a Cortolone 0.16 0.13 0.49 s Cortolone 0.17 0.10 0 .45 11-HEt 0.53 0.79 11-HAn 0 .57 0 .81 11-In (Horning and Horning, 1970; Luyten and Rutten, 1974; Pfaffenberger and Horning, 1975; 1977; Setchell and Shackleton, 1975; Setchell et al., 1975; Setchell et al., 1976; Adlercreutz and Schauman, 1976). In addition, solvolysis is often applied to completely free some androgens e.g. androsterone, from their sulphate conjugates (Janne et al., 1969; Shackleton and Honour, 1976; Pfaffenberger and Horning, 1977). 43 In general, the hydrolysis of conjugates represents the slowest and least efficient step in the entire analytical process (Shackleton and Honour, 1976). (b) Extraction and purification Steroids have been extracted from urine (and plasma) by several methods. Perhaps the earliest and still rrost popular, is by direct extraction with large volumes of solvents of suitable polarities. Of these ethyl acetate, dichloromethane and chloroform are most coillIOC>nly used for the polar corticosteroids (Horning and Horning, 1970; Luyten and Rutten, 1974; Pfaffenberger and Horning, 1975; 1977). Stillwell et al (1973) describe a method for extracting polar steroids such as estriol and cortisol from urine and plasma into a small volume of solvent, following the addition of potassium or a:mrnonium carbonate wlµ.ch acts essentially as a "salting-out" agent. Many workers now combine or replace the bulk extraction procedure with chromatography on a non-specific neutral resin (such as "Amberlite XAD-2") eluting with a polar solvent (Luyten and Rutten, 1974; Axelson and Sjovall, 1974; Setchell and Shackleton, 1975; Setchell et al., 1975; Setchell et al., 1976 ; Shackleton and Honour, 1976). Further purification , with the possibility of fractionating the steroids into groups, has been achieved by chromatography in miscible solvents on neutral or ion-exchange, lipophilic derivatives of "Sephadex" e.g. LH-20, "Lipidex" (which contains hydroxyl alkyl groups) and D.E.A.P. "Sephadex" (diethyl aminohydroxyl propyl) LH-20, alone or in combination (Axelson and Sjovall, 1974; Sjovall, 1975; Setchell et al., 1976; Shackleton and Honour, 1976). Often, solvent systems are chosen so: that \ the steroids elute with the solvent front in a small volume while interfering compounds and steroids of widely differing polarities remain on the solumn. {c) Derivative formation Steroids containing hydroxyl and ketone groups, particularly members of the C-21 adrenocorticosteroid group with the 17a,21-diol-20one structure must be stabilised through derivative formation to prevent loss of the side chain by thermal decomposition during G.C. analysis. Olemical modification of reactive groups may also help to minimise the ir adsorption 44 to the glass and support material. A variety of methods are available, including the cyclic boronate, acetonide, and methylene deoxy procedures reviewed by Sakauchi and Horning (1971). By far the nost popular however, would appear to be the formation of methoxime-trimethylsilyl ethers (Gardiner and Horning, 1966; Horning and Horning, 1970; Sakauchi and Horning, 1971; Vollmin, 1971; Thenot and Horning, 1972; Axelson and Sjovall, 1974; Pfaffenberge r and Horning, 1975; Shackleton and Honour, 1976). Demisch and Steib (1968) showed that the trimethylsilyl (T.M.S.) ether derivatives of testosterone and its metabolites gave improved stability and separation of isomers, shorter retention times, and required lower col1.llllll temperatures than the parent compounds. Reactive ketone groups are readily converted to enol-T.M.S. ethers, which are however ~eadilyhydrolysed and therefore are not always entirely s~tisfactory for analytical purposes (Sakauchi and Horning, 1971). It is therefore, preferable to deactivate reactive ketone groups prior to silylation. The 0-methyloximes, described by Fales and Luukkainen (1965) are suitable for this purpose and are less prone to decomposition than the N,N-dimethyl hydrazones (Vanden Heuval and Horning, 1963). Most workers use the now "classic" method of Thenot and Horning (1972) where methoxyamine hydrochloride is used to derivatise reactive ketone groups prior to trirnethylsilylation. The highly hindered 11-ketone group does not form a methoxime, but neither does it show any significant enol-ether formation. Silylation procedures vary, depending on the reactivity of the groups involved, and a number of silyl donors and reaction conditions are in ;use. Thenot and Horning showed that the least reactive of the adrenocorticosteroids, a cortol, could be persilylated in 2 hours at 100°c, when the powerful silyl donor trimethylsilyl imidazole was used in the presence of pyridine hydrochloride, a by- product of the methoxime reaction. It has also been shown that the concentration of methoxyamine HCl used affects the ratio of syn- and anti- isomers produced, and problems may arise if its concentration is not kept constant (Pfaffenberger and Horning, 1975; Shackleton and Honour, 1976). Devaux et al {1971) have denonstrated the successful use of benzyloxirne ketone derivatives which improve the resolution of steroids containing reactive ketones from those with non-reactive ones. It does however, res