AN INVESTIGATION OF POTATO ALKALOID LEVELS IN
THE NORMAL POPULATION AND THEIR RELATIONSHIP TO FOETAL
NEURAL TUBE DEFECTS
Submitted in fulfilment of the requirements for the degree: Doctor of Philosophy of the University of Surrey by:
MICHAEL HUGH HARVEY
October 1988 ProQuest Number: 10798521
All rights reserved
INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a com plete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest
ProQuest 10798521
Published by ProQuest LLC(2018). Copyright of the Dissertation is held by the Author.
All rights reserved. This work is protected against unauthorized copying under Title 17, United States C ode Microform Edition © ProQuest LLC.
ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346 CONTENTS
Page
SECTION 1 INTRODUCTION
1.1 Introduction 1
1.2 Chemistry of the Solanum Alkaloids 3
1.3 Biochemistry of the Solanum Alkaloids 10
1.4 Distribution of Glycoalkaloids Within the Potato 14
1.5 Factors Affecting the Formation of Glycoalkaloids 17
1.5.1 Light 17 1.5.2 Temperature and Storage 18 1.5.3 Mechanical Injury 20 1.5.4 Climatic Conditions 21 1.5.5 Cooking and Potato Processing 21 1.5.6 Genetic Control 23 1.5.7 Control of Tuber Glycoalkaloid Content 25
1.6 Physiological Functions of Potato Alkaloids ' , 28
1.7 Analysis of Potato Glycoalkaloids 30
1.7.1 Extraction of Glycoalkaloids from the Potato 30 1.7.2 Colorimetric Quantification 32 1.7.3 Thin-Layer Chromatography 33 1.7.4 Gas Chromatography 34 1.7.5 High Performance Liquid Chromatography 35 1.7.6 Immunoassay Methods 38 1.7.7 Other Methods of Glycoalkaloid Analysis 40
1.8 Pharmacology and Toxicology of Potato 41 Glycoalkaloids
1.8.1 Glycoalkaloid Poisoning in Man and Domestic 41 Anim als 1.8.2 Pharmacology of Potato Glycoalkaloids 44 1.8.3 Pharmacokinetics of Potato Glycoalkaloids 47 1.8.4 Toxicity of Glycoalkaloids in Experimental 49 Anim als 1.8.5 Teratogenic Effects of Potato Glycoalkaloids 52 1.8.6 Aims of the Present Study 52 Page
SECTION 2 RADIOIMMUNOASSAY OF POTATO ALKALOIDS
2.1 Introduction 54
2.2 Materials and Methods 64
2.2.1 M aterials 64 2.2.1.1 Assay Buffer 64 2.2.1.2 Immunogen 64 2.2.1.3 Antiserum 64 2.2.1.4 Label 65 2.2.1.3 Standard 65 2.2.1.6 Alkaloid-Free Serum/Urine/Saliva 66 2.2.1.7 Dextran-Coated Charcoal 66 2.2.1.8 Scintillation Vials 66 2.2.1.9 Scintillation Fluid 67
2.2.2 Methods 67 2.2.2.1 Original Radioimmunoassay Method for Serum 67 Solanidine 2.2.2.2 Modified Radioimmunoassay Method for Serum 69 Solanidine 2.2.2.3 Radioimmunoassay for Serum Total Alkaloid 71 71 2.2.2.4 Radioimmunoassay Method for Salivary Solanidine and Total Alkaloid 2.2.2.3 Radioimmunoassay Method for Urinary Solanidine 72 and Total Alkaloid 2.2.2.6 Summary of Potato Alkaloid Radioimmunoassay 72 2.2.2.7 Quality Control 78
2.3 Results 81
2.3.1 Development and Validation of the Radioimmunoassay 81 for Potato Alkaloids 2.3.1.1 Radioactive 8-Counting 81 2.3.1.2 Label 82 2.3.1.3 Antiserum 85 2.3.1.4 Establishment of Total Alkaloid Assay 96 2.3.1.5 Optimisation of Sample Volume 96 2.3.1.6 Preparation of the Standard Curve 101 2.3.1.7 Assay Incubation Times 104 2.3.1.8 Detection Limit 106 2.3.1.9 Precision 108 2.3.1.10 Label Immunoreactivity 108 2.3.1.11 Assay Recovery 108 2.3.1.12 Non-Specific Binding 113 2.3.1.13 Zero Binding 113 2.3.1.14 50% Intercept 113 Page
SECTION 3 NEURAL TUBE DEFECTS
3.1 Introduction 114
3.2 Potato Alkaloids and Neural Tube Defects 124
3.2.1 Introduction 124 3.2.2 M aterials and Methods 128 3.2.2.1 Collection of Specimens 128 3.2.2.2 Analytical Methods 129 3.2.3 Results 129 3.2.4 Discussion 140
3.3 Zinc and Neural Tube Defects 146
3.3.1 Introduction 146 3.3.2 Materials and Methods 149 3.3.2.1 Collection of Specimens 149 3.3.2.2 Analytical Methods 149
3.3.3 Results 151 3.3.4 Discussion 158
3.4 Folate and Neural Tube Defects 161
3.4.1 Introduction 161 3.4.2 Analytical Method 161 3.4.3 Results 162 3.4.4 Discussion 163
SECTION 4 ALKALOID CONCENTRATIONS IN SELECTED POPULATIONS
4.1 Alkaloid Concentrations in the Normal Population 165
4.1.1 Introduction 165 4.1.2 Materials and Methods 165 4.1.2.1 Collection of Samples 165 4.1.2.2 Analytical Methods 166
4.1.3 Results 167 4.1.4 Discussion 167
4.2 Toxicological Effects of Excessive Intake of 178 Potato Glycoalkaloids in Man
4.2.1 Introduction 178 4.2.2 Materials and Methods 178 4.2.2.1 Case History 178 4.2.2.2 Glycoalkaloid Analysis 179
4.2.3 Results 179 4.2.4 Discussion 180 Page
SECTION 5 POTATO ALKALOIDS IN HUMAN SALIVA
5 J Introduction 186
5.2 Materials and Methods 193
5.2.1 Collection of Specimens 193 5.2.2 Analytical Methods 193
5.3 Results 195
5.4 Discussion 204
SECTION 6 AFFINITY CHROMATOGRAPHY
6.1 Introduction 206
6.2 Materials and Methods 213
6.2.1 M aterials 213 6.2.2 Analytical Methods 214 6.2.2.1 Preparation of Anti-Solanum Alkaloid Immunosorbent 214 Using Cyanogen Bromide Activated Sepharose 4B 6.2.2.2 Preparation of Solanidine Free Serum, Urine and 216 Saliva 6.2.2.3 Preparation of Porous Glass and Controlled Pore 218 Ceramic Immunosorbents 6.2.2.4 A ffinity Extraction of Serum Samples and Removal 218 of Alkaloid Cross-Reactants 6.2.2.5 Scale Up of Controlled Pore Ceramic Column for 219 A ffinity Chromatography for 24 Hour Urine Samples
6.3 Results 221
6.4 Discussion 225
SECTION 7 HPLC AND TLC OF POTATO ALKALOIDS
7.1 Introduction 226
7.2 Materials and Methods 235
7.2.1 M aterials 235 7.2.2 Analytical Methods 235 7.2.2.1 Hydrolysis and TLC Separation of Potato Alkaloids 235 7.2.2.2 Spectral Analysis of Potato Alkaloids 237 7.2.2.3 HPLC of Potato Alkaloids 237 7.2.2.4 Development of HPLC Method 238
7.3 Results 243
7.3.1 Spectral Analysis 243 7.3.2 TLC of Potato Alkaloids in Serum and Urine 243 7.3.3 HPLC Analysis of Potato Alkaloids in Serum and Urine 243
7.4 Discussion 254 Page
SECTION 8 PHARMACOKINETICS OF POTATO ALKALOIDS
8.1 Introduction 256 259 8.2 Materials and Methods 259 8.2.1 M aterials 25q 8.2.2 Methods 25q 8.2.2.1 Preparation of the Potato Meal 25g 8.2.2.2 Pharmacokinetic Studies 26o 8.2.2.3 Analysis of Total Glycoalkaloid Content of Potatoes 8.2.2.4 Analysis of Alkaloid Content of Serum and Urine 261
8.3 Results 262
8.3.1 Total Glycoalkaloid Content of Potato Samples 262 8.3.2 Analysis of Serum and Urine from Pharmacokinetic 262 Study
8.4 Discussion 278
SECTION 9 GENERAL DISCUSSION AND CONCLUSIONS
9.1 Methodology 281
9.2 Alkaloid Concentrations in Normal Populations 285
9.3 Neural Tube Defects 287
9.4 Further Studies 289
ACKNOWLEDGEMENTS 291
REFERENCES 292
APPENDIX 307 SUMMARY
Radioimmunoassay methods were developed for the measurement of solanidine and total alkaloid in serum, urine and saliva.
Solanidine and total alkaloid concentrations were measured in a group of normal pregnant women and a group of women pregnant with a foetus subsequently shown to be affected by a neural tube defect (NTD). Serum alkaloid concentrations were, contrary to expectations, lower in the NTD group, casting doubt on the theory suggesting a link between potato alkaloids and NTD serum zinc and folate concentrations were also measured in the two groups and no significant differences were noted.
Solanidine and total alkaloid concentrations were measured in the serum, urine and saliva of healthy volunteers and reference ranges established.
Total alkaloid concentrations were higher than solanidine in serum and urine but similar in saliva, indicating solanidine as the predominant salivary alkaloid.
Alkaloid concentrations measured in a case of potato poisoning and in subjects on high alkaloid intake had serum, urine and salivary solanidine and total alkaloid concentrations much greater than the reference ranges.
A pharmacokinetic study was carried out to obtain body fluid alkaloid concentrations following a potato load. The glycosides were found to be rapidly absorbed before metabolism to solanidine;urinary excretion accounted for a very small percentage of the to ta l dose.
A ffinity chromatography allowed concentration of potato alkaloids in urine and serum for qualitative analysis of metabolites.
A high performance liquid chromatography (HPLC) method was developed and used in conjunction with thin layer chromatography (TLC) to separate the principle alkaloids found in urine and serum. As well as the glycosides, a- solanine and a-chaconine, the presence of the aglycone solanidine was established.
Additionally, several of the lower glycosides were found in serum and urine. ABBREVIATIONS
AFP Alpha Fetoprotein
ASB Anencephaly and Spina B ifida Cystica
BSA Bovine Serum Albumin
B/Bo Percentage counts bound in zero tube c.p.m. Counts per minute
CNS Central nervous system
CV Coefficient of variation d.p.m. Disintegration per minute
ELISA Enzyme-linked immunosorbent assay
GC Gas chromatography
GI tract Gastrointestinal tract
HPLC High performance liquid chromatography i.v. Intravenous i.p. Intraperitoneal
NSB N on-specific binding
NTD Neural tube defect ng/ml Nanograms per millilitre nmol/1 Nanomoles per litre
RIA Radioimmunoassay
SD Standard deviation
SEM Standard error of the mean
TGA Total glycoalkaloid
TA Total alkaloid
TLC Thin layer chromatography
UV Ultraviolet v/v Volume/volume FIGURES AND TABLES
Page
CHAPTER 1
Figure I The Five Main Groups of Solanum Alkaloids 4 Figure 2 The Structure of Solanidine and the a-Glycosides 6 Figure 3 Proposed Biosynthetic Pathway of Glycoalkaloid 11 Production Figured Hypothetical Formation of Solanidine from Cholesterol 12 Figure 3 Anatomical Distribution of Glycoalkaloids Within 15 the Potato Tuber Figure 6 The Spectral Response of Potato Tubers to Light of 19 -2 -1 Equal Numbers of Quanta (0.23p mol m~ , S" , 500h)
Table 1 Nutritional Content of Common Potato (Solanum 2 Tuberosum L). Table 2 Structural Relationship Among the Glycosides of 7 Solanidine Table 3 Physical Constants of the Major Potato Alkaloids 8 Table 4 Distribution of Glycoalkaloids in the Potato Plant 15 Table 3 Cases of Suspected Glycoalkaloid Poisoning in Man 43 Table 6 Toxicity of Glycoalkaloids in Experimental Animals 50
CHAPTER 2
Figure 1 Classical Immunoassay 56 Figure 2 Summary of the Assay Procedure for the Determination 73 of Solanidine Figure 3 Summary of the Assay Procedure for the Determination 74 of Total Alkaloid Figure 4 Composite Standard Curve for Solanidine 76 Figure 5 Composite Standard Curve for Total Alkaloid 77 Figure 6 Calculation of the Specific Activity of Tritiated 84 Solanidine Label Figure 7 Antiserum Dilution Curve for HP/S/RG2-1A. 87 Page
Figure 8 Displacement Curve for HP/S/RG2-1A 88 Figure 9 Parallelism Curve of Normal Serum, Urine and Saliva 92 With Solanidine Standards for Antiserum HP/S/RG2-1A Figure 10 P attern of the Immune Respnse in Soay Sheep RG14 94 Immunised with a-Solanidine-Bovine Serum Albumin Figure 11 Antiserum Dilution and Displacement Curves for 95 G/S/RG14-1A Figure 12 Parallelism Curve of Normal Serum, Urine and Saliva 98 with Solanidine Standards for Antisera G/S/RG14-1A Figure 13 Comparison of Radioimmunoassay Standard Curves in 102 Different Matrices Figure 14 Zero-Binding and Non-Specific Binding at Different 105 Incubation Times Figure 13 Charcoal Separation Studies 107 Figure 16 Precision Dose Profile for Serum, Urine and Salivary 109 Alkaloids Figure 17 Quality Control Chart for Serum Urine and Salivary 111 Alkaloids
Table 1 T itre and Displacement of Sheep Anti-Solanidine 89 Antisera HP/S/RG2 Series Table 2 Cross-Reactions of Various Steroids and Steroidal 90' Alkaloids with Anti-Solanidine Antiserum Table 3 Comparison of the T itre and Displacement of Sheep 93 Anti-Solanidine Antisera HP/S/RG2-IA, G/S/RG14-1A and G/S/RG14-2A Table 4 Cross-Reaction of Various Steroids and Steroidal 97 Alkaloids with Anti-Solanidine Antiserum Batch G/S/RG14-1A
Table 5 Comparison of Assay Parameters at Different Sample 100 Volumes for Serum and Saliva
CHAPTER 3
Figure 1 Decrease in the Incidence of Anencephaly and Spina 11.6 Bifida in England and Wales, Scotland and Northern Ireland from 1972-1982. Page
Figure 2 Anencephalus: Rates per 100,000 Total Births 1976- 117 1980. Counties and Metropolitan Districts of England and Wales Figure 3 Spina Bifida: Rates per 10,000 Total Births 1976- 118 1980. Counties and Metropolitan Districts of England and Wales Figure 4 Areas of Collection of Specimens Used in this Study 130 Figure 3 Serum Solanidine Concentrations in Pregnant Women 132 Carying Either a Normal Fetus or one with NTD Figure 6 Serum Total Alkaloid Concentrations in Pregnant 134 Women Carrying a Normal Fetus or one with NTD Figure 7 Distribution of Potato Alkaloid Concentrations in 141 Pregnant Women Figure 8 Distribution of Potato Alkaloid Concentrations in 142 Normal Non-Pregnant Women in Winter and Summer Figure 9 Standard Curve for the Estimation of Serum Zinc 152 Figure 10 Precision Dose Profile for Serum Zinc 153 Figure 11 Serum Zinc Concentrations in Pregnant Women 154 'Carrying a Normal Fetus Figure 12 Serum Zinc Concentrations in Pregnant Women 155 Carrying a Foetus with an NTD Figure 13 Zinc Concentrations Throughout Pregnancy 157
Table 1 Comparison of Serum Solanidine Concentrations on 137 Women Pregnant with a Normal Fetus and Those With a Fetus Affected by an NTD Table 2 Comparison of Serum Total Alkaloid Concentrations 138 in Women Pregnant with a Normal Foetus and Those With a Foetus Affected by an NTD Table 3 Regression Analysis Data for Solanidine and Total 139 Alkaloid on Maternal AGe, AFP Concentration and Age of Specimen Table 4 Distribution of Normal and NTD Samples Assayed for 150 Serum Zinc Table 3 Comparison of Serum Zinc Concentrations (p mol/1) in 156 Women Pregnant with a Normal Foetus and Those Affected by an NTD Page
CHAPTER 4
Figure 1 Distribution of Solanidine Concentrations in Selected 168 Groups Figure 2 Distribution of Total Alkaloid Concentrations in 169. Selected Groups Figure 3 Linear Regression Plot of Potato Intake Versus Serum 171 Solanidine Levels in Male Subjects Figure 4 Linear Regression Plot of Potato Intake Versus Serum 172 Solanidine Levels in Female Subjects
Table 1 Range of Alkaloid Concentrations in Selected 170 Populations Table 2 Comparisons of Serum Alkaloid Concentrations in 176 Swedish Subjects on a High Alkaloid Diet with Those on a Normal Diet Table 3 Urine Alkaloid Concentrations in the Poisoned 181 Subject Table 4 Urine Alkaloid Concentrations in Normal Subjects 182
CHAPTER 5
Figure 1 A Diagrammatic Representation of a Typical Salivary 187 Gland Showing Electrolyte Exchange Figure 2 Summer Solanidine and Total Alkaloid Concentrations 196 in Serum and Saliva Figure 3 Distribution of Salivary Solanidine Concentrations 198 in Summer Figure 4 Distribution of Salivery Total Alkaloid Concentrations 199 in Summer Page
Table 1 Transfer of Drugs and Steroids from Blood to Saliva 189 Table 2 Factors Causing Contamination of Saliva 192 Table 3 Collection of Saliva for Glycoalkaloid Assay 194 Table 4 Normal Range of Salivary Alkaloid Concentrations 200 in UK Subjects (in Summer) Table 3 Linear Regression Data for Serum and Salivary Solanum 201 . Alkaloid Concentrations Table 6 Comparison of Alkaloid Concentrations in Swedish 203 Subjects on High Alkaloid Diets with Those in Swedish Subjects on Normal Diets
CHAPTER 6
Figure 1 Principle of A ffinity Chromatography 208 Figure 2 Apparatus for the Extraction of 24h Urine Samples 220 Figure 3 Affinity Extraction of Tritiated Solanidine from 222 Immunosorbent Using Different Solvents Figure 4 A ffinity Extraction of Tritiated Solanidine from 224 Large Scale Controlled Pore Ceramic
CHAPTER 7
Figure 1 Arrangement of Equipment for HPLC 227 Figure 2 HPLC Retention of Radioactive Solanidine 240 Figure 3 Retention Times for a-Solanine and a-Chaconine 242 at Different Hydrogen Ion Concentrations Figure 4 Spectral Scan of Solanidine, a-solanine and 244 a-Chaconine Figure 3 TLC Analysis of Hydrolysis Products of a-solanine 245 and a-Chaconine Figure 6 TLC Analysis of Serum A ffinity Chromatography 247 E xtra ct Figure 7 TLC Analysis of Urine A ffinity Chromatography 248 E x tra c t Page
Figure 8 HPLC Analysis of a Mixture of a-Solanine, 250 a-Chaconine and Solanidine Figure 9 HPLC Analysis of Serum Analysis of 50 Minute 251 Hydrolysate Figure 10 HPLC Analysis of Serum Affinity Chromatography 252 E xtra ct Figure 11 HPLC Analysis of Urine Affinity Chromatography 253 E xtra ct
Table 1 R f Values for the Glycosides of Solanidine 246
CHAPTER 8
Figure 1 Concentrations of Solanidine and Total Alkaloid 263 Following Potato Avoidance Figure 2 Concentrations of Solanidine Over a 24h Period 267 Following the Potato Load Figure 3 Concentrations of Total Alkaloid Over a 24h Period 269 Following the Potato Load Figure 4 Concentrations of Solanidine and Total Alkaloid 271 in the 7-Day Period Following the Potato Load Figure 5 Solanidine Concentrations of a Volunteer Subject 274 on a Potato-Free D iet Figure 6 Urinary Concentrations of Solanidine and Total 275 Alkaloid Throughout the Pharmacokinetic Study
Table 1 Half-Life in Days of Solanidine and Total Alkaloid 265 Prior to 24h Pharmacokinetic Study Table 2 Half-Life of Solanidine and Total Alkaloid Prior 273 to 24h Pharmacokinetic Study Table 3 Percentage of Potato Alkaloid Dose Excreted in the 277 Urine (as Total Alkaloid Concentration) CHAPTER 1
INTRODUCTION CHAPTER 1
INTRODUCTION
1.1 INTRODUCTION
The common potato (Solanum tuberosum L.) makes a very important nutritional contribution to the diet of man. It is a very versatile foodstuff both in its natural form or, more importantly, in a processed state as found in a very wide variety of products. Its worldwide production, based on its ability to be grown under a variety of climatic conditions, ease of storage, high yield and relatively low cost, ranks it third to wheat and rice with an annual production of nearly 300m tonnes. Nutritionally, it provides an excellent source of carbohydrate energy and on a dry weight basis it has a protein content of 10% which is close to that found in wheat. It makes a significant contribution to dietary calcium, phosphorus and potassium, as well as many vitamins, notably vitamin C and thiamine. A normal serving of lOOg of boiled potatoes provides roughly one third of the daily vitamin C requirement. Table
1 illustrates the wide range of nutritionally valuable compounds found in the potato^. This important food source also contains glycosidic steroidal alkaloids which have toxicological implications in man and animals and which form the basis of the study in this thesis. Table 1
Nutritional Content of Common Potato (Solanum tuberosum L)
Figures Based on lOQq Edible Portions
Water (g) 79.8
Protein (g) 2.1
Fats (g) 0.1
Carbohydrates (g) 17.7
Energy (kcal) 76
Vitamins Elements
a + 3-carotene trace Sodium (mg) 3
Thiamine (mg) 0.11 Potassium (mg) 410
Riboflavin (mg) 0.04 Calcium (mg) 8
B6(m g) 0.25 Magnesium (mg) 24
Niacin (mg) 1.2 Iron (mg) 0.8
Free Folic Acid (yg) 9 Copper (mg) 0.16
Pantothenic Acid (mg) 0.3 Zinc (mg) 0.87
Ascorbic Acid (mg) 20 Phosphorus (mg) 53 a-Tocopherol (mg) 0.06 Sulphur (mg) 35
B iotin (yg) 0.1 Chloride (mg) 79 1.2 CHEMISTRY OF THE SOLANUM ALKALOIDS
Alkaloids are a group of naturally occurring organic compounds containing nitrogen. The name meaning 'alkali-like' derives from their basic nature and ability to form salts with acids.
Solanum steroid alkaloids have been isolated from over 350 species of the plant families Solanaceae and Liliaceae (2,3). The steroidal alkamines all possess the 27 carbon skeleton of cholestane and belong to one of five different groups (Figure 1).
(1) The solanidanes, hexacyclic tertiary bases with a fused indolizidine
moiety, eg, solanidine.
(2) The spirosolanes, eg, solasodine.
(3) 22,26 - epiminocholestanes, eg, solacongestidine.
(A) Alkaloids with an a-epiminocyclohemiketal moiety, eg
solanocapsine.
(5) The 3-aminospirostanes with a spiroketal moiety, eg, jurubidine.
Much of the interest in the Solanum steroidal alkaloids has stemmed from the importance of alkaloids in the industrial manufacture of hormonal steroids (4). In the 1960's, 75% of the world's steroid production was obtained from diosgenin, a sapogenin extracted from yams. More recently, shortage of diosgenin has led to experimentation into the use of solasodine (5,6) derived from a variety of Solanum species.
In commercially available potato tubers destined for human consumption, 95% or more of the total glycoalkaloid (TGA) fraction consists of Figure 1
The Five Main Groups of Solanum Alkaloids
The Solanidanes (Solanidine)
HO'
The Sprirosolanes (Solasodine)
The 22, 26-Epiminocholestanes (Solacongestidine)
HN
The a-Epiminocyclohemiketal Alkaloids (Solanocapsine)
h2n
The 3-Aminospirostanes (Jurubidine) a-solanine and a-chaconine (7). A normal potato contains insignificant
amounts of the glycoalkaloids; however, as discussed below, certain factors
can dramatically increase these levels. The glycosides a-solanine and a-
chaconine are derivatives of the aglycone solanidine (a member of the
solanidane group), each containing three sugar residues (Figure 2). The ratio
of a-chaconine to a-solanine is approximately 2:1 (8). Identification of the sugar residues released by partial hydrolysis has shown that a-solanine and a-
chaconine have branched sugars named 8-soltriose and B-chactriose respectively. The three sugars contained in each are rhamnose (Rha), glucose
(Glc) and galactose (Gal) for a-soianine and rhamnose (Rha), rhamnose (Rha) and glucose (Glc) for a-chaconine (9). The structural linkages are given in
Table 2, as well as those of another five minor glycosides of solanidine isolated from the potato. The existence of only one 3 form of solanine compared with two for chaconine is explained by the weakness of the a linkage between rhamnose and galactose in the soltriose.
The physical properties of the glycoalkaloids a-solanine, a- chaconine and the aglycone solanidine are given in Table 3 (10). The differences in these characteristics due to addition of sugar residues are particularly important when discussing the fate of the alkaloids in man and animals, and their degree of toxicity.
In addition to the glycosides of solanidine, a number of other similar alkaloids, members of the other groups mentioned above, have been found in the potato. The presence of these alkaloids has been due to cross-breeding of tuber species to produce varieties with such traits as cold hardiness and resistance to disease and insect pests. Figure 2.1
The Structure of Solanidine and the a-Glycosides
CH3
ch 3
Solanidine R = OH a-glycosides R = 3-soltriose or 3-chactriose (below)
Figure 2.2
Structure of the Triose Residues of a-SoIanine and a-Chaconine
0 .0 0 .0
OH OH OH OH
B-Soltriose B-Chactriose Table 2
Structural Relationships Among the Glycosides of Solanidine
Solanidine (Figure 2) R = H
a-Solanine R = O-Rha (1 2)-0-Glc-(l 3)-Gal*
3-Solanine R = 0 -G lc -(l 3)-Gal*
y-Solanine R = O -G al*
oC-Chaconine R = O-Rhad 2)-0-Rha(l 4)-Glc*
B^-Chaconine R = 0 -R h a (l 2)-G lc*
32“Chaconine R = 0-Rha(l 4)Glc*
y-Chaconine R = O -G lc*
Rha = o4-L-rhamnopyranosyl Glc = 3-D-glucospyranosyl Gal = 3-D-galactopyranosyl
* Point of attachment of sugar moiety to solanidine Table 3
Physical Constants of the Major Potato Alkaloids
Formula Mol Wt (—C) mp Solubility
SOLANIDINE ^27^43^*^ 397.6 218-219 Insoluble in water and ether. Slightly soluble in ethanol. Soluble in chloroform a-SOLANINE ^45*^73^15^ 283 Insoluble in water, ether and chloroform. Soluble in ethanol, dilute acids and amyl alcohol a-CHACONINE C45H73014N 852.1 243 As for a-solanine Tomatidenol (tomatid-5-en-3B-ol) and consequently the glycosides
a- and 3-solamarine have been found in commercially available potato tubers
(11). These represent the spirosolane group of alkaloids (Groups II above).
Other alkaloids of this group found in the potato include solasonine,
solamargine and tomatine (8,11). The presence of the solanidanes (Group I)
demissine, the leptines and commersonine in wild potato species has also led to
incorporation in commercial tubers (11, 12,13).
Solanidine glycosides are also found in a number of other edible
plant species, the aubergine (S.melongera), green and red peppers (C.annum)
(14), which must be considered as contributors of dietary alkaloid additional to
the potato. Other alkaloids are found in the tomato (Lycopersicon eculentum
M ill.) the aglycone tomatidine and its glycosides, and in the yam (Dioscorea
Spp), the alkaloid dioscorine (9). These also have toxicological implications in
man, but their consideration is outside the scope of this study, apart from
th e ir possible interference in potato alkaloid assays. 1.3 BIOCHEMISTRY OF THE SOLANUM ALKALOIDS
The biosynthesis of all steroidal compounds, including hormones, sterols, sapogenins and alkaloids is thought to occur by a common pathway.
This route from acetate via mevalonate culminating in cholesterol is shown in
Figure 3. Evidence for the existence of this pathway for the production of potato alkaloids has been provided mainly by radioactive tracer studies.
Uptake of radioactive acetate by potato sprouts produced labelling of carbon atoms in the aglycone under normal illumination)"*** Labelling of both the aglycone and sugar residues of the glycosides in the dark (13). Cholesterol has been isolated from Solanum tuberosum fed with mevalonic acid-2-^C (9) and labelled cycloartenol, lanosterol and cholesterol have all been metabolised to solanidine when applied to potato parts (16,17).
The first step from cholesterol to the alkaloids is shown to be the production of 26-hydroxycholesterol (18). Hydroxylation at the C22 P0Slti°n> addition of -NH 2 with closure of the ring led to the intermediate etioline, prior to formation of solanidine (Figure 4) (19). The source of the nitrogen atom is still debatable, but it has been postulated to come from the amino acid
L-arginine (9).
The final steps in the synthesis of the glycoalkaloids a-solanine and a-chaconine involve glycosylation of solanidine via the intermediate a and 8 forms. Potato sprouts have yielded a crude enzyme preparation catalysing the conversion of solanidine to a-chaconine by UDP- U— C glucose (20). This enzyme system is know to glycosylate a sterically unhindered 38-hydroxy group of the 5a-H or series. This and the in vitro synthesis of £f, 8 and a solanine and chaconine by potato tissue support the theory of a stepwise synthesis of the glycosides from solanidine (21). Figure 3
Proposed Biosynthetic Pathway of Glycoalkaloid Production
Acetic Acid
Mevalonic Acid
A -Isopentenyl Pyrophosphate
Farnesyl Pyrophosphate
Squalene
2,3-Oxi dos qualene
Cycloartenol
Lanosterol
Lophenol
Cholesterol
L-Arginine 26-Hydroxycholesterol
Solanidine
©C,B,Y Solanine eC,B y Chaconine Figure 4
Hypothetical Formation of Solanidine from Cholesterol
OH
Acetate MVA Cholesterol - - HO I HO Dormantinol Dormantinone HO HO t
N—' -OAc -OH
HO HO Etiolidine HO Etioline Verazlne
-OH
HO Teinemine HO Solanidine Enzymes capable of hydrolysing the glycosides in potatoes have also been demonstrated, but do not produce a simple reversal of glycosylation.
Hydrolysing enzymes rhamnosidase, glucosidase and galactosidase have been shown to be present. Solanidine has been produced from a-solanine and a-chaconine via B-solanine and B-2-chaconine without production of>f- solanine and^-chaconine (22,23). 1.4 DISTRIBUTION OF GLYCOALKALOIDS WITHIN TbE POTATO
Much discussion has occurred regarding the level of glycoalkaloids which should safely be permitted in commercially available tubers. It has been suggested that potato total glycoalkaloid content should not exceed 20mg
TGA/lOOg fresh weight of tuber. This value is not based on any toxicological considerations, but merely that tubers with high TGA content were found to taste bitter and cause a burning sensation in the throat (24,23). It was initially thought that the phenolic content of the potato might be responsible, but experiments concluded that bitterness correlated with glyco alkaloids rather than phenolics (24).
The glycoalkaloids are not found uniformly distributed throughout the tuber. The major concentrations of the glycoalkaloids are found in the outer layer of the tuber, that is, the peel and just below the skin (26). Peeling of potatoes would largely remove this alkaloid but would also destroy some of the important nutritional value by reducing the ascorbic acid content. The glycoalkaloids are found in the parenchyma cells of the periderm and cortex, in the vascular regions and in areas of high metabolic activity such as the eyes
(9). Concentrations decrease from the outside inwards with very little found in the pith (Figure 3). However, glycoalkaloids can diffuse throughout the tuber when concentrations are high (27).
Glycoalkaloids are found in most other tissues of the potato plant.
Highest concentrations are found in the sprouts, flowers and leaves (Table
4)(28). Their effects cannot be disregarded as these tissues have occasionally been eaten by humans and frequently by animals. Figure 5
Anatomical Distribution of Glycoalkaloids Within the Potato Tuber
Periderm
bud
Premedullary zone
X see below
Pith xylem
cortex
alkaloid
cortex pith branch
Enlargement of X
3 *
xylem Table 4
Distribution of Glycoalkaloids in the Potato Plant
Potato Part mg TGA/lOOg Fresh Weight of Tuber
Sprouts 200-400
Flowers 300-500
Stems 3
Leaves 40-100
Whole Tuber 1.2-20 Peel 30-60 Flesh 1.2-5.0
Bitter Tuber Peel 155-200 Flesh 25-80 1.5 FACTORS AFFECTING THE FORMATION OF GLYCOALKALOIDS
The economic importance of the potato has led to extensive
programmes to improve the yield and quality of the crop. This has led to the
discovery of a variety of factors which can influence the glycoalkaloid
content.
1.5.1 Light
When potatoes are exposed to light, a green pigmentation develops
due to increased chlorophyll levels in the outer layers. Light also has the
effect of increasing glycoalkaloid concentrations in the outer layers which can
diffuse into the flesh. Much experimental work has been carried out on this
phenomenon but comparison between experiments is difficult due to the wide
range of conditions used. It is reasonable to postulate that during growth,
harvesting, post-harvesting handling, storage and marketing, potato tubers will
be exposed to varying amounts of both natural and artificial light.
A fte r 6h of exposure of harvested potatoes to intense sunlight, total
glycoalkaloid content increased from 5 to 20mg/100g (29). Under near
freezing conditions following harvest, 'Netted Gem 1 tubers were found to have total glycoalkaloid levels as high as 45mg/100g after 72h exposure to sunlight
(30).
Several workers have investigated the effects of artificial light, though the results are often contradictory. Glycoalkaloid levels in four different varieties were found to peak after four days exposure to 75 foot candles of white fluorescent light (31). On investigating the effect of light intensity and colour, it was found that higher light intensity increased glycoalkaloid synthesis whereas coloured light (blue, green and red) had little influence (32). Early work of Connor (33) had found that the blue end of the spectrum encouraged glycoalkaloid synthesis, the red-yellow end increased chlorophyll synthesis without altering the glycoalkaloid content. This indicated that the greening of the potato tuber did not necessarily lead to a high glycoalkaloid content. In contrast, Zitnak found a significant increase in glycoalkaloid concentration following exposure of tubers to an infrared light source (13,000/5) for up to 16 days (9). Recent work (34) investigated the spectral responses for chlorophyll and glycoalkaloid synthesis using light of well defined wavelengths and equal number of quanta. The spectra were similar, with chlorophyll having maximal spectral responses at 475 and 675nm and glycoalkaloids at 430 and 650nm. The authors explained the differences with earlier work on their use of narrower band widths and a stricter experimental protocol. The spectral responses are shown in Figure 6. The conflicting data seen above needs to be reviewed after considering the many other factors influencing glycoalkaloid formation (see below). Until experiments are designed which control these other factors, no definitive data on any one influence on glycoalkaloids formation will be forthcoming.
1.5.2 Temperature and Storage
The influence of temperature both in the growing environment and in storage of potato tubers has been shown to affect glycoalkaloid concentrations. Storage of tubers at 4-8°C leads to increased glycoalkaloid concentrations compared with those of similar tubers stored at 12-15°C (12).
Using a bitterness index as a qualitative measure of glycoalkaloid concentration, tuber bitterness decreased with temperatures above 10°C (35).
In contrast when considering the effect of mechanical injury on potato tubers, control tubers stored at 21°C showed increased alkaloid synthesis compared 700 O I 600 600 (nm )
W AVELE NGTH (nm) NGTH AVELE W WAVELENGTH WAVELENGTH 500 500 chaconine c<— = □ =solanine□ ©c— a The Spectral Response of Potato Tubers to Light of Equal Numbers of Quanta (0.23ymol m~ , , S~ 500h) 400 - 400 - IQ- 5 0 - 6 0 - 3 0 - 4 0 - 20 70—t 2-Or 10 3 0
C3 - 1 >“Q O < -J -J o — Q (/) (i.o\o> U-QC Lii(/)X O I J ODCOfl.I >. J J (J.o\oE) Figure 6 with those stored at 4°C (36). In this experiment the relative humidity at both
4°C and 21°C was 60%; in other experiments humidities have not been constant with temperature.
Several authors have reported an increase in glycoalkaloid content of tubers with increased storage time (9,12). Cronk et al., (37) demonstrated an increase in total glycoalkaloid content of three different tuber varieties stored for 43 days at 10°C. Using five tuber varieties, Fitzpatrick et al., (38) showed increased glycoalkaloid content on storage which reached a maximum a fte r 12 to 18 weeks and then decreased.
1.3.3 Mechanical Injury
Potato tubers are susceptible to damage, both from pests, disease and mechanical injury during growth, harvest or post-harvesting processing.
Solanine was shown to accumulate near a healed wound on a potato tuber (39). Mechanical injury caused by slicing increased the synthesis of glycoalkaloids in peeled tubers. Salunkhe et al., (40) showed increased synthesis in potato slices after 2 days of storage in the dark at temperatures of 13°C or 24°C. The rate of formation was increased when slices were stored under high light intensity. These conditions of temperature and light are common in many potato processing plants. Wu and Salunkhe (36) demonstrated that mechanical injury which might be encountered during harvesting, treatment such as brushing, cutting, dropping, puncturing and hammering greatly increase glycoalkaloids synthesis in both the peel and flesh of the tubers. Brushing caused injury mostly in the peel while dropping had a greater effect on the flesh. Cutting the tubers resulted in the greatest glycoalkaloid accumulation. It is believed that the above mechanisms represent a physiological defence against injury. The response to damage from pests and
disease will be discussed below ( 1.6).
1.5.4 Climatic Conditions
The potato is grown all over the world, and is therefore grown under
a wide variety of climatic conditions. It has been suggested that the
glycoalkaloid content of potatoes is affected by location and climate. Sinden
and Webb (41) reported significant differences in glycoalkaloid content among
the same tuber variety when grown in 39 locations in the USA. However, this
was later attributed to abnormal growing conditions or improper handling ( 12).
Other differences originally thought due to location or climate have been
found to be caused by transport conditions, or environmental factors causing
mechanical injury such as frost or hail damage.
One factor influenced by climate which does alter potato
glycoalkaloid levels is the degree of tuber maturity. Immature tubers,
irrespective of size, tend to have higher alkaloid content than the fully mature
tubers (13). Extended day lengths, cold temperatures, short growing seasons,
and time of harvesting can all influence tuber maturity.
Small tubers tend to have higher glycoalkaloid content per unit weight, whether im m ature or not (36, 42). This is based purely on the
distribution of glycoalkaloids within the tuber (discussed above in 1.4), since smaller tubers have a higher skin and cortex to pith ratio than their larger counterparts.
1.5.5 Cooking and Potato Processing
Once the potato has been harvested, it reaches the consumer either as the unaltered tuber, ready for home cooking, or in a variety of forms from the potato processing industry. Potato processing accounts for 57% of tubers consumed in the United States (43), and represents an ever increasing array of potato products, from the easily recognisable whole potato to potato parts contained in soups and stews (44).
The effect of potato processing on the glycoalkaloid content has had limited investigation (14,43,45). Bushway and Ponnampalam (43) measured the glycoalkaloid content of 20 commercial potato products. The majority of the products contained glycoalkaloid as a-solanine and a-chaconine at less than the recommended 20mg TGA/lOOg fresh weight of tuber, although no data is available on the alkaloid content prior to processing. Products included crisps, frozen french fries, dehydrated potato, canned potato and fried peels. Highest levels were found in the fried potato peels, 139-145mg of a-solanine and a- chaconine per lOOg of product. This is obviously in conflict with the recommended lim it, though alkaloid intake will depend upon the amount of product eaten, and because of the few reported cases of potato poisoning, this can be assumed to be low if alkaloids are toxic. Potato peels have been put forward as a substitute for dietary fibre in bread since they lack the high phytate content of wheat bran (46), but the glycoalkaloid contents of the peels was not however considered. Lowest levels of glycoalkaloids were found in canned potatoes, 0.09-0.15mg/100g product. Subsequently analysis of the liquor of canned new potatoes showed an alkaloid content comparable with the tuber, indicating a leaching of the alkaloids during storage.
The effect of different methods of cooking on potato glycoalkaloid content indicate that they have little effect. Bushway and Ponnampalam (43) showed slight decreases in the a-solanine and a-chaconine content of baked, boiled and microwaved potatoes compared with the raw tubers, although no statistical analysis was attempted. Solanine decomposes at 260-270°C which is 70-80°C above normal frying temperature. Frying has been found to concentrate glycoalkaloids in potato chips and peels, possible due to water loss
(47). A later study (48) with three potato cultivars showed significantly decreased glycoalkaloid content for both baking and frying. This effect was more pronounced for cortex tissue than pith tissue. Although this decrease is seen, the reduction is not sufficient to remove the threat of toxicity from fresh tubers of high glycoalkaloid content.
1.5.6 Genetic Control
The most important factor affecting the glycoalkaloid content of the potato tuber is its genetic makeup.
Glycoalkaloid levels vary widely amongst commercial cultivars and are probably controlled by genes inherited from the cultivar's ancestors. It has been found that cultivars ..with a high mean glycoalkaloid content are more likely to produce excessive glycoalkaloid contents than are cultivars with low mean contents when subjected to less than ideal environmental conditions or improper handling. Thus, all the above mentioned factors affecting potato glycoalkaloid content are influenced by genetic makeup.
Many surveys of glycoalkaloid levels in commercially available tubers have been carried out. Glycoalkaloid contents of 32 varieties of
American tubers in 1946 ranged from 2-13mg TGA/lOOg (42). In a German survey of 58 varieties grown at 6 locations, glycoalkaloid contents ranged from
2-22mg TGA/lOOg (49). Davies and Blincow (45) collected potatoes from 6 areas of the UK, numbering 133 samples, representing 16 varieties. The mean glycoalkaloid values were 10.4mg TGA/lOOg for the main UK crop, 11.3mg TGA/lOOg for the UK eariies and 12.3mg TGA/lOOg for the imported earlies.
No differences noted were due to area, collection date or sample makeup.
A Swedish survey (50) of 165 samples of potato tubers gave alkaloid levels ranging from 2 to 30mg TGA/lOOg. Highest levels were found in new potatoes, and 9 out of 111 samples of these new potatoes had glycoalkaloid concentrations above 20mg TGA/lOOg although no cases of toxicity were reported.
One particular tuber variety, B5141-6 (Lenape) was found to have an average glycoalkaloid content of 27mg TGA/lOOg when grown over 10 locations in Canada (27). Under stress conditions at one location glycoalkaloid concentrations reached 65mg TGA/lOOg (51). This cultivar was removed from the market in 1970. The high glycoalkaloid concentrations found in the Lenape cultivar are thought to result from its ancestor Solanum chacoense, a wild species containing high glycoalkaloid levels. Concentrations in this wild variety are generally 5 to 100 times higher than those found in S.tuberosum
(13).
It was initially thought that high glycoalkaloid concentrations found in some wild species were of no consequence in breeding programmes, as it was felt that suppression of glycoalkaloid synthesis was a dominant trait, and most potato variants have at least two dominant alleles for low glycoalkaloid synthesis (52). Sinden and Webb (41) recently found glycoalkaloid content to be significantly different among parent and offpsring following ten tetrapoloid crosses over 2 years. Therefore once very high glycoalkaloid levels are introduced into a breeding programme, from a wild source, higher than normal glycoalkaloid concentrations will persist amongst some of the progeny. McCollum and Sinden (53) found evidence for both dominance and non
dominance of genes for high glycoalkaloid levels from wild species, but no
evidence of complete dominance of low levels.
It is important to consider both tuber size and analytical
methodology in these studies. Wild species generally have smaller tubers than
their commercial counterparts and potato breeders need to note this in order
to get an accurate comparison of glycoalkaloid concentrations. Wild species
such as S.demissam and S.acaule contain other types of glycoalkaloids as well
as the usual solanine and chaconine. The structures of these differing
glycoalkaloids have been discussed above ( 1.2) and thus presents the obvious
problem of analytical specificity when considering glycoalkaloid content of
cross-bred tubers. Single genes appear to control the synthesis of the various
types of sugar moieties in S.chacoense (53) and recombination of genes for the
aglycone and sugars in backcross generations can result in synthesis of
glycoalkaloids not found in either parent. The synthesis of completely new
glycoalkaloids is also a possibility with advancement of genetic and cell
biology techniques (13). This again emphasises the need for specific analysis
of new potato tuber varieties, toxicological data on the new alkaloids, and the
, need to monitor large numbers of tuber accessions in a new potato breeding
programme.
1.5.7 Control of Tuber Glycoalkaloid Content
The economical importance of the potato as a food crop has led
workers to investigate ways of preventing undesirable glycoalkaloid and
chlorophyll formation.
Many methods of prevention relate to the control factors discussed
above. Inhibition of glycoalkaloid synthesis has been investigated in both whole tubers and damaged or wounded potato parts and has involved a variety of physical and chemical treatments.
The most obvious approach is the protection of tubers from light, both pre-and post-harvest and during storage and marketing.
Recommendations have been made on conditions of light, temperature and storage to keep glycoalkaloid levels low (28). The use of coloured filters and coloured packaging has been shown to result in statistically lower tuber glycoalkaloid levels (54) but did not reduce the concentrations dramatically.
Ionising radiation reduces potato tuber sprouting, but fails to reduce glycoalkaloid content at a dose of lOKrad in illuminated tubers (55). Damaged tubers and cut potato cubes were subjected to % irradiation and produced up to 92% less glycoalkaloids with radiation doses up to 200 Krad (56). These doses were 20 times higher than those used on the tubers, did not affect existing glycoalkaloid concentrations but decreased synthesis.
Other physical treatments used have included storage in modified atmospheres (55) and at subatmospheric pressure (57) and tuber storage under water (9).
Much interest has been shown in the use of chemicals as potential inhibitors of glycoalkaloid synthesis (9). Sprouting of tubers during storage leads to increased glycoalkaloid synthesis. A sprouting inhibitor iso-propyl-N-
(3-chlorophenyl)-carbamate (CIPC) was shown to inhibit wound induced alkaloid synthesis in cut potato tubers (58). Moody and Ponnampalan demonstrated significantly decreased levels of glycoalkaloids in whole tubers of two varieties treated with CIPC (9). Indolacetic acid, a plant hormone previously shown to increase potato yield and quality, decreased both glycoalkaloid and nitrate nitrogen
content when applied to three varieties of potato plants (59). Nitrate content
of plants is important because of the potential to produce carcinogenic
n itrite s .
Wu and Salunkhe (60) used hot paraffin wax to coat potato tubers,
and found inhibition of light induced glycoalkaloid and chlorophyll synthesis when wax was heated above 140°C. Several other types of oils were studied, including corn oil, peanut oil, olive oil, vegetable oil and mineral oil all with equal effectiveness. The amount of oil used could be reduced by diluting with acetone (9,12). More recently, several commercial spray lecethins have also demonstrated inhibition of glycoalkaloid synthesis (9). Detergents are another group of compounds shown to be effective in this way (9,12). Several other chemicals, ethrel, alar, phosphon, Amchem-72-A42, nemangen and telone have also been shown to reduce glycoalkaloid synthesis (61).
Many of the treatments mentioned above would prove unsuitable or uneconomical to use for the mass protection of potato tubers. In addition, many of the experiments were carried out on limited potato cultivars, and results cannot be expected to relate to all commercial cultivars, because of genetic differences. The reduction of alkaloid concentrations in commercial tubers could be achieved alternatively by breeding tubers incapable of glycoalkaloid synthesis which would circumvent all problems discussed in the previous sections. 1.6 PHYSIOLOGICAL FUNCTIONS OF POTATO ALKALOIDS
The steroidal alkaloids found in the potato are one of a number of
compounds considered to be phytoalexins, ie, compounds produced by plants in
response to infection or injury which inhibit the development of infectious
agents. Alkaloids constitute only a minor part of these phytoalexins, which
also include phenolics and terpenoids (62). The exact mechanism of the
disease resistance and stress response is very complicated and is still not fully
elucidated.
Steroidal alkaloids are known to possess pesticidal properties against the Colorado potato beetle (Leptinotarsa decemlineata) and the potato
leafhopper (Empoasca fabae) (9,63). The resistance of potatoes to the
Colorado potato beetle was first seen in several wild species (63) and was
correlated with total leaf glycoalkaloid content. Several alkaloids not
normally found in commercial tubers conferred most resistance. Leptine I was
found to be an extremely effective feeding deterrent, completely inhibiting
feeding at a ImM concentration (63). a-solanine and a-chaconine reduced
feeding rates by 50% but at a much higher concentration of 6mM. Sinden et
al., (64) found specific foliar glycoalkaloids to be much more important than
the concentration of total glycoalkaloids. The specific alkaloids most
effective were commersonine, dehydrocommersonine and the leptines. The
importance of the leptines is that they are present in the foliage of the potato
without being in the tuber (65).
The effects of the glycoalkaloids on the potato leafhopper were
similar to those on the Colorado beetle in that leptine I and tomatine proved
much more effective than the more abundant a-solanine and the a-chaconine in conferring resistance (9, 63). Investigation of the resistance to other insects and nematodes has been extrem ely lim ite d .
The resistance of the potato to various fungal pathogens has received much attention, the conclusion being that the steroidal glycoalkaloids form part of a multi-component protective system (61). In the potato, infection with incompatible races of the blight fungus, Phytophthora infestans, causes inhibition of the synthesis of steroidal glycoalkaloids and accumulation of the fungistatic sesquiterpenoid stress metabolites, rishtin and lubimin
(61,62). Arachidonic and eicosapentanoic acids from the fungus appear to inhibit the synthesis of glycoalkaloids at the level of the conversion of farnesyl pyrophosphate to squalene. However, it has been noted that certain tuber species resistant to late blight (Lenape B5141/6) have high levels of glycoalkaloids (27). In tissue culture, P.infestans has the ability to synthesise glycoalkaloids ( 66) and hydrolyse these glycosides to solanidine (67) although the significance of this in vivo is not known.
Patel et al., ( 68 ) noted depression of the growth of the fungus
Trichoderma virida in culture with increasing concentrations of a-solanine. a- solanine and a-chaconine were highly fungitoxic to Helminthosparium carbomin (9). It was found that a-chaconine was more toxic to many fungi than a-solanine and that the protonated forms were far less inhibiting than free bases (61). 1.7 ANALYSIS OF POTATO GLYCOALKALOIDS
The extensive use of the potato as a food and the possible toxicity
of the glycoalkaloids has resulted in the need for reliable methods of
glycoalkaloid analysis. Methods need to cover two areas firstly rapid sensitive
assays for analysis of tubers destined for the market place, and secondly specific methods for identification and quantitation of any new potato
alkaloids resulting from breeding programmes. In addition, because of possible
toxicity, methods need to be developed for glycoalkaloid analysis in man, this being the aim of this research programme.
Method development has led to improved extraction of the glycoalkaloids from the potato and better qualitative and quantitative procedures. The early colorimetric, thin layer chromatography (TLC) and gas chromatography methods are now being gradually replaced by high performance liquid chromatography (HPLC) and immunoassay.
1.7.1 Extraction of Glycoalkaloids from the Potato
The first step in conventional glycoalkaloid analysis involves the extraction of the glycoalkaloids from the potato tissue and, if necessary, subsequent hydrolysis to the aglycone. The extraction is based on the fact that the major glycoalkaloids occur in the form of salts which are soluble in w ater, and (as is generally the case w ith alkaloids) can be isolated from weakly acidic plant extracts by precipitation of the free base with ammonia at high pH (9). Extraction mixtures commonly used in the past have contained aqueous acetic acid and either ethanol or methanol, with or without the addition of up to 5% acetic acid (69). The early procedures usually involved a time consuming soxhlet extraction (42,70). Sachse and Bachmann (71) eliminated the soxhlet extraction step; but their procedure needed three
refluxes with boiling alcohol.
Fitzpatrick and Osman (72) used a bisolvent extraction with a
methanol/chloroform mixture, followed by the addition of aqueous sodium
sulphate to achieve phase separation, the glycoalkaloids being subsequently
hydrolysed to the aglycone with sulphuric acid. This method though much
faster than those described above (71,72) was criticised (73) for low and
variable recoveries, due to the phase separation employed and, to a lesser
extent, because of the hydrolysis step. A modification of this method (74)
dispensed with the phase separation and hydrolysis stages, concentrated the
methanol/chloroform extract and precipitated the glycoalkaloid with
ammonia. Glycoalkaloid was redissolved from the precipitate in
tetrahydrofuran-water-acetonitrile (50:30:20 v/v). More recently, Bush way et
al., (75) have further refined this method for use with freeze-dried tubers.
Samples were extracted with a solvent system of tetrahydrofuran-water-
acetonitrile-glacial acetic acid (500:300:200:10 v/v) for lOmins in a Waring
blender. After filtration and evaporation, the glycoalkaloid was precipitated overnight with ammonia and re-dissolved in tetrahydrofuran:water:acetonitrile
(50:30:20 v/v).
The above extraction methods are still rather time consuming but were necessary because of the non-specificity of the subsequent quantitative
methods then in use which required the removal of any im purities. Coxon (69) described a rapid method of extraction using aqueous hydrolysis, but this led to loss of glycoalkaloid and therefore needed a correction for recovery.
Recently, an extraction method was described (76) for a specific quantitative method. Potato samples were chopped up, and the tissue was disrupted in liquid nitrogen. After homogenisation for 2mins in methanol:water:acetic acid
(94:6:1 v/v) small aliquots were diluted in assay buffer for analysis.
After extraction of the glycoalkaloids, many methods require hydrolysis to the aglycone, usually with strong acid. As stated above, this can lead to poor recovery, especially if the aglycone is unstable. A two-phase hydrolysis procedure has been described (77) which removes the aglycone after hydrolysis into a protective organic phase.
1.7.2 Colorimetric Quantitation
C o lo rim e tric methods have been widely used and in many cases have proved reliable. However, because of the non-specificity of the end-point reaction, their accuracy is largely dependent on the degree of purity of the extracted glycoalkaloid. A further disadvantage is that they largely employ toxic and corrosive reagents.
The three classic reagents used for colorimetric determinations are the Marquis reagent (78) (a mixture of concentrated sulphuric acid and formaldehyde), Clarke's reagent (consisting of concentrated phosphoric acid and paraformaldehyde) (79) and concentrated hydrochloric acid plus antimony trichloride, described by Wierzchowski and Wierchowski (80). The main limitation of all three reagents is that they will detect only A5-unsaturated glycoalkaloids and their use will therefore exclude several saturated glycoalkaloids such as those derived from demissidine and tomatidine, which can be present as a result of cross-breeding. Fitzpatrick and Osman (72) described a non-aqueous titration methods capable of measuring saturated glycoalkaloids. The procedure involved titration of the basic nitrogen of the aglycone with a solution of 10% phenol in methanol containing the indicator bromophenol blue which changed from blue to a yellow end-point. Pure glycoalkaloids could also be titrated, and this allowed the construction of direct standard curves.
1.7.3 Thin Layer Chromatography
Thin layer chromatography provides a useful tool for qualitative separation of glycoalkaloids. Several authors have also described quantitative methods (81,82) for glycoalkaloid analysis though they are often not as accurate as other methods of quantitation.
The majority of TLC methods have been performed on silica gel and the most useful solvent systems have been mixtures of chloroform:methanol:
1% ammonia, chloroform:ethanol: 1% ammonia and ethyl acetate:pyridine:water (69). A wide range of methods have been described which has been recently reviewed (9) and in which identification was achieved by the classic colorimetric reagents.
Zitnak (9) detected 15 substances in potato cultivars, although many were not identified. Hunter et al., (83) used eight solvent systems and determined the R. values for 26 alkaloids. Spots were detected with sulphuric acid, producing many variations in colour, and a range of mobilities were reported depending upon the solvent system used. Recently, Filadefi and
Zitnak (84) described a method of mild acid hydrolysis (terminated by ammonium hydroxide) of the triglycosides a-solanine and a-chaconine. The subsequent TLC separation used a chloroform:methanol: 1% aqueous ammonia
(2:2:1 v/v) solvent system and detection with antimony trichloride exposed the full range of di- and mono-glycosides, as well as the aglycone. Quantitation of glycoalkaloids by TLC has often proved difficult,
but several promising methods have recently been developed (69). Cadle et
al., (82) used a chloroform:95% ethanol: 1% ammonium hydroxide solvent
(2:2:1 v/v) and antimony trichloride to detect a-solanine and a-chaconine whose spots were quantified using a Kontes densitometer. Using a similar
solvent system and detection with Dragendorffs reagent, quantification has
been made with a Zeiss spectrophotometer at 380nm (83). Probably the most promising method has been the use of optical brighteners of the
stilbenedisulphonic acid type. These optical brighteners were found to be sensitive and specific for glycoalkaloids and gave fluorescing spots which were stable when kept in the dark for several months (81, 86 ).
1.7.4 Gas Chromatography
Gas chromatography (GLC) has been used to separate and quantify several different glycoalkaloids. The main disadvantage of this technique is the need to first derivatise the glycoalkaloids, and the high column temperatures used in separation with consequent short column life, a major fault when carrying out routine analyses. Herb et al., (87) produced permethylated alkaloids by reacting the extracted glycoalkaloids with dimesyl sodium and methyl iodine in dimethyl sulphoxide solution. The derivatives were chromatographed on a 3% OVI column at 330°C. This allowed separation of mixtures containing solanidine, solasodine, tomatodine, 8 -chaconine, a- chaconine, a-solamarine, B-sol amarine, demissidine, commersonine and tom atine. Osman and Sinden ( 88 ) found that the above method could not separate solanidine and demissidine glycoalkaloids containing identical sugars.
They described hydrolysis conditions which converted the A3 unsaturated solanidine-like compounds to a diene whilst the demissidine-like compounds remained unchanged and direct gas chromatography then produced a separation. A GC method for determination of solanine:chaconine ratios described by Siegfried (89) was dependent upon the glucose/galactose ratio in the alkaloids, and utilised their trimethylsilyl derivatives.
A direct method for the measurement of a-solanine and a-chaconine made use of the formation of a complex between the glycoalkaloids and stigmasterol. After the separation of free stigmasterol by a solvent extraction, the glycoalkaloid complex was measured by GLC. Quantification was achieved using standard curves prepared from pure a-solanine. Although simple and easy to carry out, use of this method was limited by its lengthy analysis time.
Several authors have described direct gas chromatography of the )\ aglycones (69, 90, 77). Recently Van Gelder (77) determined the total steroidal alkaloid composition of Solanum species by high resolution gas chromatography. This required no derivatisation and used one glass column packed with 10% SE-30 on chromasorb W HP and one WCOT-fused silica column, optimum resolution being obtained at 280°C.
The combination of gas chromatography and mass spectroscopy has helped in the elucidation of glycoalkaloid structure. Recent application of field desorption (FD), mass spectroscopy and fast atom bombardment (FAB) mass spectroscopy to underivatised alkaloids (70) have provided information both on molecular weights and structure of glycoalkaloids.
1.7.5 High Performance Liquid Chromatography
The advances in technique of high performance liquid chromatography (HPLC) over recent years has led several workers to investigate this method for glycoalkaloid analysis. Though HPLC may provide fast and accurate analysis, it again requires extraction of the alkaloid in order to prevent column contamination and preserve column life. In this respect, the use of guard columns, or solid phase sample preparation may prove helpful.
Detection of the glycoalkaloids need to be carried out at wavelengths less than
210nm to achieve maximum sensitivity which limits the spectrophotometers and solvent systems which can be used.
Hunter et al., (91) separated tomatidine, solanidine, solasodine, rubijervine, veratramine and jervine using a Porasil A column eluted with acetonezhexane (2:1) but detection of the alkaloid in the fractions by TLC proved messy and time-consuming. An improvement was made to the above system (92) by eluting with n-hexane:methanol:acetone (18:1:1) from a silica column. This allowed detection of tomatidine, tomatidenol and solasodine by their UV absorbance at 213nm. Crabbe and Fryer (93) used two types of bonded columns, pbondapak C^g and a carbohydrate column, to separate the alkaloids solasodine, solasonine, solamargine and solasodiene. The reverse phase C-^g columns used m obile phase m ixtures of methanol, w ater and acetonitrile:w ater at pH7-7.5. The normal phase carbohydrate column used methanolzisopropanol and isopropanohcyclohexane solvent mixtures and in both systems alkaloids were detected using a UV spectrophotometer at 205nm.
Limited separations of the above alkaloids could be achieved with each system tried, but no system achieved separation of all the mixtures components. The use of a carbohydrate column added the disadvantage of its susceptibility to irreversible contamination and shorter life span.
Similar chromatography columns were used by Bushway et al., (94) to separate the major glycoalkaloids extracted from potatoes. Three column types were used; a pbondapak C^g, a ybondapak NH 2 and a carbohydrate column. Solvent systems consisted of mixtures of tetrahydrofuran:water:acetonitrile and compounds were monitored at either
208, 215 or 225nm. The pbondapak C^g column was unable to distinguish a- solanine and a-chaconine, but could separate these from B-chaconine. The
NH2 column separated 8 -chaconine, a-chaconine and a-solanine. a-solanine and a-chaconine from potatoe tissue extracts were separated without interference from other potato constituents on the carbohydrate column, and could be quantified at 215nm. This methodology was further developed by
Bushway (95) to allow chromatographic determination of all the metabolites of the potato glycoalkaloids. A carbohydrate column with a solvent system of tetrahydrofuran:water:acetonitrile (55:8:37) separated a-solanine and a- chaconine and their hydrolysis products tf-chaconine, 'tf-solanine, 8 ^ and chaconine and 82 solanine. Again, the unreliability of the carbohydrate column was the m ajor drawback.
Morris and Lee (96) described the use of radially compressed cartridges of normal phase silica and reverse phase Cg and C^g to analyse potato glycoalkaloids. These cartridges are robust and provide a cheap alternative to conventional columns. The mobile phase was acetonitrile:water with the addition of small amounts of ethanolamine which improved the speed of elution. The glycoalkaloid were detected at 200nm, which providing the detection equipment can operate at this wavelength and pure solvents are used, gives much greater sensitivity. However, no single column could separate all of the minor glycoalkaloids. A similar system was used for separation of the glycoalkaloids of Eastern Black Nightshade (97) namely, a- solamargine, 8 -solamargine, B-solasonine, a-chaconine and a-solanine.
Recently, Hellenas (98) has developed a simplified procedure for determination of a-solanine and a-chaconine. This reduced sample preparation time by using
Sep-Pak C^g cartridges eluted with acetonitrile water, instead of lengthy solvent extraction. Analysis was carried out on a column packed with ODS- hypersil and using a solvent system of acetonitrile:water:ethanolamine at pH4.5.
HPLC has also been utilised to prepare milligram quantities of certain potato glycoalkaloids which can be further purified for use as standards (99).
1.7.6 Immunoassay Methods
Immunoassay methods have the potential to provide sensitive, specific, precise, rapid and relatively inexpensive methods of glycoalkaloid analysis. The specificity of the antibody used in immunoassay for glycoalkaloids minimises the sample preparation and hence removes the need for lengthy extraction procedures. The ability of immunoassays to analyse large sample numbers providaya perfect tool for mass screening of potato cultivars prior to commercial distribution. Whether there is a need to produce specific antisera for each glycoalkaloid found in solanum species is unresolved at present. Until more is known about the individual toxicity of these glycoalkaloids an antiserum of broad specificity would be of most use. The possibility of linking the separating power of HPLC with the sensitivity of immunoassays could provide a further tool for glycoalkaloid analysis.
The first immunoassay method to be described for potato glycoalkaloids was a radioimmunoassay for the aglycone solanidine ( 100).
Antiserum was raised in rabbits against a solanidine-bovine serum albumin conjugate synthesised by coupling the hemisuccinate derivative of solanidine with the protein by the mixed anhydride method. The label was tritiated
5,6(^H-dihydro)-solanidine with pure solanidine as the standard. The method had a detection limit of 150pg and correlated well with the method of
Fitzpatrick and Osman (72). The antiserum proved to be the major disadvantage of this method as it cross-reacted only 28% with a-solanine and for total glycoalkaloid analysis the glycoalkaloids needed to be hydrolysed to the aglycone. This 2h hydrolysis increased analysis time and potentiated analytical error. In addition, the antiserum titre of only 1:25 prevented its widespread distribution to other workers.
More recently an antiserum has been produced which recognises solanidine, its 3-hydroxy derivatives, a-solanine, a-chaconine, and also demmisdine (76, 101). The antiserum was produced by immunising rabbits with an a-solanine-BSA conjugate prepared by the periodate cleavage method.
This antiserum was used to develop an enzyme-linked immunosorbent assay,
ELI5A (76). Microtitre plates were coated with solanine-keyhole limpet haemocyanin, which competed with the glycoalkaloid in the sample for the antibody. Detection was achieved with a second antibody linked to alkaline phosphatase which could be visualised colorimetrically.
Sensitivity was 150pg at an antiserum dilution of 1/3,000 and 2pg at a dilution of 1/20,000. This method provided for the first time, a rapid screen for potato tuber glycoalkaloid concentration without the disadvantages of the use of radioactivity and liquid scintillation counting seen in the conventional radioimmunoassay described above. The ELISA method correlated well with both colorim etric (102, 103) and HPLC methods (98). The same antiserum was also used to develop a radioimmunoassay,
using a tritiated label, in the first analysis of a limited number of human serum samples ( 101) for solanidine concentrations.
1.7.7 Other Methods of Glycoalkaloid Analysis
Several other methods of glycoalkaloid analysis have been described
(9). Gravimetric analysis (9) and a method utilising the reducing power of
glycoalkaloid sugars after hydrolysis (9) are largely outdated. Polarographic
determination based on the adsorption of copper-phosphate-glycoalkaloid complexes (9) and paper chromatography (9) of glycoalkaloids have also been described although again these have largely been superceded by more up-to-
date methods, such as immunoassay and HPLC. 1.8 PHARMACOLOGY AND TOXICOLOGY OF
POTATO GLYCOALKALOIDS
The interest shown in the glycoalkaloids, a minor constituent of the
potato, is a reflection of their possible toxicity, since the potato represents a
major food source. The question of toxicity results from the several cases of
potato poisoning describd in man and more frequent poisoning of farm animals
(see below). In the case of the former, the potatoes were found to contain higher than usual levels of glycoalkaloids. Although, as yet, absolute levels of glycoalkaloids causing toxic symptoms in man have not been determined, and absolute proof of glycoalkaloid involvement is lacking, a great deal of research has still been carried out to investigate the pharmacology and toxicology of potato glycoalkaloids in animals.
1.8.1 Glycoalkaloid Poisoning in Man and Domestic Animals
One of the earliest reported cases of potato poisoning in man attributed to increased glycoalkaloid concentrations was the reported poisoning of 56 soldiers in Berlin in 1899 (104). Since then, over 2000 cases of potato poisoning have been documented in man (105), 30 cases resulting in death. Harris and Cockburn (106) described poisoning of 61 adults and children with symptoms of headache, vomiting and diarrhoea. Glycoalkaloid concentrations of the toxic potatoes consumed were 0.41 parts of solanine per
1000 parts of potatoes, compared with 0.079 parts of solanine in the normal tuber. The toxic potatoes in this case appeared perfectly normal. Greened potatoes were the cause of two deaths described by Hansen (107). In an outbreak of potato poisoning in Cyprus reported by W illmott (108) the leaves and sprouts rather than the tubers were.eaten on account of the shortage of other fresh vegetables. The hazard of consumption of potato parts other than the tuber is rare in man, but much more common in domestic animals. In the cases described by Wilson (109) the family of a hotel proprietor ate jacket potatoes for three successive Sundays and became ill on each occasion.
Solanine content of the potatoes was 50mg/100g. Massive outbreaks attributed to potato poisoning caused by eating rotten potatoes occurred in
North Korea in 1952-1953 (110). In one area, 382 people were affected, of whom 22 died.
The best documented case of suspected potato poisoning was that described by McMillan and Thompson (26). The authors described an outbreak affecting 78 schoolboys who became ill 4-17h after eating lunch at school.
Seventeen were admitted to hospital, of whom three were dangerously ill, with circulatory collapse, sickness began with headache, vomiting, diarrhoea and abdominal pain, sometimes with fever. Neurological syptoms were reported in many patients, including apathy, restlessness, drowsiness, mental confusion, incoherence, stupor, hallucinations, dizziness, trembling and visual disturbances. The outbreak was traced to potatoes stored from the previous term, which when peeled had a solanidine content of 33.3mg/100g. Potato left over from the meal had an excessive anticholinesterase activity in vitro equivalent to 20-30mg of solanidine per lOOg of boiled, peeled potato. A summary of the above findings is given in Table 5.
Morris and Lee (105) reviewed cases of potato poisoning and gave an estimate of the toxic and lethal doses of glycoalkaloids in man. They calculated the toxic dose to be 2-5mg/Kg and the lethal dose to range from 3-
6mg/Kg. These estimates must be viewed with caution, since many of the outbreaks occurred when only inadequate methodology for glycoalkaloid analysis was available. It is probably necessary to wait until glycoalkaloid concentrations in body fluids of normal and poisoned subjects have been measured to ascertain the true values. Table 5
Cases of Suspected Glycoalkaloid Poisoning in Man
Estimated Glycoalkaloid Symptoms Onset of Symptoms Reference Content of Potatoes
41mg/100g H,D,V,A ±h-3h 106
N K A,V,R Up to 2 days 107
38mg/100g G,F,H,N - 104
27-48.9mg/100g N,G,H,A,F,V 12h 108
50mg/100g V,A,D 9h 109
33mg/100g (peeled tubers) V,D,A,H,F,N 4-17h 26
N = Neurological symptoms G = Gastroenteritis A = Abdominal pain V = Vomiting D = Diarrhoea H = Headache R = Respiration difficulties F = Fever Treatment for potato poisoning necessitates the replacement of fluid and electrolyte loss, and in some cases, the use of anticonvulsants ( 111).
Before treatment there is the need to differentiate poisoning from bacterial; viral or other entities. This may prove difficult and in practice many mild, and some severe cases of potato poisoning would be diagnosed as gastroenteritis unless potatoes were obviously and specifically implicated.
Poisoning of domestic animals by greened, sprouted and rotten potatoes is well documented (112). The animals being fed potatoes not considered fit for human consumption. Three clinical syndromes of poisoning in cattle and pigs have been described by Volker (113): (i) a nervous form seen as stupor, depression and paresis, (ii) a gastric form characterised by salivation, vomiting and diarrhoea, and (iii) an exanthematous form. In horses, cattle, pigs and poultry, the nervous type appears to predominate ( 112).
A recent case (114) describes poisoning of horses by rotten potatoes.
Total glycoalkaloid concentrations of the rotten potatoes contained 7.3ppm compared with 5.5ppm of total glycoalkaloid in freshly greened potatoes. In this case both cattle and deer eating the rotten potatoes appeared unaffected.
Suggesting horses are more susceptible to potato poisoning. Post-mortem examination revealed little abnormality and tissues examined contained undetectable quantities of glycoalkaloids.
1.8.2 Pharmacology of Potato Glycoalkaloids
The best documented pharmacological action of potato glycoalkaloids is their anticholinesterase activity. Orgell et al., (113) demonstrated the presence of a cholinesterase inhibitor in aqueous extracts of potato tissue, the peel containing much greater concentrations than the flesh. Abbot et al., (116) correlated this anticholinesterase activity with the solanine content of the potato, this was further clarified when solanine and solanidine were shown differentially to inhibit the serum cholinesterase of individual persons of the 'usual', 'intermediate' and 'atypical' phenotypes (117).
Importantly, the inhibition of the atypical phenotype was the least, suggesting less susceptibility to solanine poisoning. Patel et al., (68) showed that solanine was a weak to moderate inhibitor of both serum and red cell cholinesterase and concluded that repeated doses of solanine would have little noticeable effect resulting from cholinesterase inhibition. In rat brain tissue, a- chaconine showed 33% inhibition of nuclear fraction cholinesterase activity,
35% inhibition of m itochondrial fractio n and 33% inhibition of microsomal fraction. Brain acetylcholinesterase activity of rats given intraperitoneal doses (10, 30,60mg/kg) o f a-chaconine was 79%, 55% and 18% o f the control group (118). Heart and plasma acetylcholinesterase activity did not show the dose-related inhibition of brain tissue. The differential inhibition was shown to be due to the differing susceptibility of isoenzyme types.
Nishie et al., (119, 120) investigated the effect of the glycoalkaloids a-solanine, a-chaconine and solanidine on a variety of in vitro systems. At doses of lO pg/m l a-solanine and a-chaconine showed a positive isotropic effect on isolated frog ventricle, the potency being equivalent to that of the structurally related cardiac glycoside K-strophanthoside. The beating of cultured rat heart cells was terminated by a-solanine, tomatine and the cardiac glycoside, K-strophanthin at doses of 80yg/ml, 40pg/ml and 160pg/ml respectively (121). The aglycone solanidine had only 1/5 of the activity of the glycosides. With plain muscle, the glycoalkaloids showed the ability to contract strips of guinea pig ileum at doses 500 times greater than those o f acetylcholine (119, 120). The Ames test was used to investigate the mutagenicity of a- solanine, with negative results (122). The infectivity of herpex simplex virus type I in tissue culture was inhibited by prior incubation with aqueous suspensions of a-chaconine (123). The aglycone solanidine and a-solanine differing only in its sugar residues from a-chaconine proved ineffective. The authors proposed insertion of the glycoalkaloid in the viral envelope as the mode of inhibition.
The to x ic ity (see below) and pharmacology of the potato glycoalkaloids have been investigated in several experimental animals (9, 12).
% In unanaesthetised rabbits, lethal doses of a-solanine (30-50mg/kg by intraperitoneal injection) produced tachycardia, tachypnea and terminal bradypnea. Death appeared to be caused by central nervous system depression with initial disappearance of the EEG followed by cessation of respiration and loss of ECG signals (119). Similar results were seen with a-chaconine (120).
Rabbits fed on greened potatoes, with an estimated intake of 64mg of total glycoalkaloid per day, showed enlarged heart and liver when compared to normals (124). Blood samples collected after 30 days showed increased cholesterol and sugar, with decreased protein. There was an increase in plasma calcium with decreased sodium and potassium. Glycogen content of the liver and kidney decreased, but heart tissue glycogen increased. The hyperglycaemic effect of the glycoalkaloids has also been reported elsewhere
(125). It is likely that the decrease in plasma Na+ and K+ is due to losses through vomiting and diarrhoea. The cause of the hyperglycaemia has not been elucidated but could be due to alterations of liver metabolism, or decreased glucose usage. Dalvi (125) has shown an increase in the liver enzymes serum glutamic oxoloacetic acid transaminase (SGOT) and serum glutamic pyruvic transaminase (SGPT) but decreased activities of acetylcholinesterase and microsomal enzymes including cytochrome P-450.
The decrease in cytochrome P-450 has also resulted in increased phenobarbital
induced sleeping time in experimental animals (125).
1.8.3 Pharmacokinetics of Potato Glycoalkaloids
The pharmacokinetics of potato glycoalkaloids has been studied in
man and experimental animals. In each case, the absorption, distribution and
excretion of the glycoalkaloids has been followed using a radioactive labelled
form of the compound in question. Nishie et al., (119) followed the fate of g tritia te d solanine, administered orally at doses of 3 x 19 ppm/kg (5mg/kg) and g intraperitoneally at doses of 1.5 x 10 ppm/kg to male Fischer rats. The
solanine was tritiated at the carbon atoms adjacent to the nitrogen atom and
the double bond. After the oral dose, 72% was lost in the faeces and 6% in the
urine in the first 24h, indicating poor gastrointestinal absorption. Total losses
of ra d io a c tiv ity in the faeces was 84% of the in itia l dose. The remaining
radioactivity reached peak tissue concentrations after 12h. The descending
order of tissue concentration in various organs was spleen, kidney, liver, lung,
fat, heart, brain and blood. After the intraperitoneal administration, faecal
and urinary loss was less than that after the oral route. At higher doses
(15mg/kg), urine and faecal excretion was dramatically decreased with
corresponding increase in tissue levels (liver, spleen, kidneys and intestines).
The major metabolite in the faeces following administration was solanidine.
Intraperitoneal injection also led to the production of solanidine as well as two
unidentified metabolites thought to be partial hydrolysis products of a- solanine
Tritiated a-chaconine administered to rats (126) showed a similar fate as a-solanine. After oral ingestion, faecal elimination was rapid (80% within 48h). Tissue levels of radioactivity were highest in the liver, intermediate in the kidney, spleen and lung,and low in the blood, brain, fat and heart. When the labelled a-chaconine was administered intraperitoneally, radioactivity was largely eliminated in the urine. Again, the metabolites formed were solanidine and two compounds of intermediate polarity.
In hamsters, tritiated a-chaconine given orally was well absorbed from the gastrointestinal tract, just less than 22% of the label being excreted via urine and faeces in 7 days (118, 127). During the initial 72h only 4% of the radioactivity was excreted in the faeces, in marked contrast with the rat studies (126). Tissue concentrations of radioactivity peaked at 12h, with highest levels in the lungs, liver, spleen, skeletal muscle, kidney and pancreas
(118). After intraperitoneal injection, tissue concentrations of radioactivity were much higher. Major excretory metabolites were identified as the aglycone, solanidine, and unchanged a-chaconine. The concentrations of other chloroform soluble metabolites were found to increase with time.
Claringbold et al., (128) investigated the disposition of tritiated solasodine, an aglycone of the spirosolane type, injected into the circulation of man and the Syrian hamster. In the hamsters, 22-30% of the dose had been excreted after 8 days. Animals were sacrificed after 14, 22 and 62 days.
Tritiated label, identified as solasodine and alkaloid conjugate or metabolites were still detectable in the liver and adipose tissue after 62 days indicating prolonged body retention. In the two human subjects, the tritium was largely associated with the erythrocytes which after 8 days contained twice as much radioactivity as the plasma. Urinary and faecal excretion accounted for 7% and 20% of the radioactivity and had virtually ceased within 5 days. The same authors (129) administered tritiated solanidine intravenously to human volunteers. The solanidine was tritiated across the
nitrogen atoms as mixed perchlorates. Again, as with tritiated solasodine, the
erythrocytes contained much greater concentrations of radioactivity than the
plasma. Three phases of plasma clearance were seen having half-lives of 2-
5mins, 120-300mins and 70-150h. Excretion rates were low; after 24h only 1-
4% of the tritium was found in the urine, 1-3% in the faeces. Solanidine has
also been detected in post-mortem liver (129).
The data described in the above experiments show major differences depending on whether the tritiated alkaloid is administered to rats (119, 126) or given to hamsters and humans (127, 128, 129). This conflict is not surprising in view of the variety of alkaloids studied, the differing modes of administration, and the varied time periods used for sampling, as well as the different species being used.
1.8.4 Toxicity of Glycoalkaloids in Experimental Animals
Laboratory tests of potato glycoalkaloid toxicity have been undertaken using a number of species of experimental animals. A variation in species susceptibility has been seen but the most striking characteristic is the marked difference in toxicity depending upon the route of administration (9).
Table 6 summarises some of the data on the toxicity of potato glycoalkaloids in experimental animals. a-Solanine given to mice at an oral dose of lOOOmg/kg proved non-toxic (119), whereas in contrast, intraperitoneal doses of a-solanine in the same species had an L D ^q of 42mg/kg in 7 days. Later experiments (120) showed LD^gS for a-solanine and a-chaconine at 32.3mg/kg and 34.5mg/kg respectively. Solanidine was non-toxic at an intraperitoneal dose of 500mg/kg (119). Patel et al., (68) found the LD^g in mice to be Table 6
Toxicity of Glycoalkaloids in Experimental Animals
Species A dm inistration Dose E ffe c t
Rat Intraperitoneal 67mg/kg S* 50% death (ip)
ip 60mg/kg TGA^ 50% death ip 84mg/kg C° 50% death oral 590mg/kg S 50% death
Mouse oral lOOOmg/kg S N on-toxic ip 42 - 1.8mg/kg S 50% death in 7 days ip 27.5 - 1.74mg/kg C 50% death
Rabbit ip 30mg/kg S Death in 6.5h ip 30mg/kg C 0/2 Death rate in 8-24h
Rhesus Monkey ip 20mg/kg/day S Death after 2nd injection ip 40mg/kg TGA Death in 48h
Sheep oral 500mg/kg Lethal iv 50mg/kg Lethal
* = a-solanine
S = Total glycoalkaloid o = a-chaconine 32.3mg/kg. Other species studied showed striking differences, a-solanine had
an in rats of 590mg/kg when administered by gastric intubation (130).
The LD^g for intraperitoneal injection in this species was 75mg/kg. Other
authors (131) found similar results with LD^g of 84mg/kg for a-solanine,
67mg/kg for a-chaconine and 60mg/kg for a total glycoalkaloid extract.
From the above it would appear that glycoalkaloids given orally are
more toxic to rats (LD^g 590mg/kg) than to mice (lOOOmg/kg non-toxic).
Toxicity of the alkaloids to rabbits has caused death within 24h of one or more
of the animals treated with a-solanine or a-chaconine at doses of 20-50mg/kg
intraperitoneally. Differences in the toxicity of a-solanine and a-chaconine
have been found by Sharma et al., (132). The LD^g for a-chaconine
(19.5mg/kg) was 2/3 of that of a-solanine (25mg/kg) in mice, a finding not seen
by Nishie et al., (120).
The differences seen in the toxicity of the glycoalkaloids result
mainly from the differing routes of administration. Oral ingestion, the most
likely route in potato poisoning, shows low toxicity for the pure alkaloids. This
could be due to the poor absorption of the pure components from the
gastrointestinal tract (102). It has been suggested (26) that other factors
might be involved in the toxicity of the alkaloids. Saponins (present in
potatoes) have a well known emulsifying action that can lead to promotion of
gastrointestinal absorption. The formation of complexes between the alkaloids
and free sterols has also been demonstrated (133) and could explain the
membrane-lytic action of the glycoalkaloids. A factor not considered in the
above experiments is the possible toxicity of the minor alkaloids such as the leptines and commersonines which have an active defence against certain insects, and the sesquiterpenoid stress metabolites possible infection inhibitors
(61,63). Finally, the pH of the contents of the gut could affect both the absorption and to x ic ity of the alkaloids due to their possession of the te rtia ry nitrogen group.
1.8.3 Teratogenic Effects of Potato Glycoalkaloids
The question of the teratogenic effects of potato alkaloids arose from a theory suggesting that the birth defects, anencephaly and spina bifida cystica (ASB) could result from women eating imperfect potatoes (134). The hypothesis'was based on the similarity of the annual occurrence of peak incidence of potato blight (caused by P.infestans) and that of full-term births of children with ASB corrected for the nine month gestation period. It was further stated that the incidence of ASB could be reduced by preventing consumption of these potatoes by pregnant women. The most likely teratogens in the potato were thought to be the steroidal alkaloids. The hypothesis has led to much research which w ill be discussed in detail in Chapter 3.
1.8.6 Aims of the Present Study
1. Development of methods for measuring both solanidine and total
alkaloid (to include a-solanine and a-chaconine and their
metabolites) in body fluids by radioimmunoassay.
2. To use the above methods to establish reference ranges for potato
alkaloids in the normal healthy populations.
3. To investigate the theory linking potato alkaloids and neural tube
defects in pregnant women. To investigate the pharmacokinetics of potato alkaloids by following changes in solanidine and total alkaloid concentrations in subjects given a potato load.
The development of HPLC and TLC methods to establish the metabolism of the alkaloids in the body from the naturally occurring forms in the potato tuber. CHAPTER 2
RADIOIMMUNOASSAY OF POTATO ALKALOIDS CHAPTER 2
RADIOIMMUNOASSAY OF POTATO ALKALOIDS
2.1 INTRODUCTION
The technique of radioimmunoassay was devised by Yalow and Berson to measure the hormone insulin (133). They described a method of analysis which could be used to determine the concentration of a very wide range of analytes in the fields of medicine, veterinary and food science. As well as this classical radioimmunoassay technique, other types of immunoassay have been developed, eg, im m unom etric assays, fluorescent immunoassays, enzyme- linked immunosorbent assays, chemiluminescent assays, as w ell as many other variations.
The major advantages offered by immunoassays over other biological techniques are sensitivity, specificity, simplicity and universal applicability.
The classical immunoassay described by Yalow and Berson (135) involves the use of a labelled form of the antigen being measured, unlabelled antigen and an antibody raised against the antigen, as well as a method of separating free and antibody-bound forms of the antigen at the end of the incubation stage. This particular immunoassay system relies on the use of a limited amount of antibody in contrast to most other immunological analytical methods which use an excess of antibody. The labelled antigen is allowed to react with a fixed amount of specific antibodies in the presence of a standard solution or sample containing the antigen. The unlabelled antigen competes with the labelled antigen for the limited number of binding sites. When equilibrium has been reached, the portion of labelled antigen which has bound to antibody is separated from that which remains free in solution and the concentration of the labelled antigen in the free or bound fraction is determined! Concentrations of antigen in the samples can be determined by interpolation from a standard curve, obtained by incubating known concentrations of the antigen being measured. Figure I gives a diagrammatic representation of this system. The more unlabelled antigen present, the less labelled antigen w ill be bound by the antibody (136).
The labelled antigen, or label, is a vital part of an immunoassay system in that it allows quantitation. In classical immunoassay, it is the antigen that is labelled and the first type of labels used were radioisotopes. Since then, many other types of labels have been used such as enzymes, fluorescent and luminescent molecules.
The salient feature of radioisotopes is their ability to emit sub-atomic particles or energy which can then be measured by an appropriate detection system. The most commonly used radioisotopes are "^1, ^^1, and ^4C.
The advantages of the iodine labels are their high specific activity and emission of gamma particles, which can be efficiently and easily detected and 125 quantified. However, the half-life of I is only 60 days, giving more disintegrations in a given time, but having a shorter shelf-life than tritium
(^H) and ^4C which have much longer half-lives, that of ^4C being 5730 years.
They are both 3-emitters which have the disadvantages of lengthy and expensive sample preparation before counting by liquid scintillation spectrophotometers. Incorporation of the radioisotope into the antigen also presents different problems for iodine nucleides compared with tritium pH) and 14C. Tritium and ^4C nucleides can be usually incorporated directly into the antigen molecule, but this requires specialised facilities. Iodination of 125 125 antigens involves oxidation of sodium I to yield elemental I. The classic procedure for radioiodination is the chloramine-T method of Greenwood et al., Figure 1
Classical Immunoassay
(with a Limited Concentration of Antibody)
Ag AgAb Ag
Ab + Ag* Ag*Ab Ag* Bound Free Fraction Fraction Ag* = labelled antigen Ag = unlabelled antigen Ab = antibody AgAb, Ag*Ab = antibody-bound antigen (137). A more recent, gentler method which had a greater chance of retaining the immunoreactivity of the antigen uses an enzyme, lactoperoxidase (138). To iodinate an antigen directly, the molecule must possess either a tyrosine or histidine residue. Molecules not possessing such moieties can still be iodinated, however, if they are covalently linked to a structure which mimics them. These compounds are known as radioiodination tags. Care must be taken not to link the tag to the antigen in the same position as that used to prepare the immunogen. If this happens, the antibody will recognise the labelled antigen more strongly than the unlabelled antigen. This is not usually a problem for and ^4C labelled antigens which have internal labels. The quality of the antibody is perhaps the most important factor in determining a sensitive immunoassay system. The two types of antibody used in immunoassay are monoclonal and polyclonal. Monoclonal antibodies have been a more recent addition to the immunoassay repertoire. They are produced by immunisation of mice with the specific antigen, subsequent removal of the spleen and fusion of the stimulated lymphocytes with myeloma cells. The hybridomas formed are cultured and clones selected for the production of specific antibodies. As the cloned hybridoma cultures are derived from a single cell, the specificity of the antibody is constant. Also the stability of frozen cells provides an almost permanent supply of antibody. Disadvantages of monoclonal antibodies are the difficulties encountered in their production, the numerous cloning steps to find the specific antibody secreting hybridoma and the fact that many monoclonal antibodies lack normal precipitating behaviour. The first classical immunoassays utilised polyclonal antibodies. These are produced in a variety of animals by induction of the immune response. This is initiated by injecting the animal with an emulsion containing the antigen! In the case of large molecules such as proteins, their size is sufficient to cause a response on their own account. Smaller molecules, such as steroids, are incapable of eliciting an immune response on their own and need to be first conjugated to a larger molecule, such as albumin. These small, non-immunogenic molecules are known as haptens. The emulsion injected also contains an adjuvant, the most common one being Freund's complete adjuvant, which consists of a mixture of mineral oil, detergent and killed mycobacteria. This adjuvant mixture aids the immune response by stimulating macrophage activity and acting as a slow release depot. The most widely used route of immunisation is subcutaneously, usually at three or four sites on the animal. After the priming injection, there is great diversity in the various immunisation schedules with regard to the timing of boosters. One approach is to leave the animal until the circulating specific antibody has fallen to a steady level. This requires continual monitoring of the antiserum bleeds. This rationale should also be used for bleeds taken after boosting (139). The choice of animal for immunisation depends upon the nature of the immunogen and the amount of antiserum needed. Sheep and goats show a good response to small haptens and give a good yield of antisera. Rabbits produce a poorer response to small haptens and give only small volumes of antisera (139). An immunisation programme produces a number of bleeds of antisera which must a ll be assessed fo r th e ir particular use in an immunoassay system. The choice of a particular antiserum depends upon three factors, its titre, avidity and specificity. The first consideration of the antiserum is its titre providing the avidity and specificity of the various antisera bleeds are the same, the bleed selected is usually the one with the highest titre. The titre of an antiserum is determined by incubating various concentrations of antibody with a fixed amount of labelled antigen, thus constructing an antiserum dilution curve. The titre is defined conventionally as that dilution of antiserum which binds 50% of the added labelled antigen. It is not necessarily the antiserum dilution which is used in the derivation of a standard curve in classical immunoassay, since dilutions which bind both higher and lower amounts of label have been used successfully in immunoassay systems. The affinity of the antibody is the energy with which its antigen combining sites bind the specific antigen. The antibody-antigen reaction obeys the law of Mass Action, such that the antibody can both combine with the antigen as well as dissociate from it: Ag + Ab AqAb The association constant, Ka, is the same as the affinity and is defined by: Kg = [*9Ab) where AgAb is the molar concentration of antigen-antibody complex and Ag , Ab the molar concentrations of the antigen and antibody respectively. The a vid ity of an antiserum can be assessed in d ire ctly by setting up two antiserum dilution curves. The first is a standard antiserum dilution curve involving the incubation of a constant amount of label with increasing dilutions of antiserum (see below), the second, in addition to labelled antigen and antiserum, also contains a constant amount of unlabelled antigen (usually the top standard) at each antiserum dilution. The degree of horizontal displacement of the second curve from the antiserum dilution curve is an indirect measure of the avidity of the antiserum, the greater the displacement, the greater the avidity. A more definite way of calculating the avidity is the use of a Scatchard plot (141). The bound/free ratio is plotted on a vertical axis and the concentration of the total bound antigen, both labelled and unlabelled forms, on the horizontal axis. The slope of the asymptope of zero antigen concentration is -K. The intercept on the horizontal axis gives the total number of binding sites. The avidity of the antiserum is particularly important as it affects the sensitivity, the association time and the choice of phase separation system o f the assay. The more avid the antiserum, the greater the potential sensitivity and shorter the incubation time. Because of the multiplicity of antigenic sites on any particular immunogen, or because of immunogen impurities etc, a large number of antibodies of differing specificity and affinity are produced. The heterogeneous nature of the antibodies in a bleed can affect the specificity needed in an immunoassay system. The final stage in assessing the specificity of an antibody is to perform a series of cross-reaction experiments. Traditionally, this is done by running a series of standard curves, one containing varying amounts of the specific antigen and the others containing varying amounts of a series of compounds which have a close structural similarity to the specific antigen. Cross-reaction is defined as the ratio of the weight of specific antigen required to reduce binding of the label at zero by 50% to the weight of the cross-reactant required to reduce binding of the label at zero by the same amount, multiplied by 100 (139). However, this is not the situation found in the analysis of samples where specific antigen and cross reactant usually occur together. The specificity can therefore also be assessed by determining the cross-reaction in the presence of low, medium and high levels of specific antigen. When known compounds related to the specific antigen are not available or when their existence is unknown, specificity can be determined by parallelism studies. This involves comparison of a standard curve of the specific antigen with a curve produced by dilution (with the same matrix as the standards) of a sample containing a high level of the specific antigen. If the two curves can be superimposed there is no significant cross- reaction from compounds related to the antigen in the sample. Specificity can also be affected by assay conditions. Increase in antibody concentration can increase the concentration of antibodies which will bind to cross-reacting compounds. In certain steroid assays, the specificity of the antibody is affected by incubation time (140). The next requirement for an immunoassay system is a supply of unlabelled antigen with which to construct a standard curve. Both the antigen for the standard and attachment of label must be in a purified form. In addition, purified antigen is needed to produce specific antibodies. An important consideration is the antigen-free matrix in which the standards are diluted. The antigen-free matrix is defined as those substances other than the analyte, which are present in the sample. The standards should be diluted in the antigen-free sample matrix, unless it has been demonstrated previously that there is no difference between standards diluted in assay buffer and antigen-free matrix. The matrix must be devoid of specific antigen and this can be achieved non-specifically by adsorption with charcoal, or specifically, by affinity chromatography. After the antigen and antibody have reached equilibrium, it is necessary to separate the free and antibody-bound antigen molecules so as to be able to quantitate the amount of labelled antigen in either or both of them. A variety of methods exist for achieving this separation. These include electrophoresis, chromatography, non-specific adsorption, chemical precipitation, diffusion filtration, immunoprecipitation and the use of solid phase reagents. The most commonly used system is immunoprecipitation, using a specific antibody to the constant, species-specific region of the first or primary antibody and which binds to the primary antibody-antigen complex in the reaction mixture. Many early immunoassay systems used coated charcoal as a separation reagent. This was used for small molecular weight antigens which when in the free form were adsorbed onto the coated charcoal. Once the labelled antigen, antibody, standards and separation system are established, the assay must be optimised. The range of standard concentrations used should reflect the reference range of the analyte in the biological system of interest. The amount of labelled antigen used should have a sufficient amount of activity to be efficiently detected, but should have a small enough concentration of antigen (usually lower than that of the lowest standard) so as not to reduce assay sensitivity. The reagents used in an immunoassay are diluted in a buffer system. It is usual to add a protein such as B5A to the buffer to prevent non-specific adsorption of antigen or antibody on the tubes or other apparatus used in transfer of reagents. The final step in the optimisation is the establishment of the time needed to reach an equilibrium. It is necessary to establish both the incubation time for the reaction between the antigen and antibody and also the incubation time for the separation step. These are both affected by temperature. The above illustrates the salient features of the classical immunoassay. Although immunoassay has been used to estimate the concentrations of a very wide variety of compounds, until recently this has not included the potato alkaloids. The first immunoassay to measure the potato alkaloids was a radioimmunoassay described by Vallejo et al., (100). This method was used to measure the total glycoalkaloid (TGA) content of potatoes. The antiserum was raised in rabbits against a solanidine-bovine serum albumin conjugate. The label was tritiated dihydrosolanidine with pure solanidine as the standard. The method had a detection lim it of 150pg. The disadvantages of this assay were the need to hydrolyse the giycoaikaloids to the aglycone prior to analysis for total glycoalkaloid and the low titre of the antiserum used, 1:25. Matthew et al., (101) produced an antiserum which recognised solanidine, a-solanine, a- chaconine and demisidine. That radioimmunoassay system used a tritiated solanidine label and was used to estimate the solanidine concentration of a limited number of hospital patients. The antiserum was also used to develop an ELISA method for TGA analysis of potato tubers (76). Assay sensitivity was 150pg at an antiserum dilution of 1/3000 and 2pg at a dilution of 1/20,000. The method provided a rapid screen for the measurement of glycoalkaloid in potato tubers. The radioimmunoassay method of Matthew (101) was developed further as a prelude to the present thesis (142). 2.2 MATERIALS AND METHODS 2.2.1 Materials All general chemicals apart from chloroform were Analar grade and were purchased from British Drug Houses, Poole, Dorset. They were stored at room temperature unless otherwise stated. Aristar chloroform (BDH) was used for solvent extraction of the glycoalkaloids prior to analysis and redistilled before use. All other chemicals are described below. 2.2.1.1 Assay Buffer Barbitone buffer, 0.075M, pH8.6. 15.4g sodium barbitone and 2.8g barbitone were dissolved in 1 litre of glass-distilled water and the pH adjusted to 8.6. Bovine serum albumin (BSA) (l.OOg) (Fraction V, BDH Chemicals Limited, Poole, Dorset, UK) was added and the buffer stored at 4°C. 2.2.1.2 Immunogen The immunogen was used to produce antiserum G/S/RG14. The solanine-bovine serum albumin conjugate was synthesised by the periodate cleavage method (143) and obtained from the AFRC Institute of Food Research. The emulsion was prepared and injected by Mr B A Morris (Department of Biochemistry, University of Surrey, Guildford, Surrey, UK). It consisted of 5mg of the solanine-BSA conjugate, dissolved in 0.65ml sterile water for injection to which 0.35ml C.parvum and 0.2ml BCG vaccine were added. This was emulsified with 2ml non-ulcerative Freund's Adjuvant (Guildhay Antisera Limited, Guildford, Surrey, UK). 2.2.1.3 Antiserum Sheep anti-solanidine antiserum (HP/S/RG2-IA) which was raised against a solanine-BSA conjugate synthesised by the periodate cleavage method was supplied by Guildhay Antisera Limited (Guildford, Surrey, UK). Sheep anti-solanidine antiserum (G/S/RG14-IA) was raised against the immunogen described above. 2.2.1.4 Label Crude tritiated 16,22”^H solanidine was a kind gift of Professor J H Renwick (London School of Hygiene and Tropical Medicine, University of London, London, UK). 2.2.1.5 Standard Purified solanidine was obtained from the AFRC Institute of Food Research, Norwich Laboratory (Norwich, Norfolk, UK). The material was stored in a domestic freezer and was dissolved in chloroform to give a 1 mg/ml primary stock solution. This was diluted 1:100 in methanol to give the secondary stock solution. Both were stored in a domestic freezer. Working standards were freshly prepared from the secondary stock solution for each assay by dilution with assay buffer to give concentrations of 0.1, 0.25, 0.5, 1.5, 2.5 and 5ng solanidine in lOOpl. Cross Reactivity Standards Rubijervine Demissidine a-solanidine a-chaconine Purified solids were obtained from the AFRC, Institute of Food Resarch. Ribujervine and demissidine were treated as for solanidine. Stock solutions of a-solanine and a-chaconine were prepared by weighing and dissolving in methanol. These solutions were stored in a domestic freezer. 4-Pregnen-17 a-ol-3,20-dione (17 a-hydroxyprogesterone) 5-Pregnen-3 $-ol-20-one (pregnenolone) 5-Androsten-3 3-ol-17-one (dehydroepiandrosterone) l,3,4(10)-estratriene-3,17 3-diol (17 a-oestradiol) 113, 17a, 21-Trihydroxypregn-4-ene-3, 20-dione (cortisol) The above steroids were purchased from Steraloids Limited, Croydon, England. Stock solutions were prepared by weighing and dissolving the steroid in methanol. These solutions were stored in a domestic freezer. All cross- reactivity standards were diluted in assay buffer from the stock solution. 2.2.1.6 Alkaloid Free Serum/Urine/Saliva Alkaloid free serum/urine/saliva was prepared by affinity chromatography using cyanogen bromide activated Sepharose (Pharmacia Fine Chemicals, Uppsala, Sweden), linked to solanidine antiserum, as described in Section 6.2.2. 2.2.1.7 Dextran-Coated Charcoal Norit A charcoal, 12.5g (Sigma Chemical Company, St Louis, USA), (fines removed before use) and 1.25g dextran T-40 (Pharmacial AB, Uppsala, Sweden) were mixed in 500ml of assay buffer and stirred on a magnetic stirrer before being stored at 4°C. It was stirred thoroughly before use and added to the assay m ixture while on a magnetic stirre r. 2.2.1.8 Scintillation Vials L.I.P. insert vials (L.I.P. Equipment and Services Limited, West Yorkshire, England). 2.2.1.9 Scintillation Fluid Ready-solv MP from Beckman RIIC Limited, High Wycombe, UK. 2.2.2 Methods 2.2.2.1 Original Radioimmunoassay Method for Serum Solanidine The method described in this thesis was originally the method of Harvey for the measurement of solanidine using a radioimmunoassay following solvent extraction of the serum samples to be analysed (142). The antiserum used was a rabbit anti-solanidine antiserum raised against a solanine-BSA conjugate at the AFRC Institute of Food Research, Norwich, Norfolk, UK. The tritiated solanidine label was the same as that used for all methods reported in this thesis and was diluted to give 2,400dpm in the total tubes. Extraction of Samples and Determination of Recovery The internal standard was prepared by diluting the secondary stock label solution 1 in 100 with barbitone bufferfV To each stoppered glass extraction tube (125 x 16mm) was added lOOpl of internal standard and 2ml aliquots of serum samples. One hundred pi aliquots of internal standard were also added to two plastic vials for liquid scintillation counting to give the total recovery counts. The extraction tubes were then stoppered and vortex mixed. The serum was extracted into 10ml of twice re-distilled Aristar chloroform with mechanical shaking (Luckham’s orbital shaker) for thirty minutes. The tubes were allowed to settle for five minutes and the serum layer was removed by vacuum suction. The solvent layer was pipetted in 4ml aliquots into disposable glass tubes (75 x 12mm) and evaporated to dryness at 60°C under oxygen-free nitrogen. Two hundred pi of barbitone buffer were added to the dried extracts. After vortex mixing, the tubes were allowed to stand for 15 minutes to ensure complete solution and then vortex mixed again. One hundred pi of this solution were transferred to plastic scintillation vials and 10ml of liquid scintillant added. This was to check the recovery of alkaloid from each sample during extraction. 10ml of scintillant were also added to the total recovery vials. All scintillation vials were vortex mixed and placed on a ICN Tracerlab Coru/Mat 2700 3-counter and counted for 20 minutes per vial. In addition, 3 x 2ml aliquots of solanidine-free sera were extracted and redissolved as above. These provided 18 x 50pl aliquots to add to standard, NSB and zero tubes. Radioimmunoassay Protocol 1. 50pl aliquots were removed from the redissolved serum extracts and placed in glass assay tubes (75 x 12mm). To these were added 50pl of barbitone buffer. 2. 50pl aliquots of the standard solution of solanidine (corresponding to 0.16, 0.31, 0.63, 1.25, 2.5, 5 and lOng in 50pl) were added in duplicate to'"glass assay tubes. To these were added 50pl of extracted solanidine-free sera. 3. To all standard and sample tubes were added 200pl of solanidine antiserum diluted 1 in 20,000 in barbitone buffer. 4. Stock solanidine was diluted 1 in 100 in barbitone buffer. One hundred pi of this solution were added to the standard tubes. Only 90pl were added to the sample tubes to compensate for that already added as recovery label. 5. A ll assay tubes were vortex mixed and incubated overnight at 4°C. 6. The separation of the antibody-bound from the free alkaloid was performed by addition of 1.0ml dextran-coated charcoal. The dextran-coated charcoal was kept in suspension by a magnetic stirrer and cooled to 4°C in an ice-water bath. All tubes were then vortex mixed. After standing for 10 minutes at 4°C, the tubes were centrifuged at 3000g for ten minutes in an MSE Mistral centrifuge, precooled to 4°C . After centrifugation, 900pl of the supernatant was removed to plastic scintillation vials and 10.0ml of scintillation fluid added. The vials were capped, vortex mixed and placed on an ICN Tracerlab Coru/Mat 2700 8-counter. The vials were allowed to stand for at least one hour and then each was counted for 20 minutes. C r . 2.2.2.2 Modified Radioimmunoassay Method for Serum Solanidine The modification of the solanidine assay was carried out firstly to reduce sample volume in order that both solanidine and total alkaloid assays could be carried out in duplicate. Secondly, the use of an antiserum which, although lower in titre than the original rabbit antiserum, came from a bleed of greater volume and therefore allowed better assay continuity over the period of the study. The reduction in the chloroform volume did not effect recovery (see Section 2.3.1.5) and did not cause an excessive emulsion. Extraction of Samples and Preparation for Assay The internal standard was prepared by diluting the stock label to give 2,400dpm per lOOpl. To each glass extraction tube were added 50pl of internal standard and 500pl of serum sample. This was extracted with 5ml of twice-redistilled chloroform. 50pl of internal standard was also added to two scintillation vials to give total recovery counts. The extraction was carried out by mechanical shaking (Luckham's orbital shaker) for 30 minutes. After extraction, the serum layer was removed by vacuum suction and 2ml aliquots of the chloroform layer were transferred to disposable glass tubes and evaporated to dryness at 60°C under a stream of oxygen-free nitrogen. The dried extracts were redissolved in 200pl of barbitone buffer (vortexed and left at room temperature for 15 minutes to ensure complete solution) and 50pl removed to plastic scintillation vials for calculation of recovery. 5ml of scintillation fluid was added to all scintillation vials. Of the remainder of the redissolved extract, lOOpl were added to glass assay tubes. In the preparation of the standards, samples of solanidine-free serum (not spiked), prepared by affinity chromatography (Section 6.2.2) were extracted as above. The dried extracts were redissolved by adding 200pl of the appropriate solanidine standard in barbitone buffer (corresponding to 0.1, 0.25, 0.5, 1.0, 1.5, 2.5 and 5ng in lOOpl). For the radioimmunoassay lOOpl of each standard extract was added to glass assay tubes and lOOpl of alkaloid-free serum extracts were added to zero and non-specific binding tubes (see below). Radioimmunoassay Protocol 1. To the lOOpl of redissolved extracts in the glass assay tubes were added 200pl of diluted antiserum (HP/S/RG2-IA) diluted 1:12,000 in barbitone buffer. 2. Stock tritiated solanidine was added to give 2,400dpm per lOOpl. lOOpl of this was added to all tubes except those containing the < samples and controls. To these tubes were added only 90pl of the label in order to compensate for the presence of recovery label in the sample extract. 3. A ll tubes were vortex mixed and incubated overnight at 4°C. 4. The separation of the free and bound fractions was achieved by addition of 1.0ml of ice-cold dextran-coated charcoal. All tubes were then vortexed. 3. After standing for 10 minutes at 4°C, the tubes were centrifuged at 3000g for ten minutes in an MSE Mistral centrifuge, pre-cooled to 4°C . 6. After centrifugation, 900pl of the supernatant were removed to plastic scintillation vials and 5.0ml of scintillation fluid added. The vials were capped, vortex-mixed and stored in the dark for two hours. The vials were afterwards counted on a LKB model 8-counter for 10 minutes. The results were corrected for the recovery. 2.2.2.3 Radioimmunoassay for Serum Total Alkaloid The procedure for the measurement of total alkaloid was essentially the same as that for solanidine (described in 2.2.2.2) but with the omission of the extraction step. Serum samples (lOOpl) were added to lOOpl of assay buffer. Solanidine standards (corresponding to 0.1, 0.25, 0.5, 1.0, 1.5, 2.5, 5ng alkaloid in 100yl) were made up in buffer and added to lOOyl of solanidine-free serum. The samples and standards were then taken through the radioimmunoassay protocol as described for solanidine. 2.2.2.4 Radioimmunoassay Method for Salivary Solanidine and Total Alkaloid Salivary solanidine was measured in a similar manner to serum solanidine (2.2.2.2). The sample volume for salivary solanidine was 2.0ml. Samples and controls were spiked with internal standard and extracted with 10ml re-distilled chloroform. Two 4ml samples of the chloroform layer were removed, dried down under a stream of oxygen-free nitrogen and reconstituted as above. Solanidine standards were made up in buffer and not in extracted solanidine-free saliva. Results were corrected for recovery. Salivary total alkaloid was estimated as for serum total glycoalkaloid (2.2.2.3) except that for the salivary assay, 200pl of samples and controls were used. Solanidine standards were made up in 200pl of assay buffer (corresponding to 0.1, 0.25, 0.5, 1.0, 2.5 and 5ng in lOOpl). The antiserum used for salivary alkaloid estimation was G/S/RG14-IA at a dilution of 1:40,000 in buffer. 2.2.2.5 Radioimmunoassay Method for Urinary Solanidine and Total A lkaloid Urinary solanidine was measured in the same way as serum solanidine (2.2.2.2). Samples of alkaloid-free urine were extracted and redissolved in solanidine standard in assay buffer. The measurement of urinary total aklaloid was as for serum total alkaloid (2.2.2.3). 2.2.2.6 Summary of Potato Alkaloid Radioimmunoassay A summary of the current assay procedures for the measurement of solanidine in serum, saliva and urine is given in Figure 2. The corresponding methods for total alkaloid are shown in Figure 3. In addition to the tubes for standards samples and quality controls, the following tubes were incorporated into each assay in duplicate. Figure 2 Summary of the Assay Procedure for the Determination of Solanidine SAMPLES EXTRACTION STANDARDS 500pl (serum/urine) 2ml (saliva) 500pl (alkaloid-free serum/urine) 5ml chloroform 10ml chloroform 5ml chloroform 50plrecovery label Extraction 30 mins Extraction 30 mins Duplicate 2ml Duplicate 4ml Duplicate 2ml Aliquots aliquots of solvent aliquots of solvent of solvent layer dried layer dried down layer dried down down. 0 o-free. N 0 0 2 "free. at 02 ~free. N 2 at 60°C 2 2 60°C 60°C Redissolved in Redissolved in 200pl 200pl of assay buffer of standard in buffer 50pl taken for lOOpl sample or standard determination of e xtra ct recovery 5ml scintillant 200pl antiserum count lOOpl H solanidine Incubate overnight 4°C ASSAY lm l D.C.C. C entrifuge 900pl supernatant 5ml scintillant Count Figure 3 Summary of the Assay Procedure for the Determination of Total Alkaloid SAMPLE STANDARD Saliva Serum/Urine Saliva Serum/Urine 200pl lOOpl lOOpl std in buffer Sample 200pl std in buffer lOOpl buffer lOOpl alkaloid-free m atrix 200pl antiserum lOOyl ^H-solanidine MIX Incubate overnight at 4°C lm l Dextran-coated charcoal Incubate at room temperature 10 mins MIX Centrifuge 3000rpm, 10 mins 900pl supernatant + 5ml scintillant COUNT Total Tube This gave a measure of the total amount of radioactivity added to each of the assay tubes. I t contained lOOpl of assay label and 1.3ml of barbitone buffer (1.4ml for salivary total alkaloid). This was mixed and 900pl transferred to the scintillation vial for counting (1.0ml for salivary total alkaloid). Non-Specific Binding Tube This tube gave an indication of the binding of the radioactivity by factors other than the antisera. It contained all the reagents except the antiserum and the sample/standard or control. Assay buffer was added in their place for salivary estimation or alkaloid-free matrix or extract for other assays. Zero Tube The unlabelled alkaloid (standard/sample or control) was omitted to determine the maximum binding of labelled antigen to antiserum. The antiserum was used at a dilution such that in the absence of unlabelled alkaloid, a 50% binding was obtained. Assay buffer, or alkaloid-free matrix/matrix extract was added in the place of unlabelled alkaloid. Calculation of Results Standard curves were plotted as counts bound, expressed as a percentage of total counts on the ordinate against solanidine or total alkaloid concentration on the abscissa. Composite standard curves for serum, urine and salivary solanidine are given in Figure 4, and standard curves for serum, urine and salivary total alkaloid are given in Figure 5. When comparing standard curves between assays (see Figures 4 and 5) results for counts bound Figure 4 Composite Standard Curve for Solanidine 9 0 - 80 - • Serum ■ Saliva 70 ▲ Urine 60 - 40 - 30 - 20 - 10- 10 Solanidine nmol/l Figure 5 Composite Standard Curve for Total Alkaloid 1 0 0 -1 90 - 80 - 70 • Serum ■ Saliva 60 ▲ Urine 50 40 30 2 0 10 Total Alkaloid nmol /I 78 were expressed as a percentage of the counts bound at zero antigen concentration B/Bo x 100 where b = counts bound in standard or sample and Bo = counts in zero tube. In calculating the results for the total alkaloid assays, the concentrations in the sample tubes were read directly from the standard curve and multiplied by the appropriate dilution factor. When calculating the results for the solanidine assays it was necessary to take the individual sample recoveries of solanidine during the solvent extraction stage into account. Sample recovery counts x 100 x 5 total recovery counts = percentage sample recovery. The factor of five is obtained from the dilution of the recovery label through the extraction phase. Final results (nmol/1) = solanidine concentration determined from standard curve multiplied by the appropriate dilution factor/percentage sample recovery. 2.2.2.7 Quality Control In the validation of any assay procedure used fo r routine sample analysis, it is necessary to establish a thorough quality control programme to monitor any variation in assay performance due to reagents or assay parameters. Several quality control procedures have been described for radioimmunoassays (136, 147, 148); these were used as the basis of the following quality control programme. Precision Precision can be defined as the variation in results from the mean obtained when the sample is repeatedly measured by a given method. The coefficient of variation (CV) was used to express the precision. This was determined both w ith in assay and between assays. The inter- and intra-C V was determined for each matrix analysed, in, serum, saliva and urine. It was determined for both solanidine and total alkaloid assays for each of these m atrices. Quality control materials for determination of precision are usually obtained from stocks of pooled samples. It is essential that the quality control material contains the analyte in the biological fluid for which the assay is intended. Samples spiked with purified analyte do not accurately reflect the conditions found in the subject samples. The quality control material used for the potato alkaloid assays could not be obtained from pooled samples as each sample collected was needed'for further analysis or confirmation of results. Serum for quality control was obtained from male blood donors. Donor blood was collected without anti-coagulant and spun down to retrieve the serum. By mixing donor samples quality control pools were obtained with analyte concentrations covering the standard curve. Urine and saliva quality control pools were obtained from volunteer subjects and again mixed to give appropriate analyte concentrations. U rine control m aterials were spun to remove any debris before use. Saliva control materials (similarly samples, see Section 5) were frozen, then after thawing, spun to break down mucins and remove any epithelial cell remnants. Sufficient quality control pools were prepared for use throughout the study. Pools were aliquoted into volumes sufficient for each assay and frozen in a domestic freezer until use. Intra-assay precision was determined by measuring the solanidine and total alkaloid concentrations of each quality control sample a total of ten times in a single assay. Inter-assay precision was determined by using a maximum of four and a minimum of two different quality control pools in each assay. Label Immunoreactivity As the label was obtained in the crude form and needed to be purified by TLC before use, it was necessary to determine the immunoreactivity of each batch o f label before routine assay use. This was established by incubating the diluted purified label with excess antiserum to give the maximum binding of the label to the antiserum. 2i3 RESULTS 2.3.1 Development and Validation of the Radioimmunoassay for Potato Alkaloids The original method for the radioimmunoassay of serum solanidine (2.2.2.1) (142) was the basis for the development of the other methods described above. The same buffer system, assay tubes and phase separation system were used in all the methods. The evaluation of these parameters is described in the author's MSc thesis (1983)(142). The following sections describe the development and evaluation of the aspects of methodology which have been used in the present study. 2.3.1.1 Radioactive 8-Countinq Before counting any samples, the 8-counter was optimised for the solanidine label in Ready-Solv MP liquid scintillant. Efficiency is defined as the ratio of the observed counts per minute (cpm) to the disintegrations per minute (dpm). The efficiency for the LKB 8- counter,using label in 9D0pl aqueous phase with 3.0ml scintillant was found to be 33.3%. The volume of s c in tilla n t used in the m odified assay (2.2.2.2) was half that used in the original solanidine assay (2.2.2.1) (142) and sig nifica ntly reduced the cost of scintillant used. This increased the ratio of aqueous to solvent phase but did not significantly reduce the efficiency of counting as a more modern machine was used. In liquid scintillation counting there is some reduction of counting efficiency due to the interference in the energy transfer process. This quenching causes a spectral shift towards zero energy and is a result of three basic phenomena (144)i The first type results from photons being absorbed by coloured constituents (colour quenching). The second (photon quenching) is caused by a non-homogeneous mixture which can reduce interaction between 8-energy and solvent and solute. Finally, certain compounds, eg, alcohols or water, compete with the solute molecules for the excited solvent molecules (fluorescence quenching). Because of the low specific activity of the tritiated solanidine and hence the long counting time involved, it was impractical to determine the quench correction for each sample. Before counting, the samples were mixed thoroughly and left to stand in order to reduce the quenching effect. Plastic scintillation vials were chosen so as to give low background counts. The small 6ml vials used also had thick walls which prevent leaching of solvent and the narrow diameter of the neck reduced evaporation. The method of preparing the sample for counting may result in the emission of photons not originating from the interaction of ionising radiation of the sample. This can result in chemiluminescence or phosphorescence. Chemiluminescence occurs when chemical energy is released as light and, in the absence of coincident counting techniques, can cause a high background. Phosphorescence results from the photo-activation of the sample/scintillant by light and can re-occur by repeated exposure to light. Both these phenomena were minimised by leaving the vials stored in the dark for two hours before counting. 2.3.1.2 Label The crude tritiated solanidine used in this research was a gift from Professor J H Renwick (London School of Hygiene and Tropical Medicine, University of London, London, UK). The solanidine was labelled by reduction of A ^ ^ and A^^-dehydrosolanidine perchlorates with tritiated sodium borohydride in methanol (129X The crude label was purified by two- dimensional TLC on 20cm x 20cm glass plates coated with silica gel (G25, Camlab, Cambridge, UK). A sample (20yl) of the crude label was applied 3cm in from the corner of the plate and chromatographed in a solvent mixture of chloroform/methanol (10:1 v/v). After drying, the plate was rotated through 90° and chromatographed in a mixture of toluene/ethyl acetate/triethylamine (20:20:1 v/v). After drying, the plate was scanned in both dimensions using a Berthold radiochromatogram TLC scanner (Lb 2723). The spot with the highest radioactivity was located. The silica from this spot was scraped from the plate and extracted with 2 x 1ml volumes of chloroform. The chloroform extracts and silica were stored in a domestic freezer until further use. Aliquots of the chloroform extract were diluted with barbitone buffer to give 2,400 dpm per lOOpl fo r use in each assay. The specific activity of the label was calculated by preparing two standard curves. The first was constructed as follows: A series of tubes were set up, each containing increasing concentrations of solanidine standard (corresponding to 0.1, 0.25, 0.5, 1.0, 2.5 and 5.0ng in lOOpl). To these were added 200gl of antiserum (HP/S/RG2-IA, 1:12000) and lOOpl of label (2,400dpm per lOOpl). The second standard curve contained ^creasing aliquots of labelled antigen, starting with tubes containing 24,000dpm per lOOyl. To these were added 200pl of antiserum (HP/S/RG2-IA, 1:12,000) and lOOpl of buffer. In addition, as series of total and NSB tubes were set up for each label concentration. After incubation, separation and counting (as for methods in Section 2.2.2 above), a standard curve was plotted as shown in Figure 6. Superimposed on this were the percentage bound counts for the tubes containing only the increasing concentration of label. The average counts per Figure 6 Figure SO V ozcoro (/nzcoo 30 - Calculation of Specific A ctivity of ctivity A Specific of Calculation Tritiated Solanidine Label Solanidine Tritiated LAOD t e b /tu g n ALKALOID Standard • Label ng of cold solanidine was 39,000 counts per 1000 seconds. This was equivalent to 39 x 10^cps/mg. Assuming a counting efficiency of 33.3% (see above) this 7 9 represented 11.6 x 10 dps/mg = 6.96 x 10 dpm/mg. Since lpCi = 2.2 x lO^dpm. Specific Activity - 3.2mCi/mq Because of the low specific activity of the label the amount added was a compromise between the weight of antigen (which should not be more than th a t in the lowest standard (0.1 ng)) and the number of counts produced. Each total vial was counted to accumulate a minimum of 10,000 counts. The figure of 10,000 counts is required because of counting imprecision which at this value gives a CV of 1%, but with 100 counts the CV is 10% (145). 2.3.1.3 Antiserum Titre, Avidity and Specificity of Antisera HP/5/RG2-IA, C, D, E, F The antiserum used in the original method for the determination of solanidine (142) was a rabbit antiserum raised against a solanine-BSA conjugate. The species used for immunisation had obvious limitations on the volume of antisera produced. In this study, a number of other antisera were evaluated. These antisera were produced using the same batch of immunogen, but in a different species, ie, sheep. The advantages of using sheep is the greater volume o f antisera produced and the fa c t that sheep seem to respond better to haptens whereas rabbits give a good responsive to large molecular weight compounds (139). The antisera HP/S/RG2-IA, IC, ID, IE, IF were produced by the Department of Biochemistry, University of Surrey, Guildford, Surrey. Antiserum dilution curves were prepared for each bleed of antisera. Dilutions of each antiserum were prepared in barbitone buffer to give the following range (initial dilutions): 1:300, 1:100, 1:5000, 1:10,000, 1:20,000, 1:50,000, 1:100,000, 1:500,000. An assay was set up for each antiserum with lOOpl label, 200yl antiserum (at each dilution) and lOOpl of buffer. Thus unlabelled solanidine was omitted. After incubation, separation and counting, a graph was drawn for each bleed, plotting the percentage of total counts bound against antiserum dilution. The antiserum dilution curve for HP/S/RG2- IA is given in Figure 7. A line was drawn parallel to the abscissa from the point on the ordinate which represents 50% of the maximum bound label. The titre of the antiserum is given by the dilution at the intersection of this line with the antiserum dilution curve. The titre of each bleed is given in Table 1. In addition to the antiserum dilution curves, displacement curves were also obtained for each bleed. A set of tubes were prepared as for the antiserum dilution curves. A set of tPbes similar to those of the antiserum dilution curve was prepared which contained lOOpl of solanidine standard (lOng/tube) instead of lOOpl of assay buffer. This allowed a curve to be constructed which was parallel to each antiserum dilution curve, but displaced to the left of it. The degree of displacement is an indirect measure of the avidity of the antiserum, the greater the displacement, the greater the avidity. Figure 8 gives the dilution and displacement curves for HP/S/RG2-IA. The displacement of each of the other bleeds is given in Table 1. Figure 7 Antiserum Dilution Curve for HP/S/RG2-IA 8 0- 7 0 - /oV 6 0 - 5 0 - 4 0 - 3 0 - 20 IQ- 50 100 3 Antiserum Dilution X 10 Figure 8 Figure 7 Q. 3 C O 00 U)~ 3 C O O - 01rhOH /o 90 80 40 70 50 30 20 Displacement Curve for HP/S/RG2-IA for Curve Displacement nieu Dlto 10 Dilution Antiserum Dilution • Displacement ■ ._3 50 100 Table 1 Titre and Displacement of Sheep Anti-Solanidine Antiserum HP/S/RG2 Series Bleed T itre Di HP/S/RG2-IA 1:12,000 78 HP/S/RG2-IC 1:9,000 78 HP/S/RG2-ID 1:5,000 68 HP/S/RG2-IE 1:3,500 65 HP/S/RG2-IF 1:750 75 Each of the antiserum dilution curves and displacement curves was plotted on log 3 cycles x 10th, \ and 1 inch graph paper to allow a comparison of the displacement to be made. Taking into account the titre and avidity, HP/S/RG2-IA was chosen fo r the routine assays. The s p e cificity o f an assay can be defined as 'the degree to which an assay responds to substances other than that fo r which the assay was designed' (136) The extent to which a given compound can interfere is often referred to as the cross-reaction. Cross-reaction studies were carried out by preparing a series of standard curves. One set contained increasing concentrations of the specific antigen, while others contained increasing concentrations of the compounds which could possibly cross-react with the antibody. Table 2 Cross-Reaction of Various Steroids and Steroidal Alkaloids with Anti-Solanidine Antiserum, Batch HP/5/RG2-IA Steroid/Alkaloid Percentage Cross-Reaction Solanidine 100 a-solanine 100 a-chaconine 100 Demissidine 100 Rubijervine 2.0 17 a-hydroxyprogesterone <.001 Pregnenolone <.001 Dehydroepiandrosterone <.001 17 3-oestradiol <.001 Cortisol <.001 Cross reactivities measured in this study are given in Table 2. Only antiserum HP/S/RG2-IA was assessed fo r cross-reactivity. The cross- reactivities of the antiserum with main potato alkaloids solanidine, a- chaconine and a-solanine were 100%. This allowed the antiserum to be used for both the solanidine and total alkaloid assays. The cross-reactivity for demissidine was also 100% but this compound, although structurally related to solanidine, is present in the potato in amounts too small to affect solanidine or total alkaloid estimation. Rubijervine was added over a concentration range of 1 to lOOng/tube, ten times higher than the alkaloids solanidine, a-solanine and a-chaconine and had a cross-reactivity of 2.0%. All other endogenous steroids likely to be present in body fluids were added in concentrations present in serum. Every steroid tested showed cross-reactivities of less than 1%. In order to check for parallelism, a serum sample containing a high level of total alkaloid was diluted with alkaloid-free serum. After addition of label (lOOpl) and diluted antiserum (200yl), incubation and separation, these dilutions were compared with the standard curve. The results are shown in Figure 9 and illustrate parallelism of the diluted serum and the standard curve. Similar curves were produced for urine and saliva and these also exhibited parallelism, indicating negligible interference from any unknown compounds. Production and Evaluation of Anti-Solanidine Antisera from G/5/RG14 On initial measurement of the alkaloid content of saliva (Section 2.3.1.5) it was found that levels were much lower than those found in serum or urine. In order to increase the sensitivity of the immunoassay for salivary alkaloids, it was decided to try and produce another antiserum of greater a vid ity than HP/S/RG 2-IA. The sensitivity of an assay can be affected by the precision of the assay, the concentration of label and antiserum, avidity of antiserum and sample volume. The concentration of the label, as mentioned above, was limited by its specific activity. The volume of sample used must be limited by the availability of sample and the fact that the greater the sample volume, the greater change of non-specific matrix effects. For the above reasons, it was the avidity and dilution of the antiserum which was used to increase sensitivity. The animal used fo r imm unisation was a dark Soay sheep. The immunogen used was the same batch of solanine and bovine serum albumin conjugate synthesised by the periodate cleavage method (143) at the AFRC Institute of Food Research, Norwich and used for the immunisation of HP/S/RG2. The emulsion was prepared by Mr B A Morris, Department of Biochemistry, University of Surrey, Guildford. It consisted of 5mg of the Figure 9 Parallelism Curve of Normal Serum, Urine and Saliva With Solanidine Standards for Antiserum HP/S/RG2-IA 100 80 - serum urine saliva B,o DILUTED SAMPLE 50 - 40 - 30 50 70 90 110 130 ALKALOID nmol / 1 solanine-BSA conjugate dissolved in 0.65ml sterile water for injection to which 0i35ml Ciparvum and 0i2ml BCG vaccine were addedi This was emulsified with 2ml non-ulcerative Freund's adjuvant (Guildhay Antisera Limited, Guildford, Surrey, UK). After a pre-immunisation bleed, the immunogen was injected intramuscularly at six sites, 0.5ml per site. The animal was bled at 4, 7, 9, 11, 12, 16, 18 and 21 weeks after priming. The antiserum titre of each bleed was determined and once the titre had reached a steady state at 21 weeks, a booster injection was given. Large bleeds were then taken at 8 and 10 days post-boost. The antiserum titre of each bleed is shown in Figure 10. The avidity of each bleed of antiserum was assessed indirectly from antiserum dilution and displacement curves as described above. The maximum titre and displacements were found in the bleeds 8 and 10 days post-boost. Table 3 compares the titre and displacement of these two bleeds with that of Table 3 Comparison of Titre and Displacement of Sheep Anti-Solanidine Antisera HP/S/RG2-IA, G/S/RG14-IA and G/5/RG14-2A Bleed T itre Displacement (mn) HP/S/RG2-IA 1:12,000 78 G/S/RG14-IA 1:40,000 82 G/5/RG14-2A 1:28,000 62 HP/S/RG2-IA. Figure 11 gives the antiserum dilution curve and the displacement curve for the 8-day post-boost bleed G/S/RH14-IA. This was the antiserum subsequently chosen for the salivary immunoassay on the basis of avidity and titre . fl/Veeks) 6 8 10 12 14 16 18 20 2 TIME SINCE IMMUNIZATION with a-Solanine-Bovine Serum Albumin 32 28 40 36 Pattern of the Immune Response in Soay Sheep G/S/RG14 Immunised Figure 10 H H CC HI Figure 11 Figure 0L3CO0J W'tSCOO — 03 r-«- O H 90 50 - 0 4 Antiserum Dilution and Displacement Curves for G/S/RG14-IA for Curves Displacement and Dilution Antiserum nieu Dlto X 10 X Dilution Antiserum 3 100 Specificity studies were carried out on this bleed as described above for HP/S/RG2-IA. Table 4 gives the cross-reaction of the various steroids and steroidal alkaloids with the G/S/RG14-IA antiserum. Again, solanidine, a- solanine and a-chaconine exhibit 100% cross-reactivity. Demissidine also shows 100% cross-reactivity, but all other substances tested are negligible. Parallelism studies are shown in Figure 12 for saliva, as well as for urine, and serum. All curves were superimposable on the solanidine standard curve. 2.3.1.A Establishment of Total Alkaloid Assay The original immunoassay fo r serum solanidine (142) used a chloroform extraction step to remove the hydropht llic glycosides, a-solanine and a-chaconine. In order to assess the body fluid concentrations of a- solanine and a-chaconine, a direct immunoassay was developed. In this assay, since the antiserum cross-react equally well with solanidine and its glycosfdes, the extraction step was omitted and neat serum was added directly to the assay tubes to obtain a measurement of the total alkaloid concentration. The assay procedures for serum, urine and salivary total alkaloid are given above (Sections 2.2.2.3, 2.2.2.4 and 2.2.2.5). 2.3.1.5 Optimisation of Sample Volume The volumes of serum and chloroform used in the original solanidine assay were 2ml and 10ml respectively. This large serum volume proved wasteful since, in many cases, duplicate analysis of samples was impossible. In addition, serum from the same sample was also needed for total alkaloid determination. The sample volume was reduced to 500pl without any detriment to the assay. However, the chloroform volume could only be Table 4 Cross-Reaction of Various Steroids and Steroidal Alkaloids with Anti-Solanidine Antiserum, Batch G/S/RG14-IA Steroid/Alkaloid Percentage Cross-Reaction Solanidine 100 a-Solanidine 100 a-Chaconine 100 Demissidine 100 Rubijervine 1.0 17 ct-Hydroxyprogesterone <.001 Pregnenolone <.001 Dehydroepiendrosterone <.001 17 3-Oestradiol <.001 Cortisol <.001 3 12 Parallelism Curve of Normal Serum, Urine and Saliva With Solanidine Standards for Antisera G/S/RG14-IA 0 0 - 90 - 80 - 70 - A DILUTED SAMPLE 60 - 50 - 40 30 - 2 0 10 AW ALKALOID nmol/l reduced by half since at volumes below 5ml, emulsion problems were encountered during the extraction step. Data on recovery and sensitivity at different sample and chloroform volumes are given in Table 5. The effect on the standard curve was negligible, since the same weight of standard was added to the assay, but with decreasing amounts of alkaloid-free matrix improved sensitivity. For the total alkaloid assay, although serum concentrations were higher than those of solanidine, lOOyl of serum was used. This was equivalent to the 100y 1 of serum extract used in the solanidine assay. Experiments measuring random urine samples for solanidine and total alkaloid concentrations used identical sample volumes and gave concentrations similar to those found in serum, the only differences in the urine assays being the use of alkaloid-free urine for preparation of standards (discussed below, section 2.3.1.6). Similar experiments with saliva, however, gave salivary concentrations much lower than those found in serum or urine, many samples having concentrations below the lim it of detection of the assay, when using 500pl for the solanidine assay and lOOyl for the total alkaloid assay. For salivary solanidine determinations, the sample volume was increased to 2ml and extracted with 10ml chloroform (to prevent emulsion effects). For the total alkaloid assay, 200yl of sample and 200pl of standard were used instead of the lOOyl of sample and standard for serum and urine determinations. At salivary solanidine volumes of 500yl and 1.0ml many samples had concentrations less than the O.lng/tube standard. When multiplied by the dilution factors to give results in nmol/1 these still remained below the lowest standard, ie, <2.5nmol/l (0.1 x 10 = ng/ml x 2.5 = nmol/1). However at volumes of 2.0ml these salivary concentrations could be read from the curve and gave Table 5 Comparison of Assay Parameters at Different Sample Volumes for Serum and Saliva Serum Solanidine Sample Volume Chloroform Recovery (n=10) Sensitivity 2.0ml 10ml 7 7 .0 -6 .0 1.8nm ol/l 1.0ml 5ml 74.8 - 7.5 1.5nm ol/l 500y 1 5ml 75.2 - 8.3 1.3nm ol/l Salivary Solanidine Sample Volume Chloroform Recovery (n=10) Sensitivity 2.0ml 10ml 75.0 - 6.2 1.0nm ol/l 1.0ml 5 ml 74.2 - 4.8 l.ln m o l/1 500pl 5ml 74.0 - 5.9 l.Onmol/1 Salivary Total Alkaloid Sample Volume Sensitivity 200y 1 0.5nm ol/l lOOyl 0.8nm ol/l actual values! Comparison of the recoveries and sensitivity of different volumes are given in Table 5. 2.3.1.6 Preparation of the Standard Curves In any validation of a radioimmunoassay method, it is necessary to compare the curves produced by diluting standards in buffer with those obtained by diluting the standards in analyte-free matrix. The preparation of alkaloid-free serum, saliva and urine is described in section 6. If any matrix effect were present, this would be apparent to a greater extent in the total alkaloid assay than in the solanidine assay, since in this latter assay, the matrix is extracted with chloroform prior to the radioimmunoassay. Two standard curves were prepared for each matrix. In one set of standard curves, the assay tubes contained lOOyl of standard and lOOpl buffer. Those of the second set contained lOOpl of standard plus lOOyl of alkaloid-free matrix. From Figure 13, it can be seen that a matrix effect was present for both serum and urine but not for saliva. It is important to note that these curves were plotted as percentage total counts bound versus solanidine concentration. If the curves had been plotted as percentage of counts bound in the zero tube versus concentration, this effect of reduced binding would not have been noticed. Similar but reduced effects were seen with solanidine assays. For serum and urine assays, analyte-free m atrix was used to dilute standards for the measurement of both solanidine and total alkaloid (section 2.2.2). In the case of saliva, the lack of a matrix effect allowed the use of buffer for dilution of the standards. Figure 13 Comparison of Radioimmunoassay Standard Curves in Different Matrices (A) Standard curves for serum radioimmunoassay, A = standards diluted in alkaloid-free serum, ■ = standards diluted in assay buffer. (B) Standard curves for salivary radioimmunoassay, A = standards diluted in alkaloid-free saliva, ■ = standards diluted in assay buffer. SERUM SALIVA O (0 (0 hh ALKALOID CONCENTRATION NMOL/L ALKALOID CONCENTRATION NMOL/ 2.3.1.7 Assay Incubation Times The incubation tim e of an assay depends upon the equilibration or association time of the antigen-antibody complex. In general, the larger the molecular weight of the antigen, the larger the incubation time needed to reach equilibrium. Also the more avid the antiserum being used, the quicker equilibration is reached. The equilibration time was determined by setting up a series of NSB and zero tubes in duplicate and incubating them for different time periods before charcoal phase separation. The incubation time period covered was from 10 minutes to 48 hours. The equilibration time was determined for both antisera HP/S/RG2-IA and G/S/RG14-IA bleed. No significant differences between the two antisera wer*. noticed in the NSB or zero binding values between the lh r and 48h period (Figure 14). For convenience, an overnight incubation was used fo r routine assay. The phase separation procedure used in this assay employed dextran- coated charcoal which removes the unbound, free antigen as well as that which dissociates from the antigen-antibody complex. In the case of a strongly avid antiserum there is little dissociation of the antigen-antibody complex. Conversely, with a weakly avid antiserum, the charcoal causes appreciable dissociation of the antigen-antibody complex. Therefore, depending on the avidity of the antiserum, a longer contact time with charcoal w ill remove greater amounts of labelled antigen. The degree of dissociation and hence optimum time for charcoal separation, was estimated for both antiserum HP/S/RG2-IA and G/S/RG14-IA. A series of zero and NSB tubes was set up in duplicate and incubated overnight. The following morning, 44 48 36 (Hours) NSB ZERO TIME Zero-Binding and Non-Specific Binding at Different Incubation Times 45 35 40 30 50 Figure 14 O O 3 C (/) CQ03CT3 dextran-coated charcoal was added to all tubes and left in contact with the incubate for various lengths of time (0 minutes to 1 hour) before centrifugation. The results are shov/n graphically in Figure 15. Although in the 1 hour period the binding is only reduced to 90% of that seen at the zero time period, a negligible effect was seen after 10 minutes incubation and this tim e period was used in the routine assay. 2.3.1.8 Detection Limit The detection tim e of an assay is defined as the smallest quantity of the substance being measured which can be distinguished with confidence from zero concentration. The most widely used method to determine sensitivity defines it as 'the error at zero, taken as 3L standard deviationsdivided by the slope of the dose-dependent curve' (146). Sensitivity = Error at zero concentration Slope at Zero Concentration Using this formula, the mean sensitivity of the various assay described are given below: Serum Solanidine = 1.3nmol/l (0.5ng/ml) Urine Solanidine = 1.3nmol/l (0.5ng/ml) Salivary Solanidine = 1.0nmol/l (0.4ng/ml) Serum Total Alkaloid = l.Onmol/1 (0.4ng/ml) Urine Total Alkaloid = 1.0nmol/l (0.4ng/ml) Salivary Total Alkaloid = 0.5nmol/l (0.2ng/ml) Figure 15 Charcoal Separation Studies 100- c °/o / 95' C 0 u 90- N T S B 85- 0 U N D 1 80- N Z E R O 75- T U B E T" ~ r 10 20 X 5 0 TIME (mins) 2.3.1.9 Precision Figure 16 gives a precision dose profile for both inter- and intra- coefficient of variation for solanidine and total alkaloid in each biological fluid. Quality control charts were plotted for each QC pool (Figure 17), NSB, zero binding and 50% intercept and assay rejected if values fell outside - 2 SD of the mean value of these parameters. 2.3.1.10 Label Immunoreactivity Maximum binding = 86.5% - 2.5 (SD) n = 8. 2.3.1.11 Assay Recovery The recovery counts used in the solanidine extraction procedure are a measure of the extraction efficiency. The mean recovery was estimated from each biological fluid and was used to monitor the extraction in each assay. Serum solanidine extraction recovery = 75.6% - 8.2 (SD) n = 42 Salivary solanidine extraction recovery = 7,5.0% - 6.8 (SD) n = 42 Urine solanidine extraction recovery = 77.2% - 5.1 (SD) n = 15 The radioimmunoassay recovery was also established for serum, saliva and urine: Serum radioimmunoassay recovery = 96.8% ^ 2.6 (SD) n = 10 Salivary radioimmunoassay recovery = 101.5% - 3.2 (SD) n = 10 Urine radioimmunoassay recovery = 98.5% - 3.5 (SD) n = 10 Figure 16 Precision Dose Profile for Serum, Urine and Salivary Alkaloids ▲ = Interassay variance for serum total alkaloid A = Intrassay variance for serum total alkaloid ■ = Interassay variance for serum solanidine □ = Intrassay variance for serum solanidine • = Interassay variance for salivary total alkaloid O = Intrassay variance for salivary total alkaloid ♦ = Interassay variance for salivary solanidine O = Intrassay variance for salivary solanidine £ = Interassay variance for urine total alkaloid $ = Intrassay variance for urine total alkaloid ? = Interassay variance for urine solanidine ! = Intrassay variance for urine solanidine ALKALOID ALKALOID CONCENTRATION NMOL/L O O lu llu o —U J Z I- Ou- > < c c — <1 — OZ Figure 17 Quality Control Chart for Serum, Urine and Salivary Alkaloids X12*8* E N U 2sd- ^ Sc Domw _ A S Y Y x 8- x L X 24-7- 24-7- X 17-7- X 1*M- X 32-3- X 2- - -3 2 X X 6-5- 6-5- X sd< 2 6 « 6 x sd“ 2 2sd- 2sd 2sd- 2sd* 2sd- 2sd- 2sdi* 2sd- 2 sd-2 x x ?) en kal d Cocnrto nmol/1 oncentration C id lo a lk A Mean ) (? « • » • • « 2.3.1.12 Non Specific Binding This was checked in each assay to ensure that levels were maintained at less than 3% of the total counts. 2.3.1.13 Zero Binding This was expressed as the percentage of the total counts bound and was kept between 40 and 30%. 2.3.1.14 30% Intercept The standard solanidine concentration at which counts bound is equal to 50% of the counts bound in the zero tube: 50% intercept Serum solanidine = 25.0nmol/l - 2.6 (n = 10) Total Alkaloid = 28.0nm ol/l - 3.2 Salivary solanidine = 23.5nmol/l - 2.1 Total Alkaloid = 19.2nmol/l - 2.5 Urine Solanidine = 24.1nm ol/l - 2.6 Total Alkaloid = 25.3nmol/l - 3.0 CHAPTER 3 NEURAL TUBE DEFECTS CHAPTER 3 NEURAL TUBE DEFECTS 3.1 INTRODUCTION Neural tube defects are amongst the most severe and most common congenital malformations in man. The term 'neural tube defect' (NTD) is applied to a variety of malformations of the central nervous system which interfere with the closure of the neural tube. NTD covers anencephaly, encephalocoele, myelocoele and spina bifida, singly and in combination. In normal prenatal development, three phases occur. During the first 21 days of gestation,the foetal membranes are established and the embryonic germ layer is formed, ending with development of the intraembryonic circulation. The second embryonic phase extends into the eighth week. There is rapid growth of all systems and organs and the main features of the external body are established. The third phase, the foetal period, extends from the end of the eighth week until birth. The nervous system develops from the ectoderm, one of the three germinal layers comprising the embryonic structure at 18 days gestation. A thickened area of ectoderm, the neural plate, becomes transformed into a groove at 20 days, and by 21 days, the folds of the groove fuse at the mid-line to form the neural tube. The openings at either end of the neural tube, the anterior and posterior neuropores, close at 26 and 29 days respectively. Failure of closure of the anterior neuropore causes anencephaly and encephalocoele. A closure defect of the posterior neuropore leads to spina bifida, and meningomyocoele (149). The incidence of NTD has declined in the British Isles over the ten years from 1972-1982 (150,151)1 Some of this decrease can be attrib ute d to prenatal screening for NTD (measuring alpha-foeto protein, AFP), and subsequent early abortion, but this alone does not explain the decline. The data on births for this period for Scotland, England, Wales and Northern Ireland are given in Figure 1 (131). The incidence of anencephaly in England and Wales dropped from 14.7 per 10,000 births to 3.9 per 10,000 and of spina bifida from 18.8 to 10.4 births per 10,000 (151,152). In Scotland (153), the birth prevalence of anencephaly fell from 26 to 2 births per 10,000 and of spina bifida from 30 to 11 births per 10,000 (151). The differences in the incidence of NTD in different areas of Britain w ill be discussed below. Generally, it is agreed that NTD has a multifactorial aetiology with genetic factors rendering the foetus more susceptible to environmental influences acting in pregnancy. However, the exact aetiology still remains much of a mystery. There are several indications for the influence of genetic factors in the aetiology of NTD (154). Ethnic differences are well known. The highest rates of NTD are found in the United Kingdom, Ireland, Northern India, Egypt and Eastern Mediterranean regions. Intermediate rates are found in much of Europe and North America and the lowest rates amongst Mongolians, Negroids and Ashkenazi Jews (154,155). In the United Kingdom, the highest rates are found in the north and west of the country (Figures 2 and 3) (156). These ethnic differences can be explained only in part by genetic factors. In Boston, USA, immigrants of Irish ancestory were found to have higher rates of NTD when compared to the general population, but much lower rates than found in Ireland; suggesting the influence of environmental factors as well (157). A iue 1 Figure AFFECTED BIRTHS/ 1000 TOTAL BIRTHS 1 - Spina Bifida in England and Wales, Scotland and Scotland Wales, and England in Bifida Spina Decrease in the Incidence of Anencephaly and Anencephaly of Incidence the in Decrease otenIeadfo 1972-1982 from Ireland Northern PN BIFIDA SPINA ANENCEPHALY ENGLAND & WALES SCOTLAND N. IRELAND Figure 2 Anencephalus: Rates per 10,000 Total Births 1976-80 Counties and Metropolitan Districts of England and Wales Rates per 10,000 total births: 10.30 and over !pipiH!iii!t#KsigiiS IjijjlijiilillllplilP i i i i i KMS 100 Figure 3 Spina Bifida; Rates per 10,000 Total Births 1976-80 Counties and Metropolitan Districts of England and Wales Rates per 10,000 total births: 0 1 5 . 2 and over {fjll3.50 - 15.19 11.20 - 13.49 0 - 11.49 55* i ’ ••• Rv \ • 'M KMS 100 similar situation was seen in another American study at Washington, USA! In South Africa and Australia, Caucasian populations mainly descended from immigrants have much lower rates of NTD than their United Kingdom counterparts (153). In the Negro and Mongolian populations, the effect of migration does not seem to alter the incidence of NTD, suggesting the presence of a genetic protective factor (154,158). However, in the United Kingdom, the incidence is increased in progeny of mixed marriages (155). There appears to be a genuine sex difference in the incidence of NTD, the male to female ratio for anencephaly being about 1:3 and for spina bifida, 1:1.3 (154,155). The ratio for anencephaly is reduced to almost unity for Oriental populations (155). Family studies (153,154,159) have shown that the incidence of NTD is increased for siblings of affected children. This risk is higher in areas where the incidence of NTD is already elevated. There is also an increased risk of abortion amongst the siblings. When considering consanguinity, studies from'several countries showed excessive rates of NTD compared to those of non-consanguineous marriages. This effect may be partially environmental but it is likely to be essentially genetic. Twin studies (154,155) provide little useful information on the genetic predisposition of NTD. Data is relatively scarce, but there are indications that the proportion of both non-zygotic and dizygotic twins affected by NTD is low. The most widely accepted genetic hypothesis (155) is that the predisposition to develop NTD is polygenic, genes at a number of loci, with additive effects, being involved. The genetic aetiology of NTD must be viewed in context of the growing evidence of environmental influences in this condition. The above data on migrant studies indicate a large environmental factor. Although migrants from Ireland to the USA have higher incidences of NTD than the general USA population, this difference decreased with succeeding generation (157,160). In the UK, the fact that NTD is more frequent in the north and west of the country points to a socio-economic factor, since it is in the poorer areas of South Wales and Belfast where the incidence is up to three times higher than that of London (154). Studies in several areas (154) show that the rate of NTD progressively increases from social group I (professional and managerial) to social group V (unskilled labour). In Israel and Hungary, this trend is not apparent. One possible explanation of these social differences is diet which is discussed below. Maclean and Macleod (161) recently reviewed the data for seasonal variation of NTD. They found that embryos conceived in May and June had a greater chance of developing anencephaly than those conceived at other times of the year. In the case of spina bifida, the highest risk month was July. The highest risks for NTD births are at the extremes of the maternal age range. Birth order shows a decreased incidence for the second child, but an increased risk for subsequent deliveries (154). In the case of these factors, no specific causative agent has been suggested. Several attempts have been made to involve a food or drug-related environmental cause of NTD. The dietary component can broadly be divided into those due to deficiencies of an essential constituent and those due to the presence of toxic materials/compounds. Knox (162) compared the intake of certain foodstuffs with the incidence of NTD, both in geographical terms and seasonal differences. A negative (protective) association was found for cheese, meat and apples (interestingly, a good source of vitamins). Positive correlations were found for bread, cereals, ice cream, canned peas and varieties of cured and cooked meats. It was suggested that magnesium salts in canned peas and nitrates or nitrites in cured meats could be the causative factors. Other positive correlations for foodstuffs have been found between blighted potatoes and NTD (134) and between tea intake and anencephalus (163). Potatoes and NTD will be discussed in more detail in a subsequent section. Drugs are another group of compounds linked with NTD. Sodium valproate, an anti-convulsant, is teratogenic to rodents (164). A French study showed a 15-fold increase in the incidence of spina bifida in malformed children whose mothers were exposed to the drug (165). These results, however, were not confirmed by other groups (152). Aminopterin, a folate antagonist, is known to have caused NTD when it was under test as an abortifacient (155). Other drugs implicated in NTD have been the barbiturates and phenytoin (155). The link between dietary deficiencies and NTD has mainly centered around mineral and vitamin deficiencies. It is well established that various metallic ions are needed for normal development (166). Zinc (discussed in detail below) has been linked with central nervous system malformations (167) and is present at low levels in Middle Eastern populations, where anencephaly is common (168). Areas of England and Wales where water has a low mineral content show high rates of NTD (169). In recent years, the emphasis of investigations in this area has been the link between certain vitamin deficiencies and NTD. A wide range of developmental defects, including NTD, have been induced in most species of laboratory animals by feeding them diets deficient in one of the following vitamins: A, D, riboflavin, niacin, thiamin, folic acid and pantothenic acid (170). Vitamin A can be both teratogenic in high doses (171) or have a protective effect against NTD in splotch mice a breed particularly prone to NTD (172). Several species have shown congenital malformations due to folate deficiency or use of folate antagonists (173). In humans, the use of the folate antagonist, aminopterin (174), produced NTD when used as an abortifacient. Smithells et al., (175) measured first trimester levels of serum and red cell folate, white cell vitamin C, riboflavin saturation index and serum vitamin A. They found significantly higher values of these indices in social classes I and II when compared with classes III, IV and V. In six mothers who gave birth to infants with NTD, serum and red cell folate concentrations were found to be significantly lower than in controls. In a South Wales study (176), women who had produced a foetus with NTD reported consuming a poorer diet during the pregnancy than during their other normal pregnancies. In view of the above findings, two studies were carried out to investigate the effectiveness of preconceptional vitamin supplementation on the recurrence of NTD. In the first study (177,178) 550 women from five centres were divided into three groups: those receiving full, partial and no vitamin supplementation. The vitamin supplement was pregnavite Forte F, a multivitamin and iron preparation consisting of vitamins A, D, thiamine, riboflavin, pyridoxine, nicotinamide, ascorbic acid, folic acid, iron sulphate and calcium phosphate. The supplements were given three times daily for not less than 28 days before conception. Results showed a recurrence of NTD in 1 of 178 supplemented women, but 13 of 260 unsupplemented mothers. A double blind randomised control trial by Lawrence et al., (179), gave women four milligrams of folic acid or placebo not later than the first day of the menstrual cycle preceeding the planned pregnancy. None of 44 receiving supplements had a recurrent NTD, whereas they did occur in six of the 67 unsupplemented group. Both these studies have been criticised (152,180) the first because its lack of randomised treatment and inadequate control groups and the second due to the small number of participants involved. However the possible protective nature of vitamin supplementation against NTD cannot be disregarded. A further trial has now been started in which supplements are divided into groups of constituent vitamins in order to discover the active factor responsible. Minerals are used as the control placebo. Partipants for the various treatments have been chosen at random. However, the ethics of preventing mothers receiving possible protective vitamins is worrying (152). The argument over the aetiology of NTD is likely to continue for some tifYie. It is doubtful that the present vitamin trial will prove totally conclusive, and likely that further environmental factors will be implicated in the incidence of this condition. 3.2 POTATO ALKALOIDS AND NEURAL TUBE DEFECTS 3.2.1 Introduction In 1972, Renwick (134) put forward the theory suggesting that neural tube defects could result from the consumption of imperfect potatoes by pregnant women. The hypothesis suggested that certain varieties of potatoes infected by the blight fungus Phytophthera infestans could be responsible. This was based on the correlation between the incidence of NTD and the occurrence of potato blight in certain areas and seasons. In the United Kingdom, it was observed that the wet western regions are suitable for the fungus and have high rates of NTD. In the United States, the wetter and more temperate regions of New England have serious potato blight and raised incidence of NTD. In both the USA and Canada, blight and NTD decrease from the more temperate east, to the drier areas of the west. Evidence was put forward for several other countries, linking a low or high rate of NTD with a low or high incidence of potato blight. The known factors affecting NTD such as maternal age, parity and social class were also explained by the potato hypothesis. The high incidence of NTD in the lower social classes was attributed to increased intake of potatoes. It was speculated that the teratogenic substances in the potato could be the fungally synthesised coumarins, the potato steroidal alkaloids or the phytoalexins, rishitin and phytuberin. It was further stated that short-term potato avoidance would prevent 95% of NTD in the United Kingdom (181), and that this should be tested by a potato avoidance trial in pregnant women. The link between potato blight and NTD and the possible involvement of steroidal alkaloids could be explained by the influence of Phytophthora infestans on alkaloid production. It is known that the steroid alkaloids, solanine and solanidine, are produced by the fungus growing in a chemically-defined medium (67), and that the fungus could transform steroids, including solanine (68). Kuc (62) however, states that P.infestans suppresses the accumulation of steroid glycoalkaloids as does the accumulation of the phytoalexins. F urther, ris h itin and phytuberin do not appear to be stable and levels decrease in infected tubers after 96h (63). This hypothesis led to much discussion and was criticised by a number of authors (63,182,183). Data from Taiwan (183) suggested that the potato hypothesis would not explain the similar rates of NTD, yet much lower potato consumption compared with those in France and Sweden. Renwick initially explained this on the grounds of poor classification of cases of anencephaly and spina bifida (181,184). He further stated that the proportion of cases of NTD that could be prevented by potato avoidance would vary from country to country. The hypothesis was later modified (185) in light of a trial by Nevin and Merrett (186) of potato avoidance during pregnancy in women who had a previous infant with NTD. No significant difference in the recurrence of NTD between the potato-free and non-potato-free group was found. The failure to reduce the recurrence rate, let alone reduce it by 95% was explained by the short-term nature of the potato avoidance (185). The seasonal and regional relationships between potato blight and NTD was still stressed in the later paper (185). In addition, evidence was given for a year to year relationship where a phase lag of two years was shown between potato blight and NTD in England and Wales. It was suggested that a potato blighted in say, autumn 1900, and eaten after winter storage in May 1901 would produce a damaged foetus in February 1902. The hypothesis of 95% prevention of NTD by short term potato avoidance was discarded and in view of the work of Nevin and Merrett (186) it was proposed that the teratogenic substance could exist in body stores, which would need to be depleted before a beneficial affect was seen. More recently (187), the encouraging results of vitamin supplementation trials described above have led Renwick to propose that ' vitamins might be preventing the action of some dietary teratogen such as a- solanine and a-chaconine in the potato. The Renwick hypothesis led to a wide range of animal experiments using a variety of species (9). Poswillo (188) investigated the teratogenic potential of blighted potato concentrate (lOg/daily) fed to rats and marmosets. No defects were found in the foetal rats, but 4 of 11 marmoset foetuses in the experimental group showed gross cranial abnormalities. When repeated (189,190) using ordinary domestic potatoes, industry rejected potatoes and Erwinia-infected potatoes (known to be high in rishikin and phytuberin) fed to rats and marmosets, no foetal abnormalities were seen under any of the experimental conditions. However, in marmosets fed industry rejected potatoes, 6 of 11 offspring showed behaviour abnormalities. Interestingly, a common substance was found in blighted potatoes and the industry rejected potatoes but was not identified. Swinyard and Chaube (191) reported that gavaging pregnant rats with blighted potatoes (5mg/kg twice daily) on days 5- 11 or intraperitoneal injections of solanine or glycoalkaloids (5-1 Omg/kg), given on days 5-12 or 7-17 of gestation resulted only in minor foetal abnormalities but failed to produce NTD. Later work (131) showed eight injections of a-chaconine 5-20mg/kg or two injections of 40mg/kg resulted in a high level of foetal and maternal death but did not produce any neural tube defects. Similar injections of potato phenolic compounds chlorogenic and caffeic acid appeared to have very little effect. Pierro et al., (192) showed the harmful effects of a-chaconine on the developing nervous system of mice. Blighted potato concentrate fed to pregnant golden hamsters, rabbits and minature pigs produced no teratogenic effects in hamsters (193) but a low incidence of spina bifida was seen in the rabbits and one case of anencephaly in the litters of two pigs. One species, the Syrian hamster, did appear to show NTD after potato alkaloid ingestion in the studies of Keeler et al., (194) who showed that potato sprouts, and an alkaloid fraction prepared from them were teratogenic when given by gavage. More recent work by Renwick et al., (195) studied the teratogenic potential of three potato preparations on this species, firstly powdered potato sprouts, secondly, products of a dichotomising fractionation sequence of sprouts, and finally, the individual steroidal alkaloids, a-solanine and a-chaconine. Treatments were given by gavage at 7.5 days gestation, with the pure alkaloids also given at 8 days to allow for the possible rapid absorption. Neural tube defects were produced in a percentage of offspring from all treatment groups, and the incidence was significantly higher than in the control group. The number of cases with NTD was particularly high when the pure alkaloids were given on day 8 (>50% of live litters). Although these results are fairly convincing, the treatments also produced toxicity and death in a number o f dams. Secondly the doses used (>I00m g/kg) were nearly 10 times higher than in other teratology experiments, and represent doses 20 times higher than those considered toxic to humans (105). The conflict between species in the teratology of the potato alkaloids, a-solanine and a-chaconine, is not seen with some other alkaloids. The veratum alkaloids with a terminal furanopiperidine group consistently produced teratogenic defects in rats, rabbits, hamsters and sheep (196). Keeler et al!, (197) suggested that for teratogenic activity the basic nitrogen of an alkaloid should be accessible to the steroidal a face for activity. A function not seen in the solanidine alkaloids. The spirosolane solasodine, found in the solanum species, although structurally and configurationally similar to the veratum alkaloids failed to produce congenital defects in rats. Defects were produced in hamsters but at doses as high as 1200 to 1600mg/kg (198). The evidence for the role of the potato glycoalkaloids in the incidence of NTD is far from convincing. However, no actual measurements of potato alkaloids in pregnant women have so far been made. This study describes the results of measurements made on women pregnant with either a normal foetus or one affected by NTD, as diagnosed from AFP measurements. 3.2.2 Materials and Methods 3.2.2.1 Collection of Specimens Centres that screened for NTD by the measurement of AFP were approached for serum samples for analysis. When possible, an equal number of normals and NTD were obtained from each centre. Centres contributing specimens to the study are shown in Figure 4. Selection of sera for assay of potato alkaloids was restricted only by the sufficiency of sample remaining after the initial AFP analysis. No consideration was given to whether the women had had a previous NTD pregnancy or not. Specimens were analysed blind without knowledge of the outcome of pregnancy; this information was provided only after the analyses had been completed. It is important to note that the number and ratio of NTD to normal samples assayed from each centre did not reflect the incidence of NTD at that centre but merely the availability of suitable specimens of sufficient volume. The actual incidence of NTD in each of the areas contributing to the study is shown in the legend to Figure 4 and is based on notification of anencephalus and spina bifida births by the responsible agencies in England and Wales (Office of Population Census and Surveys) and Scotland (Common Services Agency). Originally, it was hoped to investigate an equal number of samples from NTD and normal control pregnancies but in the event, there were 210 NTD samples and 170 controls. Serum samples were stored frozen in a domestic freezer at both the collection and assay centres. 3.2.2.2 Analytical Methods The immunoassays for the measurement of solanidine and total alkaloid and the materials used are described in detail in section 2. In all assays selected internal quality control m aterials were used and both the inter- and intra-assay variations are given in section 2. 3.2.3 Results The results of the assays for solanidine and total alkaloid in all 380 serum specimens received for analysis are shown in Figures 5 and 6 respectively. The mean, median and range of both solanidine and total potato alkaloids are shown separately for each of the centres A-I. Additionally, results are subdivided into those from women pregnant with a normal or with an NTD foetus. The aggregated results for all centres are depicted as 'J'. Serum potato alkaloid concentrations varied widely between individual subjects as well as between each of the participating centres. Figure 4 Areas of collection of specimens used in this study. A = neural tube defects, A = normals. The numbers shown refer to the specimens available for assay and are not related to the incidence of NTD in the area. The incidence of NTD in the centres shown are as follows: A (Glasgow) B (Edinburgh) Rate for Scotland = 39.0 (per 10,000 total births 1976-80) C (Middlesborough) = 28.6 D (Bolton) =22.5 E (Sheffield) = 17.9 F (Nottingham) = 20.6 G (Cardiff) = 24.9 H (South East London) = 17.2 I (Ashford) = 20.8 (per 10,000 total births 1976-80) A = 3 A=10 Figure 5 Serum solanidine concentrations in pregnant women carrying either a normal foetus or one with NTD. Letters A-I represent the centres of collection as shown in Figure 4. J = all specimens collected, X = mean, I = range, H = Median. NORMAL NTD ch ro I O 1 O ~vi Ol O ui _L_ m o di IV) CD 1 0 (jl T l i o O u i m O CD ro oi i II XI 12-0 34*8 27*3 108 8-3 140 12*8 7-5 20*3 9-8 90 12*0 100 13*8 10-5 Figure 6 Serum total alkaloid concentrations in pregnant women carrying either a normal foetus or one with NTD. Letters A-I represent the centres of collection as shown in Figure 4. J = all specimens collected, X = mean, I = Range, H = Median. o o in NORMAL NTD ml O UJl oF CM in i / u i u o i LU u-l ii>v 1VJL01 aioiv>nv & in o o i o o in ml T O CM ih i CM in l>k CM o 9 'S’ o in o n o o *? CM CM d> CM in CM in Solanidine (Table 1) and total serum potato alkaloid concentrations (Table 2) in NTD and control pregnancies were compared by using Students' t-test where both sets of data were available (Centres A-C, E, G, I). Results from centre H, which provided serum samples from women with normal pregnancies only, were compared with those from Centre D which provided only NTD specimens. Comparisons were also made between the pooled or aggregated results from all normal and NTD pregnancies. The aggregated mean concentrations of solanidine was lower (p < 0.1) in the NTD group than in the control group and this was true also of results obtained from each of the individual centres. The differences were significant (p < 0.05) for only centres B, C and E. The trend for higher concentrations in the control group than those in the NTD group was also apparent for total serum potato alkaloids at centres A, B, C, E and J, but was significantly so (p < 0.05) only for centres B and E. In centres G and I, the mean total serum potato alkaloids were higher in the NTD group than those in the control group, but not significantly so. Regression analysis (Table 3) of serum potato alkaloid concentrations (solanidine and total alkaloid) on maternal age and AFP concentration failed to show any relation (p > 0.1 - < 0.5 in all cases). As mentioned in the introductory section (2.1) NTD occurs more frequently at extremes of maternal age, but this study showed that alkaloid concentrations were scattered at all levels throughout the age range. No data was available on the sex of the children or the parity of the mother, and thus it was not possible to compare these data with alkaloid concentrations. Table 1 Comparison of Serum Solanidine Concentrations on Women Pregnant with a Normal Foetus and Those with a Foetus Affected by an NTD Normal Foetus NTD t-te s t Centre n n t dof P B 33 4 2.62 35 <0.02 C 10 8 1.19 16 D 140 1.76 ZIS" <.10 H 77 E 10 13 2.71 21 <.02 F 24 - - - G 10 3 0.41 11 <.70 I 10 11 0.05 19 <.96 J 130 203 1.76 351 <.10 Table 2 Comparison of Serum Total Alkaloid Concentrations in Women Pregnant with a Normal Foetus and Those with a Foetus Affected by an NTD Normal Foetus NTD t-te s t Centre n n t dof P A 21 13 0.81 32 <50 B 35 3 2.2 36 <.05 C 11 7 1.0 16 <.40 D 141 1.61 213 <.20 H 74 E 9 13 3.8 20 <.001 F 18 - - - G 10 3 0.2 11 <.90 I 10 12 0.78 20 <.50 3 170 210 1.9 378 <.10 Table 3 Regression Analysis Data of Solanidine and Total Alkaloid on Maternal Age, AFP Concentration and Age of Specimen Solanidine Correlation Coefficient (r) t P Maternal Age -.089 -.81 >0.1 AFP -.124 -1.23 >0.1 Age of Specimen . -.220 -1.94 >0.05 Total Alkaloid Correlation Coefficient (r) t P Maternal Age -.140 -1.31 >0.1 AFP -.185 -1.48 >0.1 Age of Specimen -.221 -1.59 >0.1 The distribution of potato alkaloid concentrations for both solanidine and total alkaloid are shown as a histogram in Figure 7 for both normal pregnancies and the NTD group. Solanidine concentrations have a skew distribution with highest frequency between 0 and 10nmol/l. Total alkaloid concentrations present a more gaussian distribution, in this case, highest frequency found between 5 and 20nmol/l. The pattern for both the normal and NTD group is essentially the same. When comparing pregnant subjects with non-pregnant females shown in Figure 8 (Data for non-pregnant females is presented in more detail in section 4). Distributions of both solanidine and total alkaloid show a similar pattern to Figure 7. Peak total alkaloid concentration frequency occuring between 15-10nmol/l. The peak frequency for summer solanidine is 0-5nmol/l although the winter solanidine peak is 5- lOnmol. The large number of NTD specimens available for analysis had been collected by the various contributing centres over a prolonged period (up to three years in some instances) and stored frozen. Regression analysis of serum potato alkaloid concentration against age of specimen revealed no correlation (p > 0.1 -<0.5). This suggests that serum potato alkaloids do not deteriorate in frozen samples; this is consistent with their known stability in vitro. 3.2.4 Discussion Evidence given in section 3.1 points to neural tube defects having a multifactorial aetiology, with interaction between environmental and genetic factors. In this case, potato glycoalkaloids are one of a host of environmental teratogens which may be implicated in NTD, furthermore it must be remembered that some groups of the population may be more susceptible than others to this particular teratogen. The hypothesis that a constituent of the iue 7 Figure c n a jm ra s c N 56- 24- 32— 48- — 4 6 72— 16— 40- — 8 5 0 5 0 5 0 5 0 5 0 5 >65 65 50 45 40 35 30 25 20 15 10 5 0 -e- KAL D C CNRTO nmol/l NCENTRATION CO ID LO A LK A Concentrations in Pregnant Women Pregnant in Concentrations -O- Distribution of Potato Alkaloid Potato of Distribution T -B- t T TTL ALKALOID TOTAL NTD - ■ NORMAL = 0 NTD = • NORMAL □ = -e- -B- -e- SOLANIDINE iue 8 Figure (0 3J m oj sc z 10 14- 12 — 6 — 8 - — 5 0 5 0 5 0 5 0 5 0 55 50 45 40 35 30 25 20 15 10 5 0 e * in Norm al Non-Pregnant Women in W inter and Summer and inter W in Women Non-Pregnant al Norm in D istrib ution o f Potato A lkaloid Concentrations Concentrations lkaloid A Potato f o ution istrib D -e- LAOD OCNRTO nmol/l CONCENTRATION ALKALOID o-o t O = INTER 0=W - SOLANIDINE R E M M U •-S =UMR OA ALKALOID TOTAL l=SUMMER •o- •e- rO o potato is teratogenic and at least partially responsible for human NTD has been investigated previously only by observing the effect of short-term exclusion of potatoes from the diet of pregnant women (186). The authors concluded that the result of their experiments failed to link NTD with potato alkaloid intake, and thus support the Renwick Hypothesis (134). Direct comparison of circulating serum potato alkaloid concentrations and, by implication, potato consumption, between mothers bearing a foetus with NTD or a normal foetus has not been made, largely because suitable assays were not available. The animal work reviewed (see section 3.2.1) was largely inconclusive, the only species showing reproducible NTD after alkaloid ingestion being the Syrian hamster (193). In this study methods developed (see section 2) have allowed comparison of potato glycoalkaloid concentrations in the serum of pregnant women with a normal foetus with those carrying a foetus with an NTD. As is generally the case (section 3.1) spina bifida and anencepahly have been linked together under the general heading NTD. When, in the present study, these conditions were considered separately (data provided by centre D) no significant differences were noted between anencephaly and spina bifida (p > 0.1). Although there were noticeable differences between the mean, median and range of serum potato alkaloid concentrations in the individual participating centres, these can be largely, if not entirely accounted for by geographical differences and by the time of year when sample collections were performed. The geographical distribution of specimens collected covered many areas of England, Wales and Scotland, including areas of the North West where incidences of NTD are much higher than in the South East. All of the specimens were collected as part of the AFP screening programmes at the various centres for prenatal diagnosis of NTD. The samples were usually collected at 16 weeks of gestation. A small number were, however, collected between 15 and 22 weeks. This variable, along with maternal AFP concentrations, maternal age and duration of storage of the specimen prior to assay bear no relation to either solanidine or total alkaloid concentrations, and these factors can be discounted as possible determinants of glycoalkaloid levels in pregnancy. If the 'potato factor' of NTD causation is correct, it might reasonably have been expected that serum glycoalkaloid concentrations would have been higher in NTD mothers than those in control women at similar stages of gestation. For the individual centres and for tT»e aggregate of centres, the results show just the opposite both for solanidine and TA levels. Additionally, when comparing the two largest sample groups of 140 NTD from Bolton and 77 normals from South East London no significant difference is seen in either solanidine (p < 0.1) or total alkaloid (p < 0.2) concentrations. In view of their geographical location and pregnancy outcome the samples from Bolton would have been expected^on the basis of the Renwick theoryj^have higher alkaloid concentrations than those from the South East. It can be argued, with some justification, that since the specimens were collected after 16 weeks gestation, or later, the results do not necessarily reflect the situation pertaining at 4-5 weeks gestation when closure of the neural tube normally occurs. There is, however, no evidence to support the suggestion that potato alkaloids can be stored in the body for long periods and then released suddenly into the circulation at the onset of pregnancy (129). Although there are major metabolic changes at the onset of pregnancy these are related to increased production of endogenous steroids. Also the volume of distribution of the maternal circulation is much larger than that of the foetus and therefore the percentage of steroids entering the foetal circulation is relatively low. If solanum alkaloids were released from the body store in large enough amounts to cross the placenta in significant concentrations one wouleL expect NTD pregnancies to have raised alkaloids compared with normals, and this elevation might still be evident at 16 weeks. Research indicates a relatively long half-life for potato alkaloids in non- pregnant subjects (see Section 8). The results obtained are, however, entirely consistent with an alternative hypothesis which suggests that some nutritional deficiencies, especially of water-soluble vitamins of which potatoes are an important source m (1), can play a significant role in the causation of NTD (175,178). On this basis, the lower level of T A in mothers with NTD foetuses could have been the result of reduced potato consumption leading to a reduced intake of vitamins contained in the potato. In conclusion, within the limits of this study, no evidence was found to support the theory that potato alkaloids or the ingestion of potatoes contribute to the aetiology of NTD. 313 ZINC AND NEURAL TUBE DEFECTS 3.3.1 Introduction Zinc is present in the body in relatively large amounts (24-45mmol being present in a 70kg man). Most of the circulating zinc is contained in the erythrocytes (73-85%) with 12-22% present in the plasma. In the plasma 30- 40% is bound to p-macroglobulin, the remainder is loosely bound to albumin. Zinc is a key component of many enzyme systems, including carbonic anhydrase, alkaline phosphatase, carboxypeptidase and collagenase. The pancreas is rich in zinc where it is thought to play a role in insulin delivery. It seems to be important in growth, bone development and sexual maturation (166). A variety of clin ica l conditions has been seen to result from zinc deficiency (167). Deficiency can occur in old age, pregnancy and lactation. Clinically, deficiency can result from alcoholism, liver disease, gastrointestinal disorders, neoplastic disease and renal disease (167). Manifestations of zinc deficiency depends upon the degree of the deficiency. In moderate deficiency they include growth retardation and delayed puberty in adolescents, hypogonadism in males, rough skin, poor appetite, mental lethargy, delayed wound healing and abnormal dark adaptation. In severe cases, dermatitis, alopecia, diarrhoea, emotional disorders, weight loss, hypogonadism in males and, if unrecognised, fatality, have all been recorded. The effects of zinc deficiency are thought to occur due to its role in protein and DNA synthesis, its effect on end-organ testicular function and its importance in many metallo-enzyme systems. In zinc deficiency occurring during pregnancy there is evidence that it can lead to a variety of congenital malformations, including NTD. The effects of zinc deficiency in animals was shown as early as 1959 when chickens fed a zinc deficient diet produced chicks which were weak and died within four days (199). In a more extensive study 50% of the chick embryos were found to be grossly deformed (200). Defects in the closure of the neural tube were observed in early chick embryo explants cultivated in a zinc deficient medium (201). In rats severe zinc deficiency led to disruption of the oestrus cycles and complete failure to breed (200). Less severe deficiency led to 98% of the full- term foetuses showing gross congenital malformations, roughly 50% occurring as NTD (200). The duration of the zinc deficiency, either from birth or just prior to mating did not affect the incidence of these malformations. Frequent occurrence of central nervous system malformations in the offspring of zinc deficient rats were also described by Warhang and Petering (202). A recent review by Hurley (200) documents the wide range of malformations in the offspring of zinc deficient animals. These include deformities of the skeleton, nervous system, lung and pancreas. Sandstead specifically reviewed the effect of zinc deficiency in brain development and function (203). The author described reduction of DNA synthesis in the malformed foetuses. Nutritional deficiency of zinc in humans is fairly widespread throughout the world. It has been reported in Egypt, Iran, Morocco, Portugal, Turkey and Yugoslavia (167). The deficiency is mainly dietary, since in many of these countries, diets are based on cereals and pulses with little intake of animal proteins. Consumption of refined cereals high in phylate renders zinc unavailable for absorption (167), as does the practice of geophagia in Iran and Turkey, forming insoluble complexes with zinc. In humans, the serum zinc concentration decreases throughout pregnancy (204,205). This is thought to be a physiological effect since the decline is not influenced by zinc intake during gestation (206). In zinc deficient pregnancies, many complications have been seen. In a study of Asian women (207) zinc levels were shown to be lower than in a paired group of Europeans. The Asian group had impaired intra-uterine growth, but the authors concluded that this was not linked to zinc status. In other studies (200,203) mothers with low plasma zinc concentrations had more complications at delivery, retarded foetal growth, increased blood pressure and increased incidence of foetal malformations. Recently, it has been suggested that zinc depletion and intra-uterine growth retardation might occur by affecting placental or umbilical prostaglandin synthesis (208). The link between zinc and NTD in humans was suggested by Hurley (200), based on the animal experiments described above. Sever and Emanual (168) put forward the epidemiological evidence that areas of zinc deficiency such as Iran and Egypt, also had high incidences of NTD. A study in Turkey showed a statistical difference in maternaTzinc concentrations of the mothers of anencephalic babies compared with those of normal mothers (209). Jameson (204) reported lower zinc concentrations in women with abnormal deliveries compared with a normal group. The lowest zinc concentrations were found in mothers of infants with congenital malformations. Further evidence is seen in women with acrodermatitis enteropathica, a genetic disease of zinc metabolism showing severe zinc deficiency. In the few patients who survived to bear children, three out of seven pregnancies resulted in congenital malformations of the skeletal (1) and central nervous system (2), a figure much higher than is normally expected. In view of the above evidence, linking zinc deficiency and NTD, a study was caried out to measure zinc concentrations in the sera of women producing a normal foetus and in those who gave birth to an infant with NTD. 3.3.2 Materials and Methods 3.3.2.1 Collection of Specimens Specimens were collected and stored as described in the Section 3.2.2.1. Due to the limited amount of serum available, some specimens could not be measured for both zinc and potato alkaloids. A total of 128 samples from mothers with normal babies and 150 with NTD foetus were assayed for zinc. The distribution of normals and NTD samples for each centre are given in Table 4. 3.3.2.2 Analytical Methods Serum zinc was analysed by atomic absorption spectrophotometry using a Perkin-Elmer model. The instrument was used at the following settings: Wavelength 213.9nm Slit Width 320pm Lamp Current 5mA P hoto m ultiplier Voltage 530V C hart Speed 20mm/min Chart Range 20mV A standard curve was constructed using 5mnnol zinc acetate (Spectrosol: BDH, Poole, Dorset) to give concentrations of 25, 50, 100, 150 and 200p mol/1 in glass distilled water. Standards were freshly prepared for each assay. Each of the above was further diluted 1:5 by adding 0.5ml of standard solution to 2.0ml water to give working standards of 5, 10, 20, 30 and 40p mol/1. Table 4 Distribution of Normal and NTD Samples Assayed for Serum Zinc Centre NTD Samples Normals Bolton 106 0 C a rd iff 3 9 Edinburgh 0 18 Glasgow 13 22 South East London 0 64 Middlesborough 5 7 Notti ngham 16 0 Sheffield 7 8 TOTAL 150 128 All samples were centrifuged before use and 0.5ml of serum was diluted with 2mls of water for assay. A maximum of 10 samples were run with each standard curve to avoid assay drift. A minimum of four control sera were measured in each assay. The control sera had mean values of 10.1, 18.8, 39.2 and 41.7y mol/1. The highest control was diluted 1 in 10 before assay; all other controls were diluted as for samples. All glassware used was acid washed before use to prevent contamination. 3.3.3 Results A typical standard curve is shown in Figure 9. The inter- and intra- assay precision expressed as the co e fficie n t o f variation is given in Figure 10. Figure 11 shows the results of zinc estimations for sera from normal pregnancies, and Figure 12 shows the results on sera from NTD pregnancies. In each case, the mean, median and range of the results is shown for each centre. Additionally, the aggregated results for all centres are depicted as for J. The normal range quoted in the laboratory where the samples were assayed (Robens Institute, University of Surrey, Guildford, Surrey) was 7-20p mol/1. The results show very few samples falling outside the normal range. Indeed, only 12 samples (9 normal, 3 NTD) had zinc concentrations below 7p mol/1 representing 4.3% of all samples assayed. The majority of the samples from each centre fall towards the middle of the normal range as indicated by the mean and median values. Table 5 compares the serum zinc concentrations in the normal and NTD using a Student’s t-test. Data is given for individual centres which contained both normal and NTD samples. For centres with only NTD or normal samples, (B, F, D, H) a comparison is made between centres B (normal, n = 18) and F (NTD, n = 15) and between centres H (normal, n = 64) and D (NTD, n = 105). Significant differences between the normal and NTD group is shown only with centres D vs H and for 3 (aggregated results). In Figure 9 Standard Curve for the Estimation of Serum Zinc •80— •70 O -60 •30 •10 10 20 40 50 ZINC pmol/l Figure 10 Precision Dose Profile for Serum Zinc O £ o c N E 3 i_ O CO O @ cJsp O OCDh-h- — o ■— Serum Zinc Concentrations in Pregnant Women Carrying a Normal Foetus 18-0- B 16*0— 14-0— s H E R U 120— ( M •- T- Z I 10-0— N C -T m S '0 * " 1 0 ^ 6 0— I 4-0— A Glasgow Range B Edinburgh C Middlesborough 9 Median E Sheffield 2-0 — G C ardiff w Mean H London J ALL 0 -\_ o 3 c o z - n scocimc/) 18 16 100 12-0* 14 4 6 *0— 8 2 - iue 12 Figure - * - - - 0 0 0 0 0 0 0 — — — — — — - A C D E G J H F Glasgow Middlesborough Sheffield Bolton Nottingham Cardiff L ondon ALL Serum Zinc Concentrations in Pregnant in Concentrations Zinc Serum Women Carrying a Foetus with an NTD an with Foetus a Carrying Women +- D 1 • T | Mean Median Range G H Table 5 Comparison of Serum Zinc Concentrations (pmol/l) in Women Pregnant with a Normal Foetus and Those Affected by an NTD Norm al F oet us NT D Centre X n X n t P A 9.2 22 8.5 12 -.80 <0.5 B 12.8 18 -1.33 <0.2 F 11.7 15 C 11.1 7 11.3 5 .13 <0.9 D 11.0 105 -5.14 <0.001 H 9.2 64 E 11.3 8 12.4 6 .23 <0.9 G 10.4 9 13.0 3 2.03 <0.05 J 10-f 128 11.0 146 3.0 <0.005 iue 13 Figure NORMAL COUJCCD§ ^_ £o — —\ O N-Z Zinc Concentrations Throughout Pregnancy Throughout Concentrations Zinc GESTATION (W eeks) most of the cases the mean serum zinc concentrations are higher in the NTD group compared with the normals. The reverse is the case for Centre A. Figure 13 shows serum zinc concentrations at various stages of gestation. Samples from women for who the stage of gestation at the time of sampling was known were used to plot the serum zinc concentration against the stage of gestation. Data was obtained from specimens at 13-20 weeks of pregnancy and mean zinc concentrations are seen to decrease from 16 weeks in both normal and neural tube defect pregnancies. 3.3.4 Discussion The results measuring zinc concentrations in women pregnant with a normal foetus and those pregnant with a NTD seem inconclusive in linking zinc deficiency and incidence of NTD. This is in relation to serum zinc concentrations. Indeed, over 95% of specimens assayed fell within the normal range, whether from normal or NTD groups, although the median values tend to fall in the lower half of the range. In the majority of the centres studied, the mean serum zinc concentrations are loiocjr in the control group than the NTD group but are significantly so only in the comparison between centres D and H and for the aggregated means, J. The significance of these findings must be viewed with caution as Centres D and H represent the largest numbers of samples studied (and therefore contribute mostly to the aggregated mean J) and are situated in different parts of the country. Clearly the difference between zinc concentrations in a normal group from the South East of England and an NTD group from the North West could be due to different dietary intakes. Other workers (205) have shown results similar to those in this study in that serum zinc concentrations have been higher in women with pregnancy complications than in normal control groups. Work from India (210) did not show an association between maternal serum zinc and outcome of pregnancy, and a study in the USA found an inverse relationship between serum zinc at mid-pregnancy and foetal size. In contrast, other groups have shown lower zinc concentrations in women with pregnancy problems compared with normal controls. A relationship was found in Australian women between impaired neonatal function at birth and a decline in plasma zinc between 18 and 32 weeks gestation in mothers with 18 week plasma zinc levels below the population mean. A study measuring hair zinc found higher concentrations in a group of mothers of newborn infants with spina bifida, and it has been found that in severe zinc deficiencies, hair zinc is increased (211). Meadows (212) measured serum, leucocyte and muscle zinc and found no difference in serum zinc in women with growth retarded babies compared with women with normal offspring but lower leucocyte and muscle zinc in the growth retarded group. The choice of system for measurement of body zinc status is obviously important. In this study, measurements were limited to serum zinc. The decrease in serum zinc w ith gestational age has been demonstrated in several studies (206,213) including the present one. Both normal and NTD groups show a decrease over the 15 to 20 weeks gestational age period. This decrease in serum zinc is thought to be a physiological phenomena, possibly associated with the decrease in albumin concentration, and further complicates the tentative link between zinc and NTD. The different results found in the various studies could be due exclusively to population differences and the concurrent dietary differences. To complicate this matter vitamin supplements taken by many women also contain certain minerals. The majority of vitamin supplements contain considerable amounts of folate. Men fed diets containing 3.5mg zinc daily were found to have decreased urinary and increased faecal losses when concurrently fed folate supplements (205). This is obvious conflict since folate is thought to be a protective factor against NTD (Section 3.4). In certain countries where NTD is frequent, zinc deficiency is also present and is thought to be caused by high fibre diets containing phylate which inhibits the absorption of zinc. Potatoes, which contain steroidal alkaloids claimed to be linked with NTD are good sources of fibre, but without phylate. The possible link between zinc deficiency and NTD cannot be disregarded, but the concept that reduced body zinc results in NTD seems more over simplistic and the evidence again favours a multifactorial aetiology fo r NTD. 3.4 FOLATE AND NEURAL TUBE DEFECTS 3.4.1 Introduction The suggestion that deficiencies in folic acid might be a factor responsible for the occurrence of NTD and that vitamin supplementation might prevent this condition has been discussed in detail in Section 3.1. In the present study, serum folate was estimated in a small number of samples collected for potato alkaloid estimation. Samples were selected by the availability of a sufficient volume for duplicate analysis, age of specimen and frequency of previous freeze/thawing cycles for other assays. In all 20 samples were analysed, 12 from normal pregnancies and 8 from pregnancies resulting in a foetus with NTD. 3.4.2 Analytical Method The assay was carried out by the Department of Microbiology, Lewisham Hospital, London. The following protocol was used: Organism L.casei NC1B 10463 attenuated to grow in 300pg/ml chloramphenicol. Inoculum Two bottles of maintenance medium (Difco Folic Acid Casei Medium) were innoculated with the previous week's culture of Lactobacillus casei containing chloramphenicol at 200pg/ml and lOOpg/ml respectively and incubated overnight at 37°C. The 200pg/ml culture was refrigerated for the next week's assay. The lOOpg/ml culture was centrifuged, washed and resuspended in 5ml distilled water. One litre of single strength Difco Folic Acid Casei Medium (47g/1) was prepared, brought to the boil and allowed to cool. Then 0.6g of ascorbic acid was added with two drops of Tween 80 and 0.15ml of L.casei prepared as above. Standards Stock folic acid solution (150mg/l) was diluted 1:1000 with distilled water to give a working solution of 150pg/l. From the working solution, nine assay standards were prepared to give a concentration of 18, 15, 12, 9, 6, 3, 1.5, 1.2 and 0.75 pg/1. 35pl of each test serum or standard in duplicate was added to 4ml of medium (containing antibiotic and organism) and incubated at 37°C for two days. In addition, several blanks and extra lOpg/l standards were prepared to check whether sufficient bacterial growth was present. After incubation tubes were mixed and read at 625nm in 3" x J" tubes. A standard curve was prepared and the serum folate concentration of test sample calculated. 3.4.3 Results The results of the serum folate estimations in the normal and NTD subjects are shown below: Normals NTD 3.5 4.0 6.0 6.0 7.5 6.5 8.0 9.5 10.0 10.0 11.0 13.0 11.5 18.0 12.0 >18.0 12.5 17.5 >18.0 >18.0 N = 12 N = 8 X = 11.29 X = 10.63 T = 0.294 P =<0.5 No significant difference was observed between the normal and neural tube defect groups. The normal range for serum folate established in the laboratory is 4-18yg/l. Only one result falls below this range, which is present in the normal group. 3.4.4 Discussion The results of this small study indicate that there is no significant difference in serum folate concentrations between a group of women pregnant with a foetus subsequently shown to be affected by an NTD, and a group pregnant with a healthy foetus. However, the mean of the NTD group is slightly lower than the normal group. The samples used in this study were from a larger group collected for potato alkaloid measurement. The samples were kept fo r less than 24 months on average before assay and were frozen and thawed a maximum of three times (twice for potato alkaloid assays and once for zinc estimation). The stability of folate in stored samples has received some attention. Kubasik et al., (214) examined plasma folate levels in samples stored at room temperature, refrigerated (4°C) and frozen (-6°C) for up to 24 days. Only samples stored frozen and refrigerated were stable for the duration of the experiments. A more extensive study by Voogd et al., (215) found a decrease of folate activity of up to 11% after one year of storage and up to 32% after two years. In the present study these findings must be taken into consideration, although the effect should be the same for both normal and NTD samples, and over 99% of samples had folate concentrations within the normal range. The next consideration is the use of the microbiological assay, which is subject to the presence of antibiotics and anti-metabolites in the sample. Finally, as with the alkaloid and zinc assays the folate levels in serum taken at 16 weeks gestation might not reflect the situation at the time of closure of the neural tube. Smithells et al., (175) estimated several vitamins in pregnant women, including serum folatei They found lower serum folate levels in social classes III, IV, V, compared with classes I and II. When comparing serum folate in normal pregnancies with mothers of infants with CN5 (central nervous system) malformation the normals had higher folate levels, but not significantly so. A finding similar to the present study. The clinical trials already conducted and those taking place at the moment (discussed in Section 3.1) suggest a protective affect of vitamin supplementation on the recurrence of NTD. However, if the vitamin supplementation is a part of a general increased preconceptional and antenatal care, this would obviously mask the specific vitamin affect. Indeed, in countries where the average diet is lacking in vitamins there is no suggestion of an increased incidence of NTD (180). The value of folate and vitamin supplementation in prevention of NTD cannot be disregarded but the evidence seems to be in favour of the multifactorial aetiology for NTD and not a single dietary, genetic or environmental cause. CHAPTER 4 ALKALOID CONCENTRATIONS IN SELECTED POPULATIONS CHAPTER 4 ALKALOID CONCENTRATIONS IN SELECTED POPULATIONS 4.1 ALKALOID CONCENTRATIONS IN THE NORMAL POPULATION 4.1.1 Introduction The measurement of potato alkaloids has largely been limited to their concentration in different potato cultivars (Section 1.7). Animal experiments involved administration of specific alkaloids and following toxicological studies (Section 1.8). The work carried out in humans investigated the administration of tritiated alkaloids (128, 129) although absolute concentrations were not determined. The only authors who have estimated potato alkaloid concentrations in humans measured solanidine levels in a limited group of samples obtained from hospital patients (101). To assess the possible toxicological affect of potato glycoalkaloids in humans it is first necessary to establish the normal body concentrations obtained from dietary intake. 4.1.2 Materials and Methods 4.1.2.1 Collection of Samples In order to establish a normal range for potato alkaloid concentrations in the normal healthy population, samples of blood and saliva were collected from a group of volunteers. Potato alkaloids are exogenous substances found mainly in potatoes and thus body levels (if any) would be expected to vary with dietary intake. The volunteers consisted of a group of healthy adults of various occupations from the London area of the United Kingdom, with 18 males aged 20-45 years of age and 15 females aged 19-63 years of age. All of the subjects were eating their normal diets prior to the study, with no restrictions being made on their dietary intake. Samples were collected in the morning before lunch during the summer months. Blood samples were collected by venepuncture; the serum being separated by centrifugation and stored in a domestic freezer until assayed. All subjects collected their own saliva samples, either immediately before or after blood sampling, as described in Section 5.2.1 (salivary potato alkaloids). The time of collection and as far as possible the same subjects had previously been used to establish winter potato alkaloid concentrations. In addition to the UK volunteers, a small group of specimens were collected in a similar manner from a group of Swedish subjects. The Swedish subjects consisted of five males aged 31-41 years of age and five females aged 31-67 years of age of whom seven ate normal summer diets. The remaining three subjects, two males and one female/, had an altered dietary intake, including potato varieties known to be high in glycoalkaloids, Ulster Chieftain (20-30mg TGA/lOOg fresh weight of tuber) and a new Swedish variety, SV72118 (18-30mg TGA/lOOg) (K E Hellanas, personal communication). They ate cooked, unpeeled potato (200-300g) daily for one week before sampling. In all cases the last potato meal was eaten 12-24h before sampling. 4.1.2.2 Analytical Methods All samples were assayed for both solanidine and total alkaloid as described in Section 2.1. In some cases sufficient sample was not available for both solanidine and total alkaloid analysis and this is reflected in the numbers shown in the results. In all assays selected internal quality control materials were used; both the inter-and intra-assay variations are given in Section 2. 4.1.3 Results The solanidine and total alkaloid concentrations for the normal UK subjects and the Swedish group are given graphically in Figures 1 and 2. The data for the three subjects on the high alkaloid diets are not included and will be discussed later. The range of values along with their means are given in Table 1. In addition, included for comparison are the ranges of alkaloid concentrations from normal pregnant women (Section 3), the winter concentrations of solanidine in normal UK subjects, and values measured by Mathew et al., (101) of solanidine in hospital patients. 4.1.4 Discussion When analyte concentrations are measured in body fluids it is usual to establish a reference interval or normal range in a group of healthy subjects. With many constituents, the distribution of concentration is symmetrical and has a characteristic Gaussian distribution. For other constituents, the distribution is assymetrical or skewed. For a Gaussian distribution the population is sym m etrically distributed around the ourUjki*vt.kic mean and 93% of the population is included in the range x + 2 standard deviations. However, about 5% of the population will fall outside these values but still be 'normal1. When constructing reference ranges, samples of the population are often used to derive a population range, however extreme values in a sample of 100 are likely to be much rarer than in a sample of 20. With reference to the potato alkaloids, these substances are not natural constituents of the body but are introduced by the diet. Thus the potato, being the main source of alkaloid, would be the main dietary factor influencing alkaloid concentration. Figures 3 and 4 show the relation between solanidine concentrations in males and females and an index of the monthly Distribution of Solanidine Concentrations in Selected Groups 55 50 45- 40 35- A o o A 30- 25' A o 20 o A • ••• • • • •• 15 0 0 A o A A 10 - ▲ ▲ A A A ▲ ▲ 5 ▲ i ▲ a a a a a a a a I------1------1 MALE SWEDEN FEMALE Male Female f UK W INTER UK 2 Distribution of Total Alkaloid Concentrations In Selected Groups □ 100 - 90 80 O 70 O oo o 60- A 50 A 40 OOO 30 o AA o A o A 20 • 8 D * A 10 - A A T I I MALE SWEDEN FEMALE UK UK T able 1 Range of Alkaloid Concentrations In Selected Populations (A) Normal UK Subjects (Summer) Solanidine (nmol) X n Range Male 10.9 18 <2.5 -42.5 Female 5.3 15 <2.5 - 15.0 A ll 9.1 33 <2.5 -42.5 Total Alkaloid (nmol/1) X n Range Male 36.2 18 3.5 - 77.5 Female 22.3 15 3.3 - 50.0 A ll 30.0 33 3.3 - 77.5 (B) Swedish Subjects (Summer) Solanidine (nmol/1) X n Range 10.5 7 2.5 - 42.5 Total Alkaloid (nmol/1) X n Range 42.3 7 20.8 - 105.0 (C) Normal UK Subjects (Winter) Solanidine (nmol/1) X n Range Male 27.0 30 5.3 - 56.3 Female 19.8 27 4.0 - 56.3 A ll 23.8 57 4.0 - 56.3 (D) Normal Pregnant Women X n Range Solanidine (nmol/1) 12.8 150 < 2 .5 -8 5 Total Alkaloid (nmol/1) 31.8 170 2.5 - 145 (E) Hospitalised Patients (Mathew et al) (101) Solanidine (nmol/1) X n Range Male 3.9 7 2.1 - 10.4 Female 3.0 27 1.0 - 12.5 A ll 3.2 34 1.0 - 12.5 Figure 3 Linear Regression Plot of Potato Intake Versus Serum Solanidine Levels in Male Subjects M 50---- o n t 45— h 1 y 40 ---- p 35— o t a 30— t o 25— l■ n t 20— a k e 15— 10— 5— Regression Line Dat£ 10 15 20 t 25 ^ 30 30 35 35 40 3) 4d 5C Serum Solanidine (nmol/l) Equation Y = 2.6515X + 2.326 Correlation coefficient R = 0.8796 Student's t value (13 d.f.) t = 6.6666 (p < 0.001) Figure 4 Linear Regression Plot of Potato Intake Versus Serum Solanidine Levels in Female Subjects M 40 0 n t 35— h 1 y 30— P o 25— t a t 20— 0 15— 1 n t a 10— k e 5 — 45 50 Serum Solanidine (nmol/l) Regression Line Data Equation Y = 0.775X+ 12.368 Correlation coefficient R = 0.7039 Student's t value (d.f.) t= 3.964 (p< 0.01) alkaloid intake indicated by their monthly potato consumption. A points system was used to differentiate the relative alkaloid content of various potato meals. A significant correlation was found for both males and females, but was less close for females than for males, possibly due to the greater variation in size of potato portions eaten by females compared with males. The normal ranges quoted in Table 1 are based purely on the extreme alkaloid concentrations found in normal healthy subjects, with no symptoms of alkaloid toxicity. The concentrations would be expected to have a lower lim it of zero, similar to other exogenous substances found in the body. The concentrations of solanidine and total alkaloid found in the summer months have a different range and mean for male and female subjects. This is to be expected since firstly it is likely that adult males would have a higher potato intake than females (indeed, this is bourne out in Figures 3 and 4) and secondly, slimming regimes during the holiday period would most likely be carried out more by women than men. The distribution of concentrations of solanidine and total alkaloid in all groups (Figures 1 and 2) reflects the varying dietary intake expected in a normal population. Comparing the UK subjects and the Swedish group in the summer months, the mean and range of both solanidine and total alkaloid are comparable but are slightly higher in the Swedish group. Some Swedish new potatoes on the market during the summer are considered a delicacy and tend to have a TGA content near the 20mg/100g lim it (50). In the UK group during the winter, comprising mainly the same subjects as the summer group, the solanidine concentrations are more than twice as high as those found in the UK group in the summer. Again, the mean concentration in males is higher than females, though the difference is not as marked as in the summer group. New potatoes eaten in the summer are generally smaller, and especially if not peeled, a high intake of alkaloid might then be expected if the same weight of potatoes was then eaten as in the winter. The results suggest then a higher potato and hence alkaloid intake in the winter as compared to the summer. Although the mean of the concentrations in the winter is more than twice that in the summer, the ranges are quite similar. An explanation for the lower lim it of solanidine concentrations being greater than the assay lim it of detection (ie, lowest concentration found > Onmol/1) could be the wide range of products which contain potatoes or potato by-products. Potato products in soups and stews, etc, are much more likely to be eaten in the winter months. The differences between summer and winter solanidine concentrations is unlikely to be due to assay methodology, although the assay used to estim ate the w inter concentrations used 2ml o f serum and an antiserum raised in rab bits. The later assay reduced sample volume to 500pl and used an antiserum raised in sheep to an identically prepared conjugate as that used in the rabbits. However, there were no major differnces in cross- reactivity between the two antisera (Section 2). The solanidine and total alkaloid concentrations in the group of normal pregnant women show similar means (though higher) and a much wider range than that of the UK group in summer. When compared to the females in the group the differences are quite marked. When compared with the winter group the mean solanidine concentration of the pregnant women is lower but again the upper limit of the range is higher. The samples from pregnant women come from all areas of the United Kingdom and were collected at d iffe re n t tim es o f the year (see Section 3). Finally, the alkaloid values previously estimated by Mathew et al., (101) in hospital patients. The samples were collected for routine biochemical analysis from a random group of patients in hospital, the excess serum being removed for alkaloid estimation. The solanidine concentrations in this group are much lower than any other group described. An alteration of diet in pregnancy is very probable but in hospital and in such patients an altered diet is almost certain. Thus the lowered alkaloid concentrations seen in hospitalised patients are not surprising. The data given in Table 1 shows the great diversity of alkaloid concentrations found, and hence, the wide range of dietary potato consumption. In the majority of groups studied the number of subjects was small. However the data for the normal pregnancy group numbering over 150 samples must provide a good indication of the range of alkaloid concentrations to be found in the UK. Caution must be exercised to lim it these ranges to ethnic cultures w ith sim ilar diets and to sim ilar seasons. The data for the Swedish subjects on the high alkaloid diets were not included in the normal ranges since they represent a falsely elevated dietary intake of alkaloid (Section 4.1.2.1). This is reflected by the serum solanidine and total alkaloid concentrations found in this group which are shown in Table 2. The mean solanidine and total alkaloid concentrations in the high intake group are significantly higher than the normal intake group and in the ranges shown, both lower and upper limits are higher than the upper limits of any other group. Table 2 Comparison of Serum Alkaloid Concentrations in Swedish Subjects on a High Alkaloid Diet With Those on a Normal Diet Serum Solanidine (nmol/1) n Mean Median Range t P 7 10.5 6.8 2.5 - 42.5 Norm al -7.54 <0.001 High Alkaloid 3 80.8 77.5 72.5 - 92.5 Serum Total Alkaloid (nmol/1) n Mean Median Range t P 7 42.3 32.5 20.8 - 105 Norm al -4.62 <0.01 High Alkaloid 3 125 >125 125 ->125 Although the subjects on the high alkaloid diet noted a bitter taste in the potato tubers, no symptoms of alkaloid toxicity were seen. This would suggest that either the body fluid alkaloid levels causing acute toxicity would be greatly in excess of the concentrations found in normal individuals or that the alkaloids per se are not responsible for producing these symptoms. The salivary alkaloid concentrations in the Swedish group as well as the UK summer group are given and discussed in Section 5.3 and 5,4. Comparison of serum solanidine and serum total alkaloid concentrations shows a significant correlation: n Mean Median Regression Line r t P Solanidine (nmol/1) 42 14.5 5.0 Y=1.535x + 20.1 0.916 13.3 <0.001 T otal alkaloid (nmol/1) 42 38.5 27.5 The mean total alkaloid concentrations are 2.7 times the'solanidine concentration. If the major alkaloids a-solanine and a-chaconine, were present in the body unmetabolised, this ratio would expectedly be much higher. Since in the potato the glycosides represent over 95% of the TGA fraction^ this observation suggests that the first stage of metabolism of the glycosides in man is hydrolysis of the sugar residues. The metabolism of the potato alkaloids w ill be discussed la te r (Section 8). 4.2 TOXICOLOGICAL EFFECTS OF EXCESSIVE INTAKE OF POTATO GLYCOALKALOIDS IN MAN 4.2.1 Introduction There have been many well-documented cases of potato poisoning in both man and animals (section 1.8.1). The majority of these outbreaks resulted from the consumption of either greened or rotten potatoes or potatoes stored for a long period of time in unfavourable conditions. In one case (108) the poisoning was caused by the consumption of sprouts rather than the more usual tubers. In animals poisoning has occurred from eating tubers, but more commonly other parts of the potato plant (Section 1.8.1). In most of the above cases the potatoes have contained high concentrations of potato alkaloid which have been put forward as a possible candidate for the poisoning (105). In the potato plant alkaloid concentrations tend to be lowest in the tuber, the peel having greater levels than the flesh (26). Much greater concentrations are found in the leaves, sprouts and flowers (28). Below is described a case of poisoning in man, in which body fluid alkaloid concentrations are measured: 4.2.2 Materials and Methods 4.2.2.1 Case History The subject was a 28 year old man from Papua, New Guinea, residing in Sweden. The patient consumed a meal consisting of potato leaves which were boiled in water for 10-15 minutes, squeezed to remove water and fried with butter, onion and fish. About 30 minutes after the meal the subject complained of a severe headache. After two hours he had colicky central abdominal pain and violent diarrhoea (green-coloured). The subject was at first unable to urinate. Vomiting began 5-6 hours after the meal when the subject was admitted to hospital! Body temperature was normal. 4.2.2.2 Glycoalkaloid Analysis An analysis of an identical batch of potato leaves cooked as described above showed a TGA content of 370mg (K E Hellenas, unpublished data). Samples received here in this laboratory from Sweden comprised of a random urine sample collected on the day of the incident and a whole blood sample collected seven days after the incident. The samples had been stored frozen prior to assay. Both solanidine and total alkaloid were estimated in the urine sample, the analytical methods used were those described in Section 2. On thawing the sample of whole blood it was found to be grossly haemolysed and due to the probable interference of haemoglobin in liquid scintillation counting, direct analysis of the haemolysed sample for total alkaloid was not attempted. Instead, solanidine analysis (involving chloroform extraction of the sample) was carried out. In order to compare the alkaloid concentrations in the urine sample with those in normal urine the creatinine content of the sample was measured using an 5MAII multichannel analyser (Technicon Instruments). A random urine sample from three normal subjects was also analysed. These subjects known to have a high potato intake (at least at one meal daily) gave random urine samples before lunch. The samples were analysed for solanidine and total alkaloid in the same assays as the sample from the poisoned subject. 4.2.3 Results The urine alkaloid concentrations found in the poisoned subject are shown in Table 3. The 24h urine creatinine in normal adult males is 8,800 to 17,000pmol/24h (x = 13,000pmol/24h). Assuming normal renal function in the poisoned subject, the 24-hour urine output of solanidine and total alkaloid can be estimated and are given in Table 3. The whole blood solanidine concentration seven days after the incident was 5.0nmol/l. The results from the normal subjects are shown in Table 4. The alkaloid concentrations found in the urine of the poisoned patient are found to be higher than any of the control subjects. It is necessary to compare the alkaloid concentrations as related to the creatinine concentration to compensate for any differences in urinary output. Using the mean adult male creatinine output (13,000ymol/l over 24 hours) as a standard, the urinary 24h output of solanidine in the poisoned subject is 112pg/24h compared with 78pg/24h the highest found in the control subjects. The total alkaloid concentrations show a much more striking difference. In the poisoned subject, 1200pg/24h compared with the highest value of 223pg/24h in a normal subject. This is reflected in the total alkaloid to solanidine ratio which is 10.7:1 in the poisoned subject and ranging from 1.68 - 3.04:1 in normals. This is therefore a much greater output of total alkaloid in the poisoned subject. The significance of this w ill be discussed below. I 4.2.4 Discussion The subjects chosen for the normal urine collection were deliberately selected for their high potato intake to give a fair comparison with the poisoned subject. An earlier study (McMillan, unpublished data) measured urinary solanidine on two subjects firstly after a 3-day potato abstenance, and secondly after a potato load. The first subject had a pre-load solanidine concentration of 0.39pg/24h and a post-load concentration of 2.22pg/24h. The Table 3 Urine Alkaloid Concentrations in the Poisoned Subject Total alkaloid 115nmol/l (46ng/ml) Creatinine = 502p mol/1 Solanidine 10.8nmol/ (4.3ng/ml) Total alkaloid 91.6ng/ymol creatinine solanidine 8.6ng/ymol creatinine 24h Urine solanidine 76pg/24h - 146pg/24h (x= 112pg/24h) Total alkaloid 806pg/24h - 1560pg/24h (x= 1200jjg/24h) Table 4 Urine Alkaloid Concentrations in Normal Subjects Subject 1 Total alkaloid = 33.0nmol/l (13.3ng/ml) Creatinine = 1300p mol/1 solanidine = 13.4nmol/l (6.7ng/ml) 24h Urine solanidine = 46.0 - 88.4pg/24h x= 68pg/24h 24h Urine total alkaloid = 90.0 - 173.0pg/24h x = 133pg/24h Subject 2 Total alkaloid = 16.8nmol/l (6.7ng/ml) Creatinine = 665p mol/1 solanidine = 10.0nmol/l (4.0ng/ml) 24h Urine solanidine = 53.0 - 102.0pg/24h x=78pg/24h 24h Urine total alkaloid = 89.0 - 172.0pg/24h x=131pg/24h Subject 3 Total alkaloid = 11.5nmol/l (4.6ng/ml) Creatinine = 268p mol/1 solanidine = 3.8nmol/l (1.5ng/ml) 24h Urine solanidine = 49.0 - 95.0pg/24h x=73pg/24h 24h urine total alkaloid = 150.0 - 289.0pg/24h x= 223pg/24h second subject had a pre-load concentration of 4.67yg/24h and a post-load of 30.38yg/24fu The first subject ate peeled boiled potatoes (= 300g/day) and the second, jacket potatoes consuming the skin (= 300g/day). The higher alkaloid content of the outer layers of the potato (Section 1.4) are reflected in the urinary solanidine output of the second subject. The urinary solanidine in these subjects was measured by gas chromatography/mass spectrometry after urine solvent extraction of the sample. The large difference in the urine concentrations after the potato load could reflect both the alkaloid intake due to the different potato parts eaten plus the previous dietary history (Subject 2 had a much higher pre-load solanidine concentration, 4.67pg/24h, compared with Subject 1). The urine collections made were accurate 24 hour collections and should more accurately reflect the alkaloid output than the random urines used in the potato poisoning case. However, the solanidine concentrations in subject 2 is the same order of magnitude as the normals in the potato poisoning study (even though a different method of alkaloid analysis was used) and is much lower than the poisoned subject. A further comparison between normal urinary alkaloid output and the potato poisoned subject can be made from the pharmacokinetic study to be described later (Section 8. ). After a potato abstinence of 28 days four subjects were given a potato load of 500g containing 14.3mg of total alkaloid. Urine was collected for the next 48 hours in two 24-hour collections. Solanidine concentrations in the first 24 hours ranged from 7 - 16.5yg/24h and 8 - 9.5yg/24h from 24-48h after dose. Total alkaloid ranged from 18.8 - 48.3yg/24h in the first 24 hours and 16.8 - 19.4yg/24h in the second. The alkaloid levels were again much lower than the poisoned subject and again the TGA:solanidine ra tio (Range 1.8 - 3.9) is much less. In all of the comparisons, the most striking differences between the values in the poisoned subject and those in the normals is the much higher total alkaloid concentration and higher TArsolanidine ratio. The amount of total alkaloid consumed (370mg) in the poisoned case represents a dose of 4.6mg/kg for an 80kg man. Calculations by Morris and Lee (105) gave the toxic dose for potato alkaloids as 2-5mg/kg and the lethal dose as 3-6mg/kg, similar to the 4.6mg/kg given above, although no data is available as to the amount of alkaloid absorbed, since diarrhoea was known to start two hours after the potato meal. The high concentration of total alkaloid in the urine would be expected to reflect the urine content of the glycoalkaloids a-solan! ine and a- chaconine which make up over 95% of potato glycoalkaloids. In the gut the lipid soluble aglycone solanidine from the potato or breakdown product from solan ine and chaconine in the gut would have a greater ability to cross cell membranes than the glycosides. In urine the situation would be reversed, the water soluble glycosides being more readily excreted than the lipid soluble solanidine. In the poisoned subject the high ratio of total alkaloid to solanidine in urine indicates the rapid elimination of the water soluble glycosides either unchanged as a-solanine and a-chaconine or as the intermediate metabolites, the 8- and ^-glycosides, before total hydrolysis to the aglycone can occur. In the pharmacokinetic study (Section 8) the ratio of total alkaloid to solanidine in the first 24h urine sample is higher than in the succeeding collection, again suggesting an initial rapid elimination of the water-soluble glycosides before hydrolysis to the aglycone. The whole blood solanidine concentration if* the poisoned subject is well within the range found in healttypjbjects and is indeed at the lower end of the range (see Section 4.2), 5.0nmol/l. Two possible explanations for this are firstly the fact that the specimen was collected seven days after the poisoning incident (unlike the urine collected on the same day) and secondly the use of whole haemolysed blood as a m atrix and consequent assay interference. Although the samples collected from the case of poisoning were not ideal, this is still the first time that any alkaloid concentrations have been measured in poisoning and shown to be much higher than those found in the urine of a variety of normal subjects. CHAPTER 5 POTATO ALKALOIDS IN HUMAN SALIVA CHAPTER 5 POTATO ALKALOIDS IN HUMAN SALIVA 5.1 INTRODUCTION Saliva has been used as a biological fluid for measuring the concentrations of several hormones and drugs. The technique is cheap, non- invasive and provides an easy method of multiple sampling and patient self collection. Saliva is secreted by three pairs of exocrine glands, the parotid, the submaxillary and the sublingual. Water accounts for 99% of the secreted fluid, the remaining 1% consisting of various salts, proteins and other minor compounds. The secretion of saliva is controlled by the autonomic nervous system, both the sympathetic and parasympathetic branches stimulating the glands. During sleep, very little saliva is secreted. When awake the basal secretion rate is about 0.5ml/min. Food in the mouth stimulates the rate of salivary secretion. The most potent stimuli for salivary secretion are acid solutions, ie, fruit juices and lemons, which may lead to a maximal secretion of 4ml of saliva per minute. Figure 1 illustrates the structure of a typical salivary gland showing the exchange of electrolytes between the blood and saliva. Sodium is actively pumped from blood to saliva gland endpiece, water flows between the acinar cells under osmotic pressure, bicarbonate and chloride follow to maintain the ionic balance. As the saliva moves down the ductal system, sodium is exchanged for potassium. This exchange is a limited mechanism and with increased flow rates the sodium concentration in saliva increases, as consequently does chloride, bicarbonate and pH. The variation in bicarbonate ion concentration with flow rate can alter the pH from 6.5 at low flow rates to 8.0 at high flow rates. Figure 1 A Diagrammatic Representation of a Typical Salivary Gland Showing Electrolyte Exchange Capillaries □round acinar cells C l" PRIMARY SECRETION BY K * Cl" ACINAR CELLS Na+ , Ureo • Venous blood Amino Acids SECONDARY Capillaries — around ducts MODIFICATION IN D U C TS c r Glucose A rterio l blood S aliva The structure of the salivary gland allows for three main mechanisms by which molecules can enter saliva from blood (216); these are active transport, ultrafiltration and intracelllar diffusion (Table 1). Lithium (molecular weight = 7) has a saliva/plasma ratio of more than 2, indicating an active secretory mechanism. An active transport mechanism has also been suggested fo r penicillin and a few other drugs, but no defin itive proof has been forthcoming. Small water soluble molecules can pass into saliva via the tight junctions between the acinar cells. Thus, lithium would be able to enter also by this route. The tight junctions prevent the passage of substances with molecular weights greater than 300. Urea (molecular weight = 60) and ethanol (molecular weight = 46) have saliva/plasma ratio near unity and are thus fully filtered. Sucrose, (molecular weight = 342), is lipid insoluble and largely excluded from saliva. This is also the case of lipid insoluble conjugated steroids whose only route of entry is ultrafiltration and which appear in saliva at approximately 1% of the unbound plasma concentrations (217). The major method of entry of drugs and steroids into saliva is intracellular diffusion through the capillary and acinar cell membranes. These compounds diffuse down a concentration gradient from the free fraction in the plasma. Compounds which are lipid soluble, non-ionised at physiological pH and of small molecular weight will diffuse rapidly. The salivary concentration of this type of compound is independent of flow rate. Hydrophilic molecules or lipid soluble compounds which have subsequently been metabolised to glucuronides or sulphates are largely unable to enter saliva by intracellular diffusion. This can be an advantage, eg, Transfer of Drugs and Steroids from Blood to Saliva Active Transport - Lithium Ultrafiltration - Small water-soluble molecules via tight junctions between acinar cells. Molecular weight < 300. Intra-Cellular Diffusion - Main mechanism for drugs and steroids influenced by: lipid solubility, molecular weight, and ionisation at physiological phenytoin is metabolised to an inactive parahydroxyphenyl, phenyl hydrantoin glucuronide, and thus salivary estimation using an assay measuring both drug and metabolite would estimate only the active drug. The change in salivary flow rate increases the concentration of bicarbonate and hence alters the pH. The changes in pH can alter the ionization of several types of compounds (depending on their pKa). The degree of ionization will alter the lipid solubility of the compound and its ability to enter saliva by intracellular diffusion. Thus, the salivary concentration of this type of compound is dependent upon flow rate. In the blood hormones and drugs are present in two forms, free and protein-bound. There is much evidence that the free fraction of the compound represent the biologically active part (218). Compounds bind to proteins by the formation of non-covalent bonds. These bonds to the major non-specific binding protein albumin are weak and allow rapid association and dissociation. The binding of certain hormones to their specific binding proteins such as sex hormone binding globulin and cortisol binding globulin is much stronger and thus the association and dissociation with these proteins is much slower. Because of these complex relationships, methods to measure the free-fraction of compounds in blood, such as equilibrium dialysis and ultracentrifugation have been limited. The simplest route to analyse free analyte concentrations is offered by saliva providing the compounds are not affected by flow rate, are lipid soluble, small and essentially non-ionised. Salivary estimations of a number of hormones or drugs have proved clinically useful. Salivary progesterone shows good correlation with matched serum samples and multi-sampling techniques throughout the menstrual cycle can provide diagnostic information as well as follow the success of treatment (219). Plasma ‘free' testosterone provides information of greater diagnostic value in infertile women than total hormone concentrations (219) and salivary testosterone has provided an index of the free hormone. In addition, salivary levels of aldosterone, cortisol and 17-a-OH progesterone have proved clinically useful. Drugs already mentioned which have been measured in saliva are lithium and phenytoin. Other anticonvulsants such as carbamazepine and ethosuxcimide have been estimated in saliva, these drugs are insignificantly ionized at physiological pH. The major problem with salivary analysis is contamination (Table 2). This may occur if the samples are taken too soon after oral administration of a drug or food which still remains in the mouth. Secondly, contamination can occur by handling drugs and contamination of the salivary sample with the fingers. Falsely high salivary concentrations can occur if the saliva contains blood as after vigorously brushing the teeth, or the presence of plasma exudate in the case of gingivitis. Components of the blood which have a high boundrfree ratio are most important in this type of contamination. Finally, the use of a non-specific assay which will cross-react with other constituents of saliva can lead to falsly elevated analyte levels. The steroidal potato alkaloids are structurally similar to endogenous body steroids. The aglycone solanidine has a low molecular weight and is lipid soluble. These criteria provide good reasons for the measurement of potato alkaloids in saliva in order to provide greater information on their action and metabolism in the body. Factors Causing Contamination of Saliva Presence of Drug or Food in the Mouth when Sampling Saliva Handling of Drug/Food - Contamination to Mouth Blood Contamination - Gingivitis Cross Reacting Compounds 3.2 MATERIALS AND METHODS 3.2.1 Collection of Specimens Specimens were collected from a group of 33 volunteers in the UK during the summer months and from a group in Sweden. The group from Sweden consisted of seven subjects eating their normal diets and three subjects on a high alkaloid diet. Further details of the subjects and diet are given in Section 4.1.2.1. The protocol for the collection of saliva must be strict to avoid contamination and is given in Table 3. The specimens were collected before lunch at the same time as the blood samples. By rinsing the mouth prior to sampling it was hoped to remove any food contaminants left from the previous meal and to avoid brushing the teeth which could cause bleeding. The presence of mucins in saliva increases its viscosity and causes difficulty in handling. Mucins and any epithelial cells were removed by freezing and centrifugation, leaving a clear supernatant for assay. 5.2.2 Analytical Methods The immunoassays for the measurement of solanidine and total alkaloid in saliva and the materials used are described in detail in Section 2 (Radioimmunoassay of Potato Alkaloids). In all assays selected internal quality control materials were used; both the inter and intra-assay variations are given in section 2. Table 3 Collection of Saliva for Glycoalkaloid Assay 1. Specimens Collected Before Lunch. Blood Sample Taken Immediately Before or After Saliva 2. Each Subject to Rinse Mouth with Water 30 Minutes Prior to Sample. No Further Food ro Drink Until After Collection 3. Mixed Saliva (5mls) Collected in a Universal Container 4. Sample Frozen, Thawed and Spun. Supernatant Removed and Stored at -20°C Until Assay 3.3 RESULTS Figure 2 shows the concentrations of alkaloids in saliva in all subjects tested. In addition, to compare the serum and salivary levels, the serum concentrations already shown in Section 4 are given. Subjects are divided into groups, all subjects, UK subjects, Swedish subjects and Swedish subjects minus the three on high-alkaloid diets. The winter UK solanidine concentrations are also given for comparison. The median levels of serum and salivary solanidine are similar for each of the groups, but the mean solanidine concentrations in serum are higher than the corresponding salivary levels. In contrast total alkaloid concentrations in serum have a relatively much higher mean than their salivary counterparts. Salivary solanidine has a very similar but slightly lower mean and median than salivary total alkaloid in all groups. In comparison serum concentrations of total alkaloid have a much higher mean and median, more than twice as high as those of serum solanidine. Figures 3 and 4 show the distribution of solanidine and total alkaloid for all subjects. The UK summer subjects are separated into males and females. Table 4 shows the mean and range of values for salivary solanidine and total alkaloid for UK summer subjects. As with the serum values given in Section 4, the mean and range for both solanidine and total alkaloid in saliva are lower in females than in males. A linear regression analysis (Table 5) for serum solanidine versus salivary solanidine and serum total alkaloid versus salivary total alkaloid showed a significant correlation in both cases, the correlation coefficient for total alkaloid being greater than that for solanidine. Comparison of the alkaloid levels in the three Swedish subjects who had eaten potatoes with an unusually high alkaloid content with those on a normal Figure 2 Summer Solanidine and Total Alkaloid Concentrations in Serum and Saliva A = A ll Subjects , n=43 B = UK Subjects , n=33 C = Swedish Subjects , n=10 D = Winter UK Concentrations , n=57 X = Mean 1 = Extreme Values |—| = Median Value * = Extreme Values Minus the 3 Subjects on High Alkaloid Diets SOLANI DINE TOTAL ALKALOID Tt -s o E c ego \D n © in IX n IX n o 3=j) mz-o-z>rO(/) iue 3 Figure Distribution of Salivary Solanidine Concentrations in Summer in Concentrations Solanidine Salivary of Distribution MALE • • • • K UK UK SWEDEN ■■ FEMALE AA of Salivary Total Alkaloid Concentrations in Summer 11 O 10 O 9 A 8 7 □ 6 O 5 4- □ 3* O ° □ AAA 2* O A ® g * OOOOOOO □ AAA 1- T MALE SWEDEN FEMALE UK UK Table 4 Normal Range of Salivary Alkaloid Concentrations in UK Subjects (in Summer) Solanidine (nmol/1) X n Range Male 2.8 18 <0.6-9.3 Female 2.0 13 <0.6-2.3 A ll 2.6 33 <0.6-9.3 Total Alkaloid (nmol/1) X n Range Male 3.3 18 <1.3-10.5 Female 2.2 15 <1.3-4.8 A ll 3.0 33 <1.3-10.5 o o o □ CD O V V tA CM i— I CM lA O' CM fA CN a * CO fA vo CM (D co c O 3 + c .2 X 'to < r CM co Ov r—1 0 o CO u • cn CD o 0 II II Ol. > > LA CO • LA >4 0 CM CD CM CM u 01 Os (—( CO c i I > 0 LA VO fA fA IN co DC CM CD fA "co c tn o & (0 tx 4-) c 0 c 2 tn 0 '*3 LA 0 u c 0 CD O j fA o o 2 LA CM CM CM O CO D 0 Vo LA c C" 1*4o m—4 CO 0 CO ~ IA IA rA CO < 0 tx E C" 3 0 C JO cn CM CM CM CM o O' O' O' O' tn o E v_c / CO TD £ X > X LA jd c E E I 0 3 D _> 3 tx tx n 3 0 0 O 0 6 to tn tn H tn tn diet are shown in Table 6. The salivary values, as well as the serum values given in Section 4, are shown. Student's t-te s t was used to evaluate any significant difference for each solanum alkaloid parameter measured. In every case, a significant increase was noted in the high intake group. Table 6 5.4 DISCUSSION Salivary total alkaloid and solanidine concentrations in both the UK and Swedish groups correlate well with the serum analytes, the range of salivary alkaloid concentrations representing about 10% of the serum levels for total alkaloid and 20% for solanidine. In the case of endogenous steroids salivary values are 2-10% of the plasma values (219). The establishment of a salivary assay provides a non-invasive technique, allowing simple multi-sampling for the further study of solanum alkaloids in man. Further development of the salivary assays is needed to reduce the lim it of detection, since at present, the majority of samples measured have alkaloid concentrations approaching this lim it. Measurements of hormones and drugs in saliva have been recommended as an effective alternative to serum assays for low molecular weight analytes (216). It is thought that the analyte levels in saliva represent the free non- protein bound analyte, and give a better indication of biological activity than the total analyte. From Table 5 the mean serum solanidine, 14.5nmol/l is nearly a third of the mean total alkaloid, 38.5nmol/l. This difference is not reflected for saliva where the mean salivary solanidine is 3.2nmol/l and the mean salivary total alkaloid is 3.7nmol/l giving a salivary total alkaloid to solanidine ratio of 1.2:1.0. This indicates that the salivary alkaloid is predominantly the aglycone solanidine. Solanidine is very lipid soluble and has a molecular weigh to of less than 400. Steroids found in saliva are mainly the lipid-soluble unconjugated steroids (217). They enter saliva mainly via the intracellular route, by diffusion and their concentrations are independent of salivary flow rate. Although conjugated steroids are largely excluded from saliva, small amounts do enter via the ’tight junctions' and concentrations are dependent on salivary flow rate. The glycosides a-solanine and ot-chaconine, have molecular weights greater than 800, and more important, hydrophillic sugar residues and would be expected to be almost entirely excluded from saliva. Whether the lower glycosidic hydrolysis products could diffuse into saliva is unknown, although it is reasonable to expect these could be present in the small percentage of salivary alkaloid that is not solanidine. The differences in serum solanidine and total alkaloid between the Swedish group on a normal diet and those on a high alkaloid diet, discussed in Section 4, are again seen in the salivary levels. Both salivary solanidine and total alkaloid are significantly higher in the group on the high alkaloid diet. However, both in the normal diet group and the high alkaloid intake group, the means of the salivary solanidine and salivary total alkaloid are very similar, again pointing to the predominance of solanidine in saliva. The successful modification of the serum alkaloid assays permitted the measurement of both solanidine and total alkaloid in saliva. This has helped to investigate the type of potato alkaloid which can actually enter the saliva by virtue of its differing molecular structure. CHAPTER 6 AFFINITY CHROMATOGRAPHY CHAPTER 6 AFFINITY CHROMATOGRAPHY 6.1 INTRODUCTION Chromatography involves the separation of (a number of) the components of a mixture or purification of one of the components of that mixture. This is achieved by the differential migration of the components in a mobile phase through a medium, the stationary phase. The separation procedure in practice is normally carried out on a chromatographic bed in a column. The bed comprises small particles of the medium usually packed in a tube. The liquid mobile phase is made to flow through the bed by gravity, pressure or some other mechanical means. One of the many chromatographic techniques available is adsorption chromatography. This utilises a chromatographic medium which permits a more or less specific interaction with some or all of the components of the mixture to be resolved. Chromatographic adsorptions are based on the molecular interactions between the medium and the mobile phase consisting primarily of Van der Waals forces, electrostatic forces, hydrogen bonds and hydrophobic forces (220). Affinity chromatography is a type of adsorption chromatography in which the molecule is specifically and reversibly adsorbed by a complementary binding substance (ligand) immobilised on an insoluble support (matrix). Examples of substances which can be used as ligand to purify their respective binding substances are shown below: Ligand Binding Substance Enzyme Substrate, product, inhibitor, coenzyme Hormones Binding proteins, receptors Antibodies Antigen, virus, cell Plant Lectin Cell, cell surface receptor, carbohydrate, polysaccharides Nucleic Acid Complementary base sequences, histones, nucleic acid polymerase binding protein \ In addition, any of the binding substances can be used as ligand and vice versa, providing that easy attachment to the solid phase is possible. The technique of affinity chromatography requires covalent linkage of the ligand to a solid support, which is packed into a chromatographic bed. The principle of affinity chromatography is depicted in Figure 1. The solid support matrix used in affinity chromatography can consist of a number of macromolecular structures. They are usually in the form of a gel comprising a three-dimensional lattice with internal pores containing the liquid. A number of criteria must be fulfilled for the solid support to be successful (in affinity chromatography). Firstly, the support should form a loose porous network permitting a uniform and unrestricted entry and exit of large macromolecules. The particles should be uniform, spherical and rigid; these three factors influence the resolution and flow rate of the system. Interaction of the gel itself with the substance of interest must be small in order to minimise non-specific adsorption and therefore the gel should be chemically inert. The gel must be chemically and physically stable under the conditions employed during the chromatographic process. Finally, but most importantly, the gel must have functional groups which can be chemically utilised to form covalent linkages with the ligand (220). Figure 1 The Principle of Affinity Chromatography Ligand 4 | M ✓ a s t 'A Immobilization r y * i s x y 4 Adsorbtion / y o ► 1 ✓ A Wash y Regeneration Several solid materials have been used for affinity chromatography; these include cellulose, cross-linked dextrans, agarose, croSs-linked agarose polyacrylamide gels, polyacrylamide agarose gels, porous glass and ceramics. For the purpose of this study, discussion will be limited to the properties of agarose, porous glass and ceramics. Agarose is a linear polysaccharide consisting of alternating 1,3-linked 6-O-galactopyranose and 1,4-linked 3,6-anhydro-a-L-galactopyranose residues. It is available commercially in a bead form as Sepharose (Pharmacia Fine Chemicals Limited, Uppsala). The hydroxyl group on the sugar residues can be derivatised easily for covalent attachment to the ligand. The backbone of agarose is not joined together by covalent bonds but interacts via hydrogen bonding. The most widely used Sepharose is Sepharose 4B which has an exclusion lim it in gel filtration of 20 x 10^ daltons. This enables a wide range of ligands, even large molecules, to be used without reducing capacity or flow rates (221). Controlled pore glass or ceramic represents aerogels which do not shrink on drying and thus do not lead to reduced flow rates when used under pressure. Controlled pure glass is formed when certain barosilicate glasses are heated at 700-800°C and leached with acid. The boric acid component of the glass is readily etched out and subsequent controlled treatment with caustic soda produces a range of pore sizes from 43-2300 R. Porous glass is an insoluble packing matrix which is unaffected by changes in eluant, flow rates, pH or ionic strength. The main disadvantage of porous glass and ceramic is the non-specific surface adsorption of some proteins (219). The success of affinity chromatography is based on the use of purified ligand which can be linked to a convenient solid support, the ability to adsorb specifically the substance of interest from a complex mixture, removal of unwanted substances without disrupting the ligand/substance of interest interaction and finally, elution of the substance of interest to yield a regenerated ligand/chromatographic matrix for further use. The ligand ideally should have an affinity for the binding substance in the -4 -8 range of 10” -10” M in free solution. If the affinity of the ligand is too high, problems w ill be encountered eluting the binding substance; if it is too low, a problem is encountered in attaching sufficient ligand to the matrix, a lim it imposed by the matrix itself. Another important factor is the nature of the ligand. When using a small ligand for purification of a macromolecule, it is often necessary to use a spacer arm between the matrix and ligand to prevent steric hinderance of the specific interaction. The concentration of the ligand can have a number of effects. At too high concentrations, the binding efficiency can be reduced due to steric hinderance between active sites. Secondly, substances are more strongly bound to the ligand and therefore more difficult to elute, and finally non-specific binding is increased. For Sepharose 4B, a concentration of l-10pmoles of ligand per ml of gel is recommended (221). The chemical coupling of the ligand to the matrix can be carried by a wide range of techniques. One technique favoured for attaching ligands to agarose, dextran or cellulose involves activation of the carrier with cyanogen kcv&ctes and subsequent coupling of primary amines to the activated matrix. With Sepharose 4B, cyanogen bromide reacts with the hydroxyl groups to covert them to imidocarbonate groups which reacts with nucleophiles. There is also evidence that cyanate ester groups are present in activated Sepharose. The activation procedure produces greater chemical stability by cross-linking the Sepharose. The activated groups react with the ligand to produce isourea linkages. Sepharose 4B is available in a cyanogen bromide activated form which averts the use of toxic cyanogen halides in the user laboratory. Porous glass or ceramic can be linked to ligands by treating the glass with silane coupling agents. The silane reagents have an organic functional group at one end and a silylalkoxy group at the other. Commonly primary amino groups are introduced into the porous glass. The silanol derivatives of porous glass or metal oxide groups on ceramic are susceptible to activation w ith CNBr. The final step of affinity chromatography is the elution of the specific binding substance from the ligand without destruction, or disruption of the ligand-matrix bond. For the purpose of this study, the discussion will be limited to desorption from immunoabsorbents. The antigen-antibody interaction can be disrupted in a number of ways, depending on the nature of the solid support. Agarose type supports tend to be unstable outside the pH range 4-9 but are resilient to eluants containing high concentrations of salt, urea, water-miscible organic solvents, or guanidine hydrochloride. The porous glass and ceramic supports are unaffected by changes in pH or ionic strength (219). Typical desorbing agents used for immunoabsorbents are: 1. Changes in pH, both acid, ie, proprionic acid, glycine/HCL pH2.5 and alkaline with high salt concentration, ie, NaCl adjusted to pHll.O with ammonia. 2. Polarity reducing reagents, which affect hydrophobic interactions, ie, dioxane (<10%) or ethylene glycol (<50%). 3. Dissociating agents such as urea and guanidine hydrochloride. 4. Chaiotropic ions, which disrupt the structure of water and reduce hydrophobic interactions such as SCN , CCl^COO", I~, C10^~. In this study, the technique was intended for use firstly to produce a potato alkaloid free matrix for use in radioimmunoassay methods described in Section 2 and secondly to provide a means of concentrating potato alkaloids and their possible metabolites from body fluids, urine, serum and saliva to enable identification and quantitation by other chromatographic techniques. 6.2 MATERIALS AND METHODS 6.2.1 Materials A ffinity Chromatography Matrices Sepharose 4B (CN-Br activated), Sigma Chemical Company, St Louis, USA. Porous glass and controlled pore ceramic from Dr P Kwasowski, Department of Biochemistry, University of Surrey. Coupling Buffer Bicarbonate buffer 0.1M, pH8.3 containing 0.5M sodium chloride. 8.40g sodium hydrogen carbonate 29.22g sodium chloride Dissolved in 1 litre of glass-distilled water and the pH adjusted to 8.3. Stored at room temperature. Acetate Buffer Acetate buffer 0.1M, pH4.0 containing 0.5M sodium chloride and 0.1% sodium azide. 5.8ml glacial acetic acid 1.5g sodium acetate (anhydrous) lg sodium azide Dissolved in 1 litre of glass distilled water and pH adjusted to 4.0. Stored at room temperature. Glycine Buffer Glycine buffer 0.2M, pH8.0 containing 1% sodium azide. 15.02g glycine lg sodium azide Dissolved in 1 litre of glass distilled water and pH adjusted to 8.0. Stored at room temperature. Dialysis Buffer Bicarbonate buffer 0.1M, pH8.3 containing 0.15M sodium chloride. 42.0g sodium hydrogen carbonate 43.8g sodium chloride Dissolved in 3 litres of glass distilled water and pH adjusted to 8.3 and stored at room temperature. Glycine/HCl Buffer Glycine/HCl buffer, pH2.5 7.51g glycine 7.2ml concentrated hydrochloric acid Dissolved in 1 litre glass distilled water and pH adjusted to 2.3. Stored at 4°C. Sodium Chloride/Ammonia Sodium chloride 0.15M, pHll.O 8.76g sodium chloride Dissolved in 800ml of glass distilled water and pH adjusted to 11.0 with concentrated ammonia solution. Volume then adjusted to 1 litre. Stored at 6.2.2 Analytical Methods 6.2.2.1 Preparation of Anti-Solanum Alkaloid Immunosorbent Using Cyanogen Bromide-Activated Sepharose 4B The immunosorbent of the alkaloid antibody linked to Sepharose 4B was prepared as described by the manufacturers instructions (Pharmacia) (**»)• The antiserum used was HP/S/RG2-1A, the preparation of which was described in section 2.3.1.3. Initially the total protein and albumin content of the antiserum was estimated using a SMAII multichannel analyser (Technicon Instruments Limited). The globulin content of the antiserum was calculated by subtraction using the relationship total protein - albumin = globulin for estimation of the IgG fraction of the antiserum. It is recommended that 5- lOmg of specific protein per ml of gel are used for coupling. The antiserum HP/S/RG2-1A had a globulin content of 64mg/ml representing 1ml of antiserum per 12ml of gel. The gamma globulin fraction of the antiserum was precipitated using I anhydrous sodium sulphate. 180mg of sodium sulphate was added slowly to 1ml of antiserum with vigorous mixing to prevent localised non-specific protein precipitation. The mixture was incubated in a water bath at 23°C with occasional stirring. After 1 hour, the mixture was centrifuged at 3000rpm for five minutes and the supernatant discarded. The precipitate was then resuspended and washed in 2.5ml of 18% sodium sulphate. After further centrifugation, the precipitate was dissolved in 1ml of coupling buffer. The freeze-dried gel (3.4g) equivalent to 12ml was swollen on a grade 3 sintered glass funnel by washing with ImM hydrochloric acid. A volume of 200ml acid was used per gram of gel and the washing removed to waste using a vacuum pump. The slurry was then v/ashed with 17ml of coupling buffer and immediately transferred to a 30ml stoppered measuring cylinder containing a further 24ml of coupling buffer. The gamma globulin fraction was then quickly added and the contents of the cylinder mixed using an end over end mixer for two hours at room temperature or overnight at 4°C. The coupling reactions occurs best at pH8-10 when amino groups in the ligand are predominantly unprotonated. Coupling at lower pH is less efficient. After coupling, the gel was washed with a further 36mls of coupling buffer on a scintered glass funnel and then ■ resuspended in 36mls of glycine buffer in a measuring cylinder. This was mixed end-over-end for 2 hours at room temperature or overnight at 4°C. The glycine buffer has the effect of blocking any residual active groups left after coupling. The high salt concentration in the buffer counteracts any charged groups left on the gel by the blocking agents. The gel was then transferred back to the Sintered funnel and blocking buffer aspirated. Any excess adsorbed protein was then removed by alternate washing with 35ml aliquots of coupling buffer and acetate buffer. This was continued for several cycles, finally washing the gel with coupling buffer and storing the slurry (in coupling buffer) at 4°C until further use. 6.2.2.2 Preparation of Solanidine-Free Serum, Urine and Saliva The matrix effect found in radioimmunoassays necessitates the use of analyte free material to prepare standards (see Section 2.3.1.6). In order to compare standards prepared in assay buffer with standards prepared in analyte matrix supplies of alkaloid-free serum, urine and saliva were needed. The Sepharose 4B a-solanine antibody immunosorbent was used for this purpose. Blood was obtained from male donors at Lewisham Hospital Blood Transfusion Unit. Blood was collected in glass bottles containing no anticoagulant. After collection the bottles were stored overnight at 4°C to allow retraction of the blood clot. The supernatant serum was poured off into 23ml plastic universals and centrifuged to remove any remaining blood cells. The serum was then removed and stored in a domestic freezer until use. Urine was collected from healthy volunteer subjects after abstention from potatoes. The urine was pooled, centrifuged and stored in a domestic freezer until use. Saliva was also collected from healthy volunteers after abstention from potatoes. Pooled saliva was frozen, thawed and spun to remove any mucins (see section 5.2.1). The clear supernatant was stored in a domestic freezer un til use. To 25ml of each of the above matrices were added 250yl of the oi- solanine antibody immunosorbent. This was mixed in a plastic universal at room temperature on an end-over-end mixer for 1 hour. After mixing the immunosorbent gel was separated from the alkaloid-free matrix using disposable plastic mini-chromatography columns (Pierce Chemical Company, Cambridge). The gel being trapped above a plastic sinter, and the alkaloid- free matrix freely running through the column. The alkaloid-free matrix was assayed to ensure efficient alkaloid removal and stored frozen until use. The gel was washed on the column with several volumes of glass distilled water (= 100ml) and regenerated by repeated alternate washing with coupling and acetate buffer (25ml of each). Excess fluid was removed between washing by applying positive pressure with a 5ml syringe. Finally the gel was resuspended in coupling buffer and stored at 4°C until future use. 6.2.2.3 Preparation of Porous Glass and Controlled Pore Ceramic Immunosorbents Antiserum HP/S/RG2-Ia was coupled to controlled pore glass, and controlled pore ceramic by Dr P Kwasowski, Department of Biochemistry, University of Surrey. 500yl of each immunosorbent was prepared and stored in mini disposable columns under bicarbonate buffer at 4°C until use. 6.2.2.4 Affinity Extraction of Serum Samples and Removal of Alkaloid Cross-Reactants Serum samples were collected from subjects known to have a high potato intake. 30ml of serum was obtained and a sample analysed for total alkaloid concentration (section 2.2.2.3). Samples having a total alkaloid concentration greater than 20ng/ml (50nmol/l) was used for alkaloid extraction. 50ml of the serum was mixed with 500pl of Sepharose 4B solanidine antiserum immunosorbent in a plastic 25ml universal containersand mixed end-over-end for one hour at room temperature. After mixing the slurry was poured slowly onto a mini plastic disposable chromatography column. The Sintered filter retained the immunosorbent and the serum was collected. Any remaining serum was removed by positive pressure using a 5ml plastic syringe. 5ml Aliquots of distilled water were added to the column to a total volume of 20mls. After addition of each aliquot, the column was capped, the Sepharose resuspended and the water run off. Residue water was removed by applying positive pressure. Ten 1ml aliquots of 90% methanol were added to the column and run off with the application of positive pressure. The methanol was collected, pooled and evaporated to dryness using oxygen-free nitrogen. The dried down extract was redissolved in either 200pl acetonitrile (for HPLC, see section 7.2.2.3) or 200yl of 95% ethanol containing 0.5% acetic acid (for TLC see section 7.2.2.1). 6.2.2.5 Scale Up of Controlled Pore Ceramic Column for Affinity Extraction of 24-Hour Urine Samples In order to extract the cross-reacting alkaloids from 24h urine samples it was necessary to produce a larger immunosorbent column. This was needed in view of the large volume of urine, and the concentrations of the alkaloids found in the urine (see section 4.2.3 and section 8.3.2). Samples were collected from volunteer subjects known to be on a diet with a high potato intake. 25ml of antiserum HP/S/RG2-IA was mixed with 12.5ml of saturated ammonium sulphate. The ammonium sulphate was added gradually to avoid non-specific precipitation of protein. The mixture was left stirring for 1-2 hours and then centrifuged at 3000rpm for 10 minutes. The supernatant was discarded and the pellet resuspended in bicarbonate buffer 0.1M, pH8.3, containing 0.15M sodium chloride (lOmls). Dialysis tubing was prepared by heating in distilled water containing EDTA. The resuspended antiserum was then dialysed against 5 litres of bicarbonate buffer for five days. Buffer was changed on the third and fifth day. After dialysis the IgG cut was coupled to controlled pore ceramic by Dr P Kwasowski, Department of Biochemistry, University of Surrey. The total volume of the immunosorbent was 15ml and this was poured onto a 50ml capacity glass column containing a scintered filter. The apparatus for the affinity extraction is shown in Figure 2. The 24h urine in reservoir A was run through the column under gravity and collected in reservoir B. The flow rate was lOml/min. The urine sample was run through the column twice, and the excess urine removed with positive pressure. The column was washed with 250ml of distilled water. The column flow rate was adjusted to 5ml/min and 10ml aliquots of 90% methanol added to a total volume of 100ml. The aliquots were pooled, dried down under oxygen-free nitrogen and redissolved in either 5ml of acetonitrile (for HPLC) or 5ml 95% ethanol containing 0.5% acetic acid (for TLC). Figure 2 Apparatus for Affinity Extraction of 24h Urine Sampltg* A Urine •Solid Matrix Sinter Valve 6.3 RESULTS The initial assessment of the affinity extraction of the Sepharose, porous glass and ceramic columns was carried out by measuring the total alkaloid content of serum and urine samples after extraction. In all cases 5ml of serum or urine was added to each mini column containing 250pl of each gel. After extraction the total alkaloid content for serum and urine for each type of column was below the limit of detection of the total alkaloid assay (<1.0nmol/l). In developing the method of affinity extraction, several eluting solvents were tried to remove the alkaloid from the immunosorbent. To each disposable column containing 250pl of each immunosorbent was added lOOpl of tritiated solanidine label containing 10,000cpm per lOOpl. To this was added 1.0ml water and the fluid run off the column. This was followed by two further 5ml aliquots of distilled water. To a set of the columns containing each immunosorbent was added the following solvents: glycine/HCl buffer, pH2.5; sodium chloride/ammonia 0.15M, pHll.O and 90% methanol. 1ml aliquots of each of these solvents was added to each of the columns up to a volume of 13ml. Each fraction of water and solvent was collected and the radioactivity content counted. 1ml samples of each fraction were added to 4ml of scintillation fluid and r as in section 2.2.2. In addition, samples of the immunosorbent were counted (25yl). The results are shown in Figure 3. It can be seen that methanol was the only solvent which removed the radioactive solanidine from the immunosorbent. The majority of the label (=90%) was removed with 10ml methanol and this was the protocol adopted in the methodology. The efficiency of extraction was 86% for porous glass, 89% for controlled pore ceramic and 91% for Sepharose 4B. Sepharose 4B was chosen for the mini column extractions and controlled pore ceramic for the scale-up experiments. Figure 3 A ffinity Extraction of Tritiated Solanidine From Immunosorbents Using D ifferent Solvents ELUANT (above ELUANT ) (above A similar experiment was carried out on the 15ml controlled pore ceramic column. 100,000cpm of tritiated solanidine was added to 2 litres of glass distilled water and mixed. The labelled water was added to the column as 6.2.2.5 and run off. The column was then washed with 250ml water followed by 15 x 10ml aliquots of 90% methanol. Finally a further 250ml of water was added in 50ml x 5 aliquots. Samples of each (1ml) were counted for radioactivity and the percentage of total dose given calculated. This is shown in Figure 4 and it can be seen that 87% of the radioactivity is recovered in the first 100ml of methanol. Figure 4 Affinity Extraction of Tritiated Solanidine From Controlled Pore Ceramic ML ML Methanol _ o 003ZKW 6.4 DISCUSSION The antibody prepared for use in the immunosorbent was raised against an a-solanine-BSA conjugate linked via the sugar residue. The antiserum produced would most likely be directed to the nitrogen ring of the molecule and thus recognise any metabolites with differing sugar residues. This is born out by the cross-reaction studies which show 100% cross reaction with solanidine, a-solanine and a-chaconine (section 2.3.1.3). The choice of solid phase depended on a number of factors. The cyanogen bromide activated Sepharose 4B provided a relatively cheap, easy method for preparation of immunosorbent on a small scale. The controlled pore glass and ceramic were equally as effective but proved more difficult to couple to the antisera. When considering a large scale column, the sepharose was ineffective due to low flow rate caused by the non-rigidity of the solid phase. Controlled pore glass and ceramic were both rigid and gave good flow rates, but ceramic was chosen due to its reduced cost. The scale-up method provides a simple means of concentrating the small amounts of alkaloid found in the relatively large volumes of urine allowing further analysis to be made by other methods. CHAPTER 7 HPLC AND TLC OF POTATO ALKALOIDS CHAPTER 7 HPLC AND TLC OF POTATO ALKALOIDS 7.1 INTRODUCTION Many forms of chromatography separate components of a mixture by utilising differences in their interactions with the surface of a stationary phase. This is achieved by the establishment of an equilibrium between the compounds in the mobile phase and those in the stationary phase. In thin layer chromatography (TLC) the mobile phase, usually a solvent mixture flows through either a sheet of filter paper or a thin layer of silica to which the sample has been applied. High performance liquid chromatography (HPLC) is one of the most rapidly growing and potentially largest analytical tools for the research and diagnostic laboratory. The technique allows°Tapid and effective separation of a range of biological molecules. The basic components of an HPLC system are a solvent delivery system (mobile phase), consisting of a solvent reservoir and a pump, a sample injector, a guard column, an analytical column (the stationary phase) and a detection system (Figure I). The mobile phase of an HPLC system is delivered to the stationary phase from a reservoir via a pump. The solvents used must be kept free from dust and particulate matter and degassed to remove dissolved air which can cause baseline problems. Generally solvents of high purity are needed in HPLC. Selection of the mobile phase depends upon the nature of the stationary phase and of the sample. The mobile phase can be used in a isocratic system where a pure solvent or-a fixed mixture of solvents is used to elute the analyte. Alternatively gradient elution can provide a powerful tool Figure 1 Arrangement of Equipment for HPLC Sam ple In jecto r Guard Column 3 3 c Pressure gauge PUMP D etecto r Recorder WASTE S olvent R eservoir where a continuous change of solvent composition is achieved by programming the flow of two different solvents via two pumps into a mixing chamber^ Solvents used in HPLC range from aqueous solutions containing organic modifiers or buffers to pure organic solvents. Eluants are chosen to effect separation on the basis of solubility, polarity, donor and acceptor properties, pH and ionic strength (222). A strong solvent gives overall fast elution with polar stationary phases. Water and alcohols are the strongest solvents, the hydrocarbons the weakest. With non-polar stationary phases the weak solvents give the fastest elution. Mixtures of solvents have intermediate strengths. With chemically- bonded stationary phases solvent strength depends upon the exact nature of the stationary phase. The choice of solvents is limited by a number of factors. Firstly the detection system. If a UV detector is used solvents which absorb at the wavelength of interest cannot be used, nor can solvents which have a high UV cut-off when low wavelength detection is needed (223). Changes in the mobile phase composition can cause base-line d rift in the refractive index detectors and the mobile phase must have high electrical conductivity for use with electrochemical detectors. Strong acids and halides and mixtures of carbon tetrachloride with other solvents w ill corrode stainless steel and pH below 3 or above 8 will destroy bonded column packing. Other factors such as viscosity, flash point, toxicity and ease of disposal will also off-set choice of mobile phase. As im portant as the mobile phase in HPLC is the choice of stationary phase (223). This is usually packed in stainless steel columns. Modern HPLC utilises small rigid particles of diameter less than 10pm. These particles need special methods of column packing. Earlier columns used large 30-70pm porous beads which could easily be packed but resulted in low mass transfer and are now largely used for guard columns. The smaller rigid particles have a large surface areoLrelative to volume and therefore give improved mass transfer, but this is off-set by the need for greater operating pressures to pump the mobile phase (223). The m ajority o f stationary phases are based on silica. Although alumina, graphitised and non-graphitised carbon blocks, carbon coated silica gels, and styrene-divinylbenzene copolymers have also been used (222). Silica based stationary phases can be coated with organic molecules for use in partition chromatography. This was initially carried out by surface adsorption but gave poor column reproducibility. This has been solved by covalently linking the molecules by reacting the surface hydroxyl groups on the silica. Several bonded phases of this type have been produced. The most common are the non-polar octadecylsilyl (ODS or C18). Smaller carbon chain lengths G8 and Cl as well as more polar phases with cyanopropyl, aminopropyl, phenyl, glicydyl and alkyl groups are also available (224). Non-polar stationary phases used with polar eluants are known as reversed phase systems, polar stationary phases with non-polar eluant known as normal phase systems. The more polar phases mentioned above behave like reversed phase systems with methanol- water eluants and have been used to separate polar water soluble compounds. Some of these bonded phases can also be used as weak ion exchangers. Conventional ion exchangers and molecular sieve chromatography represent two other forms of HPLC. The stationary phase for ion exchange is usually porous polystyrene beads with fixed ionic groups. Ion exchange methods are generally not as fast or efficient as other methods of HPLC and separation can better be carried out using reverse phase bonded packings. Molecular sieve chromatography needs rigid porous particles such as spherogel for work under pressure and can be used to separate compounds such as proteins, enzymes and saccharides. The internal surfaces of the stainless steel columns are polished and lined with either glass or PTFE which prevents interaction of the mobile phase with the column wall. Modern steel columns are usually between 10 and 25cm in length giving sufficient theoretical plates (see below) for most separations. More recently radi ally compressed plastic columns have been available removing any voids in the packing by external pressure on the column walls (222). For efficient HPLC separation the tubing, connectors and fittings of an HPLC system must conform to the highest specifications. Minimum dead volume is needed throughout the system with fittings being inert, non- leachable and able to stand high pressures. Solvent and sample clean-up is important since particulate matter can cause blockages in the fittings. The pumps used in HPLC need to give a constant flow of mobile phase at the required pressure. Modern pumps are microprocessor controlled, dual head reciprocating piston pumps. Again, blockages in the pumps interfere with the efficient HPLC system. Injection of relatively large sample volumes 20-200pl is carried out by means of sample loops. The sample can be cleaned up prior to injection and dissolved in the mobile phase to avoid alteration of column conditions. Finally, the detection system for HPLC. Ideally this should be sensitive, have a linear reproducible response over a wide concentration range and have a low cell volume to avoid dilution of fast peaks. The most frequently encountered detectors are spectrophotometers. These may be either a fixed wavelength or wavelength scanning detectors. Instruments now have the capability of operating at low wavelengths down to 190nm, necessitating UV transparent mobile phases. Fluorogenic detectors can be many times more sensitive than light absorption detectors. There use is limited by the number of molecules having natural fluorescence, although derivatisation of compounds to produce fluorescent compounds may overcome this. The next important type of detector is the electrochemical type. This can be used to separate molecules that do not have a suitable wavelength absorption or which are present at very low concentrations. The detectors can be up to 1000 times more sensitive than UV detectors. They can detect electro-active analytes which cause an electrochemical reaction at the surface of the working electrode. Electro-active molecules inctide those containing aldehyde, quinone, phenol, tertiary amine and sulphur groups. Mobile phases fo r use w ith electrochem ical detectors ' be free from electroactive species, stabilisers or modifiers. In addition, the mobile phase must be weakly conductive and have a pH maximising the electro-active nature of the analyte (223). Other detectors for HPLC have been used, but often present problems! Refractive index detectors are not sufficiently sensitive or specific for routine clinical analysis. The use of mass spectrometry requires collection of the column fractions and often removal of the eluate before analysis. On line mass spectrometry detection is now being developed. As well as the general detectors listed above, it is often possible to detect specific analytes in column fractions by routine chemical methodology. Whatever form of HPLC system krused the same features for efficient separation are needed. Firstly, the method should be selective giving a clear-cut separation of individual peaks, be efficient producing sharp peaks and have the capacity to separate complex mixtures in a short time. The separation of peaks is defined by the separation factor a which is represented as: ct=tR?- to ------l R, I - t o where tn and tD are the retention times of the samples and to the retention Rj R2 times of the solvent front. The efficiency N (otherwise known as the theoretical plate number) is defined as: N = 1 6 t^ where t^ is the retention time and W the peak width. As mentioned earlier most separations require between 3000 and 10,000 theoretical plates. The time taken for the separation is related to the capacity of the column. The capacity ratio: K' - lR • l0 lO Finally, the resolution of the two or more peaks, Rg, is defined as: Rs = (tR 2 " fcR ^ x 2 (wj + W2) the ratio of the separation between the peaks divided by the average peak width. Application to the Measurement of Potato Alkaloids TLC and HPLC have provided useful analytical methods for the investigation of potato alkaloids. The methods have been described in detail in section 1. TLC has been used to both q u a lita te ^ n d quantitate both the major potato glycoalkaloids and their plant metabolites, as well as the other minor alkaloids found in the potato plant (9, 70). The main differences in the methods being the choice of solvents. Most methods having relied on rather non-specific detection methods which will detect a very wide range of alkaloids and steroids (72, 78, 79, 80). HPLC of potato alkaloids has been a more recent advancement. The potato glycoalkaloids present one ojp the classic problems for HPLC. Detection needs to be carried out at wavelengths less than 210nm which lim its the type of detectors and solvents which can be used. Several types of columns have been used to separate glycoalkaloids, from reversed phase Cg and C^g silica columns (93, 9) to NH2 columns (94) and carbohydrate columns have been used to separate the glycosides (94). Because of the wavelength lim it the commonest solvents used are acetonitrile and water with some minor additives. Although many of the other methods can achieve partial separation o f many of the glycosides, no one method has been developed which can resolve the aglycone, along with a, 8 and lS glycosides of solanine and chaconine. In order to investigate the metabolism of potato alkaloids in humans an HPLC method was developed to be used in conjunction with TLC and affinity chromatography. 7.2 MATERIALS AND METHODS 1\2A M aterials Acetonitrile HPLC S grade was obtained from Rathbern Chemicals, Peebleshire, Scotland. All other solvents used were of Analar grade from British Drug Houses, Poole, Dorset. All solid chemicals were laboratory grade. Alkaloids Solanidine, a-solanine and a-chaconine were obtained from Dr M R A Morgan, AFRC Institute of Food Research, Norwich Laboratory, Norwich. 7.2.2 Analytical Methods 7.2.2.1 Hydrolysis and TLC Separation of Potato Alkaloids The method used fo r the hydrolysis and TLC separation of a - solanine and a-chaconine was an adaptation of the method of Filadelfi and Zitnak (85). 5mg samples of a-solanine^and~oFchaconine~were-dissolved-each— in 5ml of 95% ethanol containing 1% hydrochloric acid, in a glass test tube. The tubes were covered with cling film and heated in a water bath at 75°C. Every five minutes 200gl aliquots were removed from each tube over a one- hour period. Hydrolysis was terminated in each aliquot by adding 400pl of 2% aqueous ammonium hydroxide. The aliquots were then evaporated to dryness and the residues redissolved in 200pl of 95% ethanol containing 0.5% acetic acid. Thin layer chromatography was carried out on 20 x 10cm plastic silica gel G (Camlab, Cambridge, UK) plates. The solvent was prepared by shaking a 2:2:1 mixture of chloroform:methanol: and 1% aqueous ammonium hydroxide in a separating funnel until the solvent was saturated. The lower organic layer was left in the tank and allowed to equilibrate for three hours before use. 20pl of each of the hydrolysates prepared above were applied to the TLC plates (3 per plate) along with a plate containing the standards a- solanine, a-chaconine and solanidine (1.5pg of each dissolved in 95% ethanol containing 0.5% acetic acid). After development, the plates were dried and then sprayed with a solution of antimony trichloride in glacial acetic acid (25% w/w) using a fume cupboard. Spots were visualised by heating the plates in an oven at 85° for 15 minutes. A further set of 20 x 20cm TLC plates was prepared. The first contained a mixture of 20pl of the 50 minute hydrolysate of a-solanine and 20pl of the 50 minute hydrolysate of a-chaconine run alongsi3eH:he~standards a-solanine, a-chaconine and solanidine. This plate was sprayed as above and Rf values calculated for the visualised spots and compared with reference values. Each of the other plates had one (40pl) application of the s50 minute hydrolysate mixture along with five lanes containing the 200yl of the other timed hydrolysates from a-solanine and a-chaconine. After development the lanes on each plate containing the hydrolysate mixure were cut from the rest of the plate and the spots visualised as above. The plates were then reassembled and the visualised spots used to locate the unstained glycosides from the other hydrolysates. These were then scraped from the plates, the scrapings from each separate glycoside pooled and the individual glycosides eluted from the silica with methanol. This provided a set of standard glycosides for use in HPLC and further TLC analysis. The above TLC method was used to qualitate the glycoalkaloids obtained from the affinity chromatography of human serum and urine described in Section 6. 50pl samples of extract dissolved in 95% ethanol containing 0.5% acetic acid were applied to TLC plates. 7.2.2.2 Spectral Analysis of the Potato Alkaloids In order to assess the optimum wavelength for the HPLC detection of the major potato alkaloids, a-solanine, a-chaconine and solanidine, a spectral scan was run on each of these three compounds. The scans were run on a Varian 2200 spectrophotometer from a wavelength of 230nm to 190nm. Standard solutions containing solanidine (2mg/ml) in methanol, a-chaconine (1 mg/ml) in methanol and a-solanine (1 mg/ml) in methanol, prepared for radioimmunoassay (see section 2 ) were used to give lpg of each alkaloid. The methanol was evaporated to dryness and the standards re-dissolved in 5ml of a solution of acetonitrilerwater, 60:40 v/v. The acetonitrile was HPLC grade, water glass distilled and both filtered and sonicated before use. Each standard solution was scanned against a water blanks In~~addition~a~scan-was— carried out on acetonitrile, water and acetonitrile:water, 60:40 respectively, each against a water blank. 7.2.2.3 HPLC of Potato Alkaloids Apparatus The HPLC system consisted of a Waters Assoc. Lambda-Max model 481 LC spectrophotometer (Waters Assoc., Milford, Mass, USA), a Waters V6K Injector, Waters model 510 pumps, Waters model 680 automated gradient controller and a 120 servogar chart recorder. The column was a 7.8 x 30cm stainless steel column containing a microbondapak NH2 packing (Waters Assoc.). Method The solvents used were acetonitrile HPLC grade, glass distilled water and phosphoric acid analar grade. Before use all solvents were degassed by sonicating and filtered to remove any particulate matter. Filters of pore size 0.5p were used for the water and 0.45p organic filters for acetonitrile. The mobile phase was acetonitrile:water 80:20 v/v adjusted to pH4.0 with phosphoric acid. The flow rate was l.Oml/min; ambient column temperature; wavelength 200nm; attenuation 0.05 AUFS and chart speed lcm/min. Before use the column was lagged with cotton wool to maintain column temperature. Samples were injected with a 250pl Hamilton gas tight glass syringe. Samples of a-solanine, a-chaconine, solanidine, the lower glycosides and the human body fluid samples were prepared as follows: Samples of radioimunoassay standards of a-solanine, a-chaconine and solanidine in methanol were evaporated to dryness to give 2pg of each. These were redissolved in 200pi of mobile phase, filtered and lOOpl samples injected onto the column. 20pl samples of the hydFolysis~pFbducTs~Trf~rx=solanine-and-a— chaconine (see section 7.2.2.1) at the 50 minute incubation were evaporated to dryness and redissolved in 200pl of mobile phase. The samples were filtered and lOOpl injected onto the column. Samples of human serum and urine were subjected to affinity chromatography (see section 6.2.1) eluted with methanol and re-dissolved in mobile phase. A fter filtration lOOpl samples were injected onto the column. 7.2.2.4 Development of HPLC Method The column was initially purged with laboratory grade hexane (200ml) at 254nm followed by equal volumes of analar grade chloroform, HPLC grade acetonitrile and glass distilled water. All solvents were filtered before use! Once a steady baseline had been achieved with acetonitrile, the wavelength was reduced to 200nm. The initial assessment of retention times was carried out using radioactive solanidine. 20pl of tritiated solanidine (see section 2 ) containing roughly 20,000cpm was evaporated to drynes and re-dissolved in 200pl of acetonitrile. 50pl was used for total tubes and lOOpl was added to the column. One m illilitre fractions were collected from the column up to a total volume of 25mls, representing a retention time of 25 minutes at a solvent flow rate of lm l/m in. All other column conditions were as in Section 7.2.2.3. This was repeated for two different mobile phases. One containing acetonitrilerwater, 80:20 v/v, the second, acetonitrile:water, 60:40 v/v. To each 1ml fraction was added 4ml of liquid scintillant.— To-the total tubes was added 950pl of the mobile phase and 4ml of scintillant. These were then counted as described in section 2.3.1.2. Figure 2 shows a plot of percentage radioactivity counts against time. As can be seen from the chromatogram Figure 2, the radioactive peaks for both mobile phases coincides with peaks on the chromatogram. The retention times for solanidine can be seen to be five minutes fo r the 80:20 mobile phase and three minutes for the 60:40 mobile phase. The solanidine peak with the 60:40 mobile phase appeared just before or more likely in the solvent front. It was therefore decided to use the 80:20 solvent system. Figure 2 HPLC Retention of Radioactive Solanidine C h r o m a JML Solanidine t o 9 J^Lsolvent front r a m Flow Rate 1-0 ml^min W avlength 200nm 0 05 AUFS 100 9 0 6(^40 mobile phase 8 0 /2 0 80 70 60- C 50 o u n t s 30 10 10 Time (minutes) The injection of radioimmunoasssay standards of a-solanine and a- chaconine onto the column with the 80:20 mobile phase showed no peaks apart from the solvent front. The pH of the mobile phase was then altered using potassium dihydrogen phosphate. This was added to give a pH of 4.0. However, after an initial separation of a-solanine and a-chaconine spurious peaks and basline drift was encountered due to precipitation of the potassium phosphate in the column and pump values. The column was regenerated by washing first with 75ml of distilled water followed by the same volume of methanol, chloroform, methanol and a final wash with water. The pH was then adjusted using analar grade phosphoric acid. Separation of a-solanine and a-chaconine was seen using the 80:20 mobile phase at pH3.5, 4.0 and 4.5. The retention time for the different pHs for a- solanine and a-chaconine are shown in Figure 3. A t pH3.5 and 4.5 separations of a-solanine and a-chaconine are seen, but the maximum separation is achieved at pH4.0. In view of this separation, the pH of 4.0 was used in further HPLC analysis since although retention times were higher, it would allow greater separation of the lower glycosides (see section 7.3). Figure 3 Retention Times for ct-Solanine and g-Chaconine at Different pHs 2 2 - ■ R ® 2 0 — e n l 8 _ ° 1 6 - ■ T 1 4 - I m • e 1 2 - ■ • *Pj 1 0 - n t 8 - • 6 - 4 — 2 - pH 3-5 pH 4-0 pH 4 -5 • ^-CHACON INE ■ otSOLANINE 7.3 RESULTS 7.3.1 Spectral Analysis Figure 4 shows the spectral scan of the aglycone, solanidine and the glycosides, a-solanine and a-chaconine. The peak for solanidine occurs at 200nm, below this between 190-195nm a-solanine and a-chaconine have peak absorbance. However, due to the interference from sugar residues at this level the wavelength of 200nm was used for routine analysis. 7.3.2 TLC of Potato Alkaloids in Serum and Urine Figure 3 shows the TLC analysis of the various timed hydrolysates described in section 7.2.2.1. Figure 5(a) shows the increasing TLC runs of samples from hydrolysis of a-solanine and a-chaconine from O' to 50'. The 50 minute incubation is needed to show all the lower glycosides as well as solanidine. This is longer than the 45 minutes described by Filadelfi and Zitnak (85). The rf values compare well with the original paper and are given in Table 1. The presence of a-solanine, a-chaconine and solanidine is confirmed by running pure standards on the same plates as the hydrolysates, as shown in Figure 5(b). The TLC analysis of the serum and urine extracts prepared by affinity chromatography (section 6.2) are shown in Figures 6 and 7. In addition to the extracts, pure a-solanine, a-chaconine and solanidine are run on each plate plus the 50 minute hydrolysate. The urine and serum extract separations show the presence of a-solanine, a-chaconine, solanidine as well as the lower glycosides, $ 2>,tf“Solanine, and chaconine. 7.3.3 HPLC Analysis of Potato Alkaloids in Serum and Urine HPLC analysis of a mixture of pure standards of a-solanine, a- chaconine and solanidine using the system described in section 7.2.2.3 is shown Figure 4 Spectral Scan of Solanidine, a-Solanine and g-Chacnnirm A B 1 = SOLVENT S O 2= SOLANIDINE R B A 3=b^SOLANINE, t/CHACONINE N C E 190 200 220 Wavelength (nm) Figure 5 TLC Analysis of Hydrolysis Products of ct-Solanine and ct-Chaconine 5a 5b SD Solvent Front H S OL C d S 45 50 S Solanine SD Solanidine C Chaconine Table 1 R f Values for the Glycosides of Solanidine Glycoalkaloid Rf Solanidine 0.95 g'-Chaconine 0.70 32~Chaconine 0.55 32*Chaconine 0.42 a-Chaconine 0.31 -Solanine 0.66 32~Solanine 0.38 a-Solanine 0.27 Figure 6 TLC Analysis of Serum Affinity Chromatography Extract SOLVENT FRONT SD ©£ C U S 5 0 ' SERUM PURE HYDROLYSATE EXTRACT STANDARDS Figure 7 TLC Analysis of Urine Affinity Chromatography Extract SD U C 50 URINE PURE HYDROLYSATE EXTRACT STANDARDS in Figure 8. A good separation of a-solanine and a-chaconine is achieved with the solanidine peak occurring just after the solvent front. Figure 9 shows the chromatogram obtained after injecting the 50 minute hydrolysate. The a- solanine and a-chaconine peaks are small but easily identified by their retention times. The solanidine peak is also clear giving a much higher response than the glycosides which is to be expected in the sample after a long period of hydrolysis. Several other peaks are seen, some of which blend into each other giving double peaks. By injecting the individual glycoside standards prepared in section 7.2.2.1, two of these peaks were identified as 8^ chaconine and 6-solanine. The retention tim es are given in Figure 9 and it can be seen that the order of elution is solanidine, 8^ chaconine, 8-solanine, a-chaconine and a-solanine. The identity of the other peaks was unresolved due to problems encountered with column contamination when injecting the hydrolysates. The small peaks bordering the solanidine peak could represent the ^-glycosides but could not be positively identified. Figures 10 and 11 show the typical chromatograms after injection of the affinity chromatography extracts of serum and urine. Again, a-solanine, a-chaconine and solanidine peaks are seen but with relatively more a-solanine and a-chaconine to solanidine in the urine than found in the serum. 8j chaconine and 8-soian®ine peaks were also seen but no other peaks were able to be identified. uioe-31. ro i—i sf Solvent Front FLOW RATE 1-0 ML/MIN WAVELENGTH 200 NM sd Solanidine Ce(Chaconine 0-05 AUFS Se^Solanine HPLC Analysis of a Mixture of ct-SoIanine, ot-Chaconine and Solanidine Figure 8 1 -o 1 .® O r- S3 CO CO O CO o r~ > CO m O H 10 m < z z z - > o o O z m H ■n m 0 z z 1 33 z O H j HPLC Analysis of 50 Minute Hydrolysate C o HPLC Analysis of Serum Affinity Chromatography Extract C3 CO CD CO m o o r- H > z z m in •n o co o o r* X < > m o z c H z ■n z 3D m O Z H "CD o o X > K o CO o o I— z > z z m z m co a co r~o > Z z m Figure 11 HPLC Analysis of Urine Affinity Chromatography Extract CD CO m o OP H CO o O i" z > z z m CO S3 (0 o o r~ < m z H o T1 X X > O o z o H z z m CO D CO o r- > Z z CO m CO o r“ > z o z m G> o X > o o z z m ' 7.4 DISCUSSION The HPLC and TLC analysis of the potato alkaloids in serum and urine provided some information as to the possible metabolism of the glycosides, a-solanine and a-chaconine in the body. The TLC method of Filadelfi and Zitnak (84) provided simple means of obtaining the lower glycosides as standards. The TLC separation of the urine and serum affinity chromatography extracts demonstrated the presence of all the lower glycosides. This is to be expected, since the most likely first stage of metabolism of a-solanine and a-chaconine is the hydrolytic cleavage of the sugar residues eventually yielding solanidine. Further evidence for this metabolism to solanidine is seen in the pharmacokinetic study 8.3, 8.4 where serum solanidine concentrations rise more slowly than the triglycosides, and the ratio of glycosides to solanidine is much higher than that found in the potato. The HPLC study was limited by the use of only one column, since as seen in section 1.7.3, several columns and a variety of conditions have been used to separate the various glycosides and solanidine. The NH^ pbondapak column was chosen for the ability to interact ionically, as well as in the normal phase. A t the acid pH used the NH^ m oiety would be expected to be in the protonated form, possibly retaining the more hydrophillic a-solanine and a-chaconine rather than the lipophilic solanidine. The order of elution could also be explained in terms of the lower molecular weight of solanidine. The major problem with the column was the constant contamination resulting in a drifting base-line and spurious peaks after injection. The original aim of the study to identify the peak and then quantitate them by RIA proved impossible because of this column instability. However, the study did demonstrate the HPLC separation of solanidine, a-solanine, a-chaconine and the $-glycosides from the serum and urine extracts confirming the results found by TLC. This is in agreement with the work of Nishie et al., (119) and Alozic et al., (127) who demonstrated the presence of a-chaconine (127), a-solanine (119) and solanidine and several intermediate polarity metabolites in the urine of rats (119) and hamsters (127). It is quite likely that several other metabolites of a-solanine and a- chaconine exist in the body. The possession by solanidine of the A and B rings similar to many endogenous steroids and indeed, a similar metabolic pathway for their production (section 1.3) support this theory. However, the elucidation of these metabolites would need further development of the separation techniques plus methods of identification which is discussed in section 9. CHAPTER 8 PHARMACOKINETICS OF POTATO ALKALOIDS CHAPTER 8 PHARMACOKINETICS OF POTATO ALKALOIDS 8.1 INTRODUCTION Pharmacokinetics is the study of the time course of drug and metabolite levels in different fluids, tissues and excreta of the body. For the purpose of this study, the potato alkaloids can be considered in the same light as drugs, since they are not naturally occurring compounds in the body, but are part of the dietary intake. The pharmacokinetics of the potato alkaloids have been studied in both men and experimental animals. Fate and distribution studies using tritiated solanine administered to male Fisher rats showed poor absorption from the gastrointestinal tract, rapid elimination in the urine and faeces, and peak tissue concentrations r-yU) within 12 hours (119). A dose of 5mg/kg was administered orally and its fate followed over a 4-day period. By 24h, only ‘ 10% of the radioactivity remained in the gastrointestinal tract, 72% was lost in the faeces and 6% in the urine. Assuming no enterohepatic circulation of the label, this indicated poor gastrointestinal absorption. The remaining radioactivity was distributed in various tissues in the following descending order: spleen, kidney, liver, lung, fat, heart, brain and blood. Intraperitoneal administration of 5-15mg doses of tritiated solanine showed only 34% excretion in the faeces and urine. Above 15mg excretion of radioactivity rapidly dropped and slight to moderate ascites were seen in the peritoneal cavities of the animals. Similar results were seen when tritiated chaconine was administered orally to rats (126). Faecal elimination was rapid (80% within 48h). High levels of radioactivity were found in the liver, intermediate in the kidney and lung with low levels in the blood, brain, fat and heart. When labelled a- chaconine was administerd intraperitonealiy, radioactivity was largely eliminated in the urine. Another study in hamsters (127) showed good absorption of tritiated a- chaconine from the gastrointestinal tract. Less than 21% of the label was excreted via the urine and faeces in 7 days. Of this 20% was excreted in the urine with only <1% in the faeces. This is in marked contrast to the rat studies described above (119). Tissue concentrations were highest in the lungs, liver, spleen, skeletal muscle, kidney and pancreas (118). After intraperitoneal administration, tissue concentrations were much higher. The dosage used in the hamster study was lOmg/kg. In hamsters injected with tritiated solasodine, an aglycone of the spirosolane type the rate of tritium excretion in the urine was initially 9.5% in the first 24h (128). This declined with a mean half-life of 45h and after 200h the excretion rate had fallen to about 1% of the initial rate with only 26% of the dose having been excreted. Tritiated label was still detectable in liver and adipose tissue 62 days after injection, indicating prolonged body retention. The same tritiated alkaloid was injected into two human volunteers. The label was rapidly taken up by erythrocytes, after five minutes, levels in plasma and erythrocytes being similar. The levels decreased in both subjects with an initial rapid decline followed by a much slower second phase. After eight days, erythrocytes contained twice as much radioactivity as plasma. U rinary and faecal excretion accounted fo r 7% and 20% of the dose and had virtually ceased within five days. The pharmacokinetics of the aglycone, solanidine, was studied by the same authors (129). Tritiated solanidine, administered by intravenous injection showed three phases of plasma clearance, having half-lives of 2-5 minutes, 120-300 minutes and 70-150h. Initially, only 10% of the dose was found in the plasma, declining rapidly in the first phase of clearance. As with solasodine, the label was rapidly taken up by the erythrocytes. Excretion rates were low, a fte r 24h only 1-4% of the dose had been excreted in urine and 1-3% in faeces. During the following week, combined excretion rates were roughly 2% per day. The above studies illustrate the diversity in the pharmacokinetics of alkaloids when different models are used. The main differences are seen between the administration of the glycosides, a-solanine and a-chaconine, and the lipophilic aglycones, solasodine and solanidine. In all cases tritiated compounds have been used rather than the alkaloids in their natural form, namely in foodstuffs. A variety of species were used, with different routes of administration. The time course of the experiments was also inconsistent, as is the type of tissue used to evaluate uptake. In summary, it is very difficult to make a valid comparison of the experiments described. Because of this problem it was decided to carry out a pharmacokinetic study of potato alkaloids in humans following administration of the alkaloids in the form of a potato meal. This should allow the accurate estimation of alkaloid absorption, distribution and excretion to be elucidated and hopefully explain some of the differences found in the above studies. 8.2 MATERIALS AND METHODS 8.2.1 Materials Solvents All solvents used were obtained from British Drug Houses, Poole, Dorset, and were of Analar grade. Potatoes The potatoes used in this study were Cyprus new potatoes obtained from a local supermarket during the month of July. 8.2.2 Methods 8.2.2.1 Preparation of the Potato Meal The potato load was provided in the form of a meal consisting of 500g cooked in a microwave. The whole potato including the skin was eaten. Each 500g portion of potatoes consisted of an equal number of similar sized tubers. In order to make the meal more palatable, a* small amount of butter, and boiled minced beef was provided with the potatoes. Samples of the uncooked and cooked potatoes were frozen for analysis of total glycoalkaloid. Although the potato alkaloid concentrations of the butter and beef were not estimated, this can be considered negligible since this type of food was eaten in an earlier potato avoidance study (142) yet serum solanidine levels after potato avoidance still reached a minimum. 8.2.2.2 Pharmacokinetic Studies The subjects studied were four healthy volunteers; two males and two females, ages ranging from 20-28 years. One month prior to the potato dosage, each subject collected a 24h urine sample starting at 9.00am and finishing at 9i00am the following morning. The early morning urine collected prior to the start was discarded and the early morning urine at the end of the 24h period included in the collection. During this 24h period, a blood sample was taken between 11.00-12.00am following commencement of the 24h urine sample. The urine samples were stored frozen in a domestic freezer. Blood samples were centrifuged and the serum stored frozen until assay. At weekly intervals following this date, blood samples were collected from each subject between 11.00-12.00am. During the month prior to potato dosage, all subjects were put on a strict potato-free diet. This consisted of avoiding the consumption of any potatoe or potato products including processed food likely to contain potatoes or potato starch. Total alkaloid and solanidine were estimated in the blood samples to ensure that alkaloid concentrations had reached a minimum level. At 12.00am on the day of the pharmacokinetic study each subject emptied'his or her bladder and started a 24h urine collection. A pre-dose blood sample was also taken at this time. A potato meal was then eaten which consisted of 500g of whole cooked potato (see above). Blood samples were collected every hour for the next 24h. At the end of this period, each subject then commenced a potato-free diet for at least 14 days after dosage. Blood and 24h urine collections were made daily for seven days following the potato load. 8.2.2.3 Analysis of the Total Glycoalkaloid Content of Potatoes One cooked and one uncooked frozen tuber (stored from above) were selected for analysis. Each tuber was weighed and immediately chopped into small pieces while still frozen. The chopped material was transferred to liquid nitrogen and left for 5 minutes until a powdery consistency was obtained. The powder was kept frozen prior to extraction which was carried out in a Ultra- Turrax blendeK lOg of each sample of frozen powder was extracted for two minutes with 50ml of solvent (methanol:water: glacial acetic acid; 96:6:1 v/v). After extraction the mixture was centrifuged and the supernatant retained for total glycoalkaloid analysis. The alkaloid extracts were analysed for TGA content at the AFRC Food Research Institute, Norwich, using the ELISA method of Morgan et al., (76). 8.2.2.4 Analysis of Alkaloid Content of Serum and Urine The immunoassays for the measurement of solanidine and total alkaloid in serum and urine and the materials used were those described in section 2.2.2. In all assays, selected internal quality control materials were used. The intra- and inter-assay variations are given in Section 2.3.1.9. 8.3 RESULTS 8.3.1 Total Glycoalkaloid Content of Potato Samples The two samples analysed b|J ELISA had a total glycoalkaloid content as follows: Raw whole tuber = 3.15mg TGA/lOOg Cooked whole tuber =3L*8S mg TGA/lOOg The total dose received by each subject in the meal of 500g potatoes was 14.25mg of total glycoalkaloid. Assuming an average body weight of 70kg, this represented an oral dose of 0.23mg/kg. 8.3.2 Analysis of Serum and Urine from the Pharmacokinetic Study The results from the pharmacokinetic study can be divided into three phases: the pre-dose potato-free period, the 24h period following the potato load and finally the second potato-free period of seven days'following the potato load. The results are shown for each individual subject, however, samples were not obtained in every case at the specified time consequently some values for the alkaloid concentrations are not known. Figure I gives the results of the pre-load potato-free period, showing declining concentrations for both solanidine (Figure la) and total alkaloid (Figure lb). The half-life (tf) is the time taken for the plasma concentration or the amount of drug in the body to be reduced by 50%. Many drug concentrations in plasma follow a multi-exponential pattern of decline. The half-life that is usually reported is that which corresponds to the terminal log-linear rate of elimination. The half-lives of the four subjects in this study are given in Table 1 for both solanidine and total alkaloid. The half-life for Figure 1 Concentrations of Solanidine and Total Alkaloid Following Potato Avoidancae Figure 1(a) SOLANIDINE 100n 75 50 A L 2 5 - K 2 2 -5 - A 20*— L 17-5- 0 1 D 12*5 n 10 m o J, , 3 TIME (Weeks) 4 3 2 1 TIME (Weeks) TOTAL ALKALOID 0 75 50 7-5 100 12*5 O 17-5 A A 22-5 c E o Figure 1(b) Table 1 H a lf-L ife in Days of Solanidine and Total Alkaloid Prior to 24h Pharmacokinetic Study SUBJECT A B C D Solanidine 3.5 4.9 9.0 5.6 Total Alkaloid 5.3 6.7 12.5 9.1 solanidine ranges from 3.5 to 9 days and that for total alkaloid from 5.3 to 12.5 days! Figure 2 shows the serum solanidine concentration over the first 24h of the pharmacokinetic study and Figure 3 the corresponding serum total alkaloid concentrations over this period. The solanidine concentrations started to rise five hours after ingestion of the potato load and reached a steady state eight hours later. Maximum solanidine concentrations reached were 12.5 to 15.0nmol/l. The total alkaloid concentrations show a different picture. There was an initial rapid rise in total alkaloid concentrations reaching a peak four hours after the potato load. The peak concentrations achieved were 54 to 67.5nmol/l. This was followed by a trough in total alkaloid concentrations which then rose again to give a second peak eleven hours after the load. The steady state of 60-70nmol/l in the four subjects was then achieved which persisted until after 21 hours, when total alkaloid concentrations started to decline. The solanidine and total alkaloid concentrations for the seven-day period following the potato load are given in Figures 4(a) and 4(b) respectively. The decrease in serum concentrations are seen to be similar to Figure 1. The half-lives for the solanidine and total alkaloid following the 24h pharmacokinetic study are given in Table 2. The half-lives for this period were measured in only two of the four subjects, but give similar values to those found in Table 1. A similar study of potato avoidance carried out by the author (Harvey, 1983) showed a decrease in serum solanidine concentrations (Figure 5); the half-life in this subject was 8.5 days again within the range found for solanidine in Table 1. The urinary solanidine concentrations measured before, during and after the 24h pharmacokinetic study are shown in Figure 6. A wide range of Figure 2 Concentrations of Solanidine Over a 24h Period Following the Potato Load #= Subject A □ = Subject B ■ = Subject C A= Subject D <111 411 TIME (Hours) — O «=Eo — Figure 3 Concentrations of Total Alkaloid Over a 24h Period Following the Potato Load :• = Subject A □ = Subject B ■ = Subject C ▲ = Subject D <' TIME (Hours) Jqo(0 0) c E o — Figure 4 Concentrations of Solanidine and Total Alkaloid in the 7-Day Period Following the Potato Load Figure 4(a) SOLANIDINE 100 75 50 A L K A L ° 25 • 22-5 D 20 n 17-5 m o 15- ^ 12-5 10 7-5 Time (days) l O ® -yOlP^' Table 2 Half-Life of Solanidine and Total Alkaloid Prior to 24h Pharmacokinetic Study SUBJECT A B Solanidine 3.5 4.0 Total Alkaloid 5.0 5.8 Half-life in days Figure 5 Figure “N^-o3= o-Or>^r> j - 0 0 1 - 0 5 7-5 - 5 7 10 - 5 - Solanidine Concentrations of a Volunteer a of Concentrations Solanidine ujc naPtt-re Diet Potato-Free a on Subject Hre, 1983) (Harvey, ME ( ) s k e e (w E IM T tr Ptt Avoidance Potato Start 3 14 13 Figure 6 Figure SUBJECT Alkaloid Throughout the Pharmacokinetic Study Pharmacokinetic the Throughout Alkaloid Urinary Concentrations of Solanidine and Total and Solanidine of Concentrations Urinary — < * ! _ < 10—0 $ \ - 0 = E TIME both solanidine and total alkaloid concentrations are seen in the urine of the subjects prior to the 24h studyi After the potato load, highest urinary alkaloid concentrations were found during the first 24h. The percentage of the dose (14.25mg) excreted each day following the potato load are given in Table 3. The ratios of total alkaloid to solanidine are also shown in this Table for each 24h urine collection period. Before the study and for each urine collection period following the first 24h after the potato load, the ratio of total alkaloid to solanidine ranged from 1.5 to 2.2. However, in the urine collection period immediately following the potato load, the total alkaloid to solanidine ratios in the four subjects were all above 2.5, indicating a lower concentration of solanidine. Only buSo subjectsprovided 24h urine samples for the seven day period following the potato load. Because their urinary concentrations were still above the lim it of detection of the assay at the end of this period, this was extended for a further 7-14 days. The total percentage of the total potato alkaloid dose excreted in this two week period amounted to less than 1.0% (95%) 0.26% of this being excreted in the first 24h. In the three other subjects 0.23%, 0.27% and 0.3% of the total dose was excreted in the first 24h. This was calculated from the total alkaloid concentrations. Table 3 Percentage of Potato Alkaloid Dose Excreted in the Urine (as Total Alkaloid Concentration) Percentage Dose Ratio Total Subject Alkaloid/Solanidine Subject Days A B C D A B C D 0 0.34% 0.27% 0.23% 0.26% 2.9 3.9 2.6 2.7 1 0.13 0.12 0.14 2.0 1.8 2.0 2 .07 1.5 3 .06 1.8 4 .06 1.6 5 .04 1.8 6 .04 2.2 7 .03 1.6 8 .03 1.5 10 .03 1.8 11 .03 1.9 12 .02 2.1 13 .02 1.8 8.4 DISCUSSION The pharmacokinetic study demonstrates for the first time, the increase in potato alkaloid concentrations following a potato load. The effect of cooking is seen to cause a negligible decrease in potato alkaloid concentrations, as previously demonstrated (43). The method of delivery of the alkaloid represents a normal meal. The reason for the potato avoidance prior to the potato load was to ensure minimal potato alkaloid body fluid concentrations remained to demonstrate clearly the affect of a single dose. The total glycoalkaloid content of the potatoes eaten was well below the lim it for commercial tubers (20mg TGA/lOOg fresh weight of tuber) (24,23). Yet, a clear increase in both solanidine and total alkaloid concentrations in serum was observed in all subjects. The levels reached in each subject was similar, yet the pre-dose concentrations prior to the potato abstinence varied widely. The concentrations of total alkaloid found after the potato load show a rapid absorption starting within two hours and reaching a peak at four hours. The alkaloids in the potato exist mainly as a-solanine and a-chaconine, with less than 3% present as solanidine (7). In the animal studies described in Section 8.1, tritiated a-solanine and a-chaconine was poorly absorbed from the GI tract in rats, but showed good absorption in hamsters (112,126,127). a- solanine and a-chaconine have relatively large molecular weights compared to solanidine and contain hydrophilic sugar residues, pre-disposing to poor absorption. The possible mechanism for the rapid absorption seen could result either from pH changes affecting the nitrogen moiety of the molecule, or a specific transport system. In contrast, solanidine shows a slower rise in plasma concentrations, reaching a steady level at 13 hours post-loading. Solanidine, a lipophilic molecule of small molecular weight, would be expected to rapidly cross the wall of the GI tract. The slow rise in solanidine concentrations is most likely to result from metabolism from the glycosides. This could occur as a result of hydrolysis of sugar residues by intestinal flora. Alternatively, the glycosides could be metabolised in the liver, on the first pass. Biliary excretion is another possible mechanism of elimination, as many steroids are excreted in the bile to be reabsorbed later from the GI tract. The lower glycosides, 3 and Vsolanine and chaconine, which have been found in both serum and urine (section 7.3) could follow the route of either solanidine or the higher glycosides. Glycosides with only one sugar residue are much more lipophilic than the 3 and a forms, allowing biliary excretion and reabsorption. The trough and the subsequent second peak seen with the serum total alkaloid concentrations, seen to favour biliary excretion of solanidine. The trough would represent the metabolism of a-solanine and a-chaconine to the lower glycosides and solandine with biliary excretion removing them fronrT', the plasma. After reabsorption, solanidine would contribute to the total alkaloid concentrations, giving rise to the second peak. Once the steady state of both solanidine and total alkaloid serum concentrations have been reached the levels persist until near the end of the 24h period. This again indicates a relatively long half-life for the alkaloids, seen in the pre-load potato avoidance study and the period following the 24h pharmacokinetic study. The half-lives calculated compare well with the third component found for the clearance of tritiated solanidine in the study of Claringbold et al„ (129). However, the first and second components are not seen in this study. They were not seen, possibly due to the oral adm instration of the alkaloid in this study rather than the intravenous injection of the tritiated solanidine. The urinary excretion of total alkaloid represents a very small percentage of the total dosel In other studies, urinary excretion has also been low (119,129). In one study, much of the dose was excreted in the faeces, whereas in the other study suggested low excretion was due to high body burdens of the alkaloid (129). In the present study, since faecal alkaloid concentrations were not estimated, either model could apply. Initially, over the first 24h urine collection, the excretion of total alkaloid and solanidine is maximum. This decreases over the next 14 days until minimum excretion rates exist. The ratio of total alkaloid to solanidine is much higher in the first 24 hours than in the succeeding days. This is to be expected since initially a- solanidine and a-chaconine would be excreted unaltered in higher concentrations than on succeeding days when metabolism to solanidine and lower glycosides has increased. The ratio of total alkaloid to solanidine in the urine indicates ease of excretion of solanidine even though the molecule is very lipophilic. In the normal subject, eating potatoes regularly it is likely that the differences in the serum and urinary concentrations of solanidine and total alkaloid would have been missed, due to constant levels of body fluid alkaloids, masking these differences. The information obtained in this experiment indicated the absorption of a-solanidine and a-chaconine and subsequent metabolism to solanidine possibly involving biliary excretion C H A PTER 9 GENERAL DISCUSSION AND CONCLUSIONS C H A P TER 9 GENERAL DISCUSSION AND CONCLUSIONS 9.1 METHODOLOGY The majority of the results obtained in this study resulted from the measurement of alkaloid concentrations by radioimmunoassay. The first comment on the methodology must be the reagents. The limiting factor in the immunoassay was the label. This was a tritiated label with a relatively low specific activity. This had all the usual disadvantages, namely, the use of 3 counting with the penalties of increased cost due to scintillant use and increased preparation time. An attempt was made to produce a derivative of solanidine with a view to obtaining an iodinated label; however this proved fruitless due to the instability of the aglycone under chemical attack. In future studies, with the current advances in technology it should be possible to produce either an iodinated label for use in an IRMA methodology or secondly, to use a non-isotopic label. Indeed, the ELISA method of Morgan et al., (76) proved extremely useful for the analysis of glycoalkaloids in the potato. Although the sensitivity of the method to measure body fluid concentrations may be lim itin g . The antisera used in this study were specific for the potato alkaloids a-solanine, a-chaconine and solanidine, but failed to distinguish between the aglycone and the glycosides. This might be considered a drawback, since the molecules possess differing molecular weights and have different physico chemical properties. However, the antisera which were produced in the present work did allow establishment of two assays; one which, by prior extraction, measured only solanidine and the second, a direct assay, which one would expect to detect a-solanine, a-chaconine and solanidine as well as the lower glycosides and any other metabolites whose terminal ring structure closely resembled solanidine! By being able to measure the ratio of total alkaloid to solanidine, much information was obtained on the distribution of the aglycone and the glycosides in the body fluids measured and under various physiological conditions. If a need for an antiserum to distinguish between solanidine and the glycosides was established, the conjugate for immunisation would need to be attached in such a way that the A ring containing the distinguishing features remained unaltered and free to act as a distinctive epitope. Although the antisera examined exhibited parallelism it would be pertinent to test their cross-reactivity with any similar minor alkaloids in food, as soon as available. A ll bleeds of antisera used in the radioimmunoassay indicated good avidity and were of higti titre. The use of antisera raised in sheep rather than in rabbit as used for the original study (142) was due to availability. The production of a new antiserum G/S/RG14-IA provided an immunoreagent with increased avidity and titre, and enabled the establishment of a more sensitive assay for the measurement of salivary alkaloids. However, the sensitivity of the resultant assay was again limited by the specific activity of the label. The volume of this antiserum insured continuity of reagents throughout this study and any subsequent ones. The robustness of the various assay systems was indicated by the ability to measure alkaloid concentrations in a variety of body fluids, using a range of volumes, without marked detriment to their recovery, sensitivity or precision. One major criticism of these assays could be the use of a solanidine standard when measuring both solanidine and its glycosides. Although there are differences in molecular weight this is an adequate compromise since a measurement of to ta l alkaloid was needed. This aspect of the assay must be considered when comparing the values obtained in this study with any subsequent studies. In much of the published literature the alkaloids, solanidine, a- solanine and a-chaconine are collectively termed total glycoalkaloids when assayed. This is acceptable since 95% of the alkaloid content of the potato is a-solanine and a-chaconine. In body fluids, however, the glycoside content is very much reduced, uAik a much greater concentration of solanidine. In the work reported in this thesis the assay of a-solanine, a-chaconine and solanidine is termed total alkaloids rather than total glycoalkaloid. The other methodologies used for potato alkaloid analysis were affinity chromatography, TLC and HPLC. Affinity chromatography proved to be a very useful technique both for preparation of the alkaloid-free matrix for use in the radioimmunoassay method, and for concentrating the alkaloids from samples of body fluids. The glass particles used for preparation of the antibody solid phase proved to be very robust, allowing the use of the 90% methanoljneeded, to elute the alkaloids without damage to the immunosorbent. The preparative column provided a simple method of extracting the alkaloid from large volumes of urine without being time-consuming and elution with methanol allowed ease of sample concentration by evaporation. The antisera produced for the radioimmunoassay were ideal for affinity chromatography, extracting solanidine, a-solanine and a-chaconine, as well as any of the glycosidic metabolites. The potential advantage of HPLC over TLC previously seen in many other experimental situations was not observed in these studies. TLC gave a better separation than HPLC of the lower glycosidic metabolites in serum and urine with positive identification. The HPLC system was hindered by a number of factors. Firstly, the wavelength used to measure the alkaloids limits the possible elution solvents and requires care in solvent and sample preparation. Secondly, the column proved to be very prone to contamination, causing much time to be wasted in column clean-up. The HPLC hardware provided the option of gradient elution. This could have been useful in separation of all the lower glycosides and any other metabolites. Again, the low wavelength used caused excessive baseline drift when solvent composition was changed. In any future work on alkaloid analysis by HPLC, consideration must be given to the following alternatives. If a wavelength of 200nm is used, further clean-up procedure of the samples injected could be carried out using either guard columns, or silica-bonded mini extraction columns. The choice of analytical column used here was limited by cost. Ideally, two or more columns could be used giving partial separation on one system, followed by re-injection and full separation on the second. Quantitation could then be achieved by immunoassay or quantitation and characterisation of the separated alkaloids by mass spectroscopy. 9.2 ALKALOID CONCENTRATIONS IN NORMAL POPULATIONS The use of the radioimmunoassay methods to determine alkaloid concentrations in normal subjects provided reference ranges for those compounds in serum, urine and saliva. For the first time, it was possible to compare alkaloid concentrations in a case of suspected alkaloid poisoning with concentrations in subjects on their normal healthy diets. The samples available from the one case of poisoning were limited. Indeed, few cases of potato poisoning have been documented this century (26) probably due to any cases which have occurred being confused with bacteriological food poisoning. To screen all suspected cases of food poisoning for potato alkaloids would be a difficult, if not impossible task. The wide range of alkaloid concentrations found in serum gives a useful indication of the great diversity of normal potato intake. The volunteer groups investigated (excluding pregnant women) were limited to the South East of England and a very small group from Sweden. A fulher study is needed with special reference to different ethnic groups to obtain world-wide figures for potato alkaloid concentrations in humans. Alkaloid concentrations found in saliva were roughly 10% of the serum concentrations for total alkaloid and 20% for solanidine. This reflects the situation found for many other substances measured in serum and saliva (216, 217). Ideally, two standard assays are needed, one to measure salivary concentrations and serum concentrations in subjects with low potato intake and secondly one with a wider dynamic range to measure serum concentrations in the rest of the population. Salivary solanidine and total alkaloid measurements allowed assessment of the selective entry of alkaloids into saliva, solanidine and lower molecular weight, lipophMlic molecules being the principal salivary alkaloid. The pharmacokinetic study provided valuable information on the absorption and metabolism of potato alkaloids. In previous work in humans, volunteers were given an intravenous injection of radiolabelled alkaloid (128, 129) rather than in the form of a natural meal. The present study was limited by the fact that, although the loading dose was known, the faecal excretion was not determined. This posed the question whether in view of the fact that only a small percentage of the dose was excreted in the urine, was the remainder excreted in the faeces or taken up by body stores. The long half- life found suggested the latter. Also, not knowing the actual dose entering the body, prevented estimation of volume of distribution and clearance. A limited amount of information on the metabolites of the potato alkaloids was provided by the TLC and HPLC analysis of affinity chromatography extracts of urine and serum. This was due to the methodological difficulties described in section 9.1. 9.3 NEURAL TUBE DEFECTS The controversy surrounding the possible link between potato alkaloids and foetal neural tube defects (134) necessitated in depth investigation of the relationship between these metameters. This study measured alkaloid concentrations in serum collected from a large group of women pregnant with a normal foetus and a similar sized group pregnant with a foetus subsequently shown to have NTD. The results showed no significant differences in the solanidine and total alkaloid concentrations in the normal and neural tube defect groups. The conclusions reached indicated that the results were inconsistent with the Renwick Hypothesis (134). It is proposed that involuntary or voluntary avoidance of potatoes might predispose to water-soluble vitamin deficiencies and thereby increase, rather than decrease the incidence of NTD. The investigation of plasma/serum levels of zinc and the small study of folate levels in these two groups failed to show a link between their deficiencies (reflected by low serum concentrations) and the incidence of NTD. The limitation of the investigation was the timing of the serum samples. The samples collected at 14-16 weeks gestation might not reflect the alkaloid concentrations present at 4-5 weeks of pregnancy when closure of the neural tube occurs. An indeal study would measure alkaloid concentrations on or around 4 weeks of pregnancy. This would involve the collection of a very large number of specimens so as to be certain of including a sufficient number of mothers with foetuses with neural tube defects for statistical analysis. Alternatively, samples could be collected from women who have had a previous NTD pregnancy, since the chance of a recurrent NTD is higher in this group. The problems with a study of this type are numerous. Firstly, the collection of a sufficient number of specimens in women who know they are pregnant four weeks after conception would be difficult^ Possibly the only source of suitable specimens would be from women who are actually trying to become pregnant and this would bring further problems since many of these women would be in the older age group. Secondly, there is the ethical consideration of finding high alkaloid concentrations at this early stage when conventional means of NTD detection are insensitive. Although a recent report suggests a link between vitamin deficiencies and NTD (177), the weight of the evidence is still in favour of a multifactorial aetiology. A recent study (226) shows increased perinatal mortality in the area also having highest incidence of NTD, namely, the north and west of Ireland, South Wales and the north-west of England and Scotland. These areas are also inhabited by low socio-econorrjc groups, possibly indicating a poorer diet, but not specifically a particular item of the diet. 9.4 FURTHER STUDIES Little of the work described above has involved investigation of the possible toxicity of the potato alkaloids. This has been due to the study being predominantly based on investigations in human subjects. This has limited induced toxicity for ethnical reasons,and cases of accidental toxicity were found on only one occasion. The possible toxicity of the potato alkaloids is a very important consideration since the economic implications of this are great. Many of the toxicity studies carried out in animals were inconclusive and conflicting (section 1.8.4) due to different routes of alkaloid administration and the variety of species used. The pharmacological investigations of the potato alkaloids (section 1.8.2) have conclusively demonstrated their anti cholinesterase effect (115, 116, 117) which could be responsible for their toxicity. This could be further investigated using tissue culture. Embryonic nervous tissue cells would provide a good model for toxicity effects and indeed could shed some light on the possible teratogenic properties, ie, NTD. Application of potato alkaloids, potato extracts or blighted potato extracts to cells in culture and their actions on the morphology, viability and function of the cells could be followed by light and electron microscopy. The availability of alkaloid antiserum proves an ideal tool for studying the uptake and internalisation of the alkaloids within the cells by immunocytochemistry. This would indicate the sub-cellular actions of the alkaloids. Biochemical analysis of the fluid medium bathing the cells before and after application of the alkaloids could shed light on the metabolic actions at the cellular level. Immunocytochemistry using anti-alkaloid antisera could also provide a tool for elucidation of the tissue binding of the alkaloids. Animal studies have demonstrated the presence of alkaloids in various tissue (119, 127) but inter-species differences are found. Immunocytochemistry of post-mortem tissue could detect the alkaloid at both the tissue and cellular levels. This study has shown for the first time the wide range of potato alkaloid concentrations found in the normal population. The absorption and distribution of the alkaloids has been discussed, along with their metabolism. It is suggested that further studies are needed to examine their possible toxicity. The theory that potato alkaloids are a causative factor of foetal NTD has been challenged and indeed, it is likely that potatoes would have a beneficial effect in pregnant women due to their high nutritional value. ACKNOWLEDGEMENTS I would like to thank Professor V Marks, Mr B A Morris and Dr M McMillan for their constant support throughout this project. I would especially like to thank Mr B A Morris and Dr M McMillan for their invaluable help and encouragement in the preparation of the manuscript. My thanks go to Dr M R A Morgan for the gift of pure alkaloids, Professor J Renwick for tritiated solanidine, Mr K E Hellanas for the provision of the Swedish specimens, Dr P Kwasowski for the preparation of the affinity chromatography solid phase, and Dr I Kitchen for the use of the HPLC equipment. Thanks go to my motherfor her constant enthusiasm in helping me complete this project and to Miss T A Bakall for her professional approach in typing the manuscript. Finally, I would like to thank the staff of Lewisham Hospital, Department of Biochemistry, and the staff of the Department of Biochemistry, University of Surrey, for help throughout the study. REFERENCES REFERENCES 1. Geigy Scientific Tables. 8th Edn (ed: C.Lentner), Basle:J.R.Geigy S.A. Volume I, 1981, p248. 2. Schreiber, K. (1968) Steroid alkaloids: The solanum group. In: The Alkaloids, (ed: R.H.F.Marske), New York:Academic Press Inc. Volume X, 1-192. 3. Ripperger, H., Schreiber, K. (1981) Solanum steroid alkaloids. In: The Alkaloids, (ed: R.G.A.Rodrigo), New York:Academic Press Inc. Volume XIX, 81-192. 4. Kloosterman, J., (1979) Development of steroid synthesis from natural products. Chemistry in New Zealand, 146-147. 5. Segal, R., Breuer, A. and Milo-Goldzweig, I. (1978) Solasodine stability under conditions of saponin hydrolysis. Journal of Pharmaceutical Sciences, 67, 1169-1170. 6. Crabbe, P.G. and Fryer, C. (1982) Evaluation of chemical analysis for the determination of solasodine in solanum lasintatum. Journal of Pharmaceutical Sciences, 71, 1356-1362. 7. Paseshnickenko, V. and Guseva, A.R. (1956) Quantitative determination of potato qlycoalkaloids and their preparative separation. Biochemistry, USSR, 21, 606-611. 8. Gregory, P. (1984) Glycoalkaloid composition of potatoes; Diversity and biological implications. American Potato Journal, 61, 115-122. 9. Jadhav, S.J., Sharma, R.P. and Salunkhe, D.K. (1981) Naturaly occurring toxic alkaloids in foods. CRC Critical Reviews in Toxicology, 9, 21-104. 10. Clarke, E.G.C. (1968) Isolation and identification of drugs. The Pharmaceutical Press, pp541-542. 11. Osman, S.F. (1983) Glycoalkaloids in potatoes. Food Chemistry, 11, 235-247. 12. Maga, J.A. (1980) Potato glycoalkaloids. CRC Critical Reviews in Food Science and Nutrition, 12, 371-405. 13. Sinden, S.L., Sanford, L.L. and Webb, R.E. (1984) Genetic and environmental control of potato glycoalkaloids. American Potato Journal, 61, 141-156. 14. Jones, P.G. and Fenwick, G.R. (1981) The glycoalkaloid content of some edible solanaceous fruits and potato parts. Journal of the Science of Food and Agriculture, 32, 419-421. 15. Gusera, A.R. and Paseshnichenko, V.A. (1958) A study of the biogenesis of potato glycoalkaloids by the method of labelled atoms. Biochemistry, USSR, 23, 385-388. 16. Ripperger, H., Moritz, W. and Schreiber, K. (1971) Zur biosynthese von Solanum-Alkaloiden aus cycloartenol oder lanosterin. Phytochemistry, 10, 2699-2704. 17. Tschesche, R. and Hulphe, H. (1967) Biosynthese von steroid derivatem in pflanzenreich III. Bioqenese von Solanidin aus cholesterin. Z.Naturforsch., 22, 791-795. 18. Heftmann, E. and Weaver, M.L. (1974) 26-hydroxycholesterol and cholest-4-en- 3-one, the first metabolites of cholesterol in potato plants. Phytochemistry, 13, 1801-1803. 19. Robinson, T. (1981) Terpenoid and steroid alkaloids. In: The Biochemistry of the Alkaloids, 2nd Edition (eds: A.Kleinzeller, G.F.Springer and H.G.Wittmann), New York:Springer Verlag. ppl54-165. 20. Waller, G.R. and Dermer, O.C. (1981) Enzymology of alkaloid metabolism in plants and microorganisms. In: The Biochemistry of Plants, New York:Academic Press. Volume 7, pp381-402. 21. Osman, 5.F. and Zacharius, R.M. (1979) American Potato Journal, 56, 475-479. 22. Swain, A.P., Fitzpatrick, T.J., Talley, E.A., Hert, S.F. and Osman, S.F. (1978) Enzymic hydrolysis of a-chaconine and a-solanine. Phytochemistry, 17, 800-801. 23. Filadelfi, M.A. and Zitnak, A. (1982) Preparation of chaconines by enzymic hydrolysis of potato berry alkaloids. Phytochemistry, 21, 250-251. 24. Sinden, S.L. and Deahl, K.L. (1976) Effect of glycoalkaloids and phenolics on potato flavour. Journal of Food Science, 41, 520-523. 25. Zitnak, A. and Filadelfi, M.A. (1985) Estimation of taste thresholds of three potato glycoalkaloids. Canadian Institute of Food Science and Technology Journal, 18, 337-339. 26. McMillan, M. and Thompson, J.C. (1979) An outbreak of suspected potato poisoning in schoolboys. Quarterly Journal: of Medicine, 48, 227-243. 27. Zitnak, A. and Johnston, G.R. (1970) Glycoalkaloid content of B5141-6 potatoes. American Potato Journal, 47, 256-260. 28. Wood, F.A. and Young, D.A. (1974) Publication 1533 Agriculture Canada. Canada Department of Agriculture. 29. Baerug, R. (1962) Influence of different rates and intensities of light on solanine content and cooking quality of potato tubers. European Potato Journal, 5, 242-251. 30. Zitnak, A. (1961) The occurrence and distribution of free alkaloid solanidine in netted qem potatoes. Canadian Journal of Biochemistry and Physiology, 39, 1257-1265. 31. Gull, D.P. and Isenberg, T.M. (1960) Chlorophyll and solanine content and distribution in four varieties of potato tubers. Proceedings of the American Society of Horticultural Science, 75, 545-552. 32. Liljemark, A. and Widoff, E. (1960) Greening and solanine development of white potato in fluorescent light. American Potato Journal, 37, 379-389. 33. Conner, H.W. (1937) Effect of light on solanine synthesis in the potato tuber. Plant Physiology, 12, 79-88. 34. Peterman, J.B. and Morris, S.C. (1985) The spectral responses of chlorophyll and glycoalkaloid synthesis in potato tubers (Solanum tuberosum). Plant Sciences, 39, 105-110. 35. Hilton, R.J. (1951) Factors in relation to tuber quality in potatoes. II. Preliminary trials on bitterness in netted gem potatoes. Canadian Journal of Agricultural Science, 31, 61-66. 36. Wu, M.T. and Salunke, D.K. (1976) Changes in glycoalkaloid content following mechanical injuries to potato tubers. Journal of the American Society of Horticultural Science, 10, 329-331. 37. Crank, T.C., Kuhn, G.O. and McARdle, F.J. (1974) The influence of stage of maturity, level of nitrogen fertilisation and storage on the concentration of solanine in tubers of three potato cultivars. _n, 163-169. 38. Fitzpatrick, T.J., Hert, S.F., Osman, S.F. and McDermott, J.A. (1977) Potato glycoalkaloids: Increases and variations of ratios in aged slices over prolonged storage. American Potato Journal, 54, 539- 39. McKee, R.K. (1955) Host parasite relationship in the dry rot disease of potatoes. Annals of Applied Biology, 43, 147-148. 40. Salunkhe, D.K., Wu, M.T. and Jadhav, S.J. (1972) Effects of light and temperature on the formation of solanine in potato slices. Journal of Food Science, 37, 969-973. 41. Sinden, S.L. and Webb, R.E. (1972) Effect of variety and location on the glycoalkaloid content of potatoes. American Potato Journal, 49, 334-338. 42. Wolf, M.J. and Duggar, B.M. (1946) Estimation and physiological role of solanine in the potato. Journal of Agricultural Research, 73, 1-32. 43. Bushway, R.J. and Ponnampalam, R. (1981) a-chaconine and a-solanine content of potato products and their stability during several modes of cooking. Journal of Agricultural and Food Chemistry, 29, 814-817. 44. Cianci, M. (1986) The potato: Apple of the earth. Restaurant Business, 85, 267-273. 45. Davies, A.M.C. and Blincow, P.J. (1984) Glycoalkaloid content of potatoes and potato products sold in the UK. Journal of the Science of Food Agriculture, 35, 553-557. 46. Toma, R.B., Orr, P.H., D'Appolonia, B., Dintiziz, F.R. and Tabekhia, M.M. (1979) Physical and chemical properties of potato peel as a source of dietary fibre in bread. Journal of Food Science, 44, 1403-1407. 47. Sizer, C.E., Mag, J.A. and Craven, C.J. (1980) Total glycoalkaloids in potatoes and potato chips. Journal of Agriculture and Food Chemistry, 28, 578- 579. 48. Ponnampalam, R. and Mondy, N.I. (1983) Effect of cooking on the total glycoalkaloid content of potatoes. Journal of Agriculture and Food Chemistry, 31, 493-495. 49. Lepper, W. (1949) Solingehalte von 58 kartoffelsarten. Z.Lebersum Unters Forsch, 89, 264-271. 50. Johnsson, H. and Hellenas, K.E. (1983) Glycoalkaloids (TGA) in Swedish potatoes. Var Foda, 35, 299-314. 51. Sinden, S.L. and Webb, R.E. (1974) Effect of environment on glycoalkaloid content of six potato varieties at 39 locations. US Department of Agriculture Technical Bulletin, 1472, 30. 52. Russ, H. (1966) The use of wild solanum species in German potato breeding of the past and today. American Potato Journal, 43, 63-80. 53. McCollum, G.O. and Sinden, S.L. (1979) Inheritance study of tuber glycoalkaloids in a wild potato, solanum chacoense bitter. American Potato Journal, 56, 95-113. 54. Jeppsen, R.B., Wu, M.T. an dSalunkhe, D.K. (1974) Some observations in the occurrence of chlorophyll and solanine in potato tubers and their control by N -benzyladenine; ethephon and filtered lights. Journal of Food Science, 39, 1059. 55. Patil, B.C., Singh, B. and Salunkhe, D.K. (1971) Formation of chlorophyll and solanine in Irish potto tubers and their control by gamma radiation and CO2 enriched packaging. Lebensm.Wiss.Technol., 4, 123. 56. Wu, M.T. and Salunkhe, D.K. (1977) Effect of gamma-irradiation on wound induced glycoalkaloid formation in potato tubers. Lebensm.Wiss.Technol., 10, 141-144. 57. Jadhav, S.J., Patil, B.C. and Salunkhe, D.K. (1973) Control of potato greening under hypobaric storage. Food Engineering, 45, 111-118. 58. Wu, M.T. and Salunkhe, D.K. (1977) Inhibition of wound induced glycoalkaloid formation in potato tubers (Solanum tuberosum L.) by isopropyl-N(3- chlorophenyl)-carbamate. Journal of Food Science, 42, 622-624. 59. Ponnampalam, R. and Mondy, N.I. (1986) Effect of solicar application of indoleacetic acid on the total glycoalkaloids and nitrate nitrogen content of potatoes. Journal of Agriculture and Food Chemistry, 34, 686-688. 60. Wu, M.T. and Salunkhe, D.K. (1972) A research note. Control of chlorophenyl and solanine synthesis and sprouting of potato tubers by hot paraffin wax. Journal of Food Science, 37, 629-630. 61. Kuc, J. (1984) Steroid glycoalkaloids and related compounds as potato quality factors. American Potato Journal, 61, 123-139. 62. Kuc, J. (1975) Teratogenic constituents of potatoes. Recent Advances in Phytochemistry, 9, 139-150. 63. Tingey, W.M. (1984) Glycoalkalods as pest resistance factors. American Potato Journal, 61, 157-167. 64. Sinden, S.L., Sanford, L.L. and Osman, S.F. (1980) Glycoalkaloids and resistance to the Colorado potato beetle in Solanum chaconaese Bitt. American Potato Journal, 57, 331-343. 65. Kuhn, R. and Low, I. (1961) The structure of leptines. -Chem.Ber., 94, 1088- 1095. 66. Mass, M.R., Post, F.J. and Salunkhe, D.K. (1977) Production of steroid glycoalkaloids by phytophthora infestans in complex and chemically defined media. Journal o f Food Safety, 1, 107-1171 67. Holland, H.L. and Taylor, G.J. (1979) Transformations of steroids and the steroidal alkaloid solanine, by Phytophthora infestans. Phytochemistry, 18,437-440. 68. Patil, B.C., Sharma, R.P., Salunkhe, D.K. and Salunkhe, K. (1972) Evaluation of solanine toxicity. Food and Cosmetic Toxicology, 10, 395-398. 69. Coxon, D.T. (1984) Methodology for glycoalkaloid analysis. American Potato Journal, 61, 169-183. 70. Sanford, L.L. and Sinden, S.L. (1972) Inheritance of potato glycoalkaloids. American Potato Journal, 49, 209-217. 71. Sachse, J. and Bachmann, F. (1969) Z.Lebensm,Unters Forsch, 141, 262. 72. Fitzpatrick, T.J. and Osman, S.F. (1974) A comprehensive method for the determination of total potato glycoalkaloids. American Potato Journal, 51, 318-323. 73. Mackenzie, J.D. and Gregory, P. (1979) Evaluation of a comprehensive method for total qlycoalkaloid determination. American Potato Journal, 56, 27- 33. 74. Bushway, R.J., Wilson, A.M. and Bushway, A.A. (1980) Determination of total glycoalkaloids in potato tubers using a modified titration method. American Potato Journal, 57, 561-565. 75. Bushway, R.J., Bureau, J.L. and Stickney, M.R. (1985) A new efficient method for extracting glycoalkaloids from dehydrated potatoes. Journal of Agriculture and Food Chemistry, 33, 45-46. 76. Morgan, M.R.A., McNerney, R., Matthew, J.A., Coxon, D.T. and Chan, H.W-S. (1983) An enzyme-linked immunosorbent assay for total glycoalkaloids in potato tubers. Journal of the Science of Food and Aqriculture, 34, 593- 598. 77. Van Gelder, W.M.J. (1985) Determination of the total C^^-steroidal alkaloid composition of solanum species by high resolution gas chromatography. Journal of Chromatography, 331, 2875-293. 78. Alberti, B. (1932) The determination of solanines. Z.Unters Lebensmittle, 64, 260-262. 79. Clarke, E.G.C. (1958) Identification of solanine. Nature, 181, 1152-1153. 80. Wierschowski, P. and Wierzchowska, Z. (1961) Z.Chem.Analityczna Warszawa, 6, 579-585. 81. Jellema, R„ Elema, E.T. and Malingre, Th.M. (1979) Optical brighteners as thin-layer chromatography detection reagents for glycoalkaloids and steroid alkaloids in Solanum species: l.Calcofluor M2R New. Journal of Chromatography, 176, 435-439. 82. Cadle, L.S., Stelzig, Dl, Harper, K.L. and Young, R.Ji (1978) Thin-layer chromatographic system for identification and quantitation of potato tuber glycoalkaloics. 26, 1453-1454. 83. Hunter, I.R., Walden, M.K., Wagner, J.R. and Heftmann, E. (1976) Thin-layer chromatography of steroidal alkaloids. Journal of Chromatography, 118, 259-262. 84. Filadelfi, M.A. and Zitnak, A. (1983) A simple TLC standard for identification of potato glycoalkaloids. Food Science and Technology Journal, 16, 151- 153. 85. Ahmed, S-S. and Muller, K. (1978) Effect of wound damages on the glycoalkaloid content in potato tuber and chips. Lebensm.Wiss.Technol., 11, 144-146. 86. Jellema, R., Ellema, E.T. and Malingre, T.M. ( ) Fluorodensitometric determination of potato glycoalkaloids on thin-layer chromatograms. 87. Herb, S.F., Fitzpatrick, T.J. and Osman, S.F. (1975) Separation of potato glycoalkaloids by gas chromatography. Journal of Agriculture and Food Chemistry, 23, 520-523. 88. Osman, S.F. and Sinden, S.L. (1977) Analysis of mixtures of solanidine and demissidine glycoalkaloids containing identical carbohydrate units. Journal of Agriculture and Food Chemistry, 25, 955-957. 89. Siegfried, R. (1976) Z.Lebensm.Unters Forsch, 162, 253-258. 90. King, R.R. (1980) Analysis of potato glycoalkaloids by gas-liquid chromatography of alkaloid components. Journal of Association of Analytical Chemistry, 63, 1226-1230. 91. Hunter, I.R., Walden, M.K., Wagner, J.R. and Heftmann, E. (1976) High pressure liquid chromatography of steroidal alkaloids. Journal of Chromatography, 119, 223-226. 92. Hunter, I.R., Walden, M.K. and Heftmann, E. (1980) High performance liquid chromatography of Solanumand veratrum alkaloids. Journal of Chromatography, 198, 363-366. 93. Crabbe, P.G. and Fryer, C. (1980) Rapid qualitative analysis of solasodine, solasodine glycosides and solasodiene by high pressure liquid chromatography. Journal of Chromatography, 187, 87-100. 94. Bushway, R.J., Barden, E.S., Bushway, A.W. and Bushway, A.A. (1979) High performance liquid chromatographic separation of potato glycoalkaloids. Journal of Chromatography, 178, 533-541. 95. Bushway, R.J. (1982) High performance liquid chromatographic determination of the metabolites of the potato glycoalkaloids a-chaconine and a- solanine, in potato tubers and potato products. Journal of Chromatography, 5, 1313-1322. 96. Morris, 5.C. and Lee, T.H. (1981) Analysis of potato glycoalkaloids with radially compressed high performance liquid chromatographic cartridges and ethanolamine in th emobile phase. Journal of Chromatography, 219, 403-410. 97. Eldridge, A.C. and Hockridge, M.E. (1983) High performance liquid chromatographic separation of eastern black nightshade (Solanum ptychanthum) glycoalkaloids. Journal of Agriculture and Food Chemistry, 31, 1218-1220. 98. Hellenas, K-E. (1986) A simplified procedure for quantification of potato glycoalkaloids in tuber extracts by HPLC: Comparison with ELISA and a colorimetric method. Journal of the Science of Food and Agriculture, 37, 776-782. 99. Bushway, R.J. (1983) Sources of alkaloid and glycoalkaloid standards for potato breeding programs and other research. American Potato Journal, 60, 793-797. 100. Vallejo, R.P. and Ercegovich, C.D. (1978) Analysis of potato for glycoalkaloid content by radioimmunoassay (RIA). National Bureau of Standards Special Publication 319, Trace Orqanis Analysis: A New Frontier in Analytical Chemistry. Proceedings of the 9th Materials Research Symposium (US), 333-340. 101. Matthew, J.A., Morgan, M.R.A., McNerney, R., Chan, H.W-S. and Coxon, D.T. (1983) Determination of solanidine in human plasma by radioimmunoassay. Food Chemistry and Toxicology, 21, 637-640. 102. Morgan, M.R.A. and Coxon, D.T. (1987) Tolerances: glycoalkaloid in potatoes. In: Natural Roxicants in Food;Proqress and Prospects, (ed: D.H.Watson) Ellis Horwood. Chapter 7, 221-230. 103. Morgan, M.R.A., Coxon, D.T., Bramham, S., Chan, H.W-S., van Gelder, W.M.J. and Allison, M.J. (1985) Determination of the glycoalkaloid^ content of potato tubers by three methods including enzyme-linked immunosorbent assay. Journal of the Science of Food and Agriculture, 36, 282-288. 104. Pfuhl, E. (1899) Uber eine massenerkranking durch vergiftung mit stark solaninhaltigen kartoffeln. Deutsch Med.Wochensch., 25, 753-755. 103. Morris, S.C. and Lee, T.H. (1984) The toxicity and teratogenicity of solanaceae glycoalkaloids, particularly those of the potato (solanum tuberosum): A review. Food Technology in Australia, 36, 118-124. 106. Harris, F.W. and Cockburn, T. (1918) Alleged poisoning by potatoes. American Journal of Pharmacy, 90, 722-726. 107. Hansen, A.A. (1925) Two fatalcases of potato poisoninq. Science, 27, 310- 311. 108. Willmolt, S.G. (1933) An investigation of solanine poisoninq. The Analyst, 58, 431-439. 109. Wilson, G.S. (1959) A small outbreak of potato poisoning. Monthly Bulletin of Ministry of Health, Public Health Laboratory Service, 18, 207-210. 110. Reebalch, L.A. and Keem, A. (1958) A case of mass poisoning by solanine. Soviet Medicine, 22, 129-131. 111. Editorial: (1979) Solanine poisoning. British Medical Journal, 1458-1459. 112. Clarke, E.G.C. and Clarke, M.L. (1975) Poisonous plantsi In: Veterinary Toxicology, Balliere Tindall:London. p360. 113. Volker, R, (1950) Frohner's Lehbruch der Toxikoloqie fur Tierarztei 6th Edition, Enke:Stuttgart. pp339. 114. Owen, R. (1985) Potato poisoning in a horse. The Veterinary Record, 117, 246- 115. Orgell, W.H. (1963) Inhibition of lumar plasma cholinesterase jn vitro by alkaloids, glycosides and other natural substances. Lboydia, 26, 36-43. 116. Abbott, D.C., Field, K. and Johnson, E.I. (1960) Observation on the correlation of anticholinesterase effect with solanine content of potatoes. Anal;yst., 83, 375-376. 117. Harris, H. and Whittaker, M. (1962) Differential inhibition of the serum cholinesterase phenotypes by solanine and solanidine. Annals of Human Genetics, 26, 73-76. 118. Alozie, S.O. (1977) Toxicological and biochemical investigations of alpha- chaconine in potato (Solanum tuberosum L.) tubers: physiologic disposition and tissue binding and inhibition of tissue cholinesterases and isoenzymes. Health Sciences, Pathology, 5854B, 129. 119. Nishie, K., Gumbmann, M.R. and Keyl, A.C. (1971) Pharmacology of solanine. Toxicology and Applied Pharmacology, 19, 81-92. 120. Nishie, K., Norred, W.P. and Swain, A.P. (1975) Pharmacology and toxicology of chaconine and tomatine. Research Communications in Chemical Pathology adn Pharmacology, 12, 657-668. 121. Bergers, W.W.A. and Plink, G.M. (1980) Toxic effect of the glycoalkaloids solanine and tomatine oncultured neonatal rat heart cells. 6, 29-32. 122. Ness, E., Jones, P.E. and Dahle, H.K. (1984) a-solanine tested for mutagenicity with Ames test. Acta Vt.Scandinavica, 25, 145-147. 123. Thorne, H.V., Clarke, G.F. and Shuce, R. (1985) The inactivation of the herpes simplex virus by some solanacone glycoalkaloids. Antiviral Research, 5, 335-343. 124. Azim, A., Shaikh, H.A. and Ahmad, R. (1982) Effect of feeding greened potatoes on different visceral organs and blood plasma of rabbits. Journal of the Science of Food and Agriculture, 33, 1275-1279. 125. Dalvi, R.R. (1985) Comparative assessment of the effect of solanine administered orally and intraperitoneally on hepatic dysfunction in male rats. Nippon Juigaku Zasshi, 47, 657-659. 126. Norred, W.P., Nishie, K. and Osman, S.F. (1976) Excretion distribution and metabolic fate of H-a-chaconine, Research Communications in Chemical Pathology and Pharmacology, 13, 161-171. 127. Alozie, S.O., Starma, R.P. and Salunkhe, D.K. (1979) Excretion of a- chaconine- H, a steroidal glycoalkaloid from solanum tuberosum L. and its metabolites in hamsters. Pharmacological Research Communications, 11, 483-490. 128. CLaringbold, W.D.B., Few, J.D., Brace, C.J. and Renwick, J.H. (1980) The disposition and metabolism of H-solasodine in man and Syrian hamster. Journal of Steroid Biochemistry, 13, 889-895. 129. CLaringbold, W.D.B., Few, J.D. and Renwick, J.H. (1982) Kinetics and retention of solanidine in man. Xenobiotica, 12, 293-302. 130. Gull, D.D., Isenberg, F.M. and Bryan, H.H. (1970) Alkaloid toxicology of solanum tuberosum. Horticultural Science, 5, 316- 131. Chaube, S. and Swinyard, C.A. (1976) Teratological and toxicological studies of alkaloidal and phenolic compounds from solanum tuberosum. Toxicology and Applied Pharmacology, 36, 227-237. 132. Sharma, R.P., Willhite, C.C., Shup, J.L. and Salunkhe, D.K. (1978a) Acute toxicity and histopathological effects of certain glycoalkaloids and extracts of Alternaria solani or Phytophthora infestans in mice. Toxicological Letters, 3, 349-35. 133. Roddick, J.G. (1979) Complex formation between solanaceous steroidal glycoalkaloids and free sterols in vitro. Phytochemistry, 18, 1467-1470. 134. Renwick, J. (1972) Hypothesis: Anencephaly and spina bifida are usually preventable by avoidance of a specific but unidentified substance present in certain potato tubers. British Journal of Preventive and Social Medicine, 26, 67-88. 135. Yalow, R.S. and Berson, S.A. (1960) Immunoassay of endogenous plasma insulin in man. Journal of Clinical Investigation, 39, 1157-1175. 136. Chard, T. (1978) An introduction to radioimmunoassay and related techniques. In: Laboratory Techniques in Biochemistry and Molecular Biology, Volume 6^ (eds: T S Work and E.Work), North-Holland Publishing Company:Oxford. 291-534. 137. GreenMmod, F.C., Hunter, W.M. and Glover, J.S. (1963) The preparation of I-labelled human growth hormone of high specific radioactivity. Biochemical Journal, 89, 114-123. 138. Marchalonis, J.J. (1969) An enzymic method for the trace iodination of immunoqlobulins and other proteins. Biochemical Journal, 113, 299- 305. 139 Morris, B.A. (1985) Principles of immunoassay. In: Immunoassays in Food Analysis, (eds: B.A.Morris and M.N.Clifford), Elsevier Applied Science Publishers:London. 21-51. 140. Vining, R.F., Compton, P. and McGinley, R. (1981) Steroid radioimmunoassay -effect of shortened incubation time on specificity. Clinical C hem istry, 27, 910-913. 141. Scatchard, G. (1949) The attractions of proteins for small molecules and ions. Annals New York Academy of Sciences, 51, 660-672. 142.. Harvey, M.H. (1983) The Development and Evaluation of a Radioimmunoassay for the Alkaloid Solanidine in Serum and its Investigation of its Levels in Human Serum. MSc Thesis. University of Surrey. 143. Butler, V.P.jr. and Chen, J.P. (1967) Digoxin-specific antibodies. Proceedings of the National Academy of Science, USA, 57, 71-78. 144. Ross, H.H. and Rasmussen, G.T. In: Liquid Scintillation Counting - Recent Developments, (eds: P.E.Stanley and B.A. Scoggins), AcademicrNew York. 363-381. 145. Walker, H.C. (1977) An approach to immunoassay. Clinical Chemistry, 23, 384-402. 146. Ekins, R.P. (1981) Basic principles and theory. British Medical Bulletin, 30, 3-11. 147. Rodbard, D., Rayford, P.L., Cooper, J.A. and Ross, G.T. (1968) Statistical quality control of radioimmunoassays. Journal of Clinical Endocrinology, 28, 1412-1418. 148. Rodbard, D. (1974) Statistical quality control and routine data processing for radioimmunoassays. Clinical Chemistry, 20, 1255-1261. 149. Behrman, R.E. and Vaughan, V.C. (1979) Diseases of the nervous system. In: Textbook of Paediatrics, 11th Edition, (ed: W.E. Nelson), W B Saunders:New York. 1560-1563. 150. Weatherall, J.A.C. (1982) A review of some effects of recent medical practices in reducing the number of children born with congenital abnormalities. Health Trends, 14, 85-88, 151. Carstairs, V. and Cole, S. (1984) Spina bifida and anencephaly in Scotland. British Medical Journal, 289, 1182-1184. 152. Leek, I. (1983) Spina bifida and anencephaly: Fewer patients, more problems. British Medical Journal, 286, 1679-1680. 153. Leek, I. (1984) The geographical distribution of neural tube defects and oral clefts. British Medical Bulletin, 40, 390-395. 154. Carter, C.O. (1974) Clues to the aetiology of neural tube malformations. Developmental Medicine and Child Neurology, 16 (suppl.32), 3-15. 155. Leek, I. (1974) Causation of neural tube defects: Clues from epidemiology. British Medical Bulletin, 30, 158-163. 156. Office of Population Censuses and Surveys. Mapping of congenital malformations. OPCS Monitor, MB3 84/4, 1984. 157. Naggan, L. and MacMahon, B. (1967) Ethnic differences in the prevalence of anencephaly and spina bifida in Boston, Massachusetts. The New England Journal of Medicine, 277, 1119-1123. 158. Leek, I. (1972) The aetiology of human malformations and insights from epidemiology. Teratology, 5, 303-316. 159. Carter, C.O. (1976) Genetics of common single malformations. British Medical Bulletin, 32, 21-26. i6o; Nightingale, E.O., Scribanu, N., McCullough, D.C. and Quinn, C. (1975) Observations on patients with neural tube defects in a metropolitan hospital clinic: An epidemioloqy history. Developments in Medicine of Child Neurology, 17, 574-5791 161. Maclean, M.H. and MacLeod, A. (1984) Seasonal variation in the frequency of anencephalus and spina bifida births in the United Kingdom. Journal of Epidemioloqy and Community Health, 38, 99-102. 162. Knox, E.G. (1972) Anencephalus and dietary intakes. British Journal of Prevention and Social Medicine, 26, 219-223. 163. Fedrick, J. (1974) Anencephalus and tea drinking. Proceedings of the Royal Society of Medicine, 67, 356-360. 164. Brown, N.A., Kao, J. and Fabro, S. (1980) Teratogenic potential of valparic acid. Lancet, i, 660-661. 165. Bjerkedal, T., Czeizel, A. and Gouyard, J. (1982) Valpuric acid and spina bifida. Lancet, u, 1096. 166. Taylor, A. (1983) The biochemistry and toxicology of heavy metals. In: Scientific Foundations of Clinical Biochemistry, Volume II, (eds: V. Marks and D.L. Williams), William Heinemann. 167. Prasad, A.S. 91985) Clinical manifestations of zinc deficiency. Annual Review of Nutrition, 5, 341-364. 168. Sever, L.E. and Emanuel, L. (1972) Is there a connection between maternal zinc deficiency and congenital malformations of the^ central nervous system in man. Teratology, 7, 117-118. 169. Lowe, C.R., Roberts, C.J. and Lloyd, S. (1971) Malformations of the central nervous system and softness of local water supplies. British Medical Journal, 2, 357-361. 170. Kalter, H. and Warkany, J. (1959) Experimental production of congenital malformations in mammals by metabolic procedure. Physiological Reviews, 39, 69-73. 171. Cohlan, S.Q. (1954) Congenital abnormalities in the rat produced by the excessive intake of vitamin A during pregnancy. Pediatrics, 13, 556- 567. 172. Kapron-Bras, C.M. and Trasler, D.G. (1985) Reduction in the frequency of neural tube defects in splotch mice by retinoic acid. Teratoloqy, 32, 87-92. 173. Nelson, M.M. (1960) Teratogenic effect of pteroylglutamic acid deficiency of the rat. Ciba Foundation Symposium on Congenital Malformations (eds: G.W.E. Wolstenholme and M. O'Connor), ChurchilhLondon. 134. 174. Laurence, K.M. (1982) Neural tube defects: A two-pronged approach to primary prevention. Pediatrics, 70, 648-650. 175. Smithells, R.W., Sheppard, S. and Schorah, C.J. (1976) Vitamin deficiencies and neural tube defects. Archives of Disease in Childhood, 51, 944-950. 176. Laurence, K.M., James, N. and Miller, M. (1980) Increased risk of recurrence of pregnancies complicated by fetal neural tube defects in mothers receiving poor diets and possible benefits of dietary counselling^ British Medical Journal, 281, 1592. 177. Smithells, R.W., Sheppard, S. and Schorah, C.J. (1980) Possible prevention of neural tube defects by periconceptional vitamin supplementation. Lancet, i, 339-340. 178. Smithells, R.W., Sheppard, S. and Schorah, C.J. (1981) Apparent prevention of neural tube defects by periconceptional vitamin supplementation. Archives of Disease in Childhood, 56, 911-918. 179. Laurence, K.M., Nansi, J. and Miller, M.H. (1981) Double-blind randomised controlled trial of folate treatment before conception to prevent recurrence of neural-tube defects. British Medical Journal, 282, 1509- 1511. 180. Wald, N.J. and Polani, P.E. (1984) Neural tube defects and vitamins: The need for a randomised clinical trial. British Journal of Obstetrics and Gynaecology, 91, 516-523. 181. Renwick, J.H. (1972) Spina bifida, anencephaly and potato blight. Lancet, J_i, 967-968. 182. Editorial: Neural tubers. Lancet, U, 222, 1972. 183. Emanuel, I. (1972) Non-tuberous neural tube defects. Lancet, ii, 879. 184. Renwick, J.H. (1973) Anencephaly and potatoes. Lancet, i, 96-97. 185. Renwick, J.H., Possamai, A.M. and Munday, M.R. (1974) Potatoes and spina bifida. Proceedings of the Royal Society of Medicine, 67, 360-364. 186. NEvin, N.C. and Merrett, J.D. (19;75) Potato avoidance during pregnancy in women with a previous infant with either anencephaly and/or spina bifida. British Journal of Preventive and Social Medicine, 29, 111-115. 187. Renwick, J.H. (1982) Vitamin supplementation and neural tube defects. Lancet, i, 748. 188. Poswillo, D.E., Sopher, D. and Mitchell, S. (1972) Experimental induction of foetal malformation with 'blighted' potato: A preliminary report. Nature, 239, 462-464. 189. Poswillo, D.E., Sopher, D. and Mitchell, S. (1973) Further investigations into the teratogenic potential of imperfect potatoes. Nature, 244, 367-368. 190. Poswillo, D.E., Sopher, D. and Mitchell, S. (1973) Investigations into the teratogenic potential of imperfect potatoes. Teratology, 8, 339-348. 191. Swinyard, C.A. and Chaube, S. (1973) Are potatoes teratogenic for experimental animals. Teratology, 8, 349-358. 192. Plerro, L.J., Haines, J.S. and Osman, S.F. (1977) Teratogenicity and toxicity of purified a-chaconine and a-solanine. Teratology, 2, 31A-32A. 193: Sharma, R.P., Willhite, C.C., Wu, M.T. and Salunkhe, D.K. (1978) Teratogenic potential of blighted potato concentrate in rabbits, hamsters and minature swine. Teratology, 18, 55-62. 194. Keeler, R.F., Young, S., Brown, D., Stallknecht, G.F. and Douglas, D. (1978) Congenital deformities produced in hamsters by potato sprouts. Teratology, 17, 327-334. 195. Renwick, J.H., Claringbold, W.D.B., Earthy, M.E., Few, J.D. and McLean, C.5. (1984) Neural tube defects produced by Syrian hamsters by potato glycoalkaloids. Teratology, 30, 371-381. 196. Keeler, R.F. (1971) Teratogenic compounds of veratum californicum (Durand). XI. Gestational chronology and compound specificity in rabbits. Proceedings of the Society of Experimental Biology and Medicine, 136, 1174-1179. 197. Brown, D. and Keeler, R.F. (1977) Structure-activity relation of steroidal amines-teratogenic solanidine derivatives. American Chemical Society, A bstract. 198. Keeler, R.F., Young, S. and Brown, D. (1976) Spina bifida anencephaly and cranial blot produced in hamsters by the solanum alkaloid solasodine. Research Communications in Chemical Pathology and Pharmacology, 13, 723-729. 199. Turk, D.E., Sunde, M.L. and Hoekstra, W.G. (1959) Zinc deficiency experimental with poultry. Poultry Science, 38, 1256. 200. Hurley, L.S. (1981) Teratogenesis of trace element deficiency. Physiology Reviews, 61, 249-295. 201. Iniguez, C., Casas, J. and Carreres, J. (1978) Effects of zinc deficiency on the chick embryo blastoderm. Acta Anatomica, 101, 120-129. 202. Warkany, J. and Petering, h.G. (1972) Congenital manformations of the central nervous system in rats produced by maternal zinc deficiency. Teratology, 5, 319-334. 203. Sandstead, H.H. (1985) Essentiality for brain development and function. Nutrition Reviews, 43, 129-137. 204. Jameson, S., (1976) Zinc and copper in pregnancy. Correlations to fetal andmatemal complications. Acta Medica Scandanavica, Suppl.593, 5- 89. 205. Mukherjee, M.D., Sandstead, H.H. and Ratnaparkhi, M.V. (1984) Maternal zinc, iron, folic acid and protein nutriture and outcome of human pregnancy. The American Journal of Clinical Nutrition, 40, 496-507. 206. Hambridge, K.M., Krebs, N.F. and Jacobs, M.A. (1983) Zinc nutritional status during pregnancy: A longitudinal study. The American Journal of Clinical Nutrition, 37, 429-442. 207. Campbell-Brown, M., Ward, R.J. and Haines, A.P. (1985) Zinc and copper in Asian pregnancies - is there evidence for a nutritional deficiency? British Journal of Obstetrics and Gynaecology, 92, 875-885. 208. Simmer, K., Punchard, N.A., Murphy, G. and Thompson, R.P.H. (1985) Prostaglandin production and zinc depletion in human pregnancy^ Pediatric Reseach, 19, 697-700. 209. Cavdar, A.O., Arcasoy, A., Baycu, T. and Himmertoglu, O. (1980) Zinc deficiency and anencephaly in Turkey. Teratology, 22, 141. 210. Prema, K. (1980) Predictive value of serum copper and zinc in normal and abnormal pregnancy. Indian Journal of Medical Research, 71, 554-560. 211. Bergman, K.E., Makosch, M.D.G. and Tews, K.-H. (1980) Abnormalities of hair zinc concentration in mothers of newborn infants with spina bifida. The American Journal of Clinical Nutrition, 33, 2145-2150. 212. Meadows, N.J., Ruse, W. and Smith, M.F. (1981) Zinc and small babies. Lancet, ii, 1135-1137. 213. Tutle, S., Aggett, P.J., Campbell, D. and Macgillurray, I. (1958) Zinc and copper nutrition in human pregnancy: A longitudinal study in normal primigravidae and in primigravidae at risk of delivery a growth retarded baby. The Americal Journal of Clinical Nutrition, 41, 1032-1041. 214. Kubasik, N.P., Graham, M. and Harrison, E.S. (1979) Storage and stability of folate and vitamin B .? in plasma and blood samples. Clinica Chimica Acta, 95, 147-149. 1 215. Voogd, C.E., Block-Scholoten, M.G.A. and Spalburg, E.C. (1985) Stability of folic acid activity in serum. Medical Laboratory Sciences, 42, 233-236. 216. Landon, J. and Mahmood, S. (1982) Distribution of drugs between blood and ' saliva. In: Immunoassays of Steroids in Saliva, 9th Tenovus Workshop, (eds: G.F. Read, D. Riad Fahmy, R.F. Walker and K. Griffiths) Alpha Omega Publishing Company:Cardiff. 47-55. 217. Vining, R.F. and McGinley, R.A. (1982) Transport of steroids from blood to saliva. In: Immunoassays of Steroids in Saliva, 9th Tenovus Workshop, (eds: G.F. Read, D. Riad Fahmy, R.F. Walker and K. Griffiths), Alpha Omega Publishing Company:Cardiff. 56-63. 218. Read, G.F. (1986) Free steroid hormones: Significance and salivary assays. Laboratory P ractice, 5, 56-63. 219. Riad-Fahmy, D., Read, G.F., Joyce, B.G. and Walke, R.F. (19 ) Steroid immunoassays in endocrinology. In: Immunoassays for the 80's. 220. Lowe, C.R. (1979) In: An Introduction to Affinity Chromatography. North Holland Publishing Company:Oxford. 221. Pharmacia Fine Chemicals (1979) Affinity chromatography Principles and Methods. Ljungforetagon AB:Orebo. 222. Koenigsberger, R. (1981) Principles of high performance liquid chromatography. In: Therapeutic Drug Monitoring, (eds: A.Richards and V.Marks), Churchill Livingstone:New York. 223. Smith, B.L., Stoll, M.S. and Perrett, D. (1987) High performance liquid chromatographyi In: A Guide to Gas-Liquid and High Performance Liquid Chromatography, (eds: G.S.Challand and M.R.Holland). Association of Clinical Biochemists:London. 66-129. 224. Majors, R.E. (1980) Recent advances in HPLC packings and columns. Journal of Chromatographic Science, 18, 488-511. 225. This Week (1988) How Death Stalks Europe. New Scientist, 23. APPENDIX Publications Harvey, M.H., McMillan, M., Morgan, M.R.A. and Chan, H.W.-S. Solanidine is present in the sera of healthy individuals and in amounts dependent on their dietary potato consumption. Human Toxicology (1985), 4, 187-194. Harvey, M.H., Morris, B.A., McMillan, M. and Marks, V. Measurement of potato steroidal alkaloids in human serum and saliva by radioimmunoassay. Human Toxicology (1985), 4, 503-512. Harvey, M.H., Morris, B.A., McMillan, M. and Marks, V. Potato steroidal alkaloids and neural tube defects: Serum concentrations fail to demonstrate a causal relation. Human Toxicology (1986), 5, 249-253. Harvey, M.H., McMillan, M., Morris, B.A. and Marks, V. The use of saliva to measure potato steroidal alkaloids in humans. In: Immunoassays for Veterinary and Food Analysis - 1, (eds: Morris, B.A., Clifford, M.N. and Jackman, R.) Elsevier Applied Science Publishers:London (1988). ppl51-161.