IM versily Microfilms International
» t
/
I2'5 2.2 If £ U& u. u i.i IriUU 1.8
1.25 1.4 ssssssss 1.6
MICROCOPY RESOLUTION TEST CHART w. NATIONAL BUREAU OF STANDARDS Si, STANDARD REFERENCE MATERIAL 1010a (ANSI and ISO TEST CHART No. 2)
\ University Microfilms Inc 300 N. Zecb Road, Ann Arbor, M I 48106 INFORMATION TO USERS
This reproduction was made from a copy of a manuscript sent to us for publication and microfilming. While the most advanced technology has been used to pho tograph and reproduce this manuscript, the quality of the reproduction is heavily dependent upon the quality of the material submitted. Pages in any manuscript may have indistinct print. In all cases the best available copy has been filmed.
The following explanation of techniques is provided to help clarify notations which may appear on this reproduction.
1. Manuscripts may not always be complete. When it is not possible to obtain missing pages, a note appears to indicate this.
2. When copyrighted materials are removed from the manuscript, a note ap pears to indicate this.
3. Oversize materials (maps, drawings, and charts) are photographed by sec tioning the original, beginning at the upper left hand comer and continu ing from left to right in equal sections with small overlaps. Each oversize page is also filmed as one exposure and is available, for an additional charge, as a standard 35mm slide or in black and white paper format.*
4. Most photographs reproduce acceptably on positive microfilm or micro fiche but lack clarity on xerographic copies made from the microfilm. For an additional charge, all photographs arc available in black and white standard 35mm slide format.*
*For more information about black and white slides or enlarged paper reproductions, please contact the Dissertations Customer Services Department.
T T A /r.T Dissertation LJ lVll Information Service
University Microfilms International A Bell & Howell Information Company 300 N. Zeeb Road, Ann Arbor, Michigan 48106 8618759
Chandrasekaran, Appavu
MICROBIAL AND HUMAN METABOLISM OF CARDIAC GLYCOSIDES
The Ohio State University Ph.D. 1986
University Microfilms
International 300 N. Zeeb Road, Ann Arbor, Ml 48106 Microbial And Human Metabolism of Cardiac Glycosides
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University
By
Appavu Chandrasekaran, M.S.
•it it ★ ★ *
The Ohio State University
1986
Dissertation Committee: Approved by
Larry W. Robertson, Ph.D.
Richard H. Reuning, Ph.D.
Raymond W. Doskotch, Ph.D.
Duane D. Miller, Ph.D. ^srry/fa. Robertson, Adviser Department of Medicinal Chemistry and Pharmacognosy DEDICATION
To My Family
- ii - ACKNOWLEDGEMENTS
I wish to express my deep and sincere appreciation and gratitude
to my adviser Dr. Larry W. Robertson for his guidance, encouragement,
and support during the course of my graduate studies. I wish to extend my appreciation to Dr. Richard H. Reuning for his encouragement and valuable discussions. My thanks are also extended to Dr. Duane D.
Miller for his interest and encouragement.
1 wish also to specifically thank Dr. C.E. Cottrell, Dr. Raymond
W. Doskotch, and Mr. Jack Fowble for help in obtaining NMR spectra.
The assistance of Dr. Alan G. Marshall, Mr. Ron E. Shomo and Mr. C.R.
Weisenberger in obtaining MS data is also appreciated.
I would like to thank Dr. Jack L. Beal for support and encourage ment during the course of my graduate program.
I would like to express my thanks to my fellow graduate students
and postdoctoral fellows, especially Tim Driscoll, Mohammed Elsheikh,
Doug Geraets, James Hui, John Loper, Bob McClanahan and Raghu Nair,
for their friendship, suggestions and assistance.
I wish also to thank Ms. Carol A. Stewart for typing this thesis.
Finally, I would like to gratefully acknowledge my loving wife,
Thilagavathi, for her support, love and understanding.
-iii- VITA
June 17 , 1936*«« . « . Born - Munivalai, India
June 1973 - June 1976. B.Sc. Botany, University of Madras, India
June 1976 - Aug. 1978 M.Sc. Botany, University of Madras, India
Jan. 1980 - Sep. 1981 M.S. Pharmacy, University of Houston, Houston, Texas
Sep. 1981 - present Graduate Associate College of Pharmacy, The Ohio State University, Columbus, Ohio
Publications:
Robertson, L.W., Chandrasekaran, A., Reuning, R.H., Hui, J., Rawal, B.D.: "Reduction of Digoxin to 20R-Dihydrodigoxin by Cultures of Eubacterium lentum", Appl. Environ. Microbiol.,1986, in press.
Shepard, T.A., Hui, J., Chandrasekaran, A., Sams, R.A., Reuning, R.H., Robertson, L.W., Caldwell, J.H., Donnerberg, R.L.: "Digoxin and Metabolites in Urine and Feces: A Fluorescence Derivatization HPLC Technique", J. Chromatogr. (Biomed. Appl.), 1986, in press.
Major Field: Pharmacognosy
-iv- TABLE OF CONTENTS
Page
DEDICATION...... ii
ACKNOWLEDGEMENTS ...... iii
VITA ...... iv
LIST OF TABLES...... xi
LIST OF FIGURES...... xiii
LIST OF NONSTANDARD ABBREVIATIONS...... xvi
GENERAL INTRODUCTION ...... 1
PART I ...... 2
CHAPTER
I. INTRODUCTION...... 3
1.1 History...... 3 1.2 Sources of Cardiac Glycosides ...... 3 1.3 Chemical Structures and Structure Activity Relationships...... 7 1.3.1 Structure Activity Relationships: ...... 12 1.3.1.1 A-B-C-D ring system ...... 13 1.3.1.2 3-OH Group...... 14 1.3.1.3 14-Hydroxy group ...... 16 1.3.1.4 12 6-hydroxy group...... 17 1.3.1.5 Lactone r i n g ...... 18 1.3.1.6 Side-chain at C 3 ...... 19 1.3.1.7 Influence of other structural modifications ...... 21 1.4 Therapeutic U s e s ...... 22 1.5 Mechanism of Action...... 23 1.6 Metabolism of Digoxin and Digitoxin...... 24 1.7 Statement of the Problem...... 35
-v- II. RESULTS AND DISCUSSION 38
2.1 Reduction of Digoxin to 20 R -Dihydrodigoxin . . 38 2.1.1 Introduction...... 38 2.1.2 Stereochemistry of digoxin reduction products...... 39 2.1.3 Hydrolysis of the sugar side chain...... 40 2.1.4 Reduction of the sugar-hydrolyzed metabolites of digoxin...... 40 2.1.5 Control experiments ...... 45 2.1.6 Preparative-scale production of dihydro metabolites...... 45 2.1.6.1 Mixed culture experiments, ...... 46 2.1.6.2 Variation of media and stationary versus shaken cultures...... 46 2.1.6.3 Concentration study ...... 47 2.1.6.4 Tlme-course study ...... 48 2.1.7 Isolation of the reduced metabolites. . . . 48 2.1.8 Mass spectral evidence...... 49 2.1.9 Preparation of dihydrodigoxigenin from dihydrodigoxin ...... 58 2.1.10 CD and NMR spectroscopic evidence .... 58 2.1.11 Conclusion ...... 61 2.2 Identification and Measurement of Digoxin and its metabolites...... 62 2.2.1 Introduction...... 62 2.2.2 Identification of 20R-Dihydrodigoxin as the Major Reduced Metabolite ...... 66 2.3 Reductive inactivation of Digitoxin by cultures of Eubacterium lentum ...... 81 2.3.1 Introduction...... 81 2.3.2 Chromatographic (HPLC) identification of dihydrodigitoxin...... 82 2.3.3 Hydrolysis of the sugar side chain...... 82 2.3.4 Formation of reduced metabolites from DT2, DTI and DTO...... 86 2.3.5 Control experiments...... 88 2.3.6 Large-scale (mg) production...... 88 2.3.6.1 Mixed culture experiments...... 88 2.3.6.2 Concentration study...... 89 2.3.6.3 Time-course study ...... 89 2.3.7 Isolation of the m e t a b o l i t e s ...... 90 2.3.8 Mass spectral e v i d e n c e ...... 90 2.3.9 Preparation of Dihydrodigitoxigenin from Dihydrodigitoxin ...... 98 2.3.10 Determination of the stereochemistry of dihydrodigitoxigenin...... 98 2.3.10.1 Conclusions ...... 105
- vi - III. EXPERIMENTAL ...... 106
3.1 Reduction of Digoxin to 20R-Dihydrodigoxin by cultures of Eubacterlum lentum...... 106 3.1.1 Chemicals and M e d i a ...... 106 3.1.2 Microorganisms ...... 106 3.1.3 Instrumentation...... 107 3.1.4 Pure E. lentum Culture Trans formations...... 107 3.1.5 Fecal Samples...... 108 3.1.6 Extraction Procedures...... 108 3.1.7 Derivatization...... 109 3.1.8 Preparation for HPLC Chromatography. . . 109 3.1.9 H P L C ...... 110 3.1.10 Standard Curves ...... 110 3.1.11 Controls ...... Ill 3.1.12 Preparative-Scale Production of Dihydro Metabolites ...... Ill 3.1.12.1 Mixed cultures ...... Ill 3.1.12.2 Variation of media ...... 112 3.1.12.3 Concentration study ...... 112 3.1.12.4 Time-course study ...... 113 3.1.12.5 General isolation procedures. . . 113 3.1.12.6 Isolation and purification of digoxin reduction product. . . . 115 3.1.12.7 Isolation and purification of other reduced metabolites .... 116 3.1.12.8 Preparation of dihydrodigoxigenin from dihydrodigoxin ...... 116 3.1.13 Spectral Properties of Digoxigenin. . . 117 3.1.14 Spectral Properties of Dihydrodigoxigenin Obtained by Acid Hydrolysis...... 117 3.2 Identification and Measurement of Digoxin and its metabolites...... 118 3.2.1 General Experimental Procedures...... 118 3.2.2 Tests for Bacterial Overgrowth ...... 118 3.2.3 Characteristics of Subject No. 11. . . . 119 3.3 Reductive Inactivation of Digitoxin by cultures of Eubacterlum lentum...... 120 3.3.1 Chemicals and M e d i a ...... 120 3.3.2 Microorganisms...... 120 3.3.3 Instrumentation...... 120 3.3.4 Eubacterium lentum Transformations . . . 121 3.3.5 Fecal Samples...... 121 3.3.6 HPLC Analysis...... 122 3.3.7 Controls ...... 122 3.3.8 Preparative-Scale Experiments ...... 123 3.3.8.1 Mixed culture experiments ...... 123
- vii - 3.3.8.2 Concentration study ...... 123 3.3.8.3 Time-course study ...... 123 3.3.8.4 General isolation procedures 124 3.3.8.5 Isolation and purification of dlgitoxin reduction product . 124 3.3.8.6 Isolation and purification of the reduced metabolites of DT2, DTI and DTO...... 125 3.3.9 Preparation of Dihydrodigitoxigenin from Dihydroodlgitoxln ...... 126 3.3.10 Spectral Properties of Digitoxigenin 126 3.3.11 Spectral Properties of Dihydrodigitoxigenin ...... 126 3.3.12 Screening Microbes for the Transformation of DTO to DGO . . . 127 3.3.13 Preparative-scale Production of DGO from DTO...... 128 3.3.14 Spectral Properties of DGO Prepared from DTO...... 128 3.3.15 Acetylation of DGO ...... 128 3.3.16 Preparation of Dihydrodigoxigenin from Dihydrodigitoxigenin...... 129 3.3.17 Spectral Properties of DHDGO Prepared from DHDTO...... 129 PART II ...... 130 IV. INTRODUCTION 131
4.1 Polar Metabolites of Digoxin ...... 131 4.2 Conjugates...... 132 4.3 Metabolites Formed by Opening of the Lactone Ring . i ...... 135 4.4 Hydroxylated Metabolites ...... 136 4.5 Microbial Transformation - Methods for Identifying Human Metabolites ...... 139 4.6 Statement of the Problem ...... 142
RESULTS AND DISCUSSION ...... 143
5.1 Biotransformation of Digoxigenin by Streptomvces aureus ...... 143 5.1.1 Introduction ...... 143 5.1.2 Screening Microorganisms for Selection of Suitable Organisms ...... 144 5.1.3 Determination of Optimum Fermentation Conditions ...... 147 5.1.3.1 Variation of media ...... 147 5.1.3.2 Amount of digoxigenin added to culture ...... 147
- viii - 5.1.3.3 Extraction procedure ...... 148 5.1.4 Preparative-scale Fermentation ...... 148 5.1.5 Digoxigenin ...... 149 5.1.6 3-Ketodigoxigenin 2 7 ...... 151 5.1.7 3-Epidigoxigenln 2 8 ...... 152 5.1.8 6-Hydroxydigoxlgenin 5 7 ...... 154 5.1.9 C o n c l u s i o n ...... 156 5.2 Biotransformation of Digoxigenin by Syncephalastrum racemosum ...... 157 5.2.1 Introduction...... 157 5.2.2 Determination of Optimum Fermentation Procedures ...... 158 5.2.2.1 Variation of media ...... 158 5.2.2.2 Amount of digoxigenin added to c u l t u r e ...... 158 5.2.2.3 Extraction procedure ...... 158 5.2.3 Preparative-Scale Fermentation ...... 159 5.2.4 Digoxigenin...... 159 5.2.5 7-Hydroxydigoxigenin 6 0 ...... 159 5.2.6 C o n c l u s i o n ...... 161 5.3 Synthesis of 17-Hydroxydigoxigenin and Digoxigenin Disulfate ...... 162 5.3.1 Introduction...... 162 5.3.2 17-Hydroxydigoxigenin 6 2 ...... 163 5.3.3 Digoxigenin dlsulfate 4 5 ...... 164 5.3.4 C o n c l u s i o n ...... 164
VI. EXPERIMENTAL ...... 165
6.1 Biotransformation of Digoxigenin by Streptomvces aureus ...... 165 6.1.1 Chemicals and M e d i a ...... 165 6.1.2 Preparation of Digoxigenin from D i g o x i n ...... 165 6.1.3 Microorganisms...... 166 6.1.4 Instrumentation...... 166 6.1.5 Screening Experiments ...... 166 6.1.6 Determination of Optimum Fermentation Procedures ...... 167 6.1.6.1 Variation of Media ...... 167 6.1.6.2 Amount of Digoxigenin ...... 167 6.1.6.3 Extraction Procedure ...... 167 6.1.7 General Isolation Procedures ...... 168 6.1.8 Isolation and Purification of Digoxigenin Metabolites ...... 168 6.1.9 Synthesis of 3-Ketodigoxigenin and 3,12-Diketodigoxigenin 169 6.1.10 Synthesis of 3-Epidigoxigenin 28 . . . . 169 6.1.11 Preparation of 3,6-Diketodigoxigenin 5 9 ...... 169
- ix - 6.1.12 Physical Properties of DGO 5 ...... 170 6.1.13 Physical Properties of 3-Ketodlgoxigenln 27 ...... 170 6.1.14 Physical Properties of 3,12-Diketodigoxlgenln 58 170 6.1.15 Physical Properties of 3-Epidigoxigenin 28 171 6.1.16 Physical Properties of 6-Hydroxydigoxigenin 57 171 6.1.17 Physical Properties of 3,6-Diketodlgoxlgenin 59 ...... 171 6.2 BlotransformatIon of Digoxigenin by Svncephalastrum racemosum ...... 172 6.2.1 General Experimental Procedures ...... 172 6.2.2 Isolation and Purification of Digoxigenin and the metabolite 60 . « . 172 6.2.3 Physical Properties of Digoxigenin . . . 173 6.2.4 Physical Properties of 7-Hydroxydigoxigenin 60 173 6.3 Synthesis of 17-Hydroxydigoxigenin and Digoxigenin disulfate ...... 174 6.3.1 Chemicals...... 174 6.3.2 Synthesis of 17-Hydroxydigoxigenin 62 . . 174 6.3.3 Synthesis of 17-Hydroxydigoxigenin Diacetate 63 174 6.3.4 Synthesis of Digoxigenin Disulfate 45 . . 175
VII. LIST OF REFERENCES ...... 176
APPENDIX...... 188
- x - LIST OF TABLES 1. Botanical sources of cardiac Glycosides of
clinical importance...... 6
2. Summary of transformation reactions catalyzed
by Eubacterlum lentum...... 31
3. Chromatographic (HPLC) detection of reduced products
from digoxin and sugar-hydrolyzed metabolites...... 44
4. Mass spectrometic identification of digoxin and
its reduced and sugar-hydrolyzed metabolites...... 55
5. Percent molar excretion of digoxin and
metabolites in subject 11...... 67
6. Percent molar recovery of digoxin and metabolites
in human urine...... 71
7. Percent molar recovery of digoxin and
metabolites in feces...... 72
8. Subject classification based on gastrointestinal
characteristics...... 74
9. Correlation of GI classification and
dihydrodigoxin formation...... 78
10. Chromatographic (HPLC) detection of digitoxin and
its reduced and sugar-hydrolyzed metabolites...... 85
11. Mass spectrometric identification of digitoxin and
its reduced and sugar-hydrolyzed metabolites...... 95
12. Organisms screened and the results obtained
for the conversion of digitoxigenin to digoxigenin...... 102
-XI- 13. Microbial transformation of digitoxin
and digitoxigenin...... 141
14. Organisms screened and the results obtained for the
formation of polar metabolites from digoxigenin...... 146
-XII- LIST OF FIGURES 1. Structures of some cardiac glycosides...... 9
2. Structures of cardenolide and bufadlenollde...... 10
3. Structures of some sugars commonly found In
cardiac glycosides...... 11
4. Formation of 3-epidigitoxigenin from digitoxigenin...... 20
5. Major known metabolites of digoxin...... 26
6. Major routes of digitoxin metabolism...... 34
7. Reduction of digoxin and its sugar-hydrolyzed
metabolites by E. lentum...... 42
8. Chromatograms (HPLC) of derlvatized reference standards
and extracted and derlvatized sample E. lentum culture
incubated with digoxin...... 43
9. FAB and LD/FT/ICR mass spectra of digoxin...... 51
10. FAB and LD/FT/ICR mass spectra of dihydrodigoxin...... 53
11. MS Fragmentation pattern of dihydrodigoxin...... 54
12. MS Fragmentation pattern of dihydrodigoxigenin...... 57
13. CD Spectrum of dihydrodigoxigenin obtained by
bacterial reduction...... 59
14. Structures of digoxin and its major metabolites...... 65
15. Chromatograms (HPLC) of derlvatized standards (50ng each)
(I), of extracted and derlvatized urine sample with a
urine blank (1 ml) (II) and of an extracted and derlvatized
fecal sample with fecal blank (1 g) (III) from subject 11
receiving digoxin...... 68
16. Chromatogram (HPLC) of an extracted and derlvatized fecal
sample (1 g) with a fecal blank from subject 10 receiving -XIII- receiving digoxin...... 73
Chromatograms (HPLC) of derlvatized reference digitoxin
(100 ng) (I) and an extracted derlvatized sample of
E. lentum culture incubated with digitoxin...... 84
Formation of dihydro metabolites from digitoxin and its
sugar-hydrolyzed metabolites...... 87
FAB and LD/FT/ICR mass spectra of digitoxin...... 92
FAB and LD/FT/ICR mass spectra of dihydrodigitoxin...... 94
MS Fragmentation pattern of dihydrodigitoxigonin...... 97
The two possible configurations of substituted y-lactones... 98
CD Spectrum and associated structure of dihydrodigitoxigenin obtained by bacterial incubations...... 100
Microbial conversions of dihydrodigitoxigenin to dihydrodigoxigenin...... 103
Synthesis of cardenolide glucuronides...... 133
Synthesis of cardenolide sulfates...... 134
Formation of digoxigenin-acid...... 135
Synthesis of 17 a-hydroxy cardenolides...... 138
MS Fragmentation of digoxigenin...... 150
Synthesis of 3-ketodigoxigenin and 3,12-diketodigoxigenin... 153
Synthesis of 3-epidoxigenin...... 153
MS Fragmentation of 6 8-hydroxydigoxigenin...... 155
MS Fragmentation of 7 8-hydroxydigoxigenin...... 160
^-NMR Spectrum (CD3COCD3> 270 MHz) of
20R-dlhydrodigoxigenin...... 189
*H-NMR Spectrum (CD3COCD3> 270 MHz) of digoxigenin...... 190
-XIV- 36. *H-NMR Spectrum (Pyridine-d^, 500 MHz) of digoxigenin...... 191
37. *H-NMR Spectrum (CD^COCD^, 270 MHz) of 3-ketodigoxigenin...... 192
38. ^-NMR Spectrum (Pyridine-dy 500 MHz) of
3-ketodigoxigenin...... 193
39. *H-NMR Spectrum (Pyridine-d^, 500 MHz) of
3,12-diketodigoxigenin 194
40. *H-NMR Spectrum (CD^COCD^t 270 MHz) of 3-epidigoxigenin...... 195
41. 1H-NMR Spectrum (CD3C0CD3> 270 MHz) of
6 8-hydroxydigoxigenin...... 196
42. ^H-NMR Spectrum (Pyridine-d^, 500 MHz) of
6 8-hydroxydigoxigenin...... 197
43. *H-NMR Spectrum (Pyridine-d^, 500 MHz) of
3.6-diketodigoxigeni n ...... 198
44. COSY (Pyridine-d^, 500 MHz) spectrum of
3.6-diketodigoxigeni n ...... 199
45. *H-NMR Spectrum (Pyridine-d^, 500 MHz) of
7 8-hydroxydigoxigenin...... 200
46. 1H-NMR Spectrum (CD3C0CD3> 270 MHz) of
17 a-hydroxydigoxigenin diacetate...... 201
-XV- LIST OF NONSTANDARD ABBREVIATIONS
Compound Abbreviation
Digoxin DG3
Digoxigenin bis-digitoxoside DG2
Digoxigenin mono-digitoxoside DG1
Digoxigenin DGO
Dihydrodigoxin DHDG3
Dihydrodigoxigenin bis-digitoxoside DHDG2
Dihydrodigoxigenin mono-digitoxoside DHDG1
Dihydrodigoxigenin DHDGO
Digitoxin DT3
Digitoxigenin bis-digitoxoside DT2
Digitoxigenin mono-digitoxoside DTI
Digitoxigenin DTO
Dihydrodigitoxin DHDT3
Dihydrodigitoxigenin bis-digitoxoside DHDT2
Dihydrodigitoxigenin mono-digitoxoside DHDT1
Dihydrodigitoxigenin DHDTO
-xvi- GENERAL INTRODUCTION
The research described in this dissertation involves the microbial and human metabolism of cardiac glycosides. It is divided into two parts, with each part divided into component chapters. Part I describes the inactivation of cardiac glycosides, digoxin, digitoxin and their sugar-hydrolyzed metabolites, by the intestinal bacterium
Eubacterlum lentum. The stereochemistry of the dihydro metabolites of these drugs were determined using spectroscopic techniques. Part II describes the microbial and chemical transformation of digoxigenin as an approach to preparation of potential human metabolites. The structures of the metabolites of digoxigenin were determined using spectroscopic and chemical techniques.
-1- Part I
INACTIVATION OF CARDIAC GLYCOSIDES
BY EUBACTERIUM LENTUM
-2- CHAPTER I
INTRODUCTION
1.1 HISTORY
The cardiac glycosides are in general naturally occurring com pounds in plants and in the venoms of certain toads. Plants contain ing the cardiac steroids have been used as heart drugs, and as arrow and dart poisons at least since 1600 B.C. The healing effects of the sea onion, squill, are described in the Ebers Papyrus (Ca. 1600 B
.C.). Later, the Romans used squill as a diuretic, heart tonic, emet ic, and as a rat poison. The use of extracts of sea onion to enforce diuresis has been mentioned in the Greek Corpus Hippocratum (Ca. 400
B.C.). Toad skins containing cardiac glycosides have been used for various medicinal purposes, even as aids for treating toothaches and as diuretics. The powdered toad skins were used in western countries as folk medicines for dropsy until they were replaced by digitalis.
In Western Europe the digitalis plant wasmentioned in the Welsh
Pharmaceutical book "Meddygon Myddmai" in 1250 under the name "foxes glofa" as a component of ointments and internally as a treatment for headaches and spasms. In 1542 Leonhard Fuchs, a botanist and
- 3 - 4 physician, coined, in his book of herbs, the scientific name "Digital is", described this plant in detail, and produced the first illustra tions of Digitalis purpurea, D. lutea, and D. ianata. He also men tioned the internal and external applications of this plant as a diuretic and laxative (1). From 1746 to 1785 several authors men tioned the use of foxglove internally or locally for a number of unre lated diseases ranging from epileptic fits to skin ulcers.
In 1785, William Withering published the classical monograph on the action of digitalis: "An account of the Foxglove and some of its med ical uses: With practical remarks on dropsy and other diseases." He described the use of this drug in the treatment of cardiac insuffi ciency only indirectly. It is evident from his writing, "it has a power over the motion of the heart to a degree yet unobserved in any other medicine, and this power may be converted to salutary ends," that he was aware that digitalis affected the heart and had a diuretic effect only in certain forms of dropsy (2). However, John Ferrier in
1799 was the first to ascribe to digitalis a primary effect on the heart and relegate the diuretic action to a position of secondary importance (3). About seventyfive years after Withering1s publication on the use of digitalis, German and French physiologists, pharmacolo gists, and clinicians emphasized the action of digitalis on the heart.
During the early twentieth century, the drug gradually came to be looked upon as a specific in the treatment of atrial fibrillation.
Oswald Schmiedeberg, a pharmacologist in Strasburgh in 1875 isolated a fraction from digitalis leaves and called it. "digitoxin." In 1930
Sydney Smith isolated digoxin from the leaves of Digitalis lanata. 1.2 SOURCES OF CARDIAC GLYCOSIDES.
Today, the most commercially Important sources of cardiac glyco sides are two species of Digitalis - D. purpurea, the official digi talis (foxglove), and D. lanata (3). The official squill Is the dried bulb of Urginea maritime (formally Scllla). Table 1 gives the botani cal sources of cardiac glycosides of clinical importance. Table 1. Botanical Sources of Cardiac Glycosides of Clinical Importance.
Plant Source Glycoside Aglycone
Digitoxin Digitoxigenin
Digitalis purpurea Linne' Gitoxin Gitoxlgenin
(leaf) Gitalin Gitaligenin
Digitoxin Digitoxigenin
Digitalis lanata Ehrh. Gitoxin Gitoxlgenin
(leaf) Digoxin Digoxigenin
Strophanthin Strophanthidin
Strophanthus Kotnbe Oliver Cymarin Strophanthidin
(seed) Cymarol S t rophanthido1
Strophanthus gratus Ouabain Ouabagenin
(Wall et Hook.) Baillon
(seed)
Urglnea maritime Proscillaridin A Scillaridin A
(Linne') Baker
or indica Kunth
(bulb) 1.3 CHEMICAL STRUCTURES AND STRUCTURE ACTIVITY RELATIONSHIPS.
The cardiac glycosides exist in plants as precursors, called prima
ry or natural glycosides. For example, the leaves of D. lanata con tain precursors termed lanatosides A, B, and C. Clinically useful drugs are obtained from these precursors by chemical and enzymatic breakdown. For example, upon mild alkaline hydrolysis and enzymatic hydrolysis, to remove an acetyl group and a terminal glucoside residue respectively, lanatoside C yields digoxln, a clinically Important drug. Each glycoside consists of an aglycone or genin, and one to four molecules of sugar (Figure 1). The aglycones possess a substi tuted 10,13-dimethyl cyclopentanoperhydrophenanthrene nucleus with hydroxy groups at 3 and 14 positions.
CH
CH
OH HO
Substituted Cyclopentano Perhydrophenanthrene
Additional hydroxy groups at carbon 12 (e.g., digoxigenin) and car bon 16 (e.g., gitoxlgenin) may be present in some aglycones. Also attached to the steroid aglycones at the seventeenth position is an a, (J-unsaturated five-membered lactone ring (butenolide) in cardeno- lides or a diunsaturated six-membered lactone ring in bufadienolides
(Figure 2).
The genins have unusual stereochemical configurations. The fusion of A-B, B-C, and C-D ring systems of the steroid nucleus is cis-anti-trans-anti-cis. This is different from androgens and corti costeroids, in which the A-B, B-C, and C-D ring systems are trans fused. The substitutions i.e. the two methyl groups and the hydroxy groups at the 3 and 14 positions, and the unsaturated lactone ring at the 17th position all have the beta configuration.
CH CH,
OH
Conformation of a typical Cardiac Glycoside - Genin
CH CH
Conformation typical of androgens and carticosteroids CH CH
CH H
OH
1^ Rj » R2 ■ H Digitoxin
2 Rj » OH R2 - H Digoxin
2 Rj “ H R2 ■ OH Gitoxin
Figure 1. . Structures of some cardiac Glycosides
VO 10
OH CH; CH, CH, CH, OH OH HO HO 4 Dlgitoxlgenln 5 Digoxlgenln
CH, CH, OH CH
OH OH HO HO'
6 Cltoxlgcnln 7 Strophanthidin
OH I
OH HO 8 Ouabagenln
9 Bufalln CH, 10 Scillarldln A CH,
CH, CH
OH OH HO HO Figure 2. Structures of cardenolldes and Bufadlenolldes 11
Several types of sugars are found attached to the C3-hydroxy group of the aglycones (Figure 3). Digitoxose, a 2,6-dideoxyhexose is the sugar present in the clinically important agents digitoxin and digox- in. The sugars are attached to each other by a 0, 1-4 linkage (Fig ure 1).
OCH,
HO OH OH OH 6-D-Glucose a-L-Rhamnose
CH, CH,
H< HO
OH OCHF u n l 8-D-Digitoxose 6-D-Cymarose
Figure 3. Structures of some sugars commonly found in cardiac Glycosides 12
1.3.1 Structure Activity Relationships:
Before discussing the structure activity relationships of cardioac tive steroids in detail a major drawback should be mentioned. That is, this area is still somewhat poorly defined because a great number of different testing methods have been used on different animals, iso lated organs, enzyme systems etc. by different laboratories, and there is no reliable way to compare the results. Until it was discovered that all compounds that are toxic and inhibit Na+ ,K+-ATPase are not inotropic, many relatively simple features were considered essential for the cardiac steroids. Also 15 years ago, most of the structural elements which are "absolutely necessary" for cardiotonic activity seemed to be evident, but subsequent synthetic modifications in the structure of cardiac glycosides disproved some of these prerequisites for a positive inotropic effect. 13
1.3.1.1 A-B-C-D ring system.
The A-B rings are cis fused in the naturally occurring glycosides.
A change of configuration at C5 yields less active compounds (4).
This loss of activity is less pronounced in the cardenolides than in the bufadienolides (5). But there are some A-B trans fused cardeno lides such as 11 that have been shown to be more potent than digoxin
CH,
HO.
o^c P OH
CHj 11 Asclepin
The influence of a double bond at C4-CS or CS-C6 is not clear. The cis configuration of the C and D rings seems to be an important fea ture. 14a-digitoxigenin 12 and 14a-artebufogenin 13 are inactive in comparison to the corresponding 0-analogs (7,8).
CH
CH
HO OH 12 14a- digitoxigenin 14
O
13 14a-aftebufogenln
HO 1 .3 .1 .2 3-OH Group.
The 33-hydroxy group was previously regarded as an Indispensable structural requirement for the cardiac activity. Later Saito et. al. synthesized 3-deoxydigitoxigenin 14 and found an activity comparable to that of digitoxigenin on isolated frog heart (9).
.O A S-r
14 3-deoxy digitoxigenin Zurcher et al. showed a 50% loss of activity for this compound in an ATPase test (7). Epimerization at C3 of the cardenolide digitox igenin, to form the 3o-digitoxigenln I5» diminishes the effect in the
ATPase test to about 1/30 of that of digitoxigenin (7). 16
CH,
CH,
OH HO 15 3a-dlgltoxigenln
1.3.1.3 14-Hydroxy group.
A free tertiary hydroxy group at C14 is consistently present In
naturally occurring cardenolldes. This hydroxy group Is not essential
for the activity, but its substitution by hydrogen as in 16 is accom
panied by a considerable loss of activity (10).
CH
CH
HO 16 14-deoxy digitoxigenin
In an early study, dehydration of the 14-OH to give a double bond at carbons 8 and 14 as in 17 or 14 and 15 as in 1A has been shown to abolish the cardiotonic activity (11). This may be due to the loss of 17 the cis relationship of angular methyl group at carbon 13 and the hydroxy group at 14.
CH CH
CH CH
HO HO 17 4®*^anhydro 18 A^'^anhydro digitoxigenin digitoxigenin
1.3.1.4 12 8-hydroxy and 168-hydroxy groups
Additional hydroxy groups at carbon 12 (e.g., digoxlgenin) and at carbon 16 (e.g., gitoxigenin) primarily change the hydrophilic- lipophilic properties and cause more rapid elimination (12).
/C=N CH. C = C CH. CH,
OH OH HO HO
19 20 18
1.3.1.5 Lactone ring
The a,fi-unsaturated lactone side-chain has been regarded as the most essential functional group for the cardiac activity. Synthesis of compounds with open chain structures of varying electronic and steric resemblance to the lactone ring has demonstrated that activity still exists when the lactone is replaced with a 170 group that has an
a,0-unsaturated radical with an electron-rich hetero atom. For exam- ple, the substitution of the C17 side-chain in digitoxigenin, leading to the trans-acrylic acid methyl ester 19 and the trans-acrylic nitrile 20 results in active compounds (13, 14). Replacing the lac tone ring with a guanyl hydrazone group diminishes the effectiveness slightly, but the resulting derivative is still of comparable potency
(15). Saturation of the double bond in the lactone ring to give dihy dro derivatives reduces the activity by tenfold or more (16). Opening of the ring completely abolishes activity (16). It is clear that the
C17 side-chain must be 0-oriented, since a change of configuration yields inactive compounds such as 21 (9).
CH.:
CH
OH HO 21 19
1.3.1.6 Side-chain at C3.
As mentioned before, a variety of sugars are found attached to the
C3 hydroxy group of the aglycone. The sugars commonly found in cardi ac glycosides either alone or in a series of two, three, or four com bined with other monosaccharindes are glucose, rhamnose, digitoxose, and cymarose (Figure 3)., The sugars substituted at the axial
3P-hydroxy group have several effects. Complete removal of the sugars to produce the genin reduces inotropic effectiveness and results in toxicity that is equal to or greater than that of the corresponding glycoside (12). This may be due to the change in lipophilic- hydro philic balance caused by removal of the polar sugar groups. Another reason may be that in vivo, sugars attached to the C3-0H protect the
C3-0H from being rapidly metabolized to the 3-one and then reduced to the less active 3o-0H isomer (Figure 4) (12). Glycosides with only one sugar group linked to the C3 hydroxy function appear to be more active than diose, triose, and tetraose glycosides (12).
Even though the glycosides with fewer sugars than digoxin are shown to be more potent than digoxin in in vitro enzyme preparations, they may not be better agents in in vivo clinical applications due to their undesirable properties with regard to uptake, metabolism and elimina tion. For example, glycosides with one sugar are known to be more rapidly metabolized and excreted than triose glycosides (1). There fore the digitoxose sugar side chain plays a major role in the uptake, transport, distribution, metabolism and elimination. 20
CH
CH
OH HO 4 Digitoxigenin
T
CH
CH
OH
22 3 Ketodigitoxigenin
CH
CH
OH HO'' 15 3epi-digitoxigenin
Figure 4. Formation of 3 epi-digitoxigenin from digitoxigenin 21
1.3.1.7 Influence of other structural modifications.
Introduction of halogens into different positions of the steroid nucleus resulted in the retention of the cardiac activity, however it never produced compounds that were more active than naturally occur ring glycosides (15). Addition of hydroxy groups in the steroid nucleus generally retains activity (9,10,15). Cardenolides with unusual structures have been synthesized and found to be active. For example, Stache et al. synthesized the cardenolide, cardio-propellane
23, with an additional steroidal ring, and showed it to possess posi tive inotropic activity in the guinea-pig heart (17). It is also reported that the cardenolide with an open A ring, 24, has some cardi ac activity (18).
CH,
HO
OH 23 cardio-propellene ,o
CH
CH
OH O I OCH 24 22
1.4 THERAPEUTIC USES
The cardiac glycosides have a great number of effects upon the heart, but by far their most important clinical use is to treat con gestive heart failure (3). The incidence of congestive heart failure is very high, and it can occur in any type of heart disease. The digitalis glycosides are of greatest value especially in treating low output cardiac failure. They are useful regardless of whether the failure is predominantly of the left or right ventricle or involves both. By increasing the force and the rate of myocardial contraction in the failing heart, digitalis glycosides bring about increased car diac output, kinesis, reduction of edema, decreased heart size and reduced central venous pressure. This is commonly called the inotrop ic action of the cardiac glycosides.
The second use of digitalis glycosides is in the management of cer tain cardiac arrhythmias, such as atrial fibrillation, atrial flutter, and paroxysmal atrial tachycardia (3). The ventricular rate is inap propriately rapid in atrial fibrillation, and it may cause a reduction of the cardiac work capacity that may lead to heart failure. The car diac glycosides reduce the ventricular rate. In the case of atrial flutter the ventricular rate is rapid and irregular due to the atrial impulses to the A-V node. The digitalis glycosides prolong the appar ent refractory period of the A-V node. This results in a decrease in the ventricular rate. A-V nodal paroxysmal tachycardia is the most common tachyarrhythmia next to atrial fibrillation. Digitalis glyco sides by their action on the vagal tone are often successfully used to treat this disorder. 23
1.5 MECHANISM OF ACTION
The mechanism of action of the cardiac glycosides is complex and has not been completely elucidated. In a patient with congestive heart failure, the digitalis glycosides improve the function by increasing the force and efficiency of the heart muscle without caus ing a corresponding increase in oxygen consumption. This effect is accomplished by the cardiac glycosides by the mechanism which may involve some or all of the following factors: (1) Influence on the transport of Na+ and K+ across the heart muscle membrane, resulting in a loss of cell K+ ; (2) increase in available intracellular calcium concentration which may be an indirect result of the above factor; (3) direct action on the contractible proteins of the heart muscle.
It is known that cardiac steroids interact with the sodium- potassium pump located in the cell membrane of the myocardial cell, and there is considerable evidence to support that these glycosides actually inhibit sodium-potassium dependent ATPase(19,20). The sodium-potassium pump is responsible for maintaining the unequal dis tributions of Na+ and K+ ions across the cell membranes. Depolariza tion of the membrane changes the permeability of the heart cell mem branes to allow sodium to move into the cell by passive diffusion and potassium to move out. The sodium-potassium pump reverses this pro cess. Energy (ATP) is required to transport potassium ions against concentration gradient and is catalyzed by Na+ , K+-ATPase. The inter action of the cardiac steroids with the receptor site on the sodium- potassium pump leads to the inhibition of ATPase hydrolysis and the efflux and Influx of sodium and potassium ions respectively. There is
also evidence to support the theory that the inhibition of the sodium-
potassium pump by the cardiac glycosides is produced by a conforma
tional change in the pump (21).
As a result of the inhibition of the sodium-potassium pump, there
is a gradual increase in intracellular sodium and a gradual small
decrease in intracellular potassium. Hie increase in intracellular
sodium is believed to be crucially related to the positive inotropic
effect of digitalis glycosides. This is so because, according to a
hypothesis by Raker, when intracellular sodium is increased due to
inhibition of the pump by the cardiac glycosides, the exchange of
extracellular sodium for intracellular calcium is diminished and
intracellular calcium is increased (22,23). A consequence of this may
be an increase in stored calcium in the sarcoplasmic reticulum avail
able for the muscle contraction process.
Gervais et al. hypothesized that the positive inotropic effect is produced by the direct interaction of the cardiac glycosides with the
cell membrane in such a way that an increased amount of calcium is bound to the cell membrane (24). The net effect is an increase in
intracellular calcium available for the heart muscle contraction pro
cess .
1.6 METABOLISM OF DIGOXIN AND DIGITOXIN
Digoxin and digitoxin have been thought to escape extensive meta bolic degradation in the human body, and were believed to be eliminat 25
ed as unchanged drugs in the urine. However, it is now known that the
metabolism of these drugs is significant in some patients (25 - 29).
The importance of metabolism as an elimination pathway for digoxin is
revealed by the results of Lukas (28). He used a specific double iso
tope dilution derivative method and found only 25 to 55% of an oral
dose of digoxin eliminated unchanged in the excreta. In another
study, Okita et al. found that only 6% of the parent glycoside was
eliminated in the urine when digitoxin was administered orally (26).
The two known major routes of metabolism of digoxin (DG3) 2 are
illustrated in Figure 5. One route consists of cleavage of the digi
toxose sugar moieties attached to the C-3 position of the steroid
nucleus to form digoxigenin bis-digitoxoside, (DG2) 25 digoxigenin
mono-digitoxoside (DG1) 26 and digoxigenin (DGO) 5. Digoxigenin is
known to undergo further metabolism mainly by oxidation to
3-ketodigoxigenin 27, which can subsequently be reduced to
3-epidigoxigenin (3o-digoxigenin) 28 (30). There is evidence to sug gest that digoxigenin mono-digitoxoside, digoxigenin and
3-epidigoxigenin may be further metabolized by conjugation (31,32).
However, these metabolites are incompletely characterized. The bis and mono-digitoxosides of digoxigenin exhibit approximately the same cardioactivity as the triglycoside, digoxin, whereas digoxigenin and subsequent more polar metabolites are shown to be much less cardioac tive (33 - 35). 26
CH, CH ,0H
CH OH
OH
n-3 2 Digoxin (DG3) n-3 29 Dlhydrodlgoxln (DHDG3) n-2 25 Digoxigenin n«2 30 Dihydrodigoxlgenin bisdlgitoxoside (DG2) bisdlgitoxoside (DHDG2) n-1 26 Digoxigenin n-1 31 Dihydrodigoxlgenin monodigitoxoside (DG1) monodigitoxoside (DHDG1) n-0 5 Digoxigenin (DGO) n-0 32 Dihydrodigoxlgenin (DHDGO) 27 3 Keto digoxigenin (3 O-DGO) 28 3 epi digoxigenin (3 Epi-DGO)
Figure 5. Major known metabolites of digoxin. The hydrolysis of cardiac glycosides into the genins and sugar
moieties by enzymatic cleavage has been reported in the very early
literature (36,37). Haack et al. were the first to confirm this
cleavage pathway and identify bis- and mono-digitoxosides of digoxige
nin and digitoxigenin. Earlier studies which involved quantitation of
the hydrolyzed metabolites indicated that the hydrolysis pathway of
digoxin metabolism is minor with less than 15% of the total urinary
excretion of lipid soluble cardenolides due to the sum of digoxigenin bis-digitoxoside, digoxigenin mono-digitoxoside and digoxigenin, and
their enzyme hydrolyzable conjugates (39,40). In one study, Clark and
Kalman found that the hydrolyzed metabolites constituted less than one per cent of digoxin and metabolites extracted from urine in four of six subjects investigated (41) However, in the other two subjects the levels of digoxixigenin, digoxigenin mono-digitoxoside, and digox igenin bis-digitoxoside were on the average 18.5%, 3% and 4.5% of the lipid extractable cardenolides found in the urine.
Recent studies by Gault et al., underscore the importance of this hydrolysis pathway of digoxin metabolism. They have demonstrated that the hydrolysis pathway occurs at least in part within the gastrointes tinal tract (42, 43). The liver is another likely site of hydrolysis
(44,45). They found that substantial degradation of digoxin occurs by the action of hydrochloric acid in the human stomach, and this hydro lysis is stimulated by pentagastrin infusion. Under normal conditions of stomach acidity, the sugar hydrolyzed metabolites accounted for about 10% of the total radioactivity in the urine. When pentagastrin 28
was infused for 30 minutes before and 60 minutes after oral adminis
tration of digoxin, the percentage of hydrolysis products approximate
ly doubled in the urine (43). Further, digoxigenin, a weakly cardio
active metabolite was found to be the major constituent of the hydro
lysis products under these conditions. These results suggest that
hydrolysis of digoxin by stomach acid may be significant in patients with disease conditions that promote gastric acidity.
The second route of metabolism of digoxin is the reduction of the double bond in the lactone ring to form dihydro compounds, which show
10 to 100 fold lower positive inotropic effect than that of the parent glycoside (16). In 1968 Luchi and Gruber reported the presence of the reduced metabolite, dihydrodigoxlgenin (DHDG0) 32, the aglycone of dihydrodigoxin (DHDG3) 29, in the urine of a patient requiring unusu ally high doses of digoxin (46). Dihydrodigoxin was subsequently dis covered as a metabolite in human plasma and urine by Watson, Clark and
Kalman (47). In another study Clark and Kalman discovered dihydrodi- goxigenin bis-digitoxoside (DHDG2) 30 and dihydrodigoxlgenin mono digitoxoside (DHDG1) 31 as metabolites of digoxin in the urine (41).
The importance of the formation of reduced (dihydro) metabolites as a major route of metabolism for digoxin in significant minority of patients is evident from several studies. The percentage of isolated dihydrodigoxin has been found to vary widely from patient to patient.
Previous studies found 0.2% to 2% of digoxin and metabolites in urine as dihydrodigoxin in seven subjects after an oral administration of digoxin (48,49). In another study, dihydrodigoxin was found to con stitute 12 to 20% of the digoxin maintenance dose in the urine of nine
patients (50,51). Others have found a range of 1-47% (average 13%) of
the total glycosides in the methylene chloride extract of urine as
dihydrodigoxin in the 50 patients surveyed (41). Peters and his co
workers investigated the urine samples of 100 patients receiving
digoxin (27,53). The average percent of dihydro metabolites present
in the urine was 12% (range 2-52%) based on the total methylene chlo
ride extractable drug plus metabolites. Of the 100 patients, 5%
excreted over 10% and seven subjects excreted over 35% of the reduced
metabolites. In a more recent study Lindenbaum et al. found that
about 10% of digoxin patients form substantial amounts of dihydro
metabolites (the reduced metabolites constituting greater than 40% of
total urinary digoxin and metabolites) (25).
Although the saturation of the double bond between C20 and C22 of
lanatoside C was suggested to be effected by intestinal microorganisms
as early as 1961 by Herman and Repke (53), only recently has it been
demonstrated that dihydrodigoxin was formed in man in the gastrointes
tinal tract and exclusively by the action of enteric bacteria. It is evident from the results of Lindenbaum and his co-workers' studies that formation of reduced metabolites of digoxin is accomplished by
intestinal bacteria in man (25,54). The wide intersubject variability
in the extent of urinary excretion of dihydro metabolites of digoxin appears to be due to variability in the intestinal microflora. Coad ministration of an antibiotic (erythromycin or tetracycline) along with a daily dose of digoxin to three patients, who were previously 30 shown to excrete substantial amounts of reduced metabolites, resulted in nearly a doubling of the steady-state serum digoxin. It appears from the lack of excretion of dihydro metabolites after antibiotic treatment that other metabolizing organs are not involved in this reduction pathway. Administration of readily bioavailable digoxin capsule or intravenous administration of digoxin reduced the amount of dihydrometabolites excreted in the urine, while delayed absorption of slow dissolving granules of orally-administered digoxin increased the extent of reduced metabolites formation (55). Further the microor ganism Eubacterium lenturn, isolated from the feces of a patient who is known to be an excretor of reduced metabolites, was found to be capa ble of converting digoxin to dihydro metabolites (56,57). Of 28 E. lentum strains tested for reducing digoxin, only 18 strains performed this reaction, including the type strain, ATCC 25559. Of 13 species of Eubacterium examined, only Eubacterium lentum could perform the reduction.
Eubacterium lentum is a common obligatory anaerobe of the human colonic flora. It is a non-sporeforming, non-motile gram positive rod
(56). Table 2. Summary of Transformation Reactions Catalyzed by Eubacterium Species Substrates Products Reaction Type Referer 21-Hydroxy-preg-4-ene-3,20-dione Pregn-4-ene-3,20-di one Dehydroxylation 58,59 (Deoxycorti costerone) (Progesterone) 3a,21-Dihydroxy-58-pregn-20-one 3a-Hydroxy-5B-pregn-20-one Dehydroxylation 58 3a, 12a-Dihydroxy-5B-cholan-24-oic acid 12a-Hydroxy-5B-cholan-24-oi c Dehydroxylation 60 Ursodeoxycholic acid Lithocholic acid Dehydroxylation 61 Cholest-5-en-3B-ol (Cholesterol) 5B-Cholestan-3B-ol Reduction 62 Cholest-4-en-3-one 5B-Cholestan-3B-ol Reduction 62 (24R)-Ergost-5-en-3B-ol (Campesterol) (24R)-5B-Ergostan-3B-ol Reduction 62 Stigmast-5-en-38-ol ( -Sitosterol) 5B-Stigmastan-3B-o1 Reduction 62 3B-Hydroxy-androst-5-en-17-one 3B-Hydroxy-5B-androstan-17-one Reduction 62 Pregn-5-ene-3B,20B-diol 56-Pregnane-3B,20B-diol Reduction 62 Linoleic acid Trans-vaccenic acid Reduction 60 16-Dehydro-progesterone 17-Isoprogesterone Reduction 63 Progesterone 5B-Pregnadione (major) and Reduction 63 5a-Pregnadione (minor) 5B-Pregnadione 5-Hydroxy-pregn-20-one Reduction 63 3a,21-Dihydroxy-5B-pregnan-20-one 21-Hydroxy-58-pregnane-3,20- Oxidation 58 di one 3a-Hydroxy-56-pregnan-20-one 5B-Pregnane-3,20-dione Oxidation 58 Chenodeoxycholic acid Ursodeoxycholic acid Epimerization 64
LJ 32
E. lentuni and other Eubacterlum species are known to perform a
variety of transformations on xenoblotic and native molecules. Table
2 illustrates the types of reactions catalyzed by these anaerobes.
For example E. lenturn has been shown to dehydroxylate deoxycortecos- terone and tetrahydro deoxycortecosterone to yield progesterone and
3a-pregnanolone respectively (58,59). Eubacterlum species strain 144 catalyzes the reduction of 16-dehydroprogesterone to
17-isoprogesterone (63) This strain also reduces the C4-C5 double bond of progesterone to give mostly 53-pregnadione and some
5a-pregnadione, and reduces the keto group of 5f)-pregnadione to a hydroxy group (64).
The present knowledge of digitoxin metabolism is very limited due to the fact that the biotransformation is complex, involving a number of enzymatic processes. The known routes of metabolism of digitoxin are shown in Figure 6. Hepatic microsomal enzymes are known to hydroxylate digitoxin (DT3) to form digoxin (65,66). The sugar side chains of digitoxin and digoxin are hydrolyzed to yield their respec tive bis- and mono-digitoxosides and genins (67,68). The genins thus formed in turn undergo epimerization at C3. Furthermore, .evidence of conjugates and other polar metabolites of the drug have been found, but the products have not been fully characterized (65). Bodem and
Unruh have reported the presence of dihydrodigitoxin (DHDT3) in the plasma of patients on digitoxin maintenance therapy (69). It is pos sible that the sugar hydrolyzed metabolites of digitoxin may also be present in the reduced form, as previously shown for digoxin. There is a significant complication in the cardiac therapy due to
the toxic effects encountered with digitoxin and digoxin (3). The
heart, the gastrointestinal tract and the central nervous system are
the body parts most affected by these drugs. The most serious adverse
effect of digitalis glycosides is their cardio toxicity. Metabolites of cardiac glycosides may also contribute to the toxic responses observed. 34
R - I.£> /^{x
n-3 R-H n-3 Rj-H n-3 Rj-OH 35 Dihydrodigitoxin ^ Dlgitoxin 2 Digoxin (DT3) n-2 Rj-OH n“2 Rl“H 25 (DG2) 33 Digitoxigenin bisdigitoxoside n-1 R.-OH n"1 Rl“H n-0 R.-OH - monodigitoxosideDl8l5?XJ r nln^ — 5 (DGO) (DTI) n-0 Rj«H 4 Digitoxigenin (DTO) Figure 6. Major Routes of Digitoxin Metabolism 35 1.7 STATEMENT OF THE PROBLEM Formation of dlhydro metabolites is a major route of digoxin metab olism. Reduction of the C20-C22 double bond of the lactone ring of digoxin leads to the introduction of a chiral center at C20. The chi ral center at C20 of the resulting lactone can have either R or S con figuration. Standard chemical catalytic hydrogenation results in the formation of two dihydrodigoxin epimers, the 20R and 20S epiraers being formed in a ratio of 3:1 (70). Recent investigations by Reuning et al. indicate that the DHDG3 epimer formed in humans is the 20R (71). The pharmacological potency or toxicity can vary depending on the stereochemistry of the reduction products. For example, it has been reported by previous investigators that the 20R and 20S epimers of C22 methylene cardenolides such as 36 showed considerable differences in inhibiting Na+ , K+ ATPase (72). CH, CH OH HO 36 36 Since the stereochemistry of the reduced metabolites formed by incubating E. lentum with digoxin or other cardiac glycosides has not been studied before, the present investigation was designed in part to determine the stereochemistry of the product(s). Although it has been shown that E. lentum can convert digoxin to one or more reduction products, the products have only been detected and quantitated by a radioimmunoassay, which gives no information as to whether the digitoxose sugars at carbon 3 are present or hydrolyzed (56,57). Likewise, it was not known whether the bacterium could con vert the sugar-hydrolyzed metabolites to their respective dihydro products. Since gastric juice of patients who have an acidic pH can hydrolyze the sugar side chain from D63 leading to the formation of DG2, DG1 and DGO, such hydrolysis products would be available to the intestinal bacteria to perform further conversion (reduction) reac tions (43). To investigate the latter possibility, the sugar- hydrolyzed metabolites were incubated with E. lentum. It has not been investigated so far whether E. lentum is capable of converting digitoxin and its sugar-hydrolyzed metabolites to their respective dihydro metabolites. In the present project digotoxin and its sugar-hydrolyzed metabolites were incubated with E. lentum and analyzed for dihydro metabolite formation and metabolite stereochemis try. To summarize, the objectives of the present project were: (1) To study biotransformation of digoxin by pure Eubacterium lentum cul tures; (2) to determine whether reduction products are formed; (3) to 37 define stereochemistry of digoxin reduction products formed; (4) to determine whether these cultures are capable of hydrolyzing the sugar side chain; (5) to describe substrate specificity of the reduction (Is the reaction specific for digoxin or will it work on the sugar hydro lyzed metabolites of digoxin and other cardiac glycosides like digi toxin and its sugar hydrolyzed metabolites?), and (6) to determine the stereochemistry of the reduction products of digotoxin if they are formed. The stereochemistry of dihydrodigoxin formed by E. lentum was determined using an HPLC procedure using references of known stereo chemistry and spectroscopic studies. To confirm the configuration at C20 of dihydrodigitoxin, it was decided that the stereochemistry of the reduction product, if formed by E. lentum, would be related to that of dihydrodigoxin. Fusarium roseum was found to perform the nec essary 12B-hydroxylation to convert dihydrodigitoxigenin to dihydrodi- goxigenin by a screening experiment for this comparison. In order to relate the studies with pure cultures of E. lentum to conversion in the human GI tract, fecal specimens from a volunteer who is known to excrete dihydrodigoxin as a metabolite of digoxin would be incubated with digoxin, digitoxin and their sugar-hydrolized metabo lites individually and analyzed for dihydro metabolites formation. CHAPTER II RESULTS AND DISCUSSION 2.1 REDUCTION OF DIGOXIN TO 20R-DIHYDRODIGOXIN BY CULTURES OF EUBACTERIUM LENTUM 2.1.1 Introduction It has been recently shown that Eubacterlum lentum, a common anae robic bacterium of the human gut, Inactivates the cardiac glycoside digoxin by converting it into reduced metabolites (56). The analyt ical method (i.e. radioimmunoassay) used in previous studies lacked specificity for either one (20R or 20S) of two epimers that can be formed by reducing digoxin (70). Therefore the stereochemistry of the reduction products (dihydro metabolites) formed from digoxin by E. lentum cultures is not known. It is also not known whether E. lentum can hydrolyze the sugar side chain of digoxin to yield digoxigenin bis- and mono-digitoxosides, and digoxigenin, and can convert the sug ar hydrolyzed metabolites (DG2, DG1 and DGO) to their respective dihy drometabolites. In order to determine the stereochemistry of the dih ydro metabolite(s) formed and the specificity of this reaction, Eubacterlum lentum ATCC 25559 cultures were incubated with digoxin and its sugar-hydrolyzed metabolites and the products were analyzed by a recently developed specific HPLC assay. The type strain E. lentum - 38 - 39 ATCC 25559 was chosen for this study since it is readily available as an identified bacterium from American Type Culture Collection, and since it has been shown to be one of 18 strains of E. lentum that can perform this reaction (56). 2.1.2 Stereochemistry of digoxin reduction products. Incubation of digoxin (10yg/ml) with growing cultures of Eubacterl um lentum ATCC 25559 for 7 days at 37°C results in a nearly quantita tive reduction of the cardiac glycoside to its dihydro metabolite, 20R-DHDG3, as shown in Figure 7. Examination of chromatogram II in Figure 8 shows clearly that most of the starting material has been converted to a single product, with very small amounts of the starting material and one additional product also present in the culture extract. Comparison with chromatogram I and the HPLC retention times shown in Table 3 indicate that the major product is 20R-DHDG3. The 20R epimer is identical to the major product found in the urine of those persons who are capable of converting DG3 to its dihydro metabo lite (71). That the reduction proceeds with high stereospecificity in E. len tum cultures may be seen by comparing chromatograms I and II in Figure 8 . Reference commercial dihydrodigoxin when derivatized with 1-naphthoyl chloride and chromatographed according to the procedure described in section 2 .2 .12, shows two separate peaks, (peaks f and g in chromatogram I) one having a retention time of 10.2 minutes (major, 40 20R-DHDG3) and one 8.6 minutes (minor, 20S-DHDG3). The commercial DHDG3 is prepared by chemical catalytic hydrogenation, and the 20R and 20S epimers are formed in a ratio of three to one (71). The stereos- pecificity was found to be the same as that obtained with E. lentum when samples from an incubation of DG3 with feces from a volunteer, who was known to excrete the reduction product in urine, were ana lyzed. The 20R-DHDG3 accounts for greater than 99% of the products formed in both lentum and fecal incubations. 2.1.3 Hydrolysis of the sugar side chain. The HPLC chromatographic method utilized in this study is also capable of detecting and quantitating the sugar-hydrolyzed metabolites of DG3 (DG2, DG1 and DGO), and their reduction products (DHDG2, DHDG1 and DHDGO), which show characteristic retention times of their naphthoyl derivatives seen in Table 3. Examination of the chromato gram II in Figure 8 , as well as other incubations of DG3 with E^ len tum or with fecal samples shows no evidence of the hydrolysis of the deoxy sugars from position 3 of the steroidal nucleus. 2.1.4 Reduction of the sugar-hydrolyzed metabolites of digoxin. E. lentum not only converts DG3 to DHDG3, but also converts the sugar-hydrolyzed metabolites of DG3 (DG2, DG1, and DGO) to their respective dihydro metabolites. Analysis of samples from incubation of E. lentum, or feces of a known reduction product former, with DG2, DG1 or DGO resulted In formation of pairs of reduction products which were identified by co-chromatography with reference compounds. The 20 R epimer constituted greater than 99% of the products formed in each case (Figure 7). Representative retention times are shown in Table 3. Therefore the sugar-hydrolyzed digoxin metabolites are good substrates for reduction by E. lentum. This represents an important addition to the contribution of Lindenbaum et al. (56), who detected the formation of digoxin reduction products using a radioimmuno assay, which could distinguish between the unsaturated and reduced lactone ring, but told nothing about the status of the remainder of the cardiac glycoside molecule. 42 CH o h ]|Hl *1—r ~ ° 0 K CH H --0 ‘ Eubacteriuml lentum H v 2o L 2 O I 2 0 ^ UNSATURATED LACTONE REDUCED LACTONE DG3 n= 3 -►DHDG3 n=3 DG2 n= 2 -►DHDG2 n= 2 DG1 n= 1 -►DHDG1 n= 1 o c DGO n ■►DHDGO n= 0 Figure 8 . Reduction of Digoxin and Its Sugar-hydrolyzed Metabolites by Eubacterlum lentum Detector Response iue . hoaorm (HPLC) ofDerivatized Chromatograms 7.Figure eeec Sadrs (I)and Extracted an and Standards Reference ie n iue Tm i minutes in Time minutes in Time Dsrivfltizcd lentum cultureE«ofSample incubated Digoxin.with s\ I CD CD to Q. tn 0 5 10 Table 3. Chromatographic (IIPLC) detection e( reduced products fron digoxin and sugar-hydrolyzcd metabolite*.* Compound Derlvatiatcd Abbreviation Retention Time 1ml Digitoxigenin DTO S. 2 Digoxigenin DGO 5.6 20S-Dihydrodigoxigenin 20S-DHDCO 6.1 20R-Dihydrodigoxigcnin 20R-DHDC0 6.5 Digoxigenin r.eot-digitoxoside DCl 6.3 20S-Dihydrodigoxigcnln mono- dlgltososidc 20S-DHDG1 6.5 20R-Dihydrodigoxlgcnln mono- dlgitososide 20R-DHDG1 7.2 Digoxigenin bis-digitoxoside DC2 7.A 20S-Dihydrodigoxigcr.in bis* digitoxosidc 20S-DHDC2 7.6 20R-Dihydrodigoxigcnin bis- digitoxoside 20R-D!iDG2 7.9 Digoxin DC3 6.2 20S-Di!>ydrodigoxin 20S-DMDC3 6.6 20R-Dihydrodigoxin 20R-DKDC3 10.2 Lichrosorb Si 60; Hexane: Methylene chloride: Acetonitnle (5:1:1) 45 2.1.5 Control experiments Control experiments were performed to ensure that the peaks seen In the HPLC chromatograms were not due to interference from the media or the bacteria. These included extracting blank media, deriva- tizing the extract with 1-naphthoyl chloride and analyzing the deriva tives by HPLC. Also, digoxin was added to sterile medium and auto- claved E. lentum cultures, the tubes were incubated for 7 days and analyzed for metabolites. Blank E. lentum cultures were incubated with solvent (DMF, 0.16ml) and analyzed for interfering peaks. The controls did not show any conversion, and there were no interfering peaks from the microorganism or culture media which had retention time near either of the dihydrodigoxin epimers. 2.1.6 Preparatlve-scale production of dihydro metabolites. Along with sensitivity of HPLC, the limitations must also be appre ciated. Even though the controls (blanks) did not produce any inter fering peaks near the metabolites, it is still important to confirm the metabolites by isolating sufficient quantities where possible to identify by spectroscopy and other physical and chemical methods. To confirm the stereochemical properties of the reduction products formed by E. lentum it was decided to prepare these metabolites in large (mg) quantities. The separation of the parent unsaturated compounds from their respective dihydro metabolites by chromatography is a long, time consuming process. Therefore, the culture conditions were opti mized so that the starting materials could be quanitatively converted 46 to reduced metabolites in milligram quantities, thus avoiding the sep aration processes. 2.1.6.1 Nixed culture experiments The use of a facultative anaerobe as an oxygen scavenger in mixed culture experiments has been reported by Bokkenheuser et al. (58,59). They used Escherichia coli to provide the necessary anaerobic atmos phere for the growth of E. lentum in large Erlenmeyer flasks in their corticosteroid transformation (dehydroxylation) experiments. It was decided to use the same organism, E. coli, in the mixed culture exper iments because of its ready availability. E. coli ATCC 10536 by itself, when incubated with digoxin, was not able to perform the reduction reaction. When E. lentum was cultured in larger volumes in Erlenmeyer flasks, it did not grow well (cultures were not dense), and when it was incubated with digoxin, it did not produce quantitative conversion. On the other hand, total conversions were consistently achieved in the mixed culture experiments. Thus by making use of the mixed culture technique for growing the anaerobic bacterium the need to use expensive anaerobic hoods was eliminated. 2.1.6.2 Variation of media and stationary versus shaken cultures. E. lentum is known to grow poorly in many media (74). Therefore it was necessary to test the suitability of other media for its best growth and conversion. The media examined for growing E. lentum were: chopped meat carbohydrate broth (CMC), cooked meat medium (CMM), tryp- ticase soy broth(TS), thioglycollaete medium with and without dextrose 47 or indicator (TG) and Brain Heart Infusion broth with added cysteine (BHIC) (59). TG and BHIC media were chosen to be used in subsequent experiments based on their ability to give better growth of the bac terium and conversion of DG3 to DHDG3. It was found that when the thioglycollate medium with dextrose was used the conversion did not take place in high yield. Another experiment, in which one set of mixed ( E. lentum and E. Coli) cultures were incubated with digoxin in stationary (static) form and another set on a rotary shaker (250rpm), indicated that shaken cultures did not show any conversion and sta tionary cultures produced 100% transformation. Perhaps exposure of the anaerobic bacterium to oxygen prevents it from growing and hence stops the transformation, even in the presence of the oxygen scaven ger E. coli. 2.1.6.3 Concentration study In previous studies digoxin was incubated at a concentration of lOyg/ml of E. lentum culture medium and reported to be quantitatively converted to dihydro metabolites(56). Therefore in the initial exper iments of this study the same amount (lOyg/ml) of digoxin was added. Later, in order to determine the highest concentration (amount/ml) of DG3 that can be transformed 100% to DHDG3 by E. lentum, a concentra tion study was carried out. In this experiment the amount of digoxin added was varied in two-fold increments from lOyg/ml to 320yg/ml. Results from this experiment showed that total conversion to DHDG3 takes place, when DG3 was incubated with E. lentum up to a concentra tion of 80yg/ml. Above this amount only partial conversions were seen. 48 2.1.6.4 Tlae-conrse study In the initial studies digoxin was incubated with E. lentum cul tures for 7 days and then analyzed for metabolites. This 7 day incu bation period was chosen on the basis of literature references (56). Later on, in order to determine the minimum incubation time period required for the formation of at least some dihydro metabolites and for total conversion to dihydro metabolites, a time-course study was performed. In this time course study samples were analyzed every hour for 30 hours from the time of inoculation and addition of digoxin. The results clearly showed that the accumulation of dihydrodigoxin begins after approximately 16 hours of incubation and growth, with conversion being essentially complete in 24 hours. Therefore, while the original incubation studies were carried out for 7 days, it is possible to obtain quantitative conversion in 24 to 36 hours. 2.1.7 Isolation of the reduced metabolites. Results from the above experiments indicated that the optimal pro cedure for generation of dihydro metabolites involved the use of mixed cultures in BHIC medium. The optimal overall concentration of the starting material added to the cultures appeared to be 80yg/ml of medium and the optimal length of incubation for total conversion of the unsaturated parent compounds to the reduced metabolites appeared to be 48 hours. Using these conditions, preparative-scale fermenta tions were performed to isolate sufficient quantities of the reduced metabolites for stereochemical determination. The metabolites were isolated in pure form by methylene chloride extraction and silica gel 49 chromatography, followed by preparative TLC (PTLC) In some cases. Incubation of 500 mg of digoxin with mixed cultures of E. lentum and E. coll followed by extraction and purification by chromatography gave 392 mg of crystalline dihydrodigoxin. Similar incubations but in smaller scale resulted in the isolation of dihydro metabolites of the sugar hydrolyzed digoxin derivatives. The structures of the compounds were elucidated by comparison of their spectral properties with those of the unsaturated parent compounds. The spectroscopic techniques employed in structure elucidation included mass specrometry (MS), cir cular dichoism (CD), and proton NMR (*H-NMR) spectroscopy. 2.1.8 Mass spectral evidence. Mass spectrometry is one of the most important tools for identifi cation of organic molecules, but the low volatility of glycosides like digoxin poses a major problem. For example, the conventional electron impact mass spectrometry (EI/MS) does not produce a molecular ion for digoxin (51). One way of overcoming this problem is to increase the volatility by chemical derivatization. Watson et al. used the chemi cal derivatization approach to identify dihydrodigoxin in human urine (47). However, their gas chromatography/mass spectrometry (gc/ms) method could not differentiate between dihydrodigoxin and the agly- cone, dihydrodigoxigenin. This is because in the derivatization reac tion sugar residues of dihydrodigoxin were removed by hydrolysis to give a dihydrodigoxigenin derivative, and this was not different from the product obtained by derivatizing dihydrodigoxigenin itself. Therefore special ion formation techniques like fast atom bombardment (FAB), negative-ion chemical ionization or laser desorption are neces sary to obtain a detectable signal in the molecular ion region (75). A.P. Bruins used a negative ion (0H~) desorption chemical ionization technique and obtained a large pseudomolecular ion peak at m/z 779 for (M-H)+ as base peak for digoxin (76). Kistemaker et al. and Posthu- mas et al. also were able to get clear pseudomolecular ion peaks for digoxin using laser desorption-magnetic sector methods (75,77). In the present study, FAB and Laser desorption Fourier Transfor/Ion Cyclotron Resonance (LD/FT/ICR) techniques were used to identify the reduced metabolites. The FAB technique used in this study often yielded a very low or no pseudomolecular ion, whereas the LD/FT/ICR technique always produced a large pseudomolecular ion (M+K)+ , allowing the ready identification of the metabolites. Both techniques gave characteristic fragment ions for each of the parent unsaturated com pounds and corresponding dihydro metabolites. The FAB mass spectrum of reference digoxin showed a very small (M+H)+ ion at m/z 781. Successive loss of each of the sugars produced corresponding peaks at m/z 651, 621 and 391. Other characteristic peaks (m/z 373,355, and 337) due to losses of the three hydroxy groups as water molecules were also evident (Figure 9). The LD/FT/ICR spec trum of the same sample (DG3) gave a very tall pseudomolecular ion at m/z 819 for (M + K)+ , corresponding to a molecular weight of 780 for digoxin. In contrast to the FAB mass spectrum the large pseudomolecu lar ion is also the base peak in the LD/FT/ICR mass spectrum. Other fragment ions seen in the FAB mass spectrum were also observed in the LD/FT/ICR mass spectrum (Figure 9). 51 r m iTTi m i 'n in r tnT IT I I f I I I I I I M 700 730 *n n *rh W i ft i i ^ i^ riT^'T » n n n rTi i »11 ni ii i i r i i i ■ i i 300 350 400 450 500 350 10C. 113 819 (B+K) (A+K) oH 400 * 500 600 700 800 Figure 9. FAB (I) and Ld/FT/ICR (II) mass spectra of digoxin 52 The metabolite obtained from incubation of E. lentum with digoxin showed a (M+Na)+ » pseudomolecular ion in FAB, peak at m/z 805. Green wood and Snedden in their study used the characteristic fragment ion at m/e 338 of dihydrodigoxin for identifying and measuring this metabolite in human urine samples (51). This fragment ion at m/e 339 (338 in EI/MS) along with a peak at m/e 357 were clearly seen in the FAB mass spectrum of the purified metabolite (Figure 10). These frag ments arise by the loss of the sugar side chain and the loss of 2 or 3 hydroxy groups in the form of water. Figure 11 illustrates the frag mentation pattern that leads to the formation of these peaks. The LD/FT/ICR spectrum also confirmed that the isolated metabolite is dih ydrodigoxin by giving an Intense (M + K)+ , pseudomolecular ion, peak at m/z 821 (Figure 10). The mass spectra obtained by using FAB and LD/FT/ICR techniques clearly showed the metabolites isolated from the incubations of digox igenin mono-and bis-digitoxosides and digoxigenin were their respec tive dihydro compounds. Also the mass spectra obtained for each of the reduced compounds prepared by catalytic hydrogenation of the unsa turated parent drugs were essentially identical to those of the iso lated metabolites. The pseudomolecular ions and the key fragment ions observed for these compounds are listed in table 4. t m r r m i r m & ,131 Oh II O 821 (M+K) 0-1 A o-itL UJL CH, CH, OH. OH 782 +"23 - 805 OH OH CH, CH, 392 OH HO O—S,o OH CH, CH, OH .O OH CH, CH, 356 CH CH, 338 Figure II. MS Fragmentation Pattern of Dihydrodigoxin Table 4. Mai* Spectrometrlc Identification of Digoxin and lta Reduced and Sugar-hydrolyzed metabolite*. Compound (H.W.) FAB K**« Spectrum LD/FT/ICR Ma«« Spectrum Pseudomolecular Key Paeudoaolecular Key Ion (R.I.Z) Fragment* Ion (R.I.%) Fragment* DC3 (780) 781 (H*H)+ 651,521,391, 819 ( W K ) + 651,521,391 (0.2) 373,355,337 (100) DHDC3 (782) 805 (M*la)+ 357,339 821 (H+K)+ 357,339 (0.6) (100) DC2 (650) 651 (l«H)+ 503,391 689 ( H * ) + 355,337 (1.8) (100) DHDG2 (652) 675 (M-Wa)+ 339 691 (M4K)+ 561,431,339 (1.9) (100) DCl (520) 521 ( m « ) + 373,355,337 559 (H4K)+ 373,355,337 (l.«) (100) DHDC1 (522) 545 (H+K*)+ 357,339 561 (H4K)+ - (1.3) (100) DGO (390) 391 O H H ) + 373,355,337 429 ( M « ) + 373,355,337 (14.1) (100) DKDCO (392) - 357,339 431 0*HC)+ 357,339 (100) 56 Because of the higher volatility of the genins compared to their parent glycosides, the Electron Impact (El) technique can be used to produce a large number of fragment ions and thus important information necessary to characterize them. Digoxigenin and its metabolite iso lated by incubating the former compound in the E. lentum were subject ed to El ionization and 2 sets of mass spectra were obtained. Even though the molecular ions were not seen in these spectra, the fragment ions clearly showed that the purified metabolite is indeed the dihydro derivative of the parent compound. The fragment ions for the metabo lite were two mass units higher than those of digoxigenin. The mass spectrum obtained for the product by catalytic hydrogenation of DGO again could be matched almost exactly with the one for the bacterially produced metabolite. The postulated fragmentation pattern of the dih ydro metabolite (DHDGO) is illustrated in Figure 12. This pattern is 4* i postulated based on peaks appearing at m/z 374 (M -I^O), 356 (M -2 H2O), and the fragmentation patterns described as diagnostic for other similar cardiac steroids (78-80 ). 57 014 CM, HO ,o OH CH. OH CH, CH. 374 OH OH ,o CH OH CH, 219 OH CH, CH, 356 CH 302 CH, 201 CM CHjI CH, CH, 338 284 162 \ CH, 162 of Dihydrodigoxlgenl" Figure 12. MS Fr.g-nt.tion Pattern 58 2.1.9 Preparation of dihydrodigoxigenin from dihydrodigoxin Since the sugar moieties of dihydrodigoxin interfere with the interpretation of various spectra, it was necessary to hydrolyze the sugar side chain and obtain the pure aglycone for the assignment of stereochemistry. An acid hydrolysis procedure to cleave the sugar side chain followed by silica gel chromatography gave pure dihydrodi- goxigenin from dihydrodigoxin produced by E. lentum reduction. In order to determine whether the acid hydrolysis procedure used in obtaining dihydrodigoxigenin from dihydrodigoxin altered the original stereochemistry of the lactone ring by any means it was tested by hydrolyzing dihydrodigoxin obtained from a commercial source and known to have a mixture of 20R and 20S epimers in a 3:1 ratio. Dihydrodi goxigenin obtained from this test hydrolysis had a mixture of 20R and 20S epimers in the same 3:1 ratio (70). It is clear from this test that the acid hydrolysis of the dihydrodigoxin obtained from E. lentum incubations, did not affect the stereochemistry of the lactone ring. The same acid hydrolysis method was used to obtain the aglycone from each of the reduced metabolites with fewer sugars than dihydrodigoxin. 2.1.10 CD and NMR spectroscopic evidence Using HPLC, Bockbrader and Reuning were able to successfully sepa rate the two epimers of dihydrodigoxigenin prepared by catalytic hydrogenation of digoxigenin (70). Also they assigned the stereochem istry of the major component as 20R and the minor component as 20S based on circular dichroism curves and NMR spectra (70). The configu ration at C20 of the minor epimer was confirmed to be S by an indepen dent x-ray crystallographic study (81). 59 The circular dlchroistn spectrum of the 20R-DHDG0 showed a positive curve ([0 ]2 i8!B+136) and the 20S-DHDGO a negative curve ([8]2i6!="206). The CD spectra of the aglycones obtained from the metabolite dihydro digoxin by acid hydrolysis, and from the reduced metabolites with few er or no sugars showed positive curves similar to the one reported for the 20R epimer (Figure 13). 200 ISO 100 -50 -1 0 0 Wavelength (am) Figure 13. CD Spectrum of dihydrodigoxigenin obtained by bacterial reduction. 60 The 20R and 20S epimers of DHDGO also show significant differences in chemical shifts of C21 and C22 protons in the proton NMR. These differences are caused by the differences in interactions of these protons with the hydroxy group at C12 position (70). This is evident from the pronounced changes in chemical shifts that occur for the C21 protons of the 20R epimer due to proximity to the aromatic moiety, when the C12 hydroxy group is derivatized with 3,5 dinitrobenzoyl chloride (70). No such interaction is seen between the C21 protons and the dinitrobenzoate ester of the 20S epimer. The proton NMR spectra obtained in CD3COCD3 for the aglycones of the reduced metabolites showed triplets at A.00 and 4.40 ppm for the C21 protons. Comparison of these values and other regions of the spectra with those of the previously characterized reference 20R-DHDG0 clearly showed that they are essentially identical. The stereochemis try of each of the dihydro derivatives obtained by incubating DG2, DG1 and DG0 individually with mixed cultures of E. lentum and E. coll was also found to be 20R by obtaining the aglycones by acid hydrolysis of the glycosides and comparing *H-NMR of them with the previously char acterized 20R-DHDG0 epimer. 61 2.1.11 Conclusion The anaerobic bacterium Eubacterlum lentum, a common constituent of the Intestinal microflora, Inactivates dlgoxin by reducing the unsatu rated lactone ring. Reduction of the cardiac glycoside by growing cultures of E. lentum ATCC 25559 proceeds In a stereospeclfic manner and the configuration at C20 of the resulting product is determined to be R. This Is in contrast to the 3:1 ratio of 20R and 20S epimers formed in the chemical catalytic hydrogenation. E. lentum does not hydrolyze the digitoxose sugars from carbon 3 of the parent drug. However, the synthetically prepared sugar-hydrolyzed metabolites of dlgoxin (DG0,DG1 and DG2) are reduced to the corresponding dihydro metabolites. The configuration at C20 of these reduced metabolites was also found to be R. 62 2.2 IDENTIFICATION AND MEASUREMENT OF DIGOXIN AND ITS METABOLITES IN HUMANS 2.2.1 Introduction Dlgoxin, the most widely used drug for the treatment of heart dis ease, was long believed to escape extensive metabolic degradation (82). However, It Is now evident that a significant portion of this drug may be converted to Inactive metabolites in approximately 10% of patients (25). The recognized routes of dlgoxin metabolism are a hydrolytic route and a reductive route (83). The hydrolytic pathway leads to the formation of approximately equally active compounds with fewer sugars attached to the steroid nucleus and a less cardioactive genin (34,84). The reductive pathway, associated with a specific bac teria in the GI microflora, leads to the formation of inactive dihydro metabolites (Figure 14) (56,85). The degree of formation of dihydro metabolites appears to vary widely among patients. One group has reported dihydrodigoxin accounting for 1-47% of the methylene chloride extractable parent drug and metabolites in the urine (41). Another group has reported a range of 2-52% as dihydrometabolites in the lipid-extractable glycosides (27). Lindenbaum et al. estimated greater than 40% of urinary glycosides representing dihydro metabo lites in about 10% of dlgoxin patients (25). Moreover, their results indicate that the variability in extent of excretion of dihydro metabolites may be due to variability in the intestinal microflora that are responsible for formation of reduced metabolites (56). Chem ical catalytic hydrogenation of the double bond in the lactone ring of dlgoxin, which is used to prepare commercially available reference dihydrodigoxin, leads to the formation of two epimers, a 20R-epimer constituting 75% of the products and a 20S-epimer amounting to 25% (71). Reuning et al. have recently determined the stereochemistry of dihydrodigoxin excreted in humans as 20 R (71). The analytical meth ods (i.e. radioimmuno assays and chromatographic assays) used in pre vious studies for measuring dihydro metabolites lacked specificity for the 20R-dihydrodigoxin epimer. They could not distinguish between dihydrodigoxin and the sugar-hydrolyzed and reduced metabolites (56,73,83). Also some of the radioimmuno assays (RIA's) developed for quantitating dihydrodigoxin were known to have some cross reactivity with the unsaturated parent drug itself (25,27,41,83). Therefore the range of amounts of excreted dihydro metabolites reported in the ear lier literature may represent both the parent drug and a mixture of the reduced metabolites. Since the dihydro metabolites were quanti tated only in the urine in previous studies, the percentage of reduced metabolites excreted in the feces and the total amount of dihydro metabolites formed are not known. As part of the present ongoing study on the effects of aging and GI disease on digoxin metabolism, significant information concerning the identity and quantities of spe cific metabolites has recently been obtained. A specific HPLC assay has been used to identify and measure digoxin and its hydrolytic and dihydro metabolites in this present study. This assay can separate the two epimers of dihydrodigoxin from one another and from digoxin, and it can provide distinct peaks for the sugar-hydrolyzed and reduced 64 metabolites as well. The dihydro metabolites with fewer sugars (20R epimers) obtained from the E. lentum incubations described previously were used as reference compounds in the present HPLC assay. CH CH OH OH CH H --0 OH DGO n = 0 DIGOXIGENIN 21 OGI n: 1 DIGOXIGENIN MONO-DIGITOXOSIDE OG2 n = 2 DIGOXIGENIN BIS-DIGITOXOSIDE 22 DG3 n= 3 DIGOXIN 20R-DHDG3 n = 3 R= 20R-DIHYDRODIGOXIN Figure 14. Structures of Digoxin and Its Major Metabolites in 66 2.2.2 Identification of 20R-Dihydrodlgoxin as the Major Reduced Dlgoxin Metabolite In Humans Of the 11 subjects studied, one (No.11) was found to be signifi cantly different from the others in the types and quantities of metabolites excreted. Dihydrodigoxin, specifically the 20R-epimer, constituted over 50% of the dose excreted in this subject (Table 5). This epimer accounted for 24% of the dose in urine and 27% in feces (Figure 15). Other metabolites found in the urine included digoxige- nin and digoxigenin bis digitoxoside, and these represented less than 2% of the dose. Trace amounts (less than 5 ng) of these two metabo lites were also seen in the fecal samples. The other epimer of dihy drodigoxin (20S) was not present in measurable amounts in any of the samples assayed. Unchanged digoxin accounted for 45% of the dose in urine, and only trace amounts of digoxin were seen in fecal samples. 20R-Dihydrodigoxin accounted for almost all of the glycosides recov ered in feces. Nearly all of the dose (97% of dose) administered in this subject was recovered in the form of unchanged drug and metabo lites. Table 5. Percent Molar Excretion of Dlgoxin and Metabolites In Subject 11 Total Subjects DCO DG1 DC2 DG3 R-DHDC3 Percentage of Dose Recovered Urine 1.47 0.00 Trace 45.10 23.81 70.38 11 Feccs Trace 0.00 Trace Trace 26.69 26.69 Total 1.47 0.00 Trace 45.10 50.50 97.07 Urine 3.60 1.67 0.94 25.56 1.18 32.95 (0.15-15.57) (0.00-6.54) (0.00-2.97) (8.14-54.27) (0.00-6.52) (14.21-56.83) Others* Fcces 2.04 2.98 2.12 12.41 2.42 21.97 (0.00-13.31) (0.00-15.63) (0.00-11.21) (0.00-48.68) (0.00-11.24) (3.22-57.72) Total 5.64 4.65 3.06 37.97 3.60 54.92 (0.15-20.66) (0.00-15.91) (0.23-12.63) (12.08-73.26) (0.00-13.62) (29.96-87.05) *Mean of ten and range in brackets. iue 5 Crmtgas HL) fDrvtzdSadrs 5 gec) (I), (50 of (HPLC)Standardseach)ng Derivatized of Chromatograms 15.Figure DETECTOR RESPONSE I L. i I I I i _i j v and of an Extracted and Derivatized fecal sample with Fecal Blank (lg) FecalfecalBlank sample with andDerivatized Extracted of anand (II) (1 aml)Urine Sampleblank Urinewith and Derivatized Extracted II fo ujc 11(III)digoxin. receiving fromsubject I (min) E TIM __ —i —L I— i— I— 1 m O O i * o fO o o i w o z o a: o 10 V) OJ 00 ON 69 This study involved analyzing both urine and feces for digoxin and metabolites. In previous studies, dihydrodigoxin was measured only in the urine. Also, as mentioned earlier, the RIA's that were used in the earlier studies to quantitate dihydro metabolites, lacked specificity for the 20R-epimer, the only epimer found in humans (71). They may not have been able to differentiate dihydrodigoxin from other potential reduced metabolites like dihydrodigoxigenin bis- and mono- digitoxosides and dihydrodigoxigenin (73). The RIA's also showed some cross reactivity not only with digoxin but also with digoxigenin bis and mono-digitoxosides, and digoxigenin (83). Therefore the percent ages of dihydrodigoxin reported in the earlier literature may not accurately reflect the actual amounts of the reduced metabolites which are formed. The analytical method used in the present study is capable of separating the reduced metabolites with fewer or no sugar from dihy drodigoxin, digoxin and other hydrolytic metabolites. Dihydrodigox igenin was the only other reduced metabolite, apart from dihydrodigox in, found in subject 11, and that too was present in trace amounts (less than 5 ng). Since over 50% of the dose was recovered as 20R-dihydrodigoxin in this subject, there was ample opportunity for the formation of detectable amounts of other reduced metabolites. It is clear from these results that 2OR-dihydrodigoxin is the major reduced digoxin metabolite formed in humans. Also, since essentially all of the administered dose was recovered (97%), it is evident that 70 dihydrodigoxin formed in this subject is not being further metabo lized. Therefore, the excretion of 51% of digoxin dose as 20 R-dihydrodigoxin is an accurate reflection of the amount of dihydrodi goxin actually formed. The results from the present study agree with the earlier report by Lindenbaum et al. that dihydro metabolites accounts for over 40% of the administered digoxin dose in some patients (25). The fraction of administered digoxin found as 20R-dihydrodigoxin in the other ten sub jects ranged from 0 to 14% (Figure 16). The details of the metabo lites identified in these subjects and their characteristics are given in Tables 6 and 7 (87) and the ranges and mean are shown in Table 5. The unusually large percentage of dihydrodigoxin formation in subject 11 may be due to bacterial overgrowth which appears to be present in the intestine. Diagnostic tests performed for bacterial overgrowth yielded positive values for subject 11 (Table 8). TVo other subjects, who also showed signs of bacterial overgrowth, excreted a maximum of only 4% of the dose as dihydrodigoxin. Therefore bacterial overgrowth may not be the only cause for the formation of large amounts of dihy dro metabolites. Bacterial overgrowth was often found to be associat ed with hypochlorhydria (abnormally low secretion of stomach acid) in the subjects studied. Table 6. Percent Molar Excretion of Dlgoxin and major metabolites In Hunan Urine. DCO DC1 DG2 DG3 20R-DHDG3 Total Total Total Subject GI Classification a in Urine in Feces 1 NANB 2.82 0.00 0.52 16.88 6.52 26.74 3.22 29.96 2 LANB Trace 6.54 Trace 19.34 0.53' 26.41 7.30 33.71 3 LANS Trace 1.56 2.97 16.24 1.44 22.21 12.26 34.47 5 NANB 0.15 0.08 0.00 37.52 0.00 37.75 26.28 64.03 6 LAHB (7.35)* 1.65 0.53 8.14 Trace 17.67 17.25 34.94 7 NANB 15.57 3.03 0.23 10.17 0.33 29.33 57.72 87.05 8 LANB 9.28 3.44 1.42 8.56 0.58 14.28 20.62 34.90 9 LAHB Trace 0.28 0.68 28.23 Trace 29.19 27.49 56.68 10 LANB 6.23 0.11 0.52 36.21 2.38 45.45 23.67 69.12 11 LAHB 1.47 0.00 Trace 45.10 23.81 70.38 26.69 97.07 12 NANB b 0.05 2.51 54.27 0.00 56.83 23.79 80.62 ibreviatlons: NANB - "Normal acid, normal bacteria LANB - "Low acid, normal bacteria;" LAHB - "Low acid, high bacteria." ^value could not be determined due to blank interference ()* May not reflect accurate value due to blank interference Table 7. Percent Molar Excretion of Dlgoxin and Major Metabolites In fecea ------Feces------Subject Cl Classification3 DGO DG1 DG2 DC3 20R-DHDG3 Total Total Total In feces In urine 1 NANB Trace 0.00 0.00 0.00 3.22 3.22 26.74 29.96 2 LANB 0.79 0.25 0.59 5.67 0.00 7.30 26.41 33.71 3 LANB 1.56 6.67 0.00 1.48 2.55 12.26 22.21 34.47 5 NANB 0.00 0.00 3.56 22.72 0.00 26.28 37.75 64.03 6 LAHB (13.31)* 0.00 0.00 3.94 0.00 17.25 17.67 34.94 7 NANB 1.83 7.21 0.00 48.68 0.00 57.72 29.33 87.05 8 LANB 0.00 0.00 11.21 5.83 3.58 20.62 14.28 34.90 9 LAHB 0.00 15.63 0.00 8.26 3.60 27.49 29.19 56.68 10 LANA 2.89 0.00 1.05 8.49 11.24 23.67 45.45 69.12 ULAHB Trace 0.00 Trace Trace 26.69 26.69 70.38 97.07 12 NANB 0.00 0.00 A.80 18.99 0.00 23.79 56.83 80.62 Abbreviations: NANB - "Normal acid, normal bacteria," LANB - "Low acid, normal bacteria," LAHB - "Low acid, high bacteria" ()* may not reflect value due to blank Interference 100 ng/ml o lO FECAL SAMPLE FECAL BLANK ■ i i i i I l l I_ l— I— I 0 5 10 RETENTION TIMETmin) Figure 16. Chromatogram (HPLC) of Extracted and Derivatized Standards, a Fecal Sample (lg) and a Fecal Blank from Subject 10 receiving Digoxin. Table 8. Subject Classification Based On Gastrointestinal Characteristics Acid Status Bacterial Status ^Jejunal Aspirate Culture b c aPentagastrin Serum Serum (CFU/ML) Final Subject Stimulation Castrln B-12 ^Sympotomatic eBABT Anaerobic Aerobic Classification 1 Normal 50 591 No Normal * <105 NANB 2 Low Acid 50 322 No Normal * * LANB 3 Low Acid 500 252 No Abnormal ** LANB 5 Normal 100 753 No Normal 4x10* <104 NANB 6 Low Acid 80 517 Yes Abnormal 2xl04 0 LAHB 7 Normal <50 * No Normal 0 0 NANB 8 Low Acid 75 1084 No Normal 0 0 LANB 9 Low Acid <50 1111 Yes Abnormal 5xl06 g1.6xl05 LAHB 10 Low Acid <50 236 No Abnormal l.SxlO2 7xl02 LANB 11 Low Acid 170 513 No Abnormal 5.5xl06 >105 LAHB 12 Normal 200 319 No Abnormal 0 one colony NANB *Tesc or culture results not available. NANB ■ Normal Acid, Normal Bacteria NAHB * Normal Acid, High Bacteria LANB * Low Acid, Normal Bacteria LAHB • Low Acid, High Bacteria Footnotes - - see next page a. Pentagastrln stimulation normals had normal acid output by standard criteria. Abnormal* Included both hypo and achlorhydrlc Individuals. b. Normal value <150. c. Normal values 190-810. d. Diarrhea and weight loss, consistent with bacterial overgrowth. c. Bile acldbreath test: viewed mainly as a screening test for bacterial overgrowth— may suffer from false positive and false negative results. Jejunal aspirate culture counts and sympto matology are primary determinants of the bacterial status of each subject. f. Jejunal aspirate culture results were evaluated using the following arbitrary categories: 2 Total Colony Forming Units (Cri))/ml Jejunal Aspirate (anaerobic or aerobic 0-10* Normal 10*-10^ High Normal >105 High (2) GL Simon and SL Gorbach Gastroenterology 1984;86: 174-196. g. Culture results from a duodenal aspirate obtained approximately six months prior to research study admission. Ln 76 Eubacterium lentum, the bacterium that has been shown to perform the reduction of dlgoxin to Its dlhydro metabolites, Is a poor grower In common bacteriological media, but addition of substrate amounts of arginine to the peptone based media dramatically increases the growth (74). Addition of increasing amounts of arginine produces enhanced growth of this organism but the formation of dihydro metabolites, decreases (56). This indicates that the mere presence of this bacter ium alone does not lead to the formation of large amounts of inactive metabolites. It is speculated that the diet of the patients may be a factor in determining the amounts of reduced metabolites formed. Because the types and amounts of amino acids present in the diet may vary and influence the growth of E. lentum and the reduction reaction. Other factors could not be ruled out. For example, addition of hemin, or pyruvate or presence of hydrogen in the growth medium is known to stimulate and maximize the reductases of the species Eubacterium Sp. strain 144 which catalyzes the reduction of 16-dehydroprogesterone to 17-isoprogesterone(63). Subject ll's upper intestinal aspirate, like the other ten sub ject's samples, when cultured with digoxin, was not able to convert digoxin to the dihydro metabolite. On the other hand, ten of the eleven subjects' (including ll's) fecal cultures converted digoxin almost quantitatively to 20R-dihydrodigoxin. Only one fecal sample, obtained from a subject (No. 12) who was known not to excrete dihydro metabolites, failed to convert dlgoxin to dihydrodigoxin. All the other samples, which included samples from subjects who do not excrete 77 dihydro metabolites, were able to perform this conversion (Table 9). These results indicate that even in the patients, who excrete large percentages (over 40%) of dihydro metabolites, the reduction reaction may take place in the lower intestine only. Further studies are nec essary to determine the generality of the above observations. Table 9. Correlation of Cl Classification and Dihydrodigoxin Formation Dihydrodigoxin Formation IN VIVO IN VITRO Detection Detection Fecal Jejunal Aspirate Subject GI Classification in Urine In Feces Incubation Incubation 1 Normal acid, Normal bacteria 6.52 3.22 * * 2 Low acid, Normal bacteria 0.53 0.00 * * 3 Low acid, Normal bacteria 1.44 2.55 ** 5 Normal acid, Normal bacteria 0.00 0.00 +(100) - 6 Low acid, High bacteria Trace 0.00 +(40) - 7 Normal acid. Normal bacteria 0.33 0.00 +(100) - 8 Low acid, Normal bacteria 0.58 3.58 +(100) - 9 Low acid, High bacteria Trace 3.60 +(100) - 10 Low acid. Normal bacteria 2.38 11.24 +(100) - 11 Low acid, High bacteria 23.81 26.69 +(100) - 12 Normal acid. Normal bacteria 0.00 0.00 -- * 9 Not done or not available + 9 dihydrodigoxin was formed - 9 no dihydrodigoxin was formed () 9 percent dihydrodigoxin formed 79 In the other ten subjects studied the sum of digoxin and metabo lites recovered accounted for less than 90% of the dose. In some sub jects only one third of the dose was recovered in urine and feces. Some of the samples from these subjects showed the presence of the sugar-hydrolyzed and reduced metabolites, but the amounts were far too small to account for up to two thirds of the missing dose (Table 7). It is likely that there may be additional routes of metabolism, other than hydrolysis and reduction, that predominate in these subjects. This is evident from the peaks that are repeatedly seen in HPLC chro matograms of urine and fecal samples. The retention times of these peaks do not correspond to any of the known metabolites (Tables 6 and 7). The possible metabolites that may be formed in these subjects could include water soluble conjugates and other polar metabolites that have not yet been characterized. Previous studies have indicated that the sugar hydrolyzed metabolites of digoxin may be further metabolized and excreted as sulfates or glucuronides (30,34). These metabolites may not be identified by the analytical method used in this study for two reasons. First, the solvent, methylene chloride used for the extraction of the less polar hydrolytic products and reduced metabolites, may not extract these conjugates. Second, even if they were extracted they may not get derivatized to form suitable fluorescent esters for the detector to identify them, and the deriva tives if formed may not be eluted by the less polar solvent used to separate the other metabolites. It has been shown by Heinz and Flasch that some of the dihydro metabolites formed are excreted as highly polar dihydrodigoxinic acid in cats 37 (87). The basic environment or a lactonase in the 61 bac terial flora of the subjects, who are known to be hypochlorhydric and therefore may have bacterial overgrowth, may open the lactone ring of these compounds resulting in highly polar and water soluble metabo lites (34). Hydroxylation of the parent drug and metabolites by bac teria or liver may also lead to polar metabolites. In conclusion, the identification of 20R epimer as the only dihy drodigoxin epimer present in these subjects confirmed our earlier in vitro findings that the bacteria (Eubacterium lentum) which catalyze this reduction form 20R-dihydrodigoxin from digoxin (73) and also that the 20R epimer is the only epimer found in humans (71). 81 2.3 REDUCTIVE INACTIVATION OF DIGITOXIN BY CULTURES OF EUBACTERIUM LENTUM 2.3.1 Introduction Next to digoxin, digitoxin is the most widely used inotropic agent for the treatment of certain heart diseases. Digitoxin is used as an alternative to Digoxin in the U.S. for this purpose. It is now well recognized that considerable portions of this drug are metabo lized in the body by some patients (65,66). Even though this drug has been in use clinically for almost two centuries, the chemical struc tures and quantities of many of the metabolites formed from this drug have not been fully characterized due to their complex nature. From the knowledge gained by the study of digoxin metabolism it can be safely assumed that digitoxin could also be eliminated as its dihyrdo metabolite. Bodem and Unruh have recently identified dihydro- digitoxin as a metabolite of digitoxin in the plasma of some patients (69). They used a gas chromatographic-mass spectroscopic technique to detect dihydrodigitoxin. Even though it can be safely assumed that dihydrodigitoxin is also formed in the GI tract by the intestinal bacteria it has not been determined whether they are the actual cause for the formation of this reduced metabolite. Hence it was also not known whether Eubacterium lentum can reduce digitoxin to form dihydrodigitoxin. The stereochem istry of the reduction product, if it can be formed by E. lentum, is also not known. If E. lentum can reduce digitoxin then it is possible that this organism can reduce the sugar-hydrolyzed metabolites of 82 digitoxin (DT2, DTI and DTO) to their respective dihydro metabolites, as previously shown for digoxin. This also has not been determined. There are no reports yet about the formation and the stereochemistry 0 of these reduced metabolites of digitoxin by E. lentum. In order to study the formation of dihydro metabolites, digitoxin and its sugar- hydrolyzed metabolites were incubated with cultures of Eubacterlum lent vim ATCC 25559 and the cultures were analyzed by an HPLC assay. 2.3.2 Chromatographic (HPLC) identification of dihydrodlgitoxin. Digitoxin (DT3), when incubated at a concentration of lOpg/ml for 7 days at 37°C with growing Eubacterium lentum cultures, was con verted to a major product, often in high yield (100%). This was evi dent from the HPLC retention time of the product (8.6 min), which was different from that of the starting material (7.9 min), but coincided with the major peak observed for reference dihydrodigitoxin. Chroma togram II in Figure 17 clearly shows that all of the starting material (DT3) has been converted to a single product. Comparison of the HPLC retention times of the product and the reference dihydrodigitoxin shown in Table 10 indicate that the product is dihydrodigitoxin. Incubation of digitoxin in TG medium with a fecal sample from a volun teer, who was known to excrete dihydrodigoxin as a metabolite of digoxin, also led to the formation of dihydrodigitoxin. 2.3.3 Hydrolysis of the sugar side chain. The naphthoyl esters of sugar-hydrolyzed metabolites of digi toxin, digitoxigenin bis and mono- digitoxosides and digitoxigenin (DT2, DTI and DTO) had different retention times from digitoxin (Table 10). Examination of chromatogram II in Figure 15, as well as other incubations of DT3 with E. lentum shows no evidence of formation these metabolites and therefore it is clear that E. lentum does not hydro lyze the digitoxose sugars from C3 of the steroidal moiety. iue 7 Crmtgas (HPLC)digitoxin reference derlvatized of Chromatograms 17.Figure Detector Response me n minutes in e im T E. lentum culture incubated with digitoxin. incubatedwith ' culture lentum E. (100 ng) (I) and an extracted derivatized sample ofsample (I) derivatized(100 extracted an and ng) 0 i i minutes in e Tim 84 Table 10. Chromatographic (HPLC) Detection of Digitoxin and Ita Reduced and Sugar-hydrolysed Metabolites.* Retention Time (min) of Compound Abbreviation 1-naphthoate Derivatives Digitoxin DT3 7.9 20R-Dlhydrodlgitoxin 20R-DHDT3 8.6 Digitoxigenin bisdlgltoxoaide DT2 7.1 20R-Dlhydrodlgitoxlgenln bisdlgitoxosidc 20R-DHDT2 7.6 Digitoxigenin monodlgltoxoslde DTI 6.5 20R-Dlhydrodlgltoxlgenin monodlgltoxoslde 20R-DHDT1 6.8 Digitoxigenin DTO 5.7 20R-Dlhydrodigltoxlgenln 20R-DHDT0 5.9 *Normal-phase column was llchrosorb Si 60; Mobile phase conslated of hexane-methylene chlorlde- acetonltile (5:1:1). 86 2.3.4 Formation of reduced metabolites from DT2, DTI and DTP. In order to further define the specificity of this reduction reaction of E. lentum, each of the sugar hydrolyzed metabolites of digitoxin (DT2, DTI and DTO) were incubated individually with cultures of the bacterium. HPLC analysis of the above cultures showed that the reduction of each of these compounds proceeded in high yield (often in 100%) (Figure 18). Table 10 shows the characteristic HPLC retention time of the 1-naphthoate derivatives of these reduced metabolites. Results from the above studies and the studies with the sugar- hydrolyzed metabolites of digoxin (DG2, DG1 and DGO) indicate that the enzyme which catalyzes the reduction of the unsaturated cardenolides to their dihydro derivatives (cardenolides) has specificity primarily for the unsaturated lactone ring and perhaps for the steroid nucleus, but not for the specific hydroxylation pattern or for the sugar side chain. n OH Eubacterium^ lentum H UNSATURATED LACTONE REDUCED LACTONE CO c JL D T3 ii (d ig it o x in) -----►-DHDT3 n=3 38 33 DT2 n= 2 ------►DHDT2 n = 2 39 34 DT1 n= 1 ------►DHDT1 n=1 40 o C II 4 DTO 1 -----►DHDTO nrO 41 Figure 18. Fornatlon of Dlhydro Metabolites from Digitoxin and Its Sugar-hydrolyzed Metabolites. 88 2.3.5 Control experiments. Control experiments were performed to ensure that the peaks seen In HPLC chromatograms were not Interference from the media or the bacteria. These included extracting uninoculated blank media, blank E. lentum cultures without added digitoxin and digitoxin incubated in blank media, derivatizing the extracts with 1-naphthoyl chloride and analyzing the derivatives by HPLC. The controls did not show any con version, and there were no interfering peaks from the bacteria or cul ture media which had retention time near the dihydro metabolites. 2.3.6 Large-scale (mg) production. In order to produce large quantities (mg) of this metabolite for spectroscopic identification, the culture and incubation condi tions of E. lentum with digitoxin were examined. It was decided to optimize the culture conditions for obtaining quantitative (total) conversions of the unsaturated starting materials to the metabolites in mg amounts to avoid time consuming separation processes that are necessary in the case of partial transformations. 2.3.6.1 Mixed culture experiments. As in the digoxin study, growing the facultative anaerobe Escherichia coli along with the obligate anaerobe E. lentum provided the necessary anaerobic atmosphere for the growth of the latter organ ism in large flasks (58,59). The mixed cultures produced total con versions every time they were tested. E. coli by itself, when incu bated with digitoxin, was not able to perform the reduction. This 89 once again confirmed that E. coli's role may to remove the dissolved oxygen and oxidized materials from the culture media. 2.3.6.2 Concentration study. The above experiments were conducted by adding digitoxin to the culture media at an overall concentration of 10yg of drug per ml of culture media. Later, it was decided to determine whether the culture could convert 100% of an increased amount of digitoxin. The overall concentration of digitoxin added to cultures were varied from lOyg to 320pg/ml of media. HPLC analysis of the derivatized extracts of these cultures showed quantitative conversion occurs up to a concentration of 80yg/ml. Above this amount only partial conversions were seen. 2.3.6.3 Time-course study In order to determine the minimum incubation time period required for total conversion of digitoxin to its metabolite by E. lentum cultures, a time-course study was conducted. In this study, samples were analyzed every hour for 30 hours from the time of inocu lation and incubation and addition of digitoxin. The results indicat ed that the metabolite starts to accumulate after 16 hours of incuba tion and growth, with conversion being essentially complete in 24 hours. Also it was found that when digitoxin was added to fully grown cultures (24-hour old) of E. lentum and incubated for 1 day, less than 10% of the starting material was converted to dihydrodigitoxin. 2.3.7 Isolation of the metabolites From the above experiments, it is clear that the optimum condi tions for preparation of these metabolites involved the use of mixed cultures in BHIC medium. The optimum length of incubation period for total conversion appeared to be 30 hours and the optimum overall con centration of the starting material added to the cultures appeared to be 80pg/ml. Using these conditions, preparative-scale fermentations were conducted to isolate sufficient quantities of the reduced metabo lites for structure determination. The metabolites were isolated in pure crystalline form by methylene chloride extraction and silica gel chromatography, followed by preparative TLC in some cases. The struc tures of the compounds were elucidated by comparison of their spectral properties with those of the unsaturated parent compounds and their dihydro derivatives prepared by catalytic hydrogenation. Additional transformations were conducted on some of the metabolites to confirm the stereochemical assignments. The spectroscopic techniques used in structure elucidation included mass spectrometry (MS), circular dichronism (CD), and proton nmr (*H-NMR) spectroscopy. 2.3.8 Mass spectral evidence Many authors have described the mass spectrometric characteris tics of digitoxin and its sugar-hydrolyzed metabolites (76,78,79,88). A.P. Bruins was able to obtain a clear, large pseudo molecular ion peak at m/z 763 for digitoxin using negative ion (OH") desorption chemical ionization technique (76). Apart from indicating additional peaks for successive loss of sugars it does not show any other key 91 fragments of the steroidal nucleus. Other chemical ionization tech niques that were used to obtain mass spectra of digitoxin by the same author also does not show very many additional fragments for the ster oid nucleus (88). There has been only one report on the mass spectro- metric identification of dihydrodigitoxin (69). The author of this report, as mentioned earlier, used a GC/MS technique to identify dihy drodigitoxin in the plasma of patients taking digitoxin. They used the characteristic fragment at m/z 340 to confirm the presence of dih ydrodigitoxin. This fragment could arise by loss of the side chain and two hydroxy groups as water from dihydrodigitoxin or by loss of two hydroxy groups from dihydrodigitoxigenin, the aglycone of dihydro digitoxin. They did not report whether they could obtain a clear molecular or pseudo molecular ion for the dihydrodigitoxin derivative. In the present study, Fast Atom Bombardment (FAB) and Laser Desorption/Fourier Transform/Ion Cyclotron Resonance (LD/FT/ICR) tech niques were used to identify the dihydro metabolites obtained from E. lentum cultures. The FAB mass spectrum of digitoxin showed a clear pseudomolecular ion (M+Na)+ at m/z 787, as seen in Figure 19. It also gave along with other fragment ions at m/z 373,357 and 323 the characteristic peak at m/z 339 (338 in EI-MS) that has been used by Bodem and Unruh to iden tify this drug in patients1 plasma samples (69). The LD/FT/ICR mass spectrum showed pseudomolecular ions for (M+K)+ and (M+Na+) at m/z 803 and 787 respectively along with a peak at m/z 543 corresponding to loss of two sugars (Figure 19). 92 m T y f H t i 'n 'i'r m 'i 11 i i r f i m 11 n r i i 11 i i fvt i 111 r r r r r r i WO 5?5 5W (25 650 6/5 700 725 750 775 339 357 373 II 803 o_ (M+K) 300) 400 500 10 800 Figure 19. FAB (I) and LD/FT/ICR (II) mass spectra of digitoxin. The FAB mass spectrum of the metabolite, isolated from E. len tum cultures incubated with digitoxin, showed a peak at m/z 789 (ca. 70% relative intensity) corresponding to a pseudomolecular ion of (M+Na+) for dihydrodigitoxin (molecular weight 766). The characteris tic fragment at m/z 341 (340 in EI-MS), which was used in a previous study to confirm the presence of dihydrodigitoxin in patient plasma samples (69) was present as the base peak. The LD/FT/ICR mass spec trum again confirmed the metabolite as the dihydro derivative of digi toxin by giving a tall pseudomolecular ion (M+K)+ at m/z 805 and also (M+Na)+ peak at m/z 789. The loss of one or two sugars from the side chain yielded the corresponding fragments at m/z 675 and 545 (Figure 20). Thus the mass spectral evidence clearly confirms the isolated metabolite as the dihydroderrivative of digitoxin. The metabolites isolated from the incubations of the sugar- hydrolyzed metabolites of digitoxin (DTO, DTI and DT2) were also proved to be dihydro derivatives by the evidence obtained from the FAB and LD/FT/ICR mass spectra. The mass spectra obtained for the cata lytic hydrogenation products of DT2, DTI and DTO were essentially identical to those of the metabolites isolated. The pseudomolecular ions and the key fragments obtained for these metabolites are given in Table 11. Dihydrodigitoxigenin obtained from incubations of E^ lentum with digitoxigenin was subjected to El ionization. The El mass spec trum obtained for the catalytic hydrogenetion product of DTO was essentially identical to this metabolite. Figure 21 illustrates the 94 ™ ^ ^ ■ ^ r n tTT^, ...... ^ ...... :OLi, 341 y .. ~e. *o. ».o. 373 40.0. 357 30. 2a la * j i r 'r > ^8>lr',IM iijo'1'1 ,,T (mb' * '11' * ^ | II O-, o 805 (M+K) o_ to (M+K-HgO) 3 ^ rl.>,l,,,l|,,l |...... I y m , 4 i | t ,1. I , f* A > 200 300 400 500 600 ™ 700-L - ' 800 Figure 20. FAB (I) and LD/FT/ICR (II) mass spectra of dihydrodigitoxin. Table 11. Mass Speccronetrlc Identification of Digitoxin and Its Reduced and Sugar-hydrolyzed Metabolites Compound (MW) FAB Mass Sprectrum LD/FT/ICR Mass Spectrum Pseudomolecular Key Paeucomolecular Key ion (R.I. Z) Fragments ion (R.I. Z) Fragments DT3 (764) 787 (M+Na)+ 339 803 (M+K)+ 543 (2) (100) DHDT3 (766) 789 (M+Na)+ 341 805 (M+K)+ 675,545 (0.5) (100) DT2 635 (M+H)+ 375,357,339 673 (M+K)+ 543,413 (1.6) (100) DHDT2 (636 659 (M+Na)+ 359,341 675 (H+K)+ 545,376 (1.8) (100) DTI (504 505 (M+H)+ 375 543 (H+K)+ - (1.3) (100) DHDT1 (506) 529 (M+Na)+ 377,341 545 (M+K)+ - (1.9) (100) DTO (374) 375 (M+H)+ 357,339 413 (M+K)+ 357,339 (2.1) (100) DHDTO (376) - 359,341 415 (H+K)+ 340 (100) fragmentation pattern postulated for this metabolite. This scheme is postulated on the basis of peaks appearing at m/z 358 (M_-H20), 340 (M+-2H20) and other fragments, and the patterns described for other similar cardiac genins (78-80). 97 CM CH 376 OH HO O CH CH CH CH 203 358 OH CH OH 302 O ,o CH 162 CH CH CH 340 CH 284 162 Figure 21. MS Fragmentation Pattern of Dihydrodigitoxigenin 98 2.3.9 Preparation of Dihydrodlgltoxlgenln from Dihydrodigitoxin Since the sugar side chain of dihydrodigitoxin interfere with the interpretation of various spectra, we decided to hydrolyze the digitoxose sugars from the C3 position and obtain the pure aglycone for assignment of stereochemistry. An acid hydrolysis procedure fol lowed by silica gel chromatography were used to obtain dihydrodigitox- igenin from dihydrodigitoxin isolated from lentum cultures. 2.3.10 Determination of the stereochemistry of dihydrodigitoxigenin. Optically active 5 membered 7-lactone rings have been shown to absorb uv light at approximately 214 to 219 nm due to the n-II tran sition (70). Results from previous studies have shown that associated with the n-II transition is a circular dichroism curve, whose sign and magnitude are dependent on the interactions within the lactone ring structure (70,89,90). The five atoms of the lactone group, C-C0-0-C seem to prefer co-planarity. Because of this preference, the stable conformations of the gamma lactone ring are restricted to a pair of enantiomers. In this pair, beta carbon, the fifth ring atom is oriented either above or below the lactone plane as shown in Figure 22. Y a 42 43 Figure 22. The two possible configurations of substituted y-lactones. 99 A.F. Beecham studied the CD and UV spectra of several 7-lactones derived from sugars and found the sign of CD curve of conformations 42 and 43 being associated with positive and negative cotton effects respectively (91). It is clear from this observation that the sign of the n-II cotton effect in 7-lactones is determined by the location of 3-carbon relative to the planar lactone ring system . The CD curve is positive when the conformation of gamma-lactone have the 3-carbon above the plane and negative when the 3-carbon is below the plane. In a recent study by Bockbrader and Reuning, it was found that these observations hold true for the two C-20 epimers of dihydrodigoxigenin and the positive CD curve is associated with R configuration at C20 (70). The beta carbon of the dihydrodigitoxigenin is attached to the seventeenth position. The preferred conformation for the 20R-epimer of dihydrodigitoxigenin would be for the steroidal ring system to be pseudo equatorial at the beta carbon. In other words, the bulkiness of the steroidal ring system forces the beta carbon up above the plane of the lactone ring. Thus a positive CD curve would be expected for 20R-dihydrodigitoxigenin. Figure 23 shows the CD spectrum obtained for dihydrodigitoxige nin isolated from E. lentum cultures. The positive CD curve observed for dihydrodigitoxigenin, based on previous evidence mentioned above strongly suggests that the configuration at C20 of this metabolite is R. 100 200 150 100 o -5 0 -1 0 0 200 225 250 275 300 325 350 Wavelength (am) CH CH OH Figure 23. CD Spectrum and associated structure of dihydrodigitoxigenin obtained by bacterial incubations. 101 In order to confirm the configuration at C20, the stereochemis try of dihydrodigitoxigenin was related to that of dihydrodigoxigenin. As stated earlier, key differences in the proton NMR spectra of the synthetically prepared 20R and 20S-dihydrodigoxigenin epimers were used to assign stereochemistry of the naturally occurring metabolite (70,71). These differences in the chemical shifts of the signals of C21 protons are caused by differences in interactions of these protons with the C12 underivatized and derivatized hydroxy group. The only structural difference between dihydrodigitoxigenin and dihydrodigox igenin is the absence of 12fl-hydroxy group, the important substituent for proton NMR spectrum based structure assignment in the former com pound. A chemical hydroxylation at C12 would be difficult, if not impossible. Therefore a microbial transformation technique was used to perform this reaction. Twelve organisms were screened for the con version of digitoxigenin to digoxigenin (Table 12). The twelve organ isms screened were selected on the basis of literature references and our own experience for their ability to hydroxylate organic molecules (91-97). Of the 12 organisms tested, the fungi PenciIlium citrinum and Fusarium roseum were found to perform this reaction in high yields. P. citrinum converted digitoxigenin to more than one product, i.e. along with digoxigenin, two other less polar products were also seen in TLC. Therefore F. roseum was selected to be used for the con version of DTO to DGO in large quantity, based on its ability to produce digoxigenin as the only product in about 90% yield (Figure 24). The product thus obtained and reference digoxigenin showed Table 12. Organisms screened and the results obtained for the conversion of digitoxigenin to digoxigenin. Organism screened ATCC Number Result obtained Actlnomucor elegans 6476 Aspergillus flavlpes 11013 Aspergillus nlger 9142 Aspergillus nigor 11394 Aspergillus ochraceus 18500 8otrytls a11i1 9435 Cunnlnghamella blakesleeana 8688A Cunnlnghamella elegans 9245 Cyllndrocarpon radlclcola 11011 Fusarlum roseum OSU Penlcllllum citrinum 16040 Syncephalastrum racemosum 18192 ♦ sign denotes conversion was seen and - sign denotes conversion was not seen. Fusarium roseum CH CH OH OH Figure 24. Microbial conversion of Dihydrodigitoxigenin to Dihydrodigoxigenin 104 similar physical and chemical characteristics, i.e., the TLC Rf val ues, the HPLC retention times, optical rotations and melting points were similar. In addition, mass spectral and NMR studies confirmed that the hydroxylation occurred at the 3-position of the carbon 12 of digitoxigenin. Incubation of dihydrodigitoxigenin, obtained by acid hydrolysis of dihydrodigitoxin, with shaken F\_ roseum cultures in PY medium at 25°C led to the formation of a major product. After extraction with chloroform and purification by sillca-gel chromatography, the product showed TLC Rf value and HPLC retention time (6.5 min) similar to those of 20R-dihydrodigoxigenin. The mass spectrum, circular dichroism spectrum and proton NMR spectrum were also identical to those of 2OR-dihydrodigoxigenin. The stereochemistry of each of the dihydro derivatives obtained by incubating DT2, DTI and DTO individually with mixed cultures of E. lentum and E. coli were also found to be 20R by preparative scale pro duction, acid hydrolysis of the digitoxose sugars to obtain the agly- cone, conversion to dihydrodigoxigenin and comparing the TLC and HPLC characteristics, and CD and NMR spectra of DHDGO thus obtained with those of the previously characterized 20 R-DHDGO epimer. 105 2.3.10.1 Conclusions Digitoxin, when incubated at a concentration of 10iig/ml with Eubacterlum lentum ATCC 25559 cultures, was converted quantitatively to its dihydro derivative. Mixed cultures of Eubacterium lentum ATCC 25559 and Escherichia coli ATCC 10536 were used to produce milligram quantities of this metabolite. The mass spectral analysis showed the purified metabolite to be dihydrodigitoxin, but gave no information concerning the stereochemistry at C20. Dihydrodgitoxigenin, the agly- cone of dihydrodigitoxin, was obtained by acid hydrolysis of the sugar side chain. The circular dichroism spectrum gave a positive curve for dihydrodigitoxigenin, indicating that the configuration at C20 is R. Incubation of isolated dihydrodigitoxigenin with cultures of Fusarlum roseum resulted in the 123-hydroxylation of the substrate to form dih- ydrodigoxigenin. The HPLC retention time and NMR spectrum of dihydro- digoxigenin thus obtained were identical to those of the previously characterized 20R-epimer. Incubation of digitoxigenin, and its mono- and bis- digitoxosides individually with E^ lentum also led to forma tion of their respective dihydro derivatives. The configuration at C20 of these reduced metabolites was also found to be R. CHAPTER III EXPERIMENTAL 3.1 REDUCTION OF DIGOXIN TO 2QR-DIHYDRODIGOXIN BY CULTURES OF EUBACTERIUM LENTUM 3.1.1 Chemicals and Media Reference digoxin and Its synthetically prepared sugar-hydrolyzed metabolites and dlhydro derivatives were obtained from Boehrlnger Mannheim Biochemicals and Burroughs Wellcome Co. The derivatizing reagent (1-napthoyl chloride) was purchased from Fluka (Hauppauge, N.Y.). The media for growing the bacterial and fecal samples were thioglycollate medium without dextrose or indicator (TG, Difco or BBL) and brain heart infusion broth (Difco or BBL) with added cysteine (BHIC) (58). Other culture media used were thioglycollate medium with dextrose and indicator (BBL), chopped meat carbohydrate broth (Carr Scarborough Microbiologicals Inc.), cooked meat medium (BBL) and tryp- ticase soy broth (BBL). 3.1.2 Microorganisms The Eubacterium lentum strain is the type culture obtained from the American Type Culture Collection (ATCC 25559). The cultures were maintained by serial transfer in Lee tubes (98) filled with TG medium - 106 - 107 with 1.5% agar added. Escherichia coli obtained from American Type Culture Collection (ATCC 10536) was maintained in trypticase soy (BBL) agar slants. 3.1.3 Instrumentation Fast Atom Bombardment and Electron Impact mass spectra were obtained with a Kratos MS-30 double-focusing E/B mass spectrometer at the Campus Chemical Instrumentation Center. Xenon was used as the collision gas in FAB mass spectra. Laser Desorption/Fourier Transform Ion Cyclotron Resonance mass spectra were obtained with a Nicolet FTMS-1000 spectrometer. Deuterated solvents were used for determining *H-NMR spectra with a Bruker HX-90 or Bruker HX-270 spectrometer. Circular dichroism spectra were obtained with a Jasco model J500A Spectropolarimeter. IR spectra were determined with a Beckman model 4230 Infrared Spectrometer. Optical rotations were measured with a Perkin Elmer 241 Photoelectric polarimeter. Melting points were determined in open end capillaries in a Thomas-Hoover unimelt appara tus and are uncorrected. 3.1.4 Pure E. lentum Culture Transformations. The steroids were dissolved in dimethyl formamide (DMF) and added to tubes containing 15 ml of freshly autoclaved medium (TG or BHIC) to give an overall concentration of 10yg/ml of digoxin. Equimolar con centrations of other compounds were used. The volume of DMF added was 0.16 ml or less per tube. The tubes were then inoculated with 1 ml of a 2-day old E. lentum culture. After mixing, the tubes were capped 108 tightly and Incubated at 37°C. After 7 days' Incubation the tubes were centrifuged, and a 1 ml aliquot of the supernatant was removed from each tube and extracted with methylene chloride. Following deri- vatization the extract was analyzed for the parent cardiac glycoside and the corresponding dihydro compounds (both epimers) by the chroma tographic method to be described (99). 3.1.5 Fecal Samples. Fecal specimens were obtained from a normal volunteer, who is known to be a dihydrodigoxin excretor, but who had not taken digoxin. A sterile spatula was used to add 3.0 g of fecal sample to a tube con taining 15 ml of sterile distilled water. After mixing for 15 seconds on a vortex mixer, 1 ml aliquots were transferred to tubes containing 15 ml of the culture medium (TG or BHIC) containing lOyg/ml of the compound of interest. The tubes were incubated for 7 days at 37°C. Following incubation the contents of the tubes were mixed, the tubes were centrifuged, and a 1 ml aliquot of the supernatant from each tube was analyzed for parent and dihydro compounds. 3.1.6 Extraction Procedures. To the 1 ml aliquot of the supernatant of each sample was added 1 ml of methylene chloride containing digitoxigenin (DTO) (100 ng) as an internal standard and 1 ml of plain methylene chloride. The tubes were hand shaken, centrifuged, and the aqueous layer was removed by aspiration. A 2 ml volume of an aqueous 5% sodium bicarbonate solu tion was added to the methylene chloride extract, the tube was shaken, 109 centrifuged, and the aqueous layer again aspirated. The organic phase was transferred to a clean test tube and evaporated to dryness under nitrogen prior to derivatization. 3.1.7 Derivatization To the evaporated sample was added 25 mg of 4-dimethyl aminopyri- dine (DMAP), lOyl of 1-naphthoyl chloride and lOOyl of acetonitrile. The sample was mixed well using a vortex mixer and was allowed to react for one hour at 50°C. 3.1.8 Preparation for HPLC Chromatography. The derivatized samples were evaporated to dryness under nitrogen and the excess derivatizing agent was hydrolyzed with 2 ml of an aque ous 5% sodium bicarbonate solution. After the tubes were shaken for 5 minutes, 2 ml of chloroform was added to solubilize the derivatives, and the tubes were rocked for 1 minute on a mechanical mixer. The aqueous layer was aspirated and the organic phase mixed with 2 ml of an aqueous 5% sodium bicarbonate solution. The aqueous layer was aspirated again, and the organic phase mixed with 3 ml of a 0.05 N hydrochloric acid solution containing 5% sodium chloride for 1 minute to remove any remaining DMAP. The aqueous layer was aspirated, and the acid wash was repeated three times. The chloroform layer was eva porated to dryness under nitrogen. 110 3.1.9 HPLC The high pressure liquid chromatograph was equipped with a model 110A pump (Beckman), a model FS970 (Kratos) fluorometer, a 200yl Injection loop, a dual channel recorder and a normal phase column (Llchrosorb S160, Jones Chromatography, 5ym particle size, 25 cm x 4.6 mm i.d.). The mobile phase consisted of hexane/methylene chloride/ acetonitrile (5:1:1). The samples were reconstituted with a small volume of mobile phase, and an aliquot was injected onto the column and chromatographed at a solvent flow rate of 2.2 ml/min. Figure 7 shows the separation of Internal standard (DTO), digoxin (DG3) and reference metabolites. 3.1.10 Standard Curves Samples containing 25 to 200 ng D63, DG2, DG1, DGO, and DHDG 3 in 2-propanol were pipetted into test tubes (15 ml), and the solvent was evaporated under nitrogen. To each of the samples was added 1 ml of blank broth or broth containing feces. Methylene chloride, 1 ml con taining 100 ng of internal standard (DTO) and 1 ml of plain solvent, was added to each tube and the compounds were extracted, derivatized with lOyl of 1-naphthoyl chloride, and chromatographed as described above. Standard curves were prepared by plotting the peak height rat ios against the amount of glycoside analyzed. The correlation coeffi cients obtained for each compound were: DGO R = 0.989; DG1 R = 0.970; DG2 R = 0.985; DG3 R = 0.998; 20 S-DHDG3 R = 0.996; and 20R-DHDG3 R = 0.997. Ill 3.1.11 Controls Apart from the above bacterial and fecal Incubations the following control experiments were performed. First, lentum was inoculated into each of the media used in.the above experiments and incubated for 2 days. The cultures were then autoclaved, digoxin (10yg/ml) was add ed, and the tubes were incubated for 7 days. Following incubation the tubes were analyzed for conversion of digoxin to dihydrodigoxin. Sec ond, digoxin (lOyg/ml in DMF) was added to a freshly inoculated E. lentum culture, extracted immediately (0 hour) and analyzed. Finally, uninoculated sterile media, blank E. lentum cultures with DMF (0.16 ml) alone were incubated at 37°C for 7 days, extracted and analyzed for any possible peaks at the retention times of the dihydro metabo lites (either epimer). The same types of controls, including auto- claving the media with feces, adding digoxin and analyzing for conver sion to dihydrodigoxin, were performed with fecal incubations. 3.1.12 Preparative-Scale Production of Dihydro Metabolites 3.1.12.1 Mixed cultures E. lentum grew poorly when it was incubated as pure (single organism) cultures in Erlenmeyer flasks containing BHIC medium. Also it pro duced only partial reduction when digoxin was added to these cultures. The dissolved oxygen perhaps inhibits the growth of this bacterium. Bokkenheuser et al. have used Escherichia coli, a facultative anae robe as an oxygen-scavenger in growing E. lentum in large flasks (59). When equal amounts (1 ml each) of E. lentum and E. coli cultures were 112 inoculated only partial conversion of digoxin was seen. The rapid growth of E. coli perhaps does not allow E. lentum to grow well. Therefore it was decided to add less amount of E. coli (one tenth of a ml) culture with E. lentum culture (1 ml) to the flasks. These cul tures produced 100% conversion of digoxin to dihydrodigoxin consis tently. 3.1.12.2 Variation of media Thioglycollate medium without dextrose or indicator (TG), brain heart infusion broth with added cysteine (BHIC) and trypticase soy broth (TS) were tested for their ability to grow the mixed cultures and produce total conversion of digoxin to dihydrodigoxin. BHIC and TG media consistently produced good growth of these organisms and also total conversion. Therefore BHIC was selected to be used in subse quent experiments. The difference between stationary cultures and shaken cultures in their ability to convert DG3 to DHDG3 was assessed by incubating two sets of identical cultures by the two different methods at 37° for 2 days and analyzing them by HPLC. Stationary cul tures produced total conversions whereas shaken cultures showed par tial or no conversions. 3.1.12.3 Concentration study To determine the highest concentration (overall amount per ml of medium) of DG3 that can be transformed 100% to DHDG3, the amount of digoxin added to the cultures was varied. The range of digoxin con centration added to cultures was from 10 yg to 320 yg/ml. Station 113 ary mixed cultures of E. lentum and E. coll In Erlenmeyer flasks In BHIC medium were used In this experiment. Following Incubation, cul tures were extracted and analyzed by the usual HPLC method to deter mine the extent of conversion. 3.1.12.4 Time-course study Time-course studies were performed to determine the minimum number of hours required for the starting of the accumulation of DHDG3 and total conversion of D63. Both pure E. lentum cultures in tubes and mixed cultures of E. lentum and E. coll in Erlenmeyer flasks were used for this purpose. The overall concentration of D63 added to tubes was 10 yg per ml of medium and the mixed cultures received digoxin at 80yg/ml concentration. Aliquots of duplicate culture samples were extracted, derivatized and analyzed by HPLC every hour for 30 hours from the time of inoculation and addition of digoxin. 3.1.12.5 General isolation procedures. Preparative-scale digoxin biotransformation was monitored using whole cultures or aliquots of cultures by TLC or HPLC respectively. The HPLC assay has been described before (Section 3.1.7). Silica gel 60 PF-254 (0.25 mm thickness) plates were purchased from E. Merck and they were developed in methylene chloride: Methanol (10:1). Chroma tograms were visualized by spraying with dinitrobenzoate spray reagent. Dinitrobenzoic acid produces a pink color with unsaturated lactone ring (digoxin) and no color with saturated lactone. Dinitro- benezoate spray reagent was prepared by mixing equal amounts of IN aqueous potassium hydroxide and 10% aqueous dinitrobenzoic acid. 114 Silica gel columns and preparative thin layer chromatography (PTLC) plates were used for the purification of dlhydro metabolites. Columns were wet-packed as a slurry of silica gel and solvent into suitable glass columns. The columns were plugged with cotton filters, filled half-way with solvent, and the slurry was poured into them. Columns were allowed to flow freely for 24 hours before use to ensure even packing. Extracts were dissolved in solvent, loaded onto the columns and eluted with the solvents used for packing. Fractions were collected in the specified volumes and analyzed by TLC. Chromatograms were visualized by spraying with anisaldehyde spray reagent and heating at 110°C for five minutes. Anisaldehyde spray reagent was prepared by mixing 90 ml of 95% ethanol, 5 ml of p-anisaldehyde and 5 ml of conc. sulfuric acid. Anisaldehyde spray produces a blue color with digoxin and dihydrodigoxin. PTLC plates were prepared with silica gel 60 PF-254 (E.Merck). Forty grams of silica gel were mixed with 120 ml of water until a homogeneous slurry was formed. The slurry was spread over four 20 x 20 cm glass plates at 0.75 mm thickness on a Shandon Southern Unoplan. The plates were dried at room temperature for 2 hours then placed in an oven (130°C) for 45 minutes. Maximum amount of material applied to plates was 35 mg. Material was dissolved in chloroform: Methanol (1:1) and spread as a line 2 cm from the bottom edge of plates. Plates were dried before and after development in suitable solvents. Bands were visualized by spraying one edge with anisaldehyde and heat ing that edge. 115 3.1.12.6 Isolation and purification of digoxin reduction product. Digoxin was dissolved in absolute ethanol (500 mg in 62.5 ml) and 2.5 ml of that solution was added to each of the twenty-five Erlenmey er flasks (500 ml size) containing 250 ml of BHIC medium to give an overall concentration of 80pg of DG3 per 1 ml of medium. Each flask was then seeded with 1 ml of 24-hour old E. lentum broth culture and 0.1 ml of E. coll broth culture. The cultures were incubated at 37°C for 2 days After incubation each culture was extracted twice with double the volume of methylene chloride. Formation of emulsions dur ing extractions could not be completely eliminated. Emulsions, when ever formed, were broken up by dividing them to small volumes and extracting with an excess of solvent. Evaporation of the solvent yielded a brown colored solid (921 mg). The extract was applied to a silica gel (PF 254, E. Merck) column (100 g, 50 cm x 2.5 cm) and eluted with C ^ C ^ : MeOH (10:1). A total of 235 fractions of 10 ml were collected. Dihydrodigoxin eluted in fractions 173-192 (411 mg). TLC analysis showed it to be still impure. The impure dihydro metabolite was applied to a second silica gel PF 254 column (75g, 45 x 2 cm) and eluted with C ^ C ^ : MeOH (20:1). A total of 250 12 ml fractions were collected. Fractions 212-223 from this column contained dihydrodigoxin (392 mg). Pure crystalline dihy drodigoxin (315 mg) was obtained after recrystallizing the metabolite from methanol. 116 3.1.12.7 Isolation and purification of the reduced metabolites of DG2, DG1 and DGO. Dihydrodigoxigenin and its mono-and bis-digitoxosides (DHDGO, DHDG1 and DHDG 2) were prepared by incubating the corresponding unsaturated compounds with E. lentum and E. coll mixed cultures. The culture and extraction conditions used were similar to the procedures described for dihydrodigoxin, except only 10 mg per flask (a total of 100 mg) of the parent compounds were added to the medium. The crude metabolites were purified by silica gel PF 254 columns and preparative TLC. Methylene chloride:Methanol (15:1) was the solvent system used for the separations. 3.1.12.8 Preparation of dihydrodigoxigenin from dihydrodigoxin Dihydrodigoxin (350 mg) prepared by bacterial reduction was dis solved in 130 ml of 60% methanol and divided equally into two Erlen meyer flasks (250 ml). The pH was adjusted to 1, using 10% HC1. The flasks were stoppered and shaken for 24 hours on a Gyrotory Shaker (NB S) at 250 rpm and 25°C. After adjusting the pH of reaction mixtures to 7 using 10% NaOH, the solution was extracted twice with double the volume of chloroform. Evaporation of the solvent yielded a slightly yellow colored solid (179 mg). Pure dihydrodigoxigenin (143 mg) was obtained by using silica gel column chromatography and a solvent sys tem of methylene chloride and methanol in a 10 to 1 ratio. Dihydrodigoxigenin mono- and bis- digitoxosides were hydrolyzed to give DHDGO by using a similar acid hydrolysis procedure. 117 3.1.13 Spectral Properties of Dlgoxigenln. Digoxigenin was obtained from Boehringer Mannheim Biochemicals and exhibited the following physical properties: mp 220-222°C; ir (KBr) V max 3450, 3410 (OH), 2930, 2825 (CH), 1780, 1745 (C=0 in enolide ring), 1630 (C=C in lactone rings) cm-1; 1H-NMR (CD3C0CD3, 90 MHz) 6 0.83 (3H, S, C18 Me), 0.95 (3H, S, C19 Me), 3.44 (1H, M, C17), 3.77 (lH,d,J=12, C12), 4.02 (1H, S (br), W 1/2 = 8, C3), 4.84 (2H, m, C21), 5.81 (1H, S (br), C22); UV, max (EtOH) (log E) 225 (4.31) nm; MS (probe) 70 ev m/e (rel. int.) 372 (M-H20) (15), 354 (M-2H20) (40), 336 (M-3H20) (5.0), 219 (M-C8H 1104) (90), 201 (M-CgH1305) (100), 147 (M-C12H1905) (35); [o]25D + 22.5 (C 1.1, MeOH). 3.1.14 Spectral Properties of Dlhydrodigoxigeniii Obtained by Acid Hydrolysis Dihydrodigoxigenin prepared by acid hydrolysis of dihydrodigoxin obtained from bacterial incubations exhibited the following physical properties: Mp 235-237°C; ir (KBr) V max 3490, 3410 (OH), 2910, 2830 (CH), 1775, 1740 (C=0, in lactone ring) cm-1; XH NMR (CD3C0CD3,90 MHz) 6 0.91 (3H, S, C-18 Me), 0.93 (3H, S, C-19 Me), 4.00 (1H, S, C-3), 4.00 (1H, t, J=8, C-21), 4.41 (1H, t, J=8, C-21). MS (Probe) 70 eV m/e (rel. int.) 374 (M-H20) (5.6), 356 (M-2H20) (71), 338 (M-3H20) (26), 219 (M-C8H1304) (5.7), 201 (M-C8H1505) (9.9), 147 (M - C12H2105) (48); [a]D24 = +20.1° (C 0.443), CD [0]218 + 135. 118 3.2 IDENTIFICATION AND MEASUREMENT OF DIGOXIN AND ITS METABOLITES IN HUMANS 3.2.1 General Experimental Procedures. Digoxin tablets were administered daily to 11 elderly subjects (age range 65-85) until steady state was attained. The maintenance dose of digoxin was either 0.125 mg or 0.25 mg/day. Complete urine and fecal samples were collected over three more days of dosing. Samples were frozen immediately after each voiding and stored at -20°C until assays were conducted. The specific HPLC assay used to identify and quanti tate digoxin and metabolites is described elsewhere (Section 3.1.6). Briefly, it consisted of extracting 1 ml of urine or 200 mg of feces with methylene chloride, derivatizing the extract with 1-naphthoyl chloride and separating the derivatives using HPLC. A normal phase column and a solvent system consisting of hexane, methylene chloride and acetonitrile (5:1:1) were used, and the esters of drug and metabo lites were detected and measured using a fluorometer. The assay for digoxin and metabolites is sensitive to 5 ng/ml in urine or 10 ng/200 mg in feces. 3.2.2 Tests for Bacterial Overgrowth A recently developed test referred to as the bile-acid breath test along with culturing and counting the number of colony forming units of microbes present in an aspirated sample of the contents of the proximal jejunum were used to diagnose bacterial overgrowth (presence of high number of bacteria in the upper GI tract) in the subjects. 119 The bile-acid breath test involved ingestion of cholyl-glycine- 1-^C (lOyCi) and repetitive measurement of the specific activity of expired ^CC>2 using hyamine traps over about a 4 hour time period. The specific activity of expired is known to be much higher in bacterial overgrowth patients than in subjects with normal bacterial flora (100). Secretions of upper intestine (duodenum) were obtained by allowing the subjects to swallow a dreyling tube and using mild suction. The upper intestinal sample was cultured both anaerobically and aerobically to determine the number of bacteria present in the intestine. Presence of greater than 10^ colony forming units per ml was taken as a sign of bacterial overgrowth. A portion of the aspi rate and a small (200 mg) fecal sample obtained from each subject before administration of digoxin were cultured separately with digoxin to determine their ability to covert digoxin to dihydrodigoxin. The methodology for culturing feces has already been described (Section 3.1.5), and the conditions used for culturing the upper intestinal aspirates, were similar to those used for culturing feces. 3.2.3 Characteristics of Subject No. 11. The subject 11 is a 74 year old, 99.3 kg, black male. He had a measured creatinine clearance of 110 ml/min. His predose serum digox in concentration on the first day of the study was 0.6 ng/ml. He took a 0.25 mg/day digoxin maintenance dose during the study period. The only other concurrent medication he took was daily multivitamins. He had a concurrent diagnosis of adult onset diabetes and hypertension. He was also diagnosed to be hypochlorhydric. 120 3.3 REDUCTIVE INACTIVATION OF DIGITOXIN BY CULTURES OF EUBACTERIUM LENTUM 3.3.1 Chemicals and Media Digitoxin and its sugar hydrolyzed metabolites (DT2, DTI and DTO), and digoxigenin were obtained from Boehringer Mannheim Biochemicals. The media for growing the bacterial cultures were thioglycollate medi um without dextrose or indicator (TG, Difco or BBL) and brain heart infusion broth (Difco or BBL) with added cysteine (BHIC) (58). The fungal cultures, were grown in a medium consisting of (per liter of H20): Pharmamedia, lOg; yeast extract, 5g; D-glucose, 20g; NaCl, 5g; K2HP04 , 5 g; (PY). 3.3.2 Microorganisms The Fusarium roseum culture was obtained from the Department of Plant Pathology and Weed Science, Mississippi State University, Mis sissippi State, Mississippi and is currently maintained by serial transfer on Czapek solution (Difco) agar slants in the OSU College of Pharmacy culture collection. 3.3.3 Instrumentation The HPLC system was equipped with a Beckman model 110A pump, a Rheodyne injection loop (200yl), a normal phase (Lichrosorb Si 60, Jones Chromatography) column and a Kratos model FS 970 fluorometer. Deuterated solvents were used for determining NMR spectra with a Bruk er HX-90 or Bruker HX-270 spectometer. Circular dichroism spetra were obtained with a Jasco model J500A spectropolarimeter. Electron Impact 121 (El) and Fast Atom Bombardment (FAB) mass spectra were obtained with a Kratos MS-30 double-focusing E/B mass spectrometer. Xenon was used as the collision gas In FAB mass spectra. Laser Desorption/Fourier Transform/Ion cyclotron resonance mass spectra were obtained with a Nlcolet FTMS-1000 spectrometer. 3.3.4 Eubacterlum lentum Transformations The cardenolldes were dissolved In ethanol and added to the freshly autoclaved TG or BHIC broth to give an overall concentration of 10 yg per milliliter. The medium was Inoculated with 1 ml of a 24-hour old E. lentum TG broth culture. The cultures were then Incubated at 37°C for 24 hours. The use of large volumes (15ml per tube) of freshly autoclaved medium followed by tight capping of the culture tubes pro vided a sufficiently anaerobic environment for E. lentum growth on a small scale. 3.3.5 Fecal Samples Fecal samples were obtained from a normal volunteer, who is known to excrete dihydrodigoxin, but who had not taken digoxin. The sample (3 g) was added to a tube containing 15 ml of sterile distilled water and mixed well. Aliquots (1ml) were transferred to tubes containing 15 ml of the culture medium (TG or BHIC) containing lOyg/ml of the compound of interest. The tubes were incubated for 7 days at 37°C. Following incubation the contents of the tubes were mixed, centrifuged and a 1ml aliquot of the supernatant from each tube was extracted and analyzed by a HPLC assay. 122 3.3.6 HPLC Analysis After incubation a 1ml sample of each culture was extracted with methylene chloride. A previously described HPLC method (Sections 3.1.6-3.1.10) was used to analyze the products in the extracts. In this method, the solvent is first evaporated and the extract was then reacted with 1-naphthoyl chloride to obtain fluorescent derivatives. The derivatives were redissolved in a small volume of mobile phase (hexane/methylene chloride/acetonitrile = 5:1:1) and injected onto the column. A fluorometer was used to detect the separated eluents. 3.3.7 Controls The following control experiments were performed: First, E. lentum was inoculated into each of the media used in the above experiments and incubated for 2 days. The cultures were then autoclaved, digitox in (lOyg/ml) was added, and the tubes were incubated for 7 days. Fol lowing incubation the tubes were analyzed for conversion of digitoxin to dihydrodigitoxin. Second, digitoxin (10 °g/ml in DMF) was added to a freshly inoculated E. lentum culture, extracted immediately (0 hour) and analyzed. Finally, uninoculated sterile media, blank E. lentum cultures with DMF (0.16 ml) along were incubated at 36°C for 7 days, extracted and analyzed for any possible peaks at the retention times of the dihydro metabolites (either epimer). The same types of con trols, including autoclaving the media with feces, adding digitoxin and analyzing for conversion to dihydrodigoxin, were performed with feca1 incubat ions. 123 3.3.8 Preparative-Scale Experiments 3.3.8.1 Hixed culture experiments Addition of a facultative anaerobe to E. lentum cultures when grown for digoxin transformation studies in large Erlanmeyer flasks produced enhanced growth of E. lentum and better conversions to dihydro metabo lites (Section 3.1.12.1). Escherichia coli ATCC 10526 was used once again as an oxygen scavenger in growing E. lentum cultures in large (500 ml size) Erlenmeyer flasks for digitoxin biotransformation exper iments. E. coli by itself, when incubated with digitoxin was not able to perform the reduction reaction. The mixed cultures produced total conversion of digitoxin to dihydrodigitoxin consistently. 3.3.8 .2 Concentration study Stationary E. lentum and E. coli mixed cultures in BHIC medium in 500 ml size Erlenmeyer flasks were used for this experiment. The overall amount of digitoxin added to cultures was varied from lOyg/ml to 320pg/ml. After incubating the cultures for 2 days, aliquots were extracted and analyzed for conversion to dihydrodigitoxin by the HPLC assay described before. 3.3.8 .3 Time-course study Both pure E. lentum cultures in tubes (TG and BHIC media) and mixed cultures of E. lentum and E. coli in 500 ml size Erlenmeyer flasks (BHIC medium) were used in this experiment. The overall amount of DT3 added to the culture tubes was 10 yg/ml of medium and the mixed cul tures received digitoxin at a concentration of 80 yg/ml of medium. 124 Aliquots of duplicate culture samples were extracted, derlvatized and analyzed using HPLC every hour for 30 hours from the time of Inocula tion (0 hour) and addition of digitoxin. 3.3.8.4 General Isolation procedures Preparatlve-scale digitoxin biotransformations were monitored using whole cultures or aliquots of cultures by TLC or HPLC respectively. The HPLC assay used in these experiments has been described before (Section 3.3.6). The TLC analyses were conducted using silica gel 60 PF-254 (0.25 mm thickness) plates purchased from E. Merck and they were developed in methylene chloride: methanol (10:1). Chromatograms were visualized by spraying with dinitrobenzoate spray reagent. Dini- trobenzoate produces a pink color with unsaturated lactone ring (digi toxin) and no color with saturated lactone. Silica gel columns and preparative thin layer chromatography (PTLC) were used for the purification of dihydro metabolites. Column frac tions were collected in specified volumes and analyzed by TLC. Chro matograms were visualized by spraying with anisaldehyde spray reagent and heating at 110°C for five minutes. 3.3.8.5 Isolation and purification of digitoxin reduction product Digitoxin(20 mg in 2.5 ml of absolute ethanol) was added to each of the 25 Erlenmeyer flasks (500 ml size) containing 250 ml of BHIC medi um. Each flask was then seeded with 1ml of 24-hour old E. lentum broth culture and 0.1 ml of E. coll broth culture (60). The cultures were incubated at 37°C for 48 hours. After incubation each culture 125 was extracted twice with double the volume of methylene chloride. Emulsions, whenever formed during extraction, were broken up by divid ing them to small volumes and extracting with an excess of solvent. Evaporation of the solvent yielded a brown colored solid (898 mg). The extract was applied to a silica gel (PF 254, E. Merck) column (lOOg, 50 cm x 2.5 cm) and eluted with methylene chloride: methanol (10:1). A total of 220 fractions of 10 ml size were collected. Dihy drodigitoxin eluted in fractions 152-171 (404 mg). TLC analysis of the pooled fractions showed it to be still impure. It was applied to a second silica gel PF 254 column (75g, 45 x 2 cm) and eluted with CH2CI2 . MeOH (20:1). A total of 210 fractions (15 ml size) were col lected. Fractions 174-201 from this column contained dihydrodigitoxin (383 mg). TLC analysis showed it to be pure. 3.3.8.6 Isolation and purification of the reduced metabolites of DT2, DTI and DTO. Dihydrodigitoxigenin and its mono- and bis- digitoxosides (DHDTO, DHDT1 and DHDT2) were prepared by incubating the corresponding unsatu rated compounds with E. lentum and E. coli mixed cultures. The cul ture, extraction, and purification procedures used were similar to the conditions described for dihydrodigitoxin, except only lOmg per flask (a total of lOOmg) of the parent compounds were added to the medium. The dihydro metabolites were purified using silica gel PF 254 columns and preparative TLC. Methylene chloride: Methanol (15:1) was the solvent system used for the chromatography. 126 3.3.9 Preparation of Dihydrodigitoxigenin from Dlhydroodlgitoxin Dihydrodigitoxin (350mg) prepared by bacterial reduction was dis solved in 130ml of 60% methanol and divided equally into two Erlenmey er flasks (250ml). The pH was ajusted to 1, using 10% HC1. The flasks were stoppered and shaken for 24 hours on a Gyrotory shaker (NBS) at 250 rpm and 25°C. After adjusting the pH of reaction mix tures to 7 using 10% NaOH, the solution was extracted twice with double the volume of chloroform. Evaporation of the solvent yielded a slightly yellow colored solid (183mg). Pure dihydrodigitoxigenin (154mg) was obtained by using silica gel column chromatography in a solvent system of methylene chloride and methanol in a 10 to 1 ratio. 3.3.10 Spectral Properties of Dlgitoxigenln Digitoxigenin was obtained from Boehringer Mannheim Biochemicals and exhibited the following physical properties; mp 215-217°C; ir (KBr) Vmax 3440, 3400(0H), 2930, 2830(CH), 1780, 1745 (C=0 in enolide ring), 1635 (C=C in lactone ring) cm'1;1H-NMR (CD3COCD3 , 90 MHz) 6 0.91 (3H, S, C18 Me), 0.95 (3H, S, C19 Me), 3.33 (1H, m, C17), 4.03 (1H, S (br), W 1/2=9, C3), 4.94 (2H, m, C21), 5.87 (1H, S(br), C22); UVX max (EtOH)(logepilson.) 222 (4.1) nm; MS (probe) 70 eV m/e (rel. int.) 356 (M-H20) (55), 338 (M-2H20) (38), 203 (M-CgH 1104) (22), 149 (M-C12H17Oa) (17); [a]D26 + 18.7 (C 1.3, MeOH). 3.3.11 Spectral Properties of Dihydrodigitoxigenin Obtained from E. lentum Incubations 127 Dihydrodigitoxigenin prepared by acid hydrolysis of dihydrodigitox in obtained from bacterial incubations exhibited the following physi cal properties: mp 219-221°C; ir (K Br) Vmax 3485, 3415 (OH), 2920, 2830 (CH), 1775, 1740 (C=0 in lactone ring) Cm-1; 1H-NMR (CD3COCD3 , 90 MHz) 6 0.95 (3H, S, C18 Me), 0.98 (3H, S, C19 Me), 4.01 (1H, S, C3), 4.01 (1H, t, J=7, C21), 4.23 (1H, t, J=7, C21). MS (probe) 70 eV m/e (rel. int.) 358 (M-H20) (5.4), 340 (M-2H20) (65), 203 (M-CgH^O^ (20), 149 (M-C12H 1904) (45); [o]D22 = 16.2° (C 0.52, MeOH) CD [0]215 + 125. 3.3.12 Screening Microbes for the Transformation of DTO to DGO Twelve organisms, shown in Table 8 , were screened for the ability to hydroxylate at 123 position of digitoxigenin. A two stage fermen tation procedure was used to generate cultures for these transforma tion experiments. PY was the medium used in these experiments. Stage one cultures were started by adding an aqueous spore suspension pre pared from 7-day old slant cultures into 100 ml of PY medium in a 500 ml Erlenmeyer flask. The stage I cultures were incubated at 25°C for 2 days on a rotary shaker at 250 rpm. Stage two cultures, used for biotransformation experiments, were initiated by adding 10 ml of 2-day old stage one cultures. Digitoxigenin (10 mg) in 0.4 ml of ethanol was added to each of the 1-day old stage two cultures. After 7 days incubation, the whole cultures or aliquots of the cultures were extracted with double the volume of chloroform and analyzed by TLC or HPLC respectively. Methylene chloride: methanol (20:1) was the sol vent system used in TLC analyses. The HPLC assay used has been described in sections 3.1.6-3.1.10. 128 3.3.13 Preparative-scale Production of DGO from DTO. Stage two cultures of F. roseum In PY medium at 25°C on a rotary shaker at 250 rpm were used for this experiment. Digitoxigenin (10 mg in 0.4 ml of Ethanol) was added to each of the 25 culture flasks and incubated for one week. Crystalline digoxigenin (207 mg) was obtained after extracting the whole cultures with chloroform and separating the product with silica gel column in a solvent system of methylene chlo ride: methanol (20:1). 3.3.14 Spectral Properties of DGO Prepared from DTO. The physical properties of DGO prepared by F. roseum incubations were essentially identical to the ones seen for DGO obtained from Boehring Mannheim Biochemicals. They have been described in section 3.1.13. 3.3.15 Acetylatlon of DGO Digoxigenin (15 mg) obtained from DTO by P\ roseum incubations was dissolved in 0.5 ml of pyridine and 0.5 ml of acetic anhydride was added. The mixture was evaporated to give 16.5 mg. *H-NMR (CDClg, 270 MHz) 6 0.91 (3H, S, C18 Me), 0.95 (3H, S, C19 Me), 3.29 (1H, M, C17), 4.8 (1H, dd, J=10, C12), 4.9 (2H, m, C21), 5.1 (1H, S (br), C3), 5.9 (1H, S-br, C22). Digoxigenin obtained from Boehringer Mannheim Biochemical when ace- tylated using a similar procedure gave essentially an identidal NMR spectrum. 129 3.3.16 Preparation of Dihydrodigoxigenin from Dihydrodigitoxigenin. A two stage fermentation procedure In PY medium at 25°C on a rotary shaker at 250 rpm similar to the one described for the biotransforma tion of DTO to DGO was used in this experiment. Dihydrodigitoxigenin (10 mg in 0.4 ml of ethanol) obtained from E. lentum incubations was added to 15 1-day old stage two F. roseum cultures. After incubating for one week, the cultures were extracted with double the volume of chloroform and the crude product was purified using silica gel PF-254 (lOOg) column and a solvent system of methylene chloride: methanol (20:1). Crystalline dihydrodigoxigenin (109mg) obtained from this transformation reaction was submitted for MS analysis. 3.3.17 Spectral Properties of DHDGO Prepared from DHDTO. The physical properties obtained for DHDGO prepared by F. roseum incubations of DHDTO were essentially identical to the ones seen for 20R-DHDGO. They have been described in section 3.1.14. Part II MICROBIAL AND CHEMICAL TRANSFORMATION OF DIGOXIGENIN: PREPARATION OF POTENTIAL HUMAN METABOLITES -130- CHAPTER IV INTRODUCTION 4.1 POLAR METABOLITES OF DIGOXIN It Is clear from the recent literature that digoxin undergoes extensive metabolism at least in some patients (25,27,54-57,86,101). 20R-Dihydrodigoxin may constitute a large portion of the metabolites formed in some of these patients who excrete low amounts of unchanged drugs (101). In the other patients who do not form high amounts of dihydrodigoxin or sugar hydrolyzed metabolites but still excrete low amounts of the parent drug, digoxin may be converted into one or more metabolites that have not been characterized yet (101). Recent studies by Gault et al. and Magnusson et al. give clues to the nature of digoxin metabolites whose structures have not been determined yet (34,43,102,103). Gault et al. in an extensive study on the influence of gastric pH on digoxin metabolism found that the major portions of the metabolites excreted by patients who formed high amounts of hydrolytic products (DG2,DG1 and DGO) were unextractable in chloroform and were more polar than the parent drug (34,39,43). The results from the studies by Magnusson et al. agree with those of Gault et al. (102,103). Therefore it appears that some of the unidentified metabolites of digoxin are more polar and water soluble than digoxin. - 131 - 132 4.2 CONJUGATES Sulfate and Glucuronide conjugates of sugar-hydrolyzed metabolites of digoxin make up some of the polar metabolites formed In dlgoxln patients (34,43). The exact structures of these conjugates have not been determined yet. Previous Investigators Incubated a combination of P-glucuronidase and arylsulfatase with polar metabolites and obtained hydrolysis products that corresponded to bis- and mono- digitoxosides of digoxigenin, and digoxigenin (34,43). Therefore indirect evidences obtained by enzyme hydrolysis procedures suggest that conjugates are part of metabolites formed in humans (34,43). Peterson et al. have developed recently procedures for synthesiz ing cardenolide glucuronides and sulfates (104). In these procedures, glucuronides are synthesized by preparing cardenolide glucosides first by the reaction of acetobromoglucose. The hydroxy methyl group of the glucose moiety is then oxidized in the presence of a platinum catalyst to the carboxyl group of the final glucuronic acid (Figure 25). 133 CH CH CH OH OH HOCH R o + PA ACOCH AC ACO OA AcO CH ACO CH Rj-H HOOC OH R^“OH o HO Figure 25. Synthesis of cardenolide glucuronides. 134 Sulfates of cardiac genins are prepared by direct reaction of agly- cones with chlorosulphonic acid in pyridine (104) (Figure 26). CH CH C1S0.H CH CH Pyridine OH OH HO 4 R - H DTO 44 R H Digltoxigenin Sulfate 5 R - OH DGO 45 R ODO^ Digoxigenin Disulfate Figure 26. Synthesis of cardenolide sulfates Conjugates prepared by using the above procedures can be used to determine the extraction methods, enzyme hydrolysis procedures and other physical properties of these types of compounds. The results obtained from the above studies should help in the identification of the metabolites that are actually formed in humans. 135 4.3 METABOLITES FORMED BY OPENING OF THE LACTONE RING Opening of the lactone ring of digoxin or its metabolites would lead to the formation of an additional hydroxy group and a free car- boxylic acid group. Presence of such functional groups would make the compound much more polar than digoxin and hence unextractable in chlo roform. Benthe, in a study on the biotransformation of digoxin by humans, found a fraction of the polar metabolites to be the metabolite formed by opening of the lactone ring of digoxigenin (105). He assigned the structure digoxigenin-acid 46 to this metabolite on the basis of mass spectral fragmentations of this compound (Figure 27). ,o OH HO. OH OH CH CH CH CH OH OH HO HO Digoxigenin-acid 5 DGO OH CH CH HO Figure 27. Formation of Digoxigenin-acid 136 As mentioned before, the basic environment or a lactonase in the 61 bacterial flora of the patients, who are known to be hypochlorhydric and therefore may have bacterial overgrowth, may open the lactone ring of digoxin and metabolites to form highly polar metabolites (34,86). Dihydro metabolites tend to be more susceptible to such hydrolysis of the lactone ring by basic pH than digoxin (87). This is evident from the finding of the formation of dihydrodigoxinic acid 37 as a metabo lite of digoxin in cats (87). 4.4 HYDROXYLATED METABOLITES Compounds formed by hydroxylation in the steroid nucleus of the glycosides and the genin have been suggested to be potential polar metabolites of digoxin (34,43). Results obtained from the studies on digitoxin and digitoxigenin biotransformation in animals and by liver homogenates support this hypothesis. Bulger and Stohs identified 5B-hydroxydigitoxigenin 48 as a polar metabolite resulting from the biotransformation of digitoxigenin 4 by rabbit liver homogenate (106). Incubation of digitoxigenin with rat adrenal homogenates led to the formation of 5&-hydroxydigitoxigenin 48 and 16B-hydroxydigitoxigenin (gitoxigenin) T_ (107). Other studies indicated that 5ft 48, 16& 6 , IB 49 and 12fJ 5 digitoxigenin were formed as metabolites of digitoxigenin when it was incubated with rat liver homogenates (108). 137 The major metabolite obtained from incubation of digitoxigenin 4 with rabbit liver homogenates was identified as 6&-hydroxy-3-epi digi toxigenin 50 by Bulger and his co-workers (109). 48R-H Rj«0H 56-Hydroxydlgltoxlgenln CH 49R-OH Rj«H lB-Hydroxydigitoxigenin R ^ OH HO CH CH OH HO' OH 50 68-Hydroxy-3-epi-digitoxigenin In a recent study, Carvalhas and others identified 17a-hydroxydigitoxin 5J. as a metabolite of digitoxin in the guinea pig (110). The identification of 17a-hydroxydigitoxin was confirmed by preparing this compound synthetically. CH CH, CH H OH 51 17a-Hydroxydigitoxin 138 A procedure was developed by Danieli and his co-workers for the synthesis of 17o-hydroxy derivatives 54 and 55 from digitoxigenin ace tate 52 and gitoxigenin diacetate 53 (111). Refluxing the compounds 52 and 53 in dry dioxane with equal amounts of selenium dioxide led to the formation of respective 17a-hydroxy derivatives (Figure 28). CH R CH SeO CH OH Dioxane OH AcO AcO 52 R - H 54 R « H 53 R - OH 55 R ■ OAc Figure 28. Synthesis of 17a-Hydroxy Cardenolides Saito et al. synthesized 17o-hydroxy-digitoxigenin 56 by reflux ing digitoxigenin 4 with selenium dioxide in dioxane (9). 17a-hydroxy-digitoxigenin 56 CH OH HO 139 4.5 MICROBIAL TRANSFORMATION - METHODS FOR IDENTIFYING HUMAN METABOLITES Complex biotransformations in humans leading to the formation of several metabolites make identification of all of the metabolites formed extremely difficult for some drugs. This is especially true in the case of cardenolides. The use of low amounts of maintenance dose, generally 0.25 mg or 0.125 mg of digoxin, due to toxicity problems, adds to the problem. Utilizing microbial models for the study of bio transformations would be beneficial in this case. Several authors have used microbial transformations as parallels to mammalian metabo lism in order to identify the metabolites formed in humans. The major advantage in using such approaches is in the production of larger amounts of metabolites for identification and pharmacological testing. For example, large amounts of drugs can be used in microbial transfor mation studies to produce necessary amounts of metabolites to be iso lated and this may not be possible in human studies due to toxicity problems. Once the microbial metabolites are identified, they can be used as reference materials for identification and quantitation of human metabolites using HPLC or GC. Smith and Rosazza have used such microbial biotransformation approaches to identify potential mammalian metabolites of several drugs (112). It is necessary to screen and use several microbes to produce and identify as many different metabolites as possible for the drug of interest. Finding suitable microbes for biotransformation studies can be done using two types of screening methods. One method 140 Involves screening as many microorganisms as possible selected random ly from a large collection of microorganisms. The second method involves screening a selected group of microorganisms based on litera ture references for their ability to perform desired reactions on sim ilar compounds. Smith and Rosazza have classified microorganisms on the basis of reactions carried out by them. They identify "microbial models of mammalian metabolism" as selected groups of microorganisms used together for parallel microbial and mammalian xenobiotic trans formation studies (112). Chandrasekaran and Robertson have used the second method to find microorganisms to produce digoxigenin and dihy- drodigoxigenin from digitoxigenin and dihydrodigitoxigenin respective ly (113). Iizuka and Naito reviewed microbial conversion of cardiac steroids (114,115). Digitalis has been used for medical purposes for more than two hundred years. Therefore it would be expected that the biotrans formations of cardenolides had been studied extensively. Review of the literature reveals that microbial transformation of digoxin and its sugar cleaved products (DG2, DG1 and DGO) has not been explored thoroughly and only biotransformation of digitoxin and its sugar cleaved products (DT2, DTI and DTO) has been explored to some extent. Table 13 illustrates the types of metabolites isolated by microbial conversion of cardiac steroids. It is clear from Table 13 that the most common type of reaction performed by microorganisms is hydroxyla- tion. Table 13. Microbial transformation of digitoxin and digitoxigenin Cardenolide Metabolite Reference Digitoxin 7-hydroxydigitoxin 115,116 digitoxigenin 116 16-dehydrodigitoxin 117 addition of C1oHo_0, 116 1Z ZU 0 Digitoxigenin 1-hydroxydigitoxigenin 114 1,7-dihydroxydigitoxigenin 118 6-hydroxydigitoxigenin 119 7-hydroxydigitoxigenin 114,120,121 5,7-dihydroxydigitoxigenin 118 11-hydroxydigitoxigenin 119 12-hydroxydigitoxigenin 93,114,122 16-hydroxydigitoxigenin 94 3-ketodigitoxigenin 114 7-hydroxy-3-ketodigitoxigenin 120 16-dehydrodigitoxigenin 119 142 4.6 STATEMENT OF THE PROBLEM Even though digitalis has been used for more than two hundred years now for the treatment of cardiac ailments, many metabolites of digoxin remain uncharacterized and untested for pharmacological activity. One major hurdle in the complete structure elucidation of digoxin metabo lites is lack of availability of sufficient quantities of these metabolites. Microbial transformation can be used as one way to over come this problem. Once sufficient quantities of metabolites are pro duced using microbial conversion methods, they may be used in HPLC or GC assays to identify and quantitate human metabolites. Also they may be used to understand fully the pharmacological activities or toxici- ties of metabolites. Another approach would be to synthesize suffi cient quantities of the compounds that are identified as animal metabolites using chemical methods and use them as references in spe cific assays to identify human metabolites. The goals of the present project included the following: (1) to select microorganisms for the production of digoxigenin metabolites by screening a limited number of organisms that are picked on the basis of literature; (2) to produce sufficient quantities of metabolites for structure identification by using the microorganisms selected; (3) to determine the structure of the metabolites produced; and (4) to synthesize potential digoxigenin metabolites using chemical methods. Streptomyces aureus ATCC 15437 and Syncephalastrum racemosum ATCC 18192 were selected for preparative scale production of digoxigenin metabolites based on initial screening results. CHAPTER V RESULTS AND DISCUSSION 5.1 BIOTRANSFORMATION OF DIGOXIGENIN BY STREPTOMYCES AUREUS 5.1.1 Introduction It has been suggested by previous Investigators that metabolites with additional hydroxy groups in the steroid nucleus of digoxin may make up some of the unidentified metabolites (34). Since digoxigenin represents the steroid nucleus of DG3, it was decided that microbial transformation studies of DGO would lead to identification of such metabolites. Initial screening experiments involving a limited number of microorganisms for the production of polar metabolites from DGO led to the selection of two suitable microbes. Streptomyces aureus ATCC 15437, selected as one of the two microbes for biotransformation stud ies, was found by TLC assay to convert DGO to one less polar compound and two more polar compounds than digoxigenin. Procedures were devel oped for the production of sufficient quantities of these metabolites for isolation and structure elucidation. Media, amount of digoxigenin added to cultures and extraction solvents were varied in the experi ments. Using the best conditions, a preparative-scale fermentation - 143 - was conducted and three metabolites were isolated and characterized as 3-ketodigoxigenin 27, 3-epidigoxigenln 28 and 60-hydroxydigoxigenin HO HO CH CH CH CH OH OH HO' 3ketodlgoxigenin 3epi-digoxigenin OH CH OH HO OH 57 6B-hydroxydigoxigenin 5.1.2 Screening Microorganisms for Selection of Suitable Organisms for Biotransformation Studies. TVelve microorganisms were screened for the transformation of digoxigenin to polar metabolites. These twelve organisms were select ed on the basis of literature references for their ability to produce hydroxylated metabolites of steroids and other related compounds. Ttoo of the twelve organisms screened, Streptomyces aureus ATCC 15437 and Syncephalastrum racemosum ATCC 18192, showed formation of polar metabolites in the TLC analysis. TLC analysis of Cunninghamella ele- gans ATCC 9245, Cunninghamella blakesleeana ATCC 86889 and Aspergil lus ochraceous ATCC 18500 cultures showed that only small amounts of starting material were left unchanged but no additional spots for 145 metabolites were seen. The 12 organisms screened and the results obtained are given in table 14. Streptomyces aureus was selected for initial study because of its ability to form more than one product. Once aureus was selected to be used in biotransformation study, media, amount of DGO added to the cultures and extraction procedures producing the most efficient transformation had to be found. 146 Table 14. Organisms screened and the results obtained for the formation of polar metabolites from Digoxigenin. Organism screened ATCC Number Result Obtained* Actinomucor elegans 6476 — Aspergillus flavipes 11013 - Aspergillus ochraceus 18500 - Botrytis allii 9435 - Cunninghamella blakesleeana 8688A - Cunninghamella elegans 9245 - Cylindrocarpon radicicola 11011 - Fusarium roseum OSU - Mucor rouxii 24905 - Penicillium citrinum 16040 - Streptomyces areus 15437 + Syncephalastrum racemosum 18192 + *(+) sign denotes polar metabolites were seen in TLC analysis (-) sign denotes no polar metabolites were seen in TLC analysis 147 5.1.3 Determination of Optimum Fermentation Conditions 5.1.3.1 Variation of media Initial screening experiments were performed in Pharmamedia yeast extract (PY). TLC analysis of aureus culture extract showed that even though three metabolites were formed in isolable amounts, some of the starting material remained unchanged. Therefore it was decided to vary the culture media to obtain better yields. DGO was added to stage two cultures of aureus ATCC 15437 that had been grown in Pharmamedia yeast extract (PY), Pharmamedia yeast extract with added mannitol (PYM), yeast malt extract medium (YM), yeast malt extract media with added mannitol (YMM) and nutrient medium (NL 406), which contained sucrose, mannitol, yeast extract, and other minor constitu ents. TLC analysis of extracts from each culture showed the produc tion of metabolites varied with the medium. The yields of metabolites were determined on the basis of the size and density of DGO spot in TLC analysis. Large, blue spots of DGO were seen from cultures grown in all media except PY. Therefore it was decided that PY is the best medium among the media tested for fermentation scale-up. 5.1.3.2 Amount of digoxigenin added to culture The initial screening experiments were conducted using 10 mg of digoxigenin per culture. Later, it was decided to determine whether the culture could metabolize an increased amount of digoxigenin. Stage II cultures grown in PY medium were chosen for the study. TLC analysis of extracts of cultures incubated with 10 mg, 20 mg and 30 mg 148 per 100 ml amounts of DGO showed best conversions take place when 20 mg of digoxigenin was added to each culture. 5.1.3.3 Extraction procedure The suitability of different solvents for extraction of unchanged digoxigenin and metabolites was determined. Stage II cultures grown in PY media and incubated with 20 mg digoxigenin were used. The cul tures were extracted with methylene chloride, chloroform and ethyl acetate. TLC analysis of the extracts revealed that CH2CI2 is capable of extracting chiefly the starting material and the less polar metabo lite, while CHCI3 can remove most of the unchanged digoxigenin, the less polar metabolite and some of the polar metabolites. Ethyl ace tate, however, can extract most of the starting material and metabo lites. 5.1.4 Preparative-scale Fermentation From the above experiments, it is clear that the optimal procedure for production of metabolites involved the use of PY medium. The optimal amount of DGO added to the cultures appeared to be 20 mg per culture and the appropriate solvent for extraction would be ethyl ace tate. Using the above experimental conditions, a preparative-scale fermentation was performed to isolate sufficient quantities of metabo lites 27, 28 and 57 for structure elucidation. Solvent extraction of cultures incubated with 400 mg of DGO followed by silica gel chroma tography and preparative TLC and recrystallization in some cases, 40 149 mg of crystalline 27, 22 mg of crystalline 28 and 25 mg of crystalline 51 were obtained. The structures of the compounds were determined by comparison of their physical and chemical properties with those of digoxigenin. Structures of the metabolites 27 and 28 were confirmed by comparing their properties with compounds produced by chemical syn thetic methods. The spectroscopic techniques employed for structure elucidation Included proton NMR(^H-NMR) spectroscopy, mass spectrome try (MS) and Infrared spectroscopy (IR). 5.1.5 Digoxigenin The physical properties of digoxigenin have been described else where (section 3.1.7). A fragmentation pattern for digoxigenin postu lated on the basis of peaks observed in mass spectrum and patterns described in literature is given in Figure 29. 150 Oh CM, CM. OH HO 390 O OH CH, OH CH, CM, CH. O h OH 372 316 ,o CM OH CH 219 CH CH 201 CH CH 336 Figure 29. MS Fragmentation of Digoxigenin 151 5.1.6 3-Ketodlgoxlgenin 27 The FAB/MS and EI/MS of 3-ketodigoxigenin 27 gave a molecular ion at m/z 388, indicating a loss of 2 mass units from digoxigenin, con sistent with oxidation of a hydroxy group. The *H-NMR of 27 (in CD3COCD3) revealed the disappearance of the signal for the 3a proton, seen at 4.02 ppm in digoxigenin, and appearance of a triplet at 2.8 ppm and three doublets at 2.55 ppm that would be expected for C4 and C2 protons respectively. It is clear from these data that the hydroxy group at C3 was oxidized to a 3-keto function. Other physical proper ties including melting point (250°-254°C) and optical rotation ([a]^jj + 3 8 (c = 0.8 methanol) were consistant with the ones reported in the literature (122). It was decided to synthesize this metabolite chemi cally from digoxigenin to confirm the structure assignment. Reaction of Jones reagent with digoxigenin (100 mg) in acetone at 0°C for 5 minutes followed by extraction, chromatography and recrystallization yielded two less polar compounds in unequal amounts (Figure 30). The major product (63 mg) 27, showed physical properties identical to those of the metabolite 27, thereby confirming the structure of 27. The minor product (22 mg), 58, showed a molecular ion of 386 in both FAB/MS and EI/MS, indicating a loss of 4 mass units from digoxigenin, and oxidation of 2 hydroxy groups. Other physical properties were consistant with the ones reported for 3,12-diketodigoxigenin 58, in the literature (122). 152 5.1.7 3-Epidigoxigenln 28 The FAB mass spectrum of 3-epidigoxigenln 28 gave a pseudomolecular ion at m/z 391, corresponding to a molecular weight of 390. The EI/MS of 28 was almost identical to that of DGO, showing peaks at m/z 372, 354 and 336, corresponding to the loss of 3 hydroxy groups in the form of water. Presence of other fragment ions identical to the ones observed for DGO led to the postulation of a fragmentation pattern similar to the one shown for digoxigenin in Figure 29. The proton NMR spectrum of 3-epidigoxigenin 28 showed one major difference from that of digoxigenin. The peak (broad singlet) observed for the 3a-proton of DGO at 4.05 ppm was absent in the spectrum of 3-epidigoxigenin, but a peak (multiplet) that would be consistant with a 3&-proton was seen at 3.55 ppm (W 1/2 = 20 Hz). It is clear from these data that this compound, 28, is a product of epimerization at the C3 hydroxy group. Other physical properties of this compound were identical to those reported in the literature (122). It was again decided to confirm the structure of the metabolite 28 by synthesizing 3-epidigoxigenin by a known chemical method from 3-ketodigoxigenln in 28. Reaction of sodi um borohydride with 3-ketodigoxigenin (50 mg) in methanol followed by work-up, chromatography and recrystallization, yielded two compounds in unequal amounts (Figure 31). The major product (31 mg), 28, showed physical properties identical to those of the metabolite 28, thereby confirming structure of 28 as 3-epidigoxigenin. The minor product (12 mg), 5, showed physical properties similar to those of digoxigenin. 153 OH HO CH. CH, CH Jones CH. OH Reagent OH HO 27 5 digoxigenin 3 Keto digoxigenin CH. 58 CH. OH 3,12 dlketo digoxigenin Figure 30. Synthesis of 3 Ketodlgoxigenin and 3,12 dlketo digoxigenin OH OH CH, CH, NaBH CH CH, OH OH HO 3 epldlgoxlgenin 27 3 Ketodlgoxigenin OH CH, digoxigenin OH HO Figure 31. Synthesis of 3 epldlgoxlgenin 5.1.8 6-Hydroxydlgoxlgenin 57 The FAB mass spectrum of the metabolite 57 gave a pseudomolecular ion at m/z 407 (MW 406), corresponding to a hydroxylated product of DGO (MW 390). This is again evident from the peaks at m/z 388, 370, 352 and 334 due to successive loss of 4 hydroxy groups in the form of water molecules. It is evident that the hydroxylation occurred in the B ring of digoxigenin from the peaks observed at m/z 181, 163 and 145. The mass fragmentation pattern postulated for metabolite 57 by compar ing the mass spectra of this compound, digoxigenin 5 and other similar cardenolides, is given in Figure 32 (83). The proton NMR spectrum of 52 (in CDgCODj, 270 MHz) showed a signal (doublet) at 4.49 ppm, corre sponding to a proton adjacent to a hydroxyl group. It is clear from the presence of a new signal for a proton geminal to a hydroxy group that the microbe introduced an additional secondary hydroxy group and that must be either at C6 or C7. Reaction of Jones reagent with metabolite 51 yielded the dike tone 59. Proton NMR of this compound 59, revealed the presence of C5 proton next to a carbonyl fucntion by the appearance of a signal at 155 Figure 32. MS Fragmentation Pattern of 6B-hydroxydigoxigenin 156 In the COSY spectrum of this compound the C5 proton was seen to be coupled only to the C4 protons. Thus the position of the newly intro duced hydroxy group was established as C6 . Also the MS of this com pound was in agreement with this assignment. 5.1.9 Conclusion Streptomyces aureus converts digoxigenin to 3 metabolites. The less polar metabolite is 3-ketodigoxigenin 27. The two polar metabo lites are 3 epidigoxigenin 28 and 63-hydroxy digoxigenin 57\ Determi nation of important physical properties and chromatographic behavior of these compounds could be utilized to detect the possible formation of these metabolites in humans. 5.2 BIOTRANSFORMATION OF DIGOXIGENIN BY SYNCEPHALASTRUM RACEMOSUM 5.2.1 Introduction As mentioned before, screening experiments led to the identifica tion of Syncephalastrum racemosum ATCC 18192 as one of the two microorganisms that produced polar metabolites of digoxigenin. S. racemosum cultures showed the formation of one polar metabolite in the TLC analysis and this metabolite showed different chromatographic characteristics from those of the polar metabolites isolated from Streptomyces -aureus ATCC 15437. Therefore it was decided to scale-up, isolate and identify this metabolite. Media, amounts of digoxigenin added to the culture and extraction procedures were varied in the experiment to obtain the best conditions for a scale-up. Preparative fermentation followed by extraction and chromatography led to the iso lation of a metabolite. Structure 60 is proposed for this metabolite based on initial proton NMR and Mass Spectra. OH CH CH OH HO OH 78-hydroxy digoxigenin 158 5.2.2 Determination of Optimum Fermentation Procedures 5.2.2.1 Variation of media Initial screening experiments were conducted using pharmamedia yeast extract (PY). The yield of the metabolite formed in this medium appeared low. Therefore it was decided to vary the media to choose the medium for growing S. racemosum that gives the best yield. PY, PYM, YM, YMM and NL 406 were the media examined. PY medium gave the best yield. 5.2.2.2 Amount of digoxigenin added to culture The above experiments were conducted using 10 mg digoxigenin per culture. Since the yield of the desired metabolite appeared to be low in every medium examined, it was decided to determine whether an increase in the amount of DGO added to the culture could lead to high er amount of the metabolite production. Stage II cultures grown in PY media were used for the study. TLC analysis of extracts of cultures grown with 10 mg, 20 mg and 30 mg of digoxigenin showed that the pro duction of the metabolite increased when 20 mg of DGO was added per culture. 5.2.2.3 Extraction procedure The use of different solvents for extraction of digoxigenin and metabolite, as described before in section 5.1.3.3, led to the identi fication of ethyl acetate as the best solvent for extraction of this metabolite. 159 5.2.3 Preparatlve-Scale Fermentation A preparative-scale fermentation, using Stage II cultures of Synce phalastrum racemosum grown in PY media, was performed to isolate suf ficient quantities of the metabolite 60 for structure elucidation. In this experiment, 600 mg of digoxigenin (20 mg per culture) was added to the cultures. After 7 days of incubation, pure starting material (520 mg) and crystalline metabolite 60 (31 mg) were isolated by ethyl acetate extraction and silica gel chromatography. The identity of the compounds was determined by comparison of the spectral properties of the isolated materials with those of DGO and other related cardeno- lides. 5.2.4 Digoxigenin The physical properties of digoxigenin has been described in sections 3.1.7 and 5.1.5. 5.2.5 7-Hydroxydigoxlgenin 60 The FAB mass spectrum of the metabolite 60 gave a pseudomolecular ion at m/z 407 for (M+H)+ corresponding to a molecular weight of 406. This indicates that a hydroxy group has been added to the starting material (MW 390) by the fungus. This is clear from the peaks appear ing at 388, 370, 352 and 334 that could arise by successive loss of 4 hydroxy groups in the form of water molecules. It is evident that the hydroxylation occurred in the B ring of DGO from the peaks seen at m/z 181, 163 and 145. 160 Figure 33. MS Fragmentation Pattern of 76-hydroxydlgoxlgenln 161 The mass fragmentation pattern postulated for this metabolite is giv en in Figure 33. This mass fragmentation pattern is postulated on the basis of the ones described for similar cardenolides (83). The proton NMR spectrum of 60 (in CDgCOCDg, 270 MHz) showed a broad singlet at 4.20 ppm (W 1/2 =12 Hz), corresponding to a proton adjacent to hydroxy group. 5.2.6 Conclusion Incubation of DGO with Syncephalastrum racemosum ATCC 18192 led to the formation of a polar metabolite. This metabolite was identified as 7f)-hydroxydigoxigenin 60 on the basis of its spectral properties and other physical properties. 162 5.3 SYNTHESIS OF 17-HYDROXYDIGOXIGEN1N AND DIGOXIGENIN DISULFATE 5.3.1 Introduction Carvalhas et al. have recently identified 17a-hydroxydigotoxin 51 as one of the metabolites of digitoxin 1 formed in the guinea pig (110). It is possible that similar 17a-hydroxy metabolites are also formed for digoxin 2 and its sugar-cleaved products (DG2, DG1 and DGO) in humans. Therefore it would be worthwhile to synthesize 17o-hydroxy derivatives digoxigenin for reference purposes and pharmacological testing. Procedures developed by Danieli et al. (Ill) and Saito et al. (9) were used to synthesize 17a-hydroxydigoxigenin 62 and 17a-hydroxydigoxigenin diacetate 63. Since there is strong evidence for the formation of conjugated metabolites of sugar-cleaved products (DG2, DG1 and DGO) in humans, it would be useful to synthesize these compounds for reference purposes. The use of the procedure developed by Peterson and his co-workers (104) led to the synthesis of digoxigenin disulfate 45. 5 R-H, Rj-H RO 61 R-Ac, Rj-H 62 R-H, R^-OH CH 63 R-Ac, Rj-OH 45 R-SO", Rj-H OH RO 163 5.3.2 17-Hydroxydigoxigenin 62 Refluxing 100 mg of DGO in dry dioxane with selenium dioxide (100 mg)_ for 24 hours followed by work-up, chromatography and recrystalli zation produced 65 mg of pure 17a-hydroxydigoxigenin 62. The FAB mass spectrum of this compound 62 showed pseudomolecular ion at 407 (MW 406) consistent with a hydroxylated product of DGO. The fragment ions that would be expected due to loss of 4 hydroxy groups in the form of water molecules were seen in both FAB and El mass spectrum. That the hydroxylation has occurred at C17 position is clear from the disappearance of the signal for C17o-proton usually seen at 3.5 ppm in DGO. The stereochemistry of the newly introduced hydroxy group is assigned as C17a by analogy. Refluxing 100 mg digoxigenin diacetate in dioxane with selenium dioxide (100 mg) for 24 hours followed by work-up and chromatography produced 72 mg of crystalline 17o-hydroxydigoxigenin diacetate 63. This compound was identical to the one prepared by acetylating 63 in the usual way with acetic anhydride and pyridine. Also this compound, 63, did not undergo further acetylation when a similar acetylation procedure was attempted. The FAB and El mass spectra clearly showed that this compound is a product of hydroxylated digoxigenin diacetate. The pseudomolecular ion at m/z 491 (MW 490) in the FAB mass spectrum and the disappearance of the signal for C17a proton in the NMR con firmed the identify of this compound as 17-hydroxydigoxigenin diace tate. The stereochemistry of the hydroxy group at C17 was assigned as a by analogy. 164 5.3.3 Digoxigenin disulfate 45 Reaction of chlorosulphonic acid with digoxigenin in pyridine according to the procedure developed by Peterson et al. (104) followed by work-up resulted in the production of digoxigenin disulfate 45. The mass spectra (FAB and El) of this compound, 45, gave pseudomo lecular ion (m/z 551) and other fragments consistent with the proposed structure of this compound. 5.3.4 Conclusion 17a-Hydroxydigoxi.genin 62 17a-hydroxydigoxigenin diacetate 63 and digoxigenin disulfate 45 were prepared by using chemical synthetic procedures described for similar cardenolides. The availability of these potential metabolites will be useful in the identification the several previously uncharacterized metabolites of digoxin formed in humans. CHAPTER VI EXPERIMENTAL 6.1 BIOTRANSFORMATION OF DIGOXIGENIN BY STREPTOMYCES AUREUS 6.1.1 Chemicals and Media Dlgoxin was supplied by Burroghs Wellcome Co. , Research Triangle Park, North Carolina. The media used for growing cultures were Phar- mamedia yeast extract (PY), Pharmamedia yeast extract with added man- nltol (PYM); yeast malt extract medium (YM), Yeast malt extract medium with added mannitol and NL 406 medium. The preparation of PY medium has been described Section 3.3.1. of part I. PYM was prepared by add ing 5% mannitol to PY medium. YM consisted of (per liter of water): yeast extract 3g; malt extract 3g; peptone 5g; and D-glucose lOg. YMM was prepared by adding 5% mannitol to YM medium. NL406 consisted of (per liter of water): mannitol 50g; sucrose 50g; succinic acid 5.4g; yeast extract 3g; KI^PO^ o. lg; MgSO^-Tl^O 0.3g; FeSO^-TI^O O.Olg; ZnSO^-T^O 0.0044g; and was adjusted to pH 5.4 with NH^OH. 6.1.2 Preparation of Digoxigenin from Digoxin Digoxin (lg) was dissolved in 200 ml of 60% methanol and distribut ed equally into 5 Erlenmeyer flasks (250 ml). The pH was adjusted to - 165 - 166 1, using 10% HC1. The flasks were stoppered and shaken for 24 hours on a Gyrotory shaker (New Brunswick Scl. Co.) at 250 rpm at 25°C After adjusting the pH of the reaction mixtures to 7 using 10% NaOH, the solution was extracted twice with double the volume of chloroform. Evaporation of the solvent yielded a slightly yellow colored solid (585 mg). Pure digoxigenin (481 mg) was obtained after chromatography in a solvent system of methylene chloride and methanol (20:1). 6.1.3 Microorganisms The microorganisms used in screening experiments were obtained from American Type Culture collection and are currently maintained in the OSU College of Pharmacy culture collection. 6.1.4 Instrumentation FAB and El mass spectra were obtained with a Kratos MS-30 double- focusing E/B mass spectrometer, at the Campus Chemical Instrumentation Center. Deuterated solvents were used for determining ^H-NMR spectra with a Bruker Hx-270 spectrometer at the College of Pharmacy or a Bruker 500 MHz spectrometer at the Campus Chemical Instrumentation Center. Other instruments used have been described in section 3.1.3. of part I. 6.1.5 Screening Experiments A procedure described for the screening of microorganisms for the 12&-hydroxylation of digitoxigenin in section 3.3.12 of part I was used for the identification of microorganisms that are capable of pro ducing polar metabolites from digoxigenin. In this experiment, Stage 167 II cultures grown in PY media were used. DGO (10 mg) in 0.4 ml of ethanol was added to each of the 1-day old stage two cultures. After 7 days of incubation, the whole cultures were extracted double with twice the volume of ethyl acetate and analyzed by TLC. Methylene chloride: methanol (15:1) was the solvent system used in TLC analy ses. A total of 12 organisms were screened, and they are listed in table 14. 6.1.6 Determination of Optimum Fermentation Procedures 6 .1.6.1 Variation of Media Stage II cultures grown in PY, PYM, YM, YMM and NL 406 were used in this experiment. The preparation of these media has been described in section 6.1.1.. Cultures were inoculated with 10 ml of stage I cul ture grown in PY medium. 6.1.6.2 Amount of Digoxigenin Stage II cultures of S. aureus were grown in PY medium. One-day- old Stage II cultures were incubated for an additional 7 days after adding 10, 20 or 30 mg of DGO. After incubation they were extracted and analyzed by the TLC procedure used in the screening experiments. 6.1.6.3 Extraction Procedure Stage II cultures of S. aureus grown in PY media were used in this experiment. The cultures incubated with digoxigenin at a concentra tion of 10 mg per flask were extracted with methylene chloride, chlo roform or ethyl acetate and analyzed by TLC. 168 6.1.7 General Isolation Procedures The Isolation procedures used In S. aureus experiments were similar to those described In section 3.1.12.5. 6.1.8 Isolation and Purification of Digoxigenin Metabolites Digoxigenin (500 mg) was dissolved in absolute ethanol (10 ml) and distributed evenly among 25 1-day old Stage II cultures of S. aureus. The cultures were incubated at 25°C on a Gyrotory shaker shaking at 250 rpm for 7 days. After incubation the cultures were extracted with twice the volume of ethyl acetate. The concentrated EtOAC extract (1.04 g) was applied to a silica gel column (PF 254, E. Merck, 100 g, 50 cm x 2.5 cm) and eluted with methylene chloride: methanol (20:1). A total of 235 fractions of 10 ml were collected. Metabolite 27 was eluted in fractions 52-62 (61 mg), Digoxigenin eluted in fractions 81-89 (352 mg), metabolite 28 eluted in fractions 91 - 97 (37 mg) and metabolite 57 was eluted in fractions 113 - 125 (41 mg). Crystalline metabolites 27 (40 mg) and 28 (22 mg) were obtained after preparative thin layer chromatography and recrystallization in methanol of the impure fraction. Fractions 113 - 125 were combined and the solvent was evaporated. The impure metabolite 57 was applied to a second column (Si PF 254, 32 g) and eluted with methylene chloride: methanol (20:1). Fractions 83 - 91 contained the pure metabolite 52 (25 mg). 169 6.1.9 Synthesis of 3-Ketodigoxlgenln and 3,12-Dlketodigoxlgenin Jones reagent (0.25 ml) was added to digoxingenin (100 mg) In 1 ml of acetone solution at 0°C and allowed to react. After 5 minutes the reaction mixture was poured into a separatory funnel containing 50 ml chloroform and 50 ml water. After mixing well the chloroform layer was drained into a round bottom flask. The aqueous solution was extracted with 50 ml of chloroform again. After evaporating chloro form a yellow colored solid (102 mg) was obtained. TLC of this extract showed disappearance of the starting material and formation of 2 products. PTLC in methylene chloride: methanol (20:1) gave 63 mg of pure 27 and 22 mg of pure 58. 6.1.10 Synthesis of 3-Epldigoxigenin 28 Sodium borohydride (150 mg) was added to 2 ml of methanol contain ing 50 mg of 3-ketodigoxigenin and the mixture was stirred at 25°C for 12 hours. The reaction mixture was filtered and extracted twice with 50 ml of chloroform. TLC of the extract showed disappearance of the starting material and formation of two products. PTLC of the extract in methylene chloride: methanol (20:1) gave crystalline 28 (31 mg) and digoxigenin 5 (12 mg). 6.1.11 Preparation of 3,6-Dlketodlgoxigenln 59 Jones reagent (0.25ml) was added to the metabolite 57 (lOmg) in acetone solution at 0°C and allowed to react. After 5 minutes, the reaction mixture was poured into a separatory funnel containing 50ml of water and extracted twice with 50ml of chloroform. After evaporat- 170 ing the solvent a yellow colored solid (15mg) was obtained. PTLC of the extract In methylene chloride: methanol (20:1) gave 8mg of crys talline 3,6-diketodigoxigenln. 6 .1.12 Physical Properties of DGO 5 The physical properties of digoxigenin have been described In sec tion 3.1.13. of part I. 6.1.13 Physical Properties of 3-Ketodlgoxlgenln 27 3-Ketodigoxigenin exhibited the following physical properties: MP 250-254°C; [ct]D25 + 38° (c = 0.8, methanol) 1H-NMR (pyridine-d5, 500 MHz) 6 0.89 (3H, S, C19-Me), 1.23 (3H, S, C18-Me), 2.45 (1H, ddd, J=ll,3, C2), 2.78 (1H, t, J=ll, C4), 5.1 and 5.21 (2H, dddd, J=18,4, C21), 6.17 (1H, S-broad, C12), 6.27 (1H, S, C22); EI/MS (probe) 70 ev m/z (rel. Int.) 388 (20), 370 (M-H20), 352 (M-2H20) (22), 337 (M-CH702) (10). 6.1.14 Physical Properties of 3,12-Dlketodlgoxigenln 58 3,12-Diketodigoxigenin exhibited the following physical properties: mp 265-667°C; 1H-NMR (pyridinde-d5, 500 MHz) 6 0.98 (3H, S, (19-Me), 1.27 (3H, S, C = 18-Me), 2.45 (1H, ddd, J=9,3, C2), 2.55 (1H, t, J=ll, C4), 2.74 (1H, t, J=ll, C4) , 4.48 (1H, t, C17), 5.12 and 5.22 (2H, dddd, J=18,4, C21), 6.29 (1H, S, C22); EI/MS (probe) 70 ev m/z (rel. Int.) 386 (100), 368 (M-H20) (50, 353 (M-CHjO) (20). 171 6.1.15 Physical Properties of 3-Epldlgoxlgenln 28 3-Epidigoxigenin exhibited the following physical properties: mp 259-261° C; [a]D24 + 26° (c = 0.71, methanol) ^H-NMR (CD3COCD3 , 270 MHz) 6 0.83 (3H, S, C18-Me), 0.93 (3H, S, C19-Me), 3.45 (1H, M, C17), 3.52 (1H, m, C3), 3.81 (1H, S-broad, C12), 4.8 and 5.0 (2H, dddd, J=18,4, C21), 5.81 (1H, S, C22). EI/MS (probe) 70 ev (rel. Int.) 372 (M-H20) (17), 354 (M-2H20) (35), 336 (M-3H20) (6 ), 219 (M-CgHn 04) (86), 201 (M-C8H1305) (95), 147 (M-C12Hj905) (27). 6.1.16 Physical Properties of 6 -Hydroxydigoxigenin 57 63-Hydroxydigoxigenin exhibited the following physical properties: mp 192-196°C; 1H-NMR (Pyridine-d5, 500 MHz) 6 1.15 (3H, S, C19-Me), 1.22 (3H, S, C18-Me), 3.75 (1H, M, C17), 4.42 (1H, S-broad, C3), 5.1 and 5.2 (2H, dddd, J=18,4, C21), 6.15 (1H, d, C12), 6.25 (1H, S, C22), 6.7 (1H, d, C6); EI/MS probe 70 ev (rel. Int.) 388 (M-H20) (3), 370 (M-2H20) (21), 352 (M-3H20) (30), 334 (M-4H20) (100), 235 (M-CgHn 04) (10), 217 (M0C8H13O5) (35), 199 (M-C8H 1506) (31), 181 (M-C12H1704) (20), 163 (M-C12H&s;19.05) (18), 145 (M-C12H2106) (35). 6.1.17 Physical Properties of 3,6-Dlketodlgoxlgenln 59 The derivtlve 59 exhibited the following spectral properties: 1H-NMR (Prldlne-d5, 500 MHz) 6 1.23 (3H, S, C19-Me), 1.3 (3H, S, C18-Me), 2.6 (1H, dd J=10,4, C5), 2.71 (1H, t J=9,5, C4). 172 6.2 BIOTRANSFORMATION OF DIGOXIGENIN BY SYNCEPHALASTRUM RACEMOSUM 6.2.1 General Experimental Procedures Chemicals and media, and instrumentation used in this study were similar to the ones described in sections 6.1.1 and 6.1.4. The screening experiments and experiments conducted for determination of optimum fermentation procedures were similar to the ones described for biotransformation of digoxigenin by S^ aureus in sections 6.1.5 and 6 .1.6.1 - 6.1.6.3. 6.2.2 Isolation and Purification of Digoxigenin and the Metabolite 60. Digoxigenin (600 mg) was dissolved in absolute ethanol (12 ml) and distributed evenly among 30 1-day old Stage II cultures of racemo sum grown in PY medium. The cultures were incubated at 25°C on a Gyrotory shaker shaking at 250 rpm for 7 days. After incubation the cultures were extracted twice with double the volume of ethyl acetate. The concentrated ethyl acetate extract (1.23 g) was applied to a silica gel column (PF 254, E. Merck, 100 g, 50 cm x 2.5 cm) and eluted with methylene chloride: methanol (20:1). A total of 210 fractions of 12 ml each were collected. Digoxigenin was eluted in fractions 75-88 (511 mg), and the metabolite 60 was eluted in fractions 105-116 (57 rag). Since the metabolite was found to be still impure in TLC analysis, it was applied to a second column (PF 254, E. Merch, 30 g) and was 173 eluted with methylene chloride: methanol (20:1). Fractions 97-106 (5 ml each) contained the pure metabolite 60 (31 mg). 6.2.3 Physical Properties of Digoxigenin The physical properties of digoxigenin have been described in sec tion 3.1.13. 6.2.4 Physical Properties of 7-Hydroxydlgoxlgenin 60 7 3-Hydroxydigoxigenin exhibited the following spectral properties: ^I-NMR (Pyridine-d5, 500 MHz)6 1.02 (3H, S, C19-Me), 1.22 (3H, S, C18-ME), 3.73 (1H, m, C17), 4.21 (1H, S-broad, C7), 4.31 (1H, S-broad, C3), 5.1 and 5.21 (2H, dddd, J=18,4, C21), 6.17 (1H, S-broad, C12), 6.2 (1H, S, C22); EI/MS (probe) 7 ev (rel. int.) 388 (M-H20) (10), 370 (M-2H20) (40), 352 (M-3H20) (45), 334 (M-4H20) (20), 235 (M-C8H n 04) (50), 181 (M-C12H1704) (32), 163 (M-C12H1905) (39), 145 (M-Ci2H2i06) (55). 174 6.3 SYNTHESIS OF 17-HYDROXYDIGOXIGENIN AND DIGOXIGENIN DISULFATE 6.3.1 Chemicals Selenium dioxide was obtained from Alfa products and chlorosulfonic acid was obtained from Fischer Scientific Company. 6.3.2 Synthesis of 17-Hydroxydlgoxigenin 62 Selenium dioxide (100 mg) was added to a solution of digoxigenin (100 mg) in dry dioxane (20 ml) and refluxed for 24 hr. The precipi tated selenium was removed by filtering and the filtrate was poured into water (50 ml) and extracted twice with double the volume of chlo roform. The organic extract was washed repeatedly with saturated aqueous sodium chloride and dried over sodium sulfate. Crystalline 17a-hydroxydigoxigenin (65 mg) was obtained after PTLC in a solvent system of CH2C12 : MeOH (10:1) and recrystallization from methanol. 17aHydroxydigoxigenin exhibited the following spectral properties: EI/MS (probe) 70 ev (rel. int.) 406 (3), 3.88 (M-H20) (30), 370 (M-2H20) (15), 352 (M-3H20) (12), 337 (M-CHg03) (7); 334 (M-4H2) (2), 219 (M-C8Hn 05) (22), 201 (M-CgH^Og) (65), 147 (M-C12H 1906) (55). 6.3.3 Synthesis of 17-Hydroxydigoxlgenin Diacetate 63 17o-Hydroxydigoxigenin diacetate 63 was synthesized and purified by using a similar procedure described for 17a-hydroxydigoxigenin and the product 63 exhibited the following spectral properties: ^H-NMR (CD3COCD3 , 270 MHz)6 0.96 (3H, S, C18-Me), 1.04 (3H, S, C19-Me), 4.85 (111, dd, C12), 5.05 and 5.15 (211, dddd, J=18,4, C21), 5.21 (1H, dd, C3), 6.0 (111, S-broad, C22). EI/MS (Probe) 70 ev (rel. int.) 490 (1.5), 370 (M-C3H1203) (10), 352 (M-CgHjgC^) (12). 175 6.3.4 Synthesis of Digoxigenin Disulfate 45 To a solution of digoxigenin (100 mg) in 1.2 ml of pyridine at -4°C 40pl of chlorosulfonic acid In 0.5 ml of pyridine was added and the reaction mixture was allowed to react at -4°C for 5 hr. The reaction mixture was then poured Into a separatory funnel containing 50 ml of water and extracted twice with 100 ml of ethanol: chloroform (1:1). Evaporation of the solvent yielded a yellow colored solid (89 mg). Preparative thin layer chromatography of the extract in methylene chloride: methanol (5:1) gave 25 mg of digoxigenin 5 and 21 mg of digoxigenin disulfate 45. Digoxigenin disulfate 45 exhibited the fol lowing spectral properties. EI/MS (probe) 70 ev (rel. int.) 368 (3), 354 (10), 336 (95), 321 (30), 282 (25), 231 (10). CHAPTER VII LIST OF REFERENCES 1. K. Greeff and H. Schadewaldt, in: "Handbook of Experimental Pharmacology," vol. 56/1, Springer Verlag, New York, 1981, pp. 1- 12. 2. E. Kerry, in: "Medical Classics,", Baltimore, Williams and Wil kins, 1937, p. 305. 3. B.F. Hoffman and J.T. Bigger, Jr., in: "The Pharmacological Basis of Therapeutics," Ed. by A.G. Gilman, L.S. Goodman, T.W. Rail and F. Murad, 7th Edition, New York, The McMillan Company, 1985, pp. 716-730. 4. M. Okada and T. Anjyo, Chem. Pharm. Bui1., 23, 2039 (1975). 5. N. Ishikawa, H. Tsuru, T. Shigei, T. Anjyoand M. Okada, Exper- ientia, 30, 1308 (1974). 6 . G.K. Patnaik and B.N. Dhawan, Arzneim. Forsch., 28, 1095 (1978). 7. W. Zurcher, E. Weiss-Berg and C. Tamm, Helv. 52, 2449 (1969). 8 . M.S. Ragab, H.H.A. Linde and K. Meyer, Helv. 45, 1794 (1962). - 176 - 177 9. Y. Salto, Y. Kanemasa and M. Okada, Chem. Pharm. Bull. , 18, 629 (1970). 10. T. Shigei, H. Tsuru, Y. Salto and M. Okada, Experlentla, 29, 449 (1973). 11. P. Hofer, H. Linde and K. Meyer, Helv. Chlm. Acta., 43, 1955 (1960). 12. R.H. Hammer, in: "Principles of Medicinal Chemistry," Ed. by W.O. Foye, pp. 395-429. Philadelphia, I.ea and Febiger, 1981. 13. J. Boutagy, A. Gelbart, and R. Thomas, Aust. J. Pharm. Sci. N.S., 2, 41 (1973). 14. R. Thomas, J. Boutagy and A. Gelbart, J. Pharm. Sci., 63, 1649 (1974). 15. T.W. Guntert and H.H.A. Linde, in: "Handbook of Experimental Pharmacology," 56/1, Ed. by K. Greeff, New York, Springer Ver- lag, 1981. 16. R.L. Vick, J.O. Kahn and G.H. Acheson, J. Pharmacol. Exp. Ther., 121, 330 (1957). 17. U. Stache, W. Fritsch, W. Haede and K. Radscheit, Liebigs Ann. Chem. 1461 (1977). 18. H. Tsuru, N. Ishikawa, T. Shigei, T. Anjyo and M. Okada, Jpn. J. Pharmacol., 27, 799 (1977). 178 19. G.E. Lindenmayer, Pharmac. Ther. B, 2, 843 (1976). 20. T. Akera and T.M. Brody, Pharmacol. Rev., 29, 187 (1978). 21. W.M. Hart and E.D. Titus, J. Biol. Chem. 248, 4674 (1973). 22. P.F. Baker, M.P. Blaustein, A.L. Hodgkin and R.A. Stelnhardt, J. Physiol. 200, 421 (1969). 23. P.F. Baker, Prog. Blophys. Molec. Biol., 24, 177 (1972). 24. A. Gervais, L.K. Lane and G.E. Lindenmayer, Ann. N.Y. Acad. Sci., 242, 445 (1974). 25. J. Lindenbaum, D.G. Rund, V.P. Butler, Jr., D. Tse-Eng and J.R. Saha, N. Eng. J. Med., 305, 789 (1981). 26. G. T. Okita, F.E. Kelsey, P.J. Talso, L.B. Smith and E.M. Gell ing, Circulation, 7, 161 (1953). 27. U. Peters, L.C. Falk and S.M. Kalman, Arch. Intern. Med. , 138, 1074 (1978). 28. D.S. Lukas, in: "Cardiac Glycosides," Ed. by G. Bodem and H.J. Dengler, New York, Springer, 1978. 29. I,. Storstein, Clin. Pharmacol. Ther., 21, 536 (1977). 30. H. Gault, J. Kalra, L. Lingerich and M. Dawe, Clin. Pharmacol. Ther., 31, 695 (1982). 179 31. J. Kuhlmann, U. Abshagen and N. Rietbrock, Europ. J. Pharma col. , 7, 87 (1974). 32. U. Abshagen and N. Rietbrock, Naunyn-Schmledeberg's Arch. Phar macol. , 276, 157 (1973). 33. F.I. Marcus, J.N. Ryan and M.G. Stafford, J. Lab. Clin. Med. , 85, 610 (1975). 34. M.H. Gault, L. L. Longerlch, J.C.K. Loo, P.T.H. Ko, A. Fine, S.C. Vasdev and M.A. Dawe, Clin. Pharmacol. Ther., 35, 74 (1984). 35. K.D. Gierke, P.F. Graves, D. Perrier, F.I. Marcus, M. Mayersohn and S. Goldman, J. Pharmacol. Exp. Ther., 212, 448 (1980). 36. H. Fischer, Naunyn-Schmledeberg1s Arch. Exp. Path^ Pharmak., 130,11 (1928). 37. E. Kelsey, Illinois Med. J., 100, 309 (1951). 38. E. Haack, F. Kaiser, F. Spinger, and H. Splngler, Naturwlssen- shaften, 44, 633 (1957). 39. J. 0. Magnusson, B. Bergdahl, C. Bogentoft, U.E. Jonsson and L. Tekenbergs, Blopharmaceut. Drug. Dlspos., 3, 211 (1982). 40. M.H. Gault, D. Sugden, C. Maloney, M. Ahmed and M. Tweeddale, Clin. Pharmacol. Ther., 25, 499 (1979). 41. D.R. Clark and S.M. Kalman, Drug Metab. Dlspos■, 2, 148 (1974). 180 42. H. Gault, J. Kalra, M. Ahmed, D. Kepkay and J. Barrowman, Clin. Pharmacol. Ther., 27, 16 (1980). 43. II. Gault, J. Kalra, M. Ahmed, D. Kepkay, L. Longerlch and J. Barrowman, Clin. Pharmacol. Ther., 29, 181 (1981). 44. A. Schmoldt and B. Ashendorf, Europ. J. Drug Metab. Pharmacoki- net., 5, 225, (1980). 45. U. Abshagen, H. Rennekamp, R. Kuchler and N. Rietbrock, Europ. J. Clin. Pharmacol.. 7, 177 (1974). 46. R. J. Luchi and J.W. Gruber, Am. J. Med., 45, 322 (1968). 47. E. Watson, D.R. Clark and S.M. Kalman, J. Pharmacol. Exp. Ther., 184, 424 (1973). 48. D. Sugden, M. Ahmed, C. Maloney and M.H. Gault, Clin. Res., 23, 610A (1975). 49. D. Sugden, M. Ahmed and M.H. Gault, J. Chromatogr. , 121, 401 (1976). 50. H. Greenwood, W. Snedden, R.P. Hayward and J. Landon, Clin. Chim. Acta, 62, 213 (1975). 51. H. Greenwood and W. Snedden, in: "Mass Spectrometry in Drug Metabolim," Ed. by, A. Frigerio and E.L. Ghisalberti, New York, Plenum, 1977, pp. 179-189. 52. U. Peters and S.M. Kalman, Z. Kardiol., 67, 342 (1978). 181 53. I. Herrmann and K. Repke, Abh. Dtsch. Akad. Wissenschaften Ber lin, S.115 (1968). 54. J. Lindenbaum, D. Tse-Eng, V.P. Butler and D.G. Rund, Am. J. Med., 71, 67 (1981). 55. D.R. Rund, J. Lindenbaum, J.F. Dobkin, V.P. Butler and J.R. Saha, Clin. Pharmacol. Ther., 34, 738 (1983). 56. J.F. Dobkin, J.R. Saha, V.P. Butler, Jr., H.C. Neu and J. Lin denbaum, Science, 220, 325 (1983). 57. J.F. Dobkin, J.R. Saha, V.P. Butler, Jr., H.C. Neu and J. Lin denbaum, Trans. Assoc. Americ. Physicians, 95, 22 (1982). 58. V.D. Bokkenheuser, J. Winter, P. Dehazya and W.G. Kelly, Appl. Environ. Microbiol., 34, 571 (1977). 59. V.D. Bokkenheuser, J. Winter, S.M. Finegold, V.L. Sutter, A.F.. Ritchie, W.E.C. Moore and L.V. Holdeman, Appl. Environ. Micro biol. , 37, 1001 (1979). 60. H. Eyssen and A. Verhulst, Appl. Environ. Microbiol., 47, 39 (1984). 61. B.A. White, R.J. Fricke and P.B. Hylemon, J. Lipid Research, 23, 145 (1982). 182 62. H.J. Eyssen, G.G. Parmentier, F.C. Compernolle, G. DePaw and M.P. Denef, Eur. J. Blochem., 36, 411 (1973). 63. T.L. Glass and C.Z. Burley, Appl. Environ. Microbiol., 49, 1146 (1985). 64. I.A. MacDonald, Y.P. Rochon, D.M. Hutchison and L.V. Holdeman, Appl. Environ. Michrobiol., 44, 1187 (1982). 65. L. Storstein, in: "Handbook of Experimental Pharmacology," Vol. 56/11, Ed. by K. Greeff, New York, Springer Verlag, 1981, pp. 3-30. 66. H.F. Vohringer and N. Rietbrock, in: "Cardiac Glycosides," Ed. by G. Bodem and H.J. Dengler, New York, Springer Verlag, 1978, pp. 64-73. 67. D. Perrier, M. Mayersohn and F.I. Marcus, Clinical Pharmacoki netics, 2, 292 (1977). 68. L. Storstein, Clin. Pharmacol. Ther., 21, 125 (1977). 69. G. Bodem and E.V. Unruh, in: "Cardiac Glycosides," Ed. by G. Bodem and H.J. Dengler, New York, Springer Verlag, 1978, pp. 74-84. 70. H.N. Bockbrader and R.H. Reuning, J. Pharm. Scl., 72, 271 (1983). 183 71. R.H. Reuning, T.A. Shepard, B.F.. Morrison and 11.N. Bockbrader, Drug Metab. Dispos., 13, 51 (1985). 72. D.S. Fullerton, T.M. Gilman, M.C. Pankaskie, K. Ahmed, A.11. From, W.L. Duax and D.C. Rohrer, J. Med. Chem., 20, 841 (1977). 73. L.W. Robertson, A. Chandrasekaran, R.H. Reuning, J. Hui and B.D. Rawal, Appl. Environ. Microbiol., 1986, in press. 74. J.F. Sperry and T.D. Wilkins, J. Bacteriol., 127, 780 (1976). 75. M.A. Posthumas, P.G. Kistemaker and H.I..C. Meuzelaar, Anal. Chemistry, 50, 985 (1978). 76. A.P. Bruins, Intern. J. Mass Spectrom. Ion Physics, 48, 185 (1983). 77. P.G. Kistemaker, M.M.J. Lens, G.J.Q. Van der Peyl and A.J.H. Boerboom, Advan■ Mass Spectrom., 8A, 928 (1980). 78. M.B.E. Fayez and S.A.R. Negm, J. Pharm. Sci., 61, 765 (1972). 79. P. Brown, F. Bruschweiler and G.R. Pettit, Helv. Chim. Acta, 55, 531 (1972). 80. J. Jones, I.A. Veliky and R.S. Ozubko, J. Nat. Prod. , 41, 476 (1978). 81. A. Mostad, Acta Chim. Scand., B36, 635 (1982). 82. T.W. Smith and E. Harber, N. Eng. J. Med., 289, 945 (1973). 184 83. R.H. Reuning and D.R. Geraets, In: "Applied Pharmacokinetics: Principles of Therapeutic Drug Monitoring," Ed. by W.E. Evans, J.J. Schentag and W.J. Jusko, San Francisco, Applied Therapeu tics Inc., 1986. 84. G.L. Lage and J.L. Spratt, J. Pharmacol. Exp. Ther. , 152, 501 (1966). 85. E.J. Bach and M. Reiter, Naunyn-Schmledeberg1 s Arch. Exp. Pha- thol. Pharmacol., 248, 437 (1964). 86. D. Geraets, J. Hul, A. Chandrasekaran, J. Caldwell, L. Robert son, R. Donnerberg and R. Reuning, Clin. Pharmacol. Ther., (1986) submittted. 87. N. Heinz and 11. Flasch, Naunyn-Schmledeberg1s Arch. Pharmacol., 303, 181 (1978). 88. A.P. Bruins, Advan. Mass Spectrom., 8A, 246 (1980). 89. A.F. Beecham, Tetrahedron Letters, 19, 2355 (1968). 90. A.F. Beecham, Tetrahedron Letters , 32, 3591 (1968). 91. A. Gubler and C. Tamm, Helv. Chim. Acta., 41, 297 (1958). 92. 11. Nawa, M. Uchibayashl, T. Kamiya, T. Yamano, H. Aral and M. Abe, Nature, 184, 469 (1959). 185 93. M. Okada, A. Yamada and M. Ishidate, Chem. Pharm. Bull. (Japan), 8, 530 (1960). 94. C. Tamm and A. Gubler, Helv. Chim. Acta., 41, 1762 (1958). 95. M. Schupbach and C. Tamm, Helv. Chim. Acta., 47, 2217 (1964). 96. C. Tamm and A. Gubler, Helv. Chim. Acta., 42, 473 (1959). 97. M. Schupbach and C. Tamm, Helv. Chim. Acta., 47, 2226 (1964). 98. S.Y. Lee, S.E. Moore and M.S. Mabee, Appl. Environ. Microbiol., 41, 386, (1981). 99. T.A. Shepard, J. Hui, A. Chandrasekaran, R.A. Sams, R.H. Reuning, L.W. Robertson, J.H. Caldwell and R.L. Donnerberg, J. Chroma- togr. (Biomed. Appl.), 1986. In Press. 100. H. Fromm and A.F. Hofman, The Lancet, Sept. 18 (1971). 101. A. Chandrasekaran, L.W. Robertson, D.R. Geraets, J. Hui, J.H. Caldwell, R.L. Donnerberg and R.H. Reuning, Drug Metab. Dispos. 1986 (submitted). 102. J.O. Magnusson, Br. J. Clin. Pharmac., 16, 741 (1983). 103. J.O. Magnusson, B. Bergdahl and S. Gustafsson, Arzneim. Forsch., 34, 87, (1984). 104. R. Peterson, H. Flasch and N. Heinz, Arzneim. Forsch., 27, 642 (1977). 186 105. G. Benthe, in: "international Symposium on Cardiac Glycosides," Ed. by G. Bodem and Dengler, New York, Springer Verlag, 1978, pp. 52-63. 106. W.H. Bulger and S.J. Stohs, Biochem. Pharmac., 22, 1745 (1973). 107. R.E. Talcott, W.H. Bulger and S.J. Stohs, Steroids, 21, 87 (1973). 108. R.E. Talcott and S.J. Stohs, Res. Commun. Chem. Path. Pharmac., 5, 663, (1973). 109. W.H. Bulger, S.J. Stohs and D.M.S. Wheeler, Biochem. Pharmac., 23, 921 (1974). 110. M.L. Carvalhas, M.A. Figueira, M.E. Araujo and M.R. Maya, Drug Metab. Dispos., 11, 85 (1983). 111. N. Danieli, Y. Mazur and F. Sondheimor, Tetrahedron, 23, 715 (1967). 112. R.V. Smith and J.P. Rosazza , J. Pharm. Sci.,64, 1737 (1975). 113. A. Chandrasekaran and L.W. Robertson, Appl■ Environ. Micro biol ., 1986 (Submitted). 114. H. Iizuka and A. Naito in "Microbial Transformation of Steroids and Alkaloids," State College, Penn., University Park Press, 1967, pp. 231-244. 187 115. H. Iizuka and A. Nalto In "Microbial Conversion of Steroids and Alkaloids," New York, Springer-Verlag, 1981, pp. 329-332. 116. P. Bellet and D. Gerard, Ann. Pharm. Franc., 21, 593 (1963). 117. P. Bellet and T. Van Thuong, Ann. Pharm. Franc., 23 , 57 (1965). 118. H. Ishii, Y. Nozaki, T. Okumura and D. Satoh, J. Pharm. Soc. (Japan), 81, 1051 (1961). 119. E. Titus, A.W. Murray and H.F.. Spiegel, J. Biol. Chem., 235, 3399 (1960). 120. G. Juhasaz and C. Tamm, Helv. Chim. Acta. , 44, 1063 (1961). 121. H. Ishii, Y. Nozaki, T. Okumura and D. Satoh, J. Pharm. Soc. (Japan), 81, 805 (1961). 122. C. Tamm and A. Gubler, Helv. Chim. Acta., 42, 239 (1959). APPENDIX -188- S / J I iMo- w \ J L U \) J ^ i.i 2.1 l.< PP* Figure 34. ^H-NMR Spectrum (CD^COCD^! 270 MHz) of 20R-dihydrodigoxigenin J “T PM Figure 35. *H-NMR Spectrum (CD^COCD^, 270 MHz) of digoxigenin . I l l _A ^JL • -'* t.O 7.* 7.0 $.% (.0 sfs ^ s!o 4-1 7P«l!* «!o jrr^71o /s * ?!t* I!S ' I!o A Figure 36. ^H-NMR Spectrum (Pyridine-d^, 500 MHz) of digoxigenin. 192 Figure 37. XH-NMR Spectrum (CD3COCD3> 270 MhZ) of 3-ketodigoxigenin. JUL JL u _J l UU l MA j L 0.4 1.0 7.4 7.0 S.4 S.D 4.4 4.0 ffR 4.4 4.0 5.4 5.0 7.4 7.0 I. Figure 38. H-NMR Spectrum (Pyridine-d,., 500 MHz) of 3-ketodigoxigenin. J iju L Ilk. A\U. 9.5 9.9 7.5 7.0 6.5 6.0 5.5 5.0 <.5 4.0 5.5 $.0 7.5 2 0 1.5 I 0 5 ff H Figure 39. 1H-NMR Spectrum (Pyridine-d-, 500 MHz) of 3,12-diketodigoxigenin. i j ___i M ,_J ^ / W A jIL 1 1 1 ..1 1 I 3 fi ' V ’1 B [j « *•» , r i*< ft* f“ « T f | ...... 1 1 .J . . , ftI .... . u . .. f t . Ji b. 4 i.4 4.4 1.4 I 4 PPH Figure 40. 1H-NMR Spectrum (CD3COCD3> 270 MHz) of 3-epidigoxigenin. Figure 41. ^H-NMR Spectrum (CD^COCD^, 270 MHz) of 6B-hydroxydigoxigenin. Jwllk — V 'T ' " I I . ' ' I ' ' . 1 |-r I I , 1 1 ' > , | I I , I | ' I I I ! I T . <-r-rt ■ I ■■ r-[ . ■ ■ iT T-v i-rT-r- 2.0 6.S C.Q S. 5 &.0 4. £. 4.0 J.S. J.O 2.S 2.0 1.‘, 1.0 PPH Figure 42. H-NMR Spectrum (Pyridine-d^, 500 MHz) of 68-hydroxydigoxigenin. 197 Ill I Ik A A A U ’ “ I * I r r n f T T . r*r . ' r * f-* rs- 9.S 8.0 7.0 7.0 G.S G.O Ij.O 4.0 4.0 i . t 3.0 2.0 2.0 1.0 pph Figure 43. H-NMR Spectrum (Pyridine-d^, 500 MHz) of 3,6-diketodigoxigenin. 11 J l k M A A j l JUL. Figure 44. COSY (Pyridine-d,., 500 MHz) spectrum of 3,6-diketodigoxigenin LL 1 Uii U l _ _ _ _ ■ , I . . ... | I I 1^'' | 1 ' ■ J ( ' J 1 J I ■ ' ■ ' I J ■ ■ ■ I ■ ■ • ' ! ■ ■ ■ ’ k ■ ' 1 ' i 1 7.0 7.0 1.1 (.0 S.S S.O «.* 4.0 ).( S.O 2.S 2.0 I.S 1.0 pro 200 Figure 45. ^H-NMR Spectrum (Pyridine-d,., 500 MHz) of 76-hydroxydigoxigenin. s' y y JL u ' i * 4.1 201 PP» Figure 46. H-NMR Spectrum (CD^COCD^. 270 MHz) of 17a-hydroxydigoxigenin diacetate.