GRAM-NEGATIVE ENDOTOXAEMIA by JOHN MICHAEL TUCHEK B . S c , The University of Saskatchewan, 1969 M.Sc., The University of Saskatchewan, 1978 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DOCTOR OF PHILOSOPHY DEGREE in THE FACULTY OF GRADUATE STUDIES Department of Pharmacology, Faculty of Medicine We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA JULY 1983 0 John Michael Tuchek, 1983 I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r an a d v a n c e d d e g r e e a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by t h e h e a d o f my d e p a r t m e n t o r by h i s o r h e r r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n o t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . D e p a r t m e n t o f ^Pk^^M t i ^ o /o The U n i v e r s i t y o f B r i t i s h C o l u m b i a 1956 Main Mall V a n c o u v e r , Canada V6T 1Y3 D a t e y ^ n ^ W e v ^ ( i i ) ABSTRACT Endotoxins are 1 ipopolysaccharide (LPS) complexes extractable from the outer membrane of gram negative b a c i l l i . These complexes comprise the 0-antigen of gram-negative bacteria and are used in the serological typing of these organisms. The release of endotoxins from gram-negative bacteria is believed to be a contributing factor to the high mortality rate associa-ted with gram-negative septic shock in humans. Indeed, the sequelae of gram-negative septic shock seen c l i n i c a l l y can be reproduced in experimental animals by the parenteral administration of purified endotoxin. The c l i n i -cal management of gram-negative septicaemia would, perhaps, be much more effective i f agents capable of antagonizing the actions of endotoxin in vivo could be used in combination with appropriate antibiotic therapy. The work reported in this thesis has ut i l ized both biochemical and physiological approaches to obtain more insight into the mechanism of action of E. col i endotoxin with the ultimate hope of finding an effective antagonist for this toxic substance. In vivo studies with rats and guinea pigs have indicated that elevations in plasma lysosomal enzyme act ivit ies can be used as a measure of the severity of endotoxic shock. The lysosomal enzymes examined in this study included acid phosphatase, N-acetyl-p-glucosaminidase and cathepsin D. Interestingly, when the investigation of these enzymes was extended to include patients in gram-negative septic shock, i t was found that the plasma of the patients also contained lysosomal enzyme act iv i t ies that were s ignif icantly elevated relative to controls. Furthermore, the average ( i i i ) plasma act iv i ty of cathepsin D in gram-negative septic shock patients was found to be s ignif icantly higher than that seen in the plasma of patients in other forms of shock. Thus, these preliminary studies indicate that measure-ments of plasma cathepsin D act ivity may have diagnostic and possibly prog-nostic value in c l in ica l gram-negative septic shock. Other in vivo investi-5 1 gations included tissue distribution studies with Cr-radiolabelled E. co l i endotoxin in guinea pigs. A positive correlation (r = 0.964) was found to exist between the accumulation of endotoxin in lung tissue and toxicity (determined by plasma acid phosphatase act iv i ty) . These results are consis-tent with what is known c l i n i c a l l y regarding the lung being a primary organ involved in the pathophysiology of gram-negative septic shock. 51 In vitro studies u t i l i z ing Cr-labelled E. co l i endotoxin and human erythrocyte membranes revealed that endotoxin is capable of binding to membranes in a specific manner. The binding of endotoxin to membranes was associated with measureable changes in functional properties including inhi-bition of K+-p-nitrophenylphosphatase act ivi ty and protection of red blood cel ls from hypotonic ly s i s . Studies with chemically modified endotoxins and enzymatically modified human erythrocytes suggested that the stabi l iz ing action of endotoxin involved l i p i d - l i p i d interactions between toxin and cell membrane. On the basis of these studies, certain membrane-active drugs including methylprednisolone, lidocaine and propranolol were selected for study as possible endotoxin antagonists. Of these agents, only propranolol 51 effectively antagonized the binding of Cr-endotoxin to erythrocyte membranes in v i t ro . The effectiveness of propranolol as an endotoxin anta-(iv) gonist was also demonstrated in vivo by its ab i l i ty to s ignif icantly reduce the accumulation of endotoxin in lung tissue and to lower plasma acid phosphatase act ivity in endotoxin-treated animals. However, only the d-isomer was effective in vivo whereas the racemate was poorly tolerated by endotoxin-treated animals. These studies indicate that membrane-active drugs (such as propranolol) are capable of antagonizing certain actions of endotoxin in vivo and as such may prove to be valuable adjuncts to specific antibiotic therapy in the c l in i ca l management of gram-negative septicaemia. (v) TABLE OF CONTENTS Abstract i i Table of Contents v List of Figures v i i i List of Tables xi Acknowledgements x i i Quotation xi i i CHAPTER 1. INTRODUCTION 1.1. Historical Background 1 1.2. The Gram-negative Cell Envelope 4 1.3. Isolation, Composition and Structure of Endotoxin 6 1.4. Biological Act ivi t ies of Endotoxins 18 1.5. Cl inica l Significance of Endotoxins 26 CHAPTER 2. MATERIALS AND METHODS 2.1. Membrane Preparations 33 2.1.1. Erythrocyte Ghosts 33 2.1.2. Heart Sarcolemmal Membrane 34 2.1.3. Liver Membranes 35 2.1.4. Lung Membranes 36 2.1.5. Preparation of Liver Lysosomes 36 2.2. Enzyme Assays 37 2.2.1. Membrane Bound Enzymes 37 22.214.171.124. Acetylcholinesterase 37 126.96.36.199. Nitrophenylphosphatase (NPPase) 37 188.8.131.52. Adenosine Triphosphatase (Na+,K+)-ATPase 38 184.108.40.206. Cytochrome C Oxidase 39 220.127.116.11. 5'-Nucleotidase 40 (vi) 2.2.2. Lysosomal Enzymes 40 18.104.22.168. Acid phosphatase 40 22.214.171.124. N-Acetyl-p-glucosaminidase 41 126.96.36.199. Cathepsin D 42 2.3. Haemolysis Experiments 43 2.4. Enzyme Treatments of Intact Red Blood Cells 44 2.4.1. Trypsinization 44 2.4.2. Neuraminidase Treatment 45 2.4.3. Phospholipase A 2 Treatment 45 2.4.4. Removal of Cholesterol from Intact Erythrocytes 46 2.5. Compositional Assays 48 2.5.1. Protein 48 2.5.2. S ia l ic Acid (N-Acetylneuraminic Acid) 48 2.5.3. Cholesterol Analysis 50 2.5.4. Phospholipid Analysis 50 2.5.5. Ketodeoxyoctanoic Ac.id 51 2.6. Thin Layer Chromatography 52 2.6.1. Intact Erythrocytes 52 2.6.2. Membranes and Endotoxins 54 2.6.3. Separation of Phospholipids ' 54 2.7. Detoxification of Endotoxin 55 2.7.1. Sodium Hydroxide Treatment 55 2.7.2. Sodium Periodate Treatment 55 2.7.3. Treatment with Hydroxylamine 56 2.8. Radioactive Labelling of Endotoxin 56 2.9. 5 lCr-Endotoxin Binding Studies 57 2.9.1. Membrane Preparations 57 2.9.2. Intact Erythrocytes 58 (vi i) 2.10. Liquid Sc int i l la t ion Counting 58 2.11. Preparation of Samples for Sc int i l la t ion Counting 59 2.11.1. Plasma Membranes 59 2.11.2. Intact Red Blood Cells 59 2.11.3. Tissues 59 2.11.4. Plasma 60 2.12. Animal Studies 60 CHAPTER 3. RESULTS 3.1. Some Physiological Effects of Endotoxin Administration 3.2. Binding Studies with ^Cr-Endotoxin 3.3. Effects of Endotoxin on Red Blood Cell Membranes 3.3.1. Erythrocyte Ghosts 3.3.2. Intact Red Blood Cells 3.4. Cr-Endotoxin Binding In Vivo 3.5. Study of Drugs as Possible Endotoxin Antagonists 3.5.1. Antagonists to Endotoxin Binding In Vitro 3.5.2. Effectiveness of Endotoxin Antagonists In Vivo 3.6. Effect of Endotoxin and Gentamycin on Endotoxin Toxicity In Vivo 119 3.7. Var iabi l i ty in Commercially Available Endotoxin Preparations 122 CHAPTER 4. DISCUSSION AND CONCLUSIONS 4.1. Gram-negative Septicaemia: A Formidable Medical Problem 124 4.2. Host Defense Mechanisms in Bacteraemia and Endotoxaemia 126 4.3. Pathophysiology of Endotoxaemia 131 4.4. Therapy of Gram-negative Bacteraemia 151 REFERENCES 162 61 74 82 82 84 99 106 106 107 (vi i i) LIST OF FIGURES Fi gure Page 1 E. co l i Cell Envelope 5 2 Structure of E. col i Lipopolysaccharide 14 3 Structure of Lipid A 15 4 Effect of Endotoxin and Haemorrhage on Blood Pressure and Nutritive Flow in the Rat 62 5 Effect of Native and Detoxified E. col i Endotoxins on Plasma Acid Phosphatase Act ivi ty in the Rat 65 6 Effect of E. col i Endotoxin on Plasma Lysosomal Enzyme Act iv i ty in Guinea Pigs 67 7 Effect of Varying Doses of E. col i Endotoxin on Plasma Lysosomal Enzyme Activi ty in Guinea Pigs 69 8 Comparison of Plasma Cathepsin D Activi t ies in Patients C r i t i c a l l y 111 with Sepsis and/or Shock 73 51 9 Chromatography of Cr-E. col i Endotoxin 77 51 10 Displacement of Bound Cr-E. col i Endotoxin from Human Erythrocyte Membranes by Unlabelled Endotoxin 78 11 Comparison of ^ C r - E . col i Endotoxin Binding to Human Erythrocytes and Erythrocyte Membranes 79 + ++ 12 Effect of E. col i Endotoxin on K - and Mg -p-Nitrophenyl-phosphatase Act ivi ty in Human Erythrocyte:Membranes 83 13 Antihaemolytic Effect of E. col i Endotoxin on Human Red Blood Cells 85 14 Effect of E. col i Endotoxin on Pre-lyt ic Leakage of K + in Human Erythrocytes 86 (Tx) Figure Page 15 Effects of Increasing Concentrations of E. co1i Endotoxin on the Osmotic Stabi l i ty of Rat and Human Red Cells 87 16 Temperature-Dependent Effects of Endotoxin on the Osmotic Stabi l i ty of Erythrocytes 89 17 Comparison of the Temperature-Dependent Effects of Endotoxin on the Osmotic Stabi l i ty of Erythrocytes from Various Animal Species 93 18 Temperature-Dependent Effects of Endotoxin (4 mg/10^ red cells) on Modified Human Erythrocytes 95 19 Effects of Detoxified Endotoxins on the Osmotic Stabi l i ty of Human Erythrocytes as a Function of Temperature 97 20 Fractionation of Sodium Hydroxide-Detoxified Endotoxin 98 21 Fractionation of Hydroxylamine-Detoxified Endotoxin 100 51 22 Correlation Between Accumulation of Cr-Endotoxin in Various Guinea Pig Organs with Toxicity 102 23 Accumulation of ^Cr-Labelled Native and Detoxified Endotoxins in Guinea Pig Lung Tissue 104 51 24 Clearance of Cr-Labelled Native and Detoxified Endotoxins from Guinea Pig Plasma 105 51 25 Displacement of Bound Cr-E. co l i Endotoxin from Human Eryth-rocyte Membranes with Lidocaine 108 51 26 Effect of Methyl prednisolone, Pranolium and Propranolol on Cr-E. col i Endotoxin Binding to Membranes 109 51 27 Double Reciprocal Plot of Cr-E. col i Endotoxin Binding to Erythrocyte Membranes in the Presence of Propranolol 110 (x) Figure Page 51 / 28 Double Reciprocal Plot of Cr-E. col i Endotoxin Binding (High Concentrations) to Membranes in the Presence of Propranolol 111 29 Effect of Various Drugs on Plasma Acid Phosphatase Act iv i ty in Endotoxin-Treated Rats 113 51 30 Effect of Drug Pretreatment on Accumulation of Cr-Endotoxin in Guinea Pig Lung 114 31 Effect of Drug Treatment on Mortality Rates in Mice Injected with Endotoxin 116 32 NaOH-Detoxified Endotoxin as an Endotoxin Antagonist in vivo 117 33 Periodate-Detoxified Endotoxin as an Endotoxin Antagonist in vivo 118 34 Effect of Gentamycin in Combination with Endotoxin on Toxicity in Rats 120 35 Effect of Chronic Gentamycin Treatment on Mortality in Endotoxin-Treated Mice 121 36 Correlation Between Extent of TNBS Incorporation into E. col i Endotoxin and Toxicity in Rats 123 (xi) LIST OF TABLES Table Page 1 Plasma lysosomal enzyme act iv i t ies in patients with septicaemia and/or shock 70 2 Comparative effects of experimental endotoxaemia and haemorrhage on lysosomal hydrolase act ivi t ies in plasma 75 51 3 Binding characteristics of Cr-endotoxin in intact human erythrocytes and erythrocyte ghosts 81 51 4 Effect of temperature on the binding of Cr-labelled lipopolysac-charide (serotype 026:B6, lot number 669176) by human erythrocytes 90 5 Thin layer chromatographic analysis of phospholipid profiles from normal human erythrocytes, rat erythrocytes and erythrocytes from a patient with a congenital deficiency of plasma lecithin:choles-terol acyltransferase (LCAT) 92 (xii) ACKNOWLEDGEMENTS Firs t of a l l , I would l ike to extend my warmest appreciation to Dr. M. C. Sutter for suggesting and ini t ia t ing this challenging project as well as for his f a i thfu l , moral support throughout the study. Secondly, I feel an unfathomable depth of gratitude to Dr. D. V. Godin for his guidance, untiring encouragement and supervision when the project was transferred into his laboratory. The years I have been associated with Dr. Godin were a most enriching experience for me in many ways. His unique ab i l i ty of combining professional expertise as a teacher and scientist with an unparalelled loyal friendship has provided me a matchless Ph.D. programme. The superb technical assistance of Therese Ng and Maureen Garnett have made the countless d i f f icul t ies encountered during the research project tota l ly vanquishable. In addition, their beloved friendship and moral support wi l l never be forgotten. Also, I was fortunate to have available to me the typing expertise of Jackie Bitz and Tracy Slocombe to help put the thesis into its f inal form. Maureen's generous help in this regard as well , is greatly appreciated. In addition, I would l ike to express my gratitude to the Medical Research Council of Canada for studentship support and to my sister Frances who over the years fa i thful ly sustained me through al l tr ibulations. ( x i i i ) "The search for truth is in one way hard and in another easy, For i t is evident that no one can master i t fu l ly , nor miss i t wholly, But each adds a l i t t l e to our knowledge of nature, And from al l the facts assembled, there arises a certain grandeur." Aristotle - 1 -CHAPTER 1 Introduction 1.1 Historical Background It has been recognized for more than a century that certain extracts from bacteria contain substances that can k i l l animals. Panum, a Danish pathologist, reported in 1856 that intravenous injections of an aaueous extract from putrifying tissues were lethal to dogs (1). Panum further noted that these extracts which were f i l tered , d i s t i l l e d and redissolved in water, could be boiled for eleven hours and s t i l l possess the ab i l i ty to k i l l animals. Since no microorganisms could have survived these procedures, Panum postulated a chemical theory of putrifaction and septic disease. Later on in the century, in 1888, Gaertner was the f i r s t to attribute meat poisoning to an organism (Bacteroides enterit idis) which he isolated. Furthermore, Gaertner found that boiled cultures of this organism were toxic when given to guinea pigs and rabbits (2) which supported Panum's postulate of a chemical entity present in bacteria that was responsible for septic disease. However, i t was near the turn of the century, in 1892, before the use of the term "endotoxin" was popularized by Pfeiffer and his colleagues (1,3) to denote a poison which was part of the l iv ing substance of bacteria and which was released only upon the disintegration of the bacterial ce l l s . The notation by Pfeiffer that endotoxins were released only upon lysis of the bacterial cel l was important at that time because this notation served to distinguish endotoxins from "exotoxins" which were known to be toxic substances that are synthesized and excreted by intact, multiplying bacteria. In view of the present knowledge of microbial toxins, a better distinc-tion between endotoxins and exotoxins can be made. Exotoxins are usually - 2 -proteins which are secreted by gram-positive and some gram-negative bacter-ia . They can be completely inactivated by heating at 60-80°C or they can be converted to toxoids which are inactive but s t i l l maintain the antigenicity of the active toxin. Exotoxins from different bacterial species exhibit their own unique pharmacological and biological actions on a particular cel l type or tissue and these actions can be neutralized by antitoxins. In contrast to exotoxins, endotoxins are 1 ipopolysaccharide-protein complexes that are produced exclusively by gram-negative bacteria. These complexes are heat resistant, do not form toxoids and are d i f f i cu l t to neutralize with antibodies. Another large difference between endotoxins and exotoxins is that endotoxins precipitate the same vast array of biological actions irres-pective of their bacterial species of origin (1,3,4,5). Pfeiffer 's theory that endotoxins were located in the protoplasm of bacterial cel l s and that the only way they could be released was through lysis of the bacterial cells prevailed for several decades. Doubts to Pfeiffer 's postulates began to appear when Ecker reported in 1917 that he could demonstrate the presence of a heat-stable toxin in the f i l t rates of young growing cultures where there was no sign of autolysis (6). Several decades later, investigators realized that both the toxic and O-antigenic properties of gram-negative bacteria could be ascribed to one macromolecular complex (1,7,8,9). Since the antigenic determinants were superficial struc-tures on the bacterium, this implied that endotoxins were also superficial components and that they were probably located on the cell wall of gram-neg-ative bacteria. Further proof that endotoxins were constituents of the cel l wall of gram-negative bacteria came from two independent studies in 1959 when Carey and Baron, working with Salmonella typhosa and Ribi et a l . , - 3 -working with Salmonella enter i t id i s , compared the toxic and antigenic properties of the cel l walls of these organisms with their corresponding protoplasmic extracts. These workers found that almost a l l of the toxic and antigenic reactions were contained in the cel l wall fraction of these organ-isms and not in the protoplasmic fraction (10,11). These results were further substantiated by Rudbach and co-workers in 1969 (12). Evidence that endotoxins are not only superficial components of gram-negative bacteria but that they can be readily released into the surrounding environment as well , was reported by several investigators (13,14) including Crutchly and co-workers (15) who used the term "free endo-toxin" to describe material found free in aerated l iquid cultures of several species of gram-negative bacteria and which possessed the properties of chemically extracted endotoxin (16). Crutchley postulated that "free endo-toxin" was due to a metabolic over-production of cel l wall material during vigorous growth in an aerated l iquid medium. In addition, Roberts has shown that simple heat treatment (80°C) of b a c i l l i in sodium chloride can cause release of up to one-half the total endotoxin content of cel ls (17) and Rogers demonstrated in 1971 that the release of endotoxin from E. col i could be effected by a warm water treatment with 0.1 M Tris without loss of v i a b i l i t y of the bacterial cel ls (18). These experiments established that endotoxin is a readily solubi1izable component of the gram-negative bacteri-al cel l wall rather than a cytoplasmic component (19) and validated Eckers' suspicions in 1917. It would therefore seem misleading to use the term "endotoxin" to signify a readily releasable toxic material that resides on the outer surface of gram-negative bacteria since the term (as denoted by Pfeiffer) implies that this toxin is contained within the bacterial proto-- 4 -plasm and released only upon lysis of the organism. Nonetheless, the nota-tion has prevailed and serves mainly as a label to identify these gram-nega-tive toxins from exotoxins which are predominantly produced by gram-positive organisms. 1.2 The Gram-negative Cell Envelope The cel l walls of gram-negative bacteria appear quite different from those of gram-positive bacteria when sections are observed under the elec-tron microscope. Cell walls of gram-positive bacteria consist mainly of a thick structureless layer composed mostly of peptidoglycan with some teichoic acid or polysaccharide or both (20). On the other hand, the gram-negative cel l envelope is a multilayered structure that is composed of predominantly the lipopolysaccharide somatic antigen (LPS), l i p id s , proteins and usually only a small amount of peptidoglycan (20) as shown schematically in Figure 1. Essentially, the gram-negative cell envelope is composed of two dist inct membranes; an inner cytoplasmic or plasma membrane (CM) and an outer membrane (OM), both of which demonstrate the usual double-track or bilayer appearance of membranes when observed under the electron microscope (21,22). The cytoplasmic and outer membranes are separated by an area of approximately 100 A which is referred to as the periplastic region and which consists of periplasmic space (PS) plus a peptidoglycan layer (PG) of approximately 25 A in thickness (23). Both the cytoplasmic and outer membranes are about 75 A thick (24). Some gram-negative bacteria also produce additional layers located externally to the outer membrane (25). Freeze-etching studies have confirmed this multilayer structure of the gram-negative cel l envelope (26,27). - 5 -F i g u r e 1 . E. c o l i C e l l E nvelope. A schematic r e p r e s e n t a t i o n i l l u s t r a t i n g the p o s s i b l e m o l e c u l a r a r c h i t e c -t u r e of the E. c o l i c e l l e n v e l o p e . A b b r e v i a t i o n s used a r e : LPS, l i p o p o l y -s a c c h a r i d e ; PL, p h o s p h o l i p i d ; OM, o u t e r membrane; PG, p e p t i d o a l y c a n ; PS p e r i p l a s m s space; and CM, c y t o p l a s m i c membrane. P o l y s a c c h a r i d e c h a i n s i n o n l y some of the LPS mo l e c u l e s are shown (taken from r e f . 23). - 6 -Although cytoplasmic and outer membranes of gram-negative bacteria are similar in the sense that both of these membranes are bilayers of l i p id and protein, these membranes have major functional as well as compositional differences. The cytoplasmic membrane contains enzymes for many biosynthe-t ic and transport functions including the enzymes for electron transport whereas the outer membrane is devoid of biosynthetic transport and electron transport functions (24,28). Indeed, the components of the outer membrane are formed by the cel l protoplasm and by the cytoplasmic membrane and these components are then transported outwards (24,29,30). Studies determining the compositional makeup of the cytoplasmic and outer membranes have been hindered by the technical d i f f i cu l t ie s involved in the complete separation of these two membranes for biochemical analysis. Nonetheless, i t is quite apparent that major compositional differences do exist in these two membranes (28,31). The main difference is that 50-60% of the outer membrane is composed of somatic antigen which is 1ipopolysaccharide in nature whereas the cytoplasmic membrane contains l i t t l e or no 1ipopolysaccharide (20). Also, as one may expect on the basis of the lack of enzymes, the outer membrane contains much less protein by weight (11—15%) than does the cyto-plasmic membrane (70-80%) (20,32). The phospholipid content of both membranes is s imilar. 1.3 Isolation, Composition and Structure of Endotoxin The most extensively studied component of the gram-negative cell wall has been endotoxin. The interest given to endotoxin is largely due to the fact that this macromolecular complex, which forms part of the outer membrane of the ce l l wall , contains the major somatic antigen (1ipopolysac-charide) in enteric bacteria and i s , therefore, important from the viewpoint - 7 -of immunochemistry and taxonomy in addition to the fact that endotoxin is a powerfully toxic agent (20). In light of this importance of endotoxin, various methods of extracting this component from gram-negative bacteria have been devised and this has been a significant factor in enhancing scien-t i f i c advances towards the greater understanding of both the biology and the chemistry of these bacterial macromolecules. One of the earliest methods used to extract endotoxin was a procedure described by Boivin in 1933 that employed a precipitation step with trichloroacetic acid at 4°C (33). Boivin called the extracted material "glucidolipide" as the material contained most of the somatic or 0-antigen of the intact .organism from which i t was derived. With further analysis of the toxic substance, Boivin in 1946 stated that i t consisted of a phospholi-pid and a nitrogenous component in addition to the polysaccharide moiety (1). Since the nitrogenous component consisted of polypeptides, Boivin suggested that these extracted endotoxins would be more appropriately described chemically as "glucidolipido-polypeptidiques". Extraction of gram-negative b a c i l l i with aqueous phenol has become the most popular method of obtaining endotoxin because the yields are much greater than with any other method (1). The original phenol extraction method of Palmer and Gerlough in 1940 (34) was modified and improved by Westphal and co-workers in 1952 (35). Further improvements were made in 1965 by Westphal and Oann (36). According to this method of extraction, dried bacteria are treated with a mixture of phenol and water (45/55, v/v) at 68°C for five minutes. Upon cooling, the homogenous mixture separates into an upper (water) phase and a lower (phenol) phase. Under these circum-stances, the endotoxin is found in the water phase as a complex with nucleic - 8 -acid. The phenol is removed either by dialysis or by ether extraction. Since endotoxin forms large aggregates in water, i t can be sedimented by ultracentrif ligation (105,000 x g) and thereby obtained relat ively free of nucleic acid contaminants. The yield of endotoxin that is obtained with this extraction procedure varies from 1 to 4% of the dry bacterial weight. Other effective methods of extracting endotoxins from gram-negative bacteria that have been reported include the diethylene glycol extraction procedure of Morgan (37), the aqueous ether extraction method as described by Ribi et a l . (11), the EDTA extraction procedure of Leive and co-workers (38) and a procedure that employs aqueous butanol as reported by Morrison and Leive (39). Each of these extraction methods yields a biological ly active material that is rich in 1 ipopolysaccharide with amounts of protein and l ip id that vary according to the particular extraction procedure used (40,41). Since the major constituent of endotoxin is 1ipopolysaccharide, many investigators use the terms "endotoxin" and "1ipopolysaccharide" inter-changeably. However, in a s t r ic t chemical sense, these two terms are not synonymous in that endotoxins are extracts that contain protein and loosely bound l ipids complexed with the lipopolysaccharide (1,3,42,43). The contri-bution that the various components of endotoxin make to the toxicity of the whole complex has been the subject of numerous investigations. The loosely bound l ip ids , which are cephalins and designated as " l i p i d B" by some inves-tigators (44,45) can be easily removed from the endotoxin complex by chloro-form extraction. Since the removal of these l ipids has no effect on the biological toxicity of endotoxin, they are regarded as being biological ly inert (44,45). In addition to l i p id B, endotoxins contain a firmly bound - 9 -l ip id moiety, denoted " l i p i d A" by Westphal and co-workers (46) which is covalently linked to both the polysaccharide and protein entities of the complex. Since l ip id A is covalently bound, i t can only be separated from the other components of endotoxin (polysaccharide and protein) by acid hydrolysis. Boivin and Mesrobeanu in 1933 (reported in 42,47) were the f i r s t to study the effects of acid hydrolysis on endotoxins. These invest i-gators treated endotoxin from Serratia marcescens with 0.2 N acetic acid at 100°C and found that a l i p id precipitate which they called "Fraction A" would be obtained and which maintained some remnants of toxic i ty . "Fraction B", which was the clear supernatant, was antigenic but not toxic. Later, Tai and Goebel proposed that endotoxins contained a "toxic moiety" (TM) which could account for the toxic effects of endotoxins (48). These workers found that upon acid hydrolysis for a much shorter time period (35 minutes) than that reported by Boivin and Mesrobeanu (4 hours) the toxic moiety of endotoxin was found to be associated with the protein fraction and upon alkaline alcohol hydrolysis, they found the toxic moiety associated with the polysaccharide component (48). Moreover, these workers concluded that the toxic moiety was neither protein nor polysaccharide. It was the investiga-tion reported by Westphal and Luderitz in 1954 that implicated the covalent-ly bound l i p i d ( l ip id A) as the toxic entity of endotoxin (46). These investigators found that by hydrolyzing endotoxin with 1 N HC1 they could obtain a precipitate which was soluble in chloroform and when coupled to a hydrophilic carrier such as protein or low molecular weight dextran, the combination would display up to one-fifth of the toxic potency of endotoxin (43,46). The polysaccharide-containing supernatants of the acid hydroly-zates were found to be non-toxic. A study by Wober and Alaupovic reported - 1 0 -in 1971 has also supported the proposal that l i p id A is the toxophore of the endotoxin complex (49). Other evidence that suggests that the l ip id A component of endotoxins is the toxic entity are the studies with bacterial mutants deficient in the O-antigenic polysaccharide chain. Endotoxins from these mutants are essentially composed of l i p id A only and their biological potency has been shown to be similar to polysaccharide-containing endotoxins prepared from non-mutant strains (50,51,52). However, other investigators have disputed the evidence that l i p i d A is the toxic moiety of endotoxin by drawing attention to the fact that l i p i d A preparations are much less toxic than the endotoxins from which they were derived (43,53). Also, R i b i and associates have demonstrated that endotoxins can lose their biological potency by mild acid hydrolysis (0.1 N acetic acid) before any measurable amount of covalently bound l ip id was released (43). Furthermore, they were able to demonstrate through the use of different extraction procedures that endotoxins containing as l i t t l e as 2% l ip id A were as biological ly active as those containing as much as 30% l ip id A (54). Ribi and associates proposed that structural configuration and/or physical size of the lipopolysaccharide complex may be more important determinants of biological toxici ty of endo-toxins than any particular molecular component (43). The protein component of endotoxins has received l i t t l e attention by investigators since Westphal and Luderitz implicated l i p i d A as the toxo-phore in endotoxin in 1954 which stimulated research on the l ip id moiety of endotoxins. Previous to 1954, one study, that was reported in 1942, did find the protein moiety necessary for endotoxic potency (55). However, investigations reported in the mid-1950's were in consensus that the protein moiety plays a minor i f not a benign role in the biological toxicity of the - 11 -endotoxin complex (56,57,58) and as a result , its presence in the endotoxin complex has largely been ignored. However, more recent investigations on the protein moiety of ^endotoxins have raised some questions regarding the earl ier conclusion that the protein moiety of endotoxins is biologically, inert. It is now known that the protein moiety is a peptide that is coval-ently linked to the l ip id A portion of the 1ipopolysaccharide complex of endotoxins (49,59,60,61). Morrison and associates have determined that the molecular weight of this l i p i d A-associated peptide is approximately twelve thousand (62). The exact contribution that this protein moiety makes to the biological effects of endotoxins is d i f f i cu l t to assess because f a i r ly rigorous procedures such as acid hydrolysis are required to remove this protein from the endotoxin complex and these extraction procedures are known to alter the 1 ipopolysaccharide moiety of the complex as well (42,47,49). Therefore, comparisons of biological toxici ty between protein-free endotox-ins and normal endotoxins are d i f f i cu l t to make. However, when the protein moiety is extracted from the endotoxin complex either by phenol extraction or acid hydrolysis, i t is collected not as pure protein, but protein conju-gated to l i p id A and this conjugated protein has been shown to possess some biological toxicity (49). Freedman and co-workers have shown that the protein component of endotoxins can act as an immunogen and can contribute to the biological effects of endotoxins by causing a delayed hypersensitiv-ity reaction (63). In addition, i t has been demonstrated that this l i p id A-associated protein is a potent mitogen and can e l i c i t mitogenic responses from lymphocytes, such as the spleen B lymphocytes from the C3H/HeJ mouse, that are normally unresponsive to the mitogenic effects of 1ipopolysacchar-ide or l i p i d A (59,62). - 12 -It has been known for many years that the polysaccharide component of endotoxin possesses antigenic properties and since the peripheral portion of the polysaccharide component consists of oligosaccharide repeating units that display a wide spectrum of var iab i l i ty within a single bacterial genus, this structural var i ab i l i ty has been used, as "O-antigens", in the fine serological typing of Enterobacteriaceae strains (33,64). The term "0-anti-gen" was f i r s t introduced by Weil and Felix in 1918 to distinguish this somatic antigen from flagellar antigens in flagellated strains of Proteus. Weil and Felix described the non-flagellated form of Proteus as the "0 form" (Ohne Hauch) and the flagellated strain as the "H form" (mit Hauch) (see 65). Therefore, the H form of Proteus was found to possess both f lagellar (H-antigens) and somatic or body antigens (O-antigens) whereas the non-flag-ellated strains only possessed O-antigens. In contrast to O-antigens, which are carbohydrate compounds and therefore heat resistant, the H-antigens are proteins whose antigenic effect can be destroyed by boiling (65). In addi-tion to 0- and H-antigens, some enterobacterial strains, particularly Escherichia, possess capsular or envelope antigens denoted as K-antigens (K derived from the German word "Kapsel") (33,65). Biochemically, the K-anti-gens are acid polysaccharides and are subdivided into thermolabile L- and B-antigens and the thermostable A-antigen (65). In general, strains of Escherichi a that contain K-antigens are more toxic than strains without K-antigens (66). Since v ir tual ly all of the biological act ivit ies attributed to endotoxin can also be e l ic i ted with 1ipopolysaccharide, which is the predominant component of endotoxin, studies directed towards elucidating the structural organization of endotoxin have focused on chemically pure 1ipopolysacchar-- 13 -ides instead. Basically, one can regard 1 ipopolysaccharides as amphipathic molecules with the hydrophilic portion being polysaccharide and the hydro-phobic portion comprised of l ip id A (64). The hydrophilic portion can also be considered to consist of two parts; a portion distal to the l ip id A region which consists of oligosaccharide repeating units (0-antigen polysac-charide) and a portion proximal to the l i p i d A region which is usually referred to as "core" polysaccharide (see Figure 2). The oligosaccharide repeating units usually consist of three or four different hexose units each (3,64). In the example given in Figure 2 of E. col i 0111:B4, the 0-antigen repeating unit consists of galactose, glucose, colitose and N-acetylglucos-amine (39,67). The number of oligosaccharide repeating units can be as few as two, in "semi-rough" bacterial mutants, or as many as ten in native or "smooth" strains (39,68,69). As has been previously mentioned, the 0-anti-gen can be used for serological typing of bacterial strains because each strain possesses an 0-antigen of unique composition. In contrast, the "core" polysaccharide displays much more constancy in structure and composi-tion for the many strains within a bacterial genus and variations in core structure are only seen when comparing different genuses of Enterobacteria- cea (70). The core polysaccharide contains a trisaccharide composed of a unique eight carbon sugar acid, 2-keto-3-deoxy-octulosonic acid as well as a seven carbon heptose, phosphorylethanolamine and several hexoses (see Figure 2). The hydrophobic region of 1ipopolysaccharide consists of l i p i d A which structurally is an unusual glycolipid that consists of e-l,6-linked D-glucosamine disaccharide units to which long chain fatty acids ( C 1 n - C i o ) are linked via both ester and amide linkages, as shown diagram-- 14 -P I P I Fa I EtNH i P I Fa - Glc (NH 2) - KDO - KDO - hept - hept - glc - gal - glc - /gal - glc - glc NAc Fa I Glc (NHL) I I Fa Fa I KDO I P EtNH hept I glc NAc I col I col lipid A Core Polysaccharide J V. J n O-Antigen Polysaccharide Figure 2. Structure of E. co l i Lipopolysaccharide. Chemical structure of E. col i 0111:B4 lipopolysaccharide (LPS) showing the three regions of the LPS molecule. Abbreviations used are: gal , galac-tose; glc, glucose; glc NAc, N-acetylglucosamine; c o l , colitose; hept, hep-tose; EtNH, ethanolamine; KDO, 2-keto-3-deoxy-octulosonic acid; Fa, fatty acid (taken from ref. 3). - 15 -Figure 3. Structure of Lipid A. Structure of l i p id A component of Salmonella lipopolysaccharides. Two l ip id A units, each consisting of two glucosamine molecules that contain ester and amide bonded fatty acids, are shown to be linked together with a pyrophosphate bridge. The 2-keto-3-deoxy-octulosonic acid trisaccharide represents the point of attachment of the "core" polysaccharide (taken from E. T. Rietschel et a l . , in Microbiology - 1975, D. Schlessinger (ed.), p. 307). 1 - 16 -matically in Figure 3. Al l the fatty acid chains in l ip id A are saturated (64), which probably accounts for the observation that the hydrophobic portions of 1 ipopolysaccharides display low f lu id i ty (71). The ester-bound fatty acids of l ip id A preparations from Enterobacteriacea are even numbered and straight-chained as well as being saturated and can occur as either unsubstituted or 3-hydroxy substituted fatty acids (70,72). In comparison, the amide-linked fatty acids appear to be uniformly 3-hydroxy substituted and, in Enterobacteriacea, this fatty acid invariably appears as 3-hydroxy-myristic acid (70) and consequently has been used as a specific marker for l i p i d A (37,70,73). A general observation in most bacterial lipopolysac-charides is that 55-75% of the total fatty acids linked to l ip id A are 3-hydroxy acids (70). It is of interest to note that the types of fatty acids bound to l i p i d A are a characteristic feature of a particular bacter-ial species and seem independent of the conditions, under which the bacteria were cultured or of the external supply of fatty acids (70,74). Therefore, a lipopolysaccharide monomer essentially consists of two glycosidical ly linked glucosamine molecules to which long-chain, saturated fatty acids are bonded via ester and amide groups. One of the glucosamine molecules is also ketosidically linked to a 2-keto-3-deoxyoctonate trisac-charide which, along with other sugars and phosphorylethanolamine, repre-sents the "core" polysaccharide region. This region, in turn, is glycosidi-cally linked to an oligosaccharide repeating unit that normally consists of four hexoses and forms the 0-antigen portion of the lipopolysaccharide. The molecular weight of a lipopolysaccharide monomer would theoretically be in the order of 14,000, that i s , i f one estimates approximately 2,000 for a l i p id A unit, 2,000 for the core polysaccharide and another 10,000 for an - 17 -average O-antigen oligosaccharide of ten pentasaccharide units (molecular weight of each unit would be approximately 1,000) (70). However, i t is known that under normal physiological conditions, the apparent molecular weight of 1ipopolysaccharides is in the order of one to twenty million (33,47,70). It follows then that lipopolysaccharides must aggregate to form large part icles . The factors responsible for aggregation seem to be predominantly hydrophobic forces exerted by the long-chain fatty acids in the l ip id A moiety (70) since agents such as detergents (both ionic and nonionic) and alkal i treatment, which spl i ts off the ester-linked fatty acids, can cause disaggregation and consequently a reduction in the appar-ent molecular weight (75,76,77). In addition, Olins and Warner have demon-strated that EDTA treatment can also cause 1 ipopolysaccharide disaggregation and they proposed that calcium ion binding as well as hydrophobic interac-tions were responsible for the association of 1 ipopolysaccharide molecules (78). However, even after various means are used to completely disperse aggregates of 1 ipopolysaccharide, the molecular weights of the smallest particles range from 25,000 to 40,000 (70). Since this observed molecular weight is approximately two to three times the calculated molecular weight of one 1ipopolysaccharide unit (approximately 14,000), this suggests that two or three 1ipopolysaccharide monomers are covalently linked together to form a polymer. Ultracentrifugation studies have supported the theory that lipopolysaccharides exist in polymeric or, more speci f ica l ly , trimeric units (79,80). The nature and position of crosslinking that has been proposed includes phosphodiester bonding between heptoses of core polysaccharides in adjacent 1 ipopolysaccharide chains (81) and pyrophosphodiester bonds that bridge diglucosamine units of l i p id A by a l ' , 4 linkage as shown in Figure 3 - 18 -(72). However, both of these proposed associations of 1ipopolysaccharide monomers have been challenged by Muhlradt and co-workers (82,83). Using 31 P-nuclear magnetic resonance techniques, Muhlradt and co-workers could find no evidence of the existence of covalent crosslinks involving phospho-diester and/or pyrophosphodiester bonds in either the core polysaccharide or the l ip id A moiety. Therefore, these investigators proposed that the aggre-gation of lipopolysaccharide molecules must result from ionic and hydropho-bic interactions (83). The poss ib i l i ty that lipopolysaccharide units could be linked by disulfide bridges is ruled out by the fact the lipopolysacchar-ides do not contain sulfur (12). Although the foregoing sections dealt with the molecular structure of lipopolysaccharides, no mention was made, either in the text or in Figure 3, of how the l ip id A-associated protein f i tted into the molecular stucture of endotoxin. This is mainly because it is not known how this protein is bound to l i p id A other than i t seems to be a covalent bond (49). Other compounds known to be associated with endotoxins include polyamines such as putres-cine, cadaverine, spermidine, and spermine; cations such as Na + , K , Mg + + and C a + + ; and l i p i d B (44,70). However, since these compounds can be easily removed from the lipopolysaccharide complex by techniques such as ion exchange chromatography (84) or by electrodialysis (85), they are not considered to be integral components of the lipopolysaccharide unit. 1.4 Biological Act ivi t ies of Endotoxins Although i t is known that the physiological responses to parenterally administered endotoxin vary markedly in different animal species (86,87), there is one common response in al l mammalian species to a dose of endotoxin of sufficient magnitude and that is death (1). As one example to i l lustrate - 19 -the differences in sensit ivity to endotoxin, a dose (on a mg/kg basis) that is lethal to mice is three orders of magnitude greater than a dose of endo-toxin that would be lethal to rabbits of comparable maturity (1). In spite of this variable response to endotoxaemia in different animal species, certain generalizations can be made which are useful in helping to under-stand the sequelae of endotoxaemia in humans. With this in mind, only those biological effects seen in experimental endotoxaemia that may be applicable to humans will be mentioned here. In the main, these effects comprise those that are responsible for the pyrexia and hypotension that is seen, and that can proceed to shock and death, in humans affl icted with gram-negative endo-toxaemia. One of the earliest biological responses (besides death) that was attributed to endotoxins was pyrogenicity, when in 1875 Burdon-Sanderson demonstrated that he could isolate a fever-inducing substance from decompos-ing meat (see 1). In the animal kingdom, Greisman and Hornick have shown that rabbits have approximately the same degree of sensit ivi ty to the pyro-genic properties of endotoxins as do human volunteers (88). The mechanism of fever production may involve a direct effect of endotoxin (or a portion of i t ) on the brain (89) or, as most of the evidence indicates, an indirect action of endotoxin on blood leukocytes to form endogenous pyrogens (90,91,92). Cells of the reticuloendothelial system, such as the Kupffer cells of the l i ve r , when incubated with endotoxin also produce pyrogens (93). The secretion of leukocyte pyrogen seems to involve a two-stage process. The f i r s t stage following endotoxin activation includes synthesis of the pyrogen de novo and the second stage involves secretion of the pyro-gen (94,95,96). The f i r s t stage of the process can be inhibited with RNA - 20 -synthesis inhibitors (97) while the secretion process can be inhibited by agents that bind to sulfhydryl groups (96). The mechanism by which endotox-in "activates" leukocytes to synthesize pyrogens is not real ly known but Dinarello has proposed that the activation process may involve a de-repres-sion of the genome for pyrogen synthesis in some as yet obscure manner (98). Evidence that may support this de-repression hypothesis is the fact that some tumor cell lines are known to spontaneously and constantly produce endogenous pyrogen which may be argued to indicate an unrepressed genome (99). Leukocyte pyrogens have molecular weights ranging from 13,000-16,000 (100,101,102), possess high specific biological act ivity (1 mg is equivalent to 33,000°C fever in a rabbit (97) and their principal site of action is in the preoptic area of the hypothalamus (103). However, the question whether or not these endogenous pyrogens can cross the blood-brain barrier is controversial and some investigators propose that leukocyte pyrogens may trigger the release of another endogenous pyrogen such as prostaglandin E^ which can penetrate the blood-brain barrier (104). A dramatic demonstration of the biological toxic i ty of endotoxins is the "local tissue reactivity phenomenon" or, as i t is more commonly known, the local "Shwartzman" reaction. This dermal reaction was described in 1928 by Shwartzman when he noted the development of a severe haemorrhagic, necrotic lesion in the skin of rabbits that received two injections of cell-free culture f i l t ra tes of Salmonella typhosa spaced twenty-four hours apart (105). The f i r s t injection was given intradermally while the second injec-tion was given intravenously, twenty-four hours later . Within two to six hours after the intravenous injection, Shwartzman noted this necrotic lesion in the site of the skin injection. The dermal Shwartzman reaction is auite - 21 -non-specific in that the "provoking" or intravenous injection can be endotoxin from a genus of bacterium that is completely different from that used in the dermal injection (1). Another related phenomenon is the c la s s i -cal "generalized Shwartzman reaction", sometimes known as the "Sanarelli-Shwartzman" reaction, which is characterized by bilateral renal cortical necrosis and glomerular thrombosis in the rabbit (89,106). To precipitate these lesions, two intravenous injections of endotoxin spaced twenty-four hours apart must be given to the rabbit. These lesions are largely the result of the extensive intravascular clotting that occurs after the second intravenous injection of endotoxin (107). There is some consen-sus that there are certain c l in ica l conditions that mimic this toxic response to endotoxin seen in the generalized Shwartzman reaction. These human correlates include renal cortical necrosis that often is associated with a variety of infectious and non-infectious complications of pregnancy (108), thrombotic thrombocytopenic purpura (109), sepsis (110) and post-renal homotransplantation (111). Although these pathologies that are seen c l i n i c a l l y resemble the Shwartzman phenomenon seen in endotoxin-treated animals, the precipitating cause in humans has not been definitely estab-lished to be due to endotoxins. Nonetheless, the Shwartzman-type reactions exemplify the potent effects that endotoxins can have on the mammalian cardiovascular system and since shock is a common manifestation in human gram-negative septicaemias, this has aroused an enormous amount of research into elucidating the interactions of endotoxins with the various mammalian systems, especially the cardiovascular system. Within the cardiovascular system, endotoxins can interact with both the humoral and cel lular components of the blood. The major humoral components - 22 -that interact with endotoxins include the complement and the coagulation systems. The complement system can be activated via two pathways; the classical pathway which requires the formation of an antigen-antibody complex to in i t ia te activation of C l , the f i r s t component of complement and the alternate pathway which involves the interaction of an antigen with a serum protein called properdin which in turn activates the terminal compon-ents of the complement system (C3, C5-C9) without the requirement for speci-f ic antibody (112). Endotoxins have been shown to be capable of activating the complement system via both the classical and the alternate pathways (3,89,112,113). When the complement system is activated, potent mediators of inflammation are generated such as the C3a and C5a fragments which possess anaphylatoxin act ivity and enhance vascular permeability as well as smooth muscle contraction through histamine release (114,115). Other media-tors that are released include leukotactic peptides, l y t i c - and phagocytosis-promoting factors (113,116,117). The relevance of these mediators, formed as a result of complement activation by endotoxin, to gram-negative sepsis in humans has become increasingly apparent since McCabe has demonstrated a significant decrease in levels of C3 in patients with gram-negative sepsis (118). Also, a study reported by Fearon and associates on patients with gram-negative sepsis has confirmed McCabe's earlier findings and in addi-t ion, provided evidence for activation of the alternate or properdin pathway as well , which can also lead to consumption of complement components (119). This latter finding strongly suggests the presence of endotoxins in human gram-negative septicaemia. One of the major complications of gram-negative septicaemia in man is the frequent occurrence of disseminated intravascular coagulation or DIC, as i t is more commonly presented in abbreviated form (120). This syndrome is - 23 -characterized by a haemorrhagic diathesis associated with generalized throm-bosis and culminates in circulatory collapse. The DIC syndrome is a conse-quence of an acute activation of the coagulation system that promotes the deposition of f ibr in thrombi within the vascular beds causing ischaemia and necrosis of v i ta l tissues and organs (124). In conjunction with this c lot t -ing process, the f ibr ino ly t i c system becomes activated as a result of f ibr in deposition and the excessive f ibr inolys is digests f i b r i n , fibrinogen and procoagulants creating a condition of hypocoaguabi1ity (121,125,126). Therefore, i t may be more appropriate to label this syndrome as "consumption coagulopathy" (121). The f i r s t indications that endotoxins affected the coagulation system were, as previously mentioned, when Sanarelli reported in 1924 and when Shwartzman reported in 1928 the pathological effects of two properly spaced injections of culture f i l t ra tes from gram-negative organisms into rabbits. Although these investigators did not know at that time that the reaction which they described was the result of endotoxins interacting with the coagulation system, it was later ' shown to be the case by other investigators (1,3). It is now known that after an infusion, or even after a single intravenous injection of endotoxin, coagulative changes are i n i t i a -ted which result in deposition of f ibr in in a variety of tissues including the lungs, l iver and spleen (3,127,128,129). The mechanism by which endo-toxins activate the clotting system has been studied extensively. The evidence to date indicates that endotoxins can in i t ia te coagulation via both intrinsic and extrinsic pathways (89). The intr ins ic pathway necessitates the activation of factor XII or Hageman factor and a study reported by Rodriguez-Erdmann whereby he noted that endotoxin markedly shortened the whole blood clotting time in s i l iconized glass tubes as well as that of - 24 -platelet-poor plasma implied that endotoxins could enhance clotting via the intrinsic pathway (122). Experiments reported by Nies and Melmon also suggested that Hageman factor could be directly activated by endotoxins (123) but i t wasn't until Morrison and Cochrane reported their findings that definitive proof was provided that endotoxins could directly activate the intr insic clotting system (130). Furthermore, the investigators showed that i t was the l i p i d A region of the endotoxin complex that was required to activate Hageman factor, presumably by providing the site of attachment for Hageman factor to the endotoxin complex (130). Since activated Hageman factor has the ab i l i ty to activate the plasminogen proactivator of the f ib inolyt ic system, prekallikrein of the kinin-forming system and factor XI of the intr insic clotting system, i t provides the means by which endotoxins can ini t ia te reactions in these important humoral systems that affect homeo-stasis (131,132). Activation of the coagulation system by the extrinsic pathway requires the release of tissue thromboplastin or a cell-derived procoagulant. No one has demonstrated a direct effect of endotoxin on any of the known proteins of the extrinsic coagulation pathway but Lerner and associates reported in 1968 that rabbits treated with endotoxin had significantly lower levels of factor VII which suggested that the extrinsic coagulation pathway was act i-vated (133). A more recent study by Garner and Evensen on dogs has corrob-orated Lerner's earl ier findings in rabbits that plasma levels of factor VII decrease after endotoxin treatment (134). In addition, Garner and Evensen were able to establish that factor VII deficient dogs treated with endotoxin had less thrombi and f ibr in deposits in tissues than did normal dogs treated with endotoxin, which again indicated strongly that the extrinsic coagula-- 25 -tion pathway is activated in endotoxaemia (134). The mechanism by which endotoxin activates the extrinsic coagulation pathway seems to involve the blood cel lular components as several investigators have shown that blood leukocytes and platelets perform an essential role in endotoxin-induced coagulative changes in vivo (135,136,137,138). These studies suggested that endotoxins interact with these blood cel lular components in a manner that results in the release of tissue factors and platelet factor 3 which in turn activate the clotting system. In vitro investigations have essentially confirmed these conclusions and have provided more insight into the types of cells involved. Thus, while i t is known that endotoxin can induce platelets to aggregate (139), to release serotonin, histamine and platelet factor 3 in vi tro (136,140), these platelet factors are weak procoagulant substances and mainly function as accelerators of the coagulation process (141). Hence, the cells that are principal ly responsible for releasing potent procoagulant substances that activate the extrinsic coagulation pathway in the presence of endotoxin are blood leukocytes (135,136,141) and, more specif ical ly , two separate research groups have shown quite conclusively that the monocyte was responsible for this procoagulant act ivi ty (142,143). One of these groups, Rivers and co-workers (142), concluded that the experimental evidence suggesting that the granulocyte was the source of tissue factors could be explained by the presence of less than one percent monocyte contamination. In concordance with this conclusion, Hi 11er and associates demonstrated that more endotoxin-induced tissue factor activity could be derived from a suspension of ten monocytes/yl than from a preparation of ten thousand granulocytes/yl (143). Hi 11er and co-workers also suggested that the tissue factor originated from the lysosomal fraction of the monocytes. Another - 26 -cel l type that could play some part in the activation of the extrinsic coagulation system during endotoxaemia is the endothelial c e l l . Several independent investigations have reported the occurrence of extensive endo-thel ia l cel l damage in arteries obtained from animals treated with endotoxin (144,145,146). Therefore, these observations raise the poss ibi l i ty that the extrinsic clotting system could be activated either by tissue thromboplastin leaking from damaged endothelial cel ls or by platelets adhering to the exposed underlying basement membrane and consequently activating the coagu-1 at ion system (136). Therefore, by virtue of the diverse actions endotoxins exert on the various components of blood that result in the liberation of potent pharma-cological agents and in the activation of the clotting system, it becomes more understandable how these gram-negative toxins are capable of inducing circulatory shock in animals (see schematic representation of the pathophy-siology of gram-negative shock in reference 147). 1.5 Cl inical Significance of Endotoxins Since all Enterobacteriaceae and most gram-negative bacteria produce endotoxin (28), any bacteraemia due to gram-negative organisms that is seen c l i n i c a l l y could potentially involve endotoxins as well . Bacteraemia caused by gram-negative organisms other than Salmonella was considered a rare entity until Waisbren reported twenty-nine cases, ten of which were caused by E. col i , at Minneapolis in 1951 (148). Also in 1951, Borden and Hall reported two cases of gram-negative bacteraemia that resulted from transfus-ing contaminated blood (149). These reports heralded the appearance of gram-negative b a c i l l i as the major cause of hospital-acquired infections and in twenty years following 1951, the incidence of gram-negative bacteraemia - 27 -has increased twenty-fold (150). Indications now are that approximately one percent of al l hospital patients either acquire gram-negative bacteraemia or are admitted for that reason and the fa ta l i ty rates for this disease vary from thir ty to f i f ty percent (150,151). This high incidence of gram-nega-tive infections with their associated lethal i ty can give a false impression of the virulence of these organisms. In actual fact, organisms commonly implicated in gram-negative sepsis are found as normal f lora on the skin, upper respiratory tract or gastrointestinal tract and possess only limited invasive capacity for their normal host (150). These organisms include mainly the Enterobacteriacea such as Escherichia c o l i , Klebsiel1 a, Entero- bacter, Proteus, Pseudomonas and Bacteroides. Infections from these organ-isms only result when the host's defense mechanisms are impaired or when these bacteria are introduced directly into a site susceptible to infection such as the urinary, respiratory or vascular system. As a result, these gram-negative b a c i l l i are sometimes known as "opportunistic" pathogens (152). This is in contrast to bacteraemias caused by virulent strains of gram-negative b a c i l l i that are not normally part of the f lora , such as Salmonell a, Brucel1 a, Haemophilus, Pasteurella (Yersinia) and enteropatho-genic E. c o l i . These organisms cause dist inctive c l in i ca l illnesses which are usually not considered as " typical" gram-negative bacteraemias where the c l i n i c a l features do not differ s ignif icantly regardless of the causative organism (150). Some of the reasons that have been given to explain the increased incidence of gram-negative bacteraemias include, interestingly enough, the very triumphs of medicine, that i s , that debilitated and chroni-cal ly i l l patients are kept alive for much longer periods of time than ever before and i t is these patients with altered host defenses that are very - 28 -susceptible to infection. Also, in conjunction with this , many hospital practices used presently allow bacteria to bypass host defense mechanisms and gain access to sites of infection. Venous and bladder catheters are an example. Another reason that is given to explain the high incidence of gram-negative bacteraemias is the popular and extensive use of antibiotics coupled with the propensity of the host's indigenous f lora for antibiotic resistance, the combination of which produces resistant strains of organisms (150,153). Patients that are predisposed to gram-negative bacteraemia are those whose host defense mechanisms have been impaired and the severity of the underlying host disease is the major determinant of outcome in bacteraemia (151,153,154). Thus, granulocytopenic patients form the major group of hospitalized patients that have the greatest risk of becoming infected with gram-negative b a c i l l i (151,155). It has also been noted that patients that have normal or even elevated leukocyte counts can develop serious infections which implies impaired granulocyte function (151). Impaired granulocyte function has been noted in patients with genetically inherited syndromes (156,157), neoplastic diseases (158,159), metabolic defects (160,161), immunologic disorders (162,163), surgical disorders (164) and extensive burns (165). Patients who receive prolonged treatments with corticoster-oids, immunosuppressive drugs, radiotherapy or anti-cancer drugs are also particularly susceptible to gram-negative sepsis (150,166) Other treatments that commonly develop gram-negative infections include surgery and instru-mentation of the genitourinary tract, surgery of the gastrointestinal tract and manipulation of infected wounds, especially after severe burns (166). - 29 -A major concern with gram-negative bacteraemia is that i t frequently progresses to circulatory shock. One estimate claims that 40% of patients with gram-negative sepsis go on to develop shock (167). Characterist ically, these patients develop shaking ch i l l s which are closely followed by fever. In the early stages of hypotension, some patients have peripheral vasodila-tation ("warm" shock) which then proceeds to vasoconstrictive shock, mani-fested by pallor and a cool, clammy skin. Respiratory distress with rapid shallow ventilation is often seen. Vomiting and diarrhea sometimes precede the circulatory manifestations of shock but the development of mental confu-sion which leads to stupor and coma along with reduced to absent urine flow are indications of severe shock (147,150,166,168). The syndrome of DIC occurs in approximately 5-10% of patients w-iith gram-negative bacteraemia (150). Although there are no real ly consistent differences between the shock syndromes associated with gram-positive or gram-negative infections (169), gram-positive bacteraemias tend to follow a slower stepwise progres-sion of symptoms (except meningococcaemia) whereas in gram-negative bacter-aemia, the transition from relat ively good health to prostration can occur within a few minutes to a few hours (167). Treatment of gram-negative bacteraemia essentially involves the prompt, aggressive use of antimicrobial agents since more than one-half of patients with some types of untreated bacteraemia die within 72 hours (170). Other treatments given are basically designed to support and maintain the respira-tory, circulatory and excretory systems through the use of mechanical devices, f luids and various pharmacological agents (150,166). However, the effectiveness of these treatments in terms of reducing the morbidity in gram-negative septicaemia is wanting. More recent developments in the - 30 -treatment of gram-negative bacteraemias have come from immunological stud-ies. Work in this area has indicated that "0" antibodies are protective against death in gram-negative infections (171). In this study, patients with low "0" antibody t iters against their infecting strain of bacteria had much higher mortality rates than patients with high "0" antibody t i ters (171). These "0" antibodies are IgG and IgM antibodies and they act as heat stable opsonins. However, the large number of different O-antigens in each bacterial species (E. coli has 150 serotypes) makes this method of vaccina-tion unwieldy except, perhaps, for Pseudomonas of which there are only a few serotypes (153). Also, vaccinations are ineffective in patients that are granulocytopenic or whose immune response is inhibited by drug treatment, such as anti-cancer therapy (172). A vaccine that may be more feasible is one which induces the formation of antibodies against the core polysacchar-ide and l i p i d A portions of endotoxin. As mentioned previously, the struc-ture of these two regions of endotoxin show l i t t l e var iab i l i ty in all the species in Enterobacteriaceae and therefore cross-protective immunity is possible. Studies in patients have in fact shown that protection against the sequelae of gram-negative shock was greatest in patients who had high antibody t i ters against the core glycol ipid of endotoxin (171,173). The protective mechanism of the core glycolipid antiserum may involve an antitoxin-like effect rather than opsonization (151). If this proves to be /the case, then this vaccine would also be effective in patients that have an impaired immune system. The results obtained with the immunoprophylaxis experiments strongly indicate the great portent endotoxins wield in the production of the seque-lae of gram-negative bacteraemias and also suggest that the biologically - 31 -active portion of endotoxin involves the l ip id A region of the complex. Analogous experiments with the cationic antibiotic polymyxin B have shown that this drug, which binds stoichiometrically to the l ip id A region of endotoxin (174), can also inhibit the biological actions of endotoxins (175). These results suggest that the binding of endotoxin to target cel ls may be a necessary prelude to the expression of its biological toxic i ty and agents which can form complexes with endotoxin, such as polymyxin B or core glycol ipid antiserum, prevent the toxic expressions of endotoxin by hinder-ing its ab i l i ty to interact with various ce l l s . The experiments described in this thesis have been designed for the primary purpose of testing this relationship between the binding of endotoxin to target cel ls and the mani-festation of its lethal effects. To this end, in vitro investigations were conducted in which the red blood cel l was used as a model target cel l with the aim of studying the nature of endotoxin-cel1ular interactions and for examining how certain membrane components can influence the binding of endo-toxins to plasma membranes. In this regard, various chemical and enzymatic treatments were employed to modify either the structure of the endotoxin or of certain constituents of the erythrocyte plasma membrane and the effect of these modifications on toxin-red blood cel l interactions was studied. Also, 51 in vitro binding experiments u t i l i z ing Cr radiolabeled E. co l i endotox-in were designed to test the capability of various pharmacological agents to antagonize the binding of endotoxin to human red blood ce l l s . Drugs that were found to be effective endotoxin antagonists in vitro were administered to endotoxin-treated animals and assessed for their ab i l i ty to mitigate the toxic effects of endotoxin in vivo. If a positive relationship between antagonism of endotoxin binding in vitro and a decrease in its biological - 32 -toxic i ty in vivo was found to exist, then the in vitro binding tests could possibly serve as a convenient screening test for drugs that would have the capability of preventing the in i t ia t ion of some of the deleterious biologi-cal actions of endotoxins. The use of these drugs in conjunction with anti-biotics would then provide a more effective mode of therapy for gram-negative bacteraemia than is currently employed. - 33 -CHAPTER 2 Materials and Methods 2.1 Membrane Preparations 2.1.1 Erythrocyte ghosts Erythrocyte membranes were prepared from outdated 0 + blood donated by the Red Cross blood bank. Basically, the procedure involved washing the red blood cel l s in a series of progressively hypotonic NaCl solutions according to the method of Godin and Schrier (176). More speci f ica l ly , the outdated 0 + blood was centrifuged at 800 x g for 5 minutes and both the plasma and buffy coat were removed. The red cel l s were then washed twice in 5 volumes of cold isotonic NaCl, each time centrifuging at 800 x g for 5 min and aspirating all remaining traces of the buffy coat. One hundred ml packed red cel ls were then resuspended in 10 volumes 0.08 M NaCl, stirred for 10 min at 4°C and then centrifuged at 9,000 x g for 5 min. The supernatant was discarded and the red cells were then resuspended in 10 volumes of 0.06 M NaCl and the procedure was repeated as for the 0.08 M step. Subsequent steps included washes, as above, in 0.04 M, 0.02 M, and 0.009 M NaCl with pH adjustment to 7.4 with Tris at the 0.009 M stage. The membranes were then resuspended in 3 volumes of 10 mM Tris-HCl pH 7.4 buffer, st irred for 10 min at 4°C and centrifuged at 20,000 xg for 10 min. The supernatant was discar-ded and the pellets were resusupended in sufficient double d i s t i l l ed H2O to yie ld a protein concentration of approximately 3-4 mg/ml. This suspen-sion of erythrocyte membranes was then quick frozen in acetone/dry ice and stored at -20 °C for later use. - 34 -2.1.2 Heart Sarcolemmal Membranes 1.0 gm of guinea pig ventricular tissue was minced with scissors and homogenized in 10 ml cold buffer which was composed of 1.25 M KC1-2 mM DTT-0.5 mM CaCl^-lO mM Tris pH 7.4. The tissue was homogenized with a Polytron PT-10 homogenizer (Brinkman Instruments) for 5 sec at 1/4 maximum 2 speed. The homogenate was f i l tered through a 0.5 mm nylon mesh and centrifuged at 1,200 x g for 10 min. The supernatant was discarded and the pellet was resuspended in 10 ml homogenization buffer and homogenized in a 40 ml glass homogenization tube with a tight f i t t ing teflon pestle using 5 up and down strokes with a Potter-Elvehjem homogenizer. The homogenate was then centrifuged at 500 x g for 10 min. The supernatant was discarded and the pellet was resuspended in 6 .0 ml 10% w/v sucrose buffer containing 2 mM DTT-10 mM Tris pH 8.2. The suspension was homogenized to homogeneity with a glass hand homogenizer. 1.0 ml of the homogenate was layered on top of each of six gradient tubes which contained a discontinuous sucrose gradient that started on the bottom with 2.5 ml 60% sucrose and on which was layered 2.5 ml portions of 55%, 52.5% and 50% w/v sucrose dissolved in 2 mM DTT-10 mM Tris pH 8.2. The gradient was centrifuged at 40,000 x g for 1 hr in a Beckman ultracentrifuge. The sarcolemmal fraction eaui1ibrated at the 55-60 % sucrose interface while the mitochondrial and sarcoplasmic reticular fraction equilibrated on top of the 50 % sucrose fract ion. The sarcolemmal fraction was removed with a Pasteur pipette, washed in 5 volumes of 10 mM Tris pH 7.4, and centrifuged at 30,000 x g for 15 min. The pellet was resuspended in double d i s t i l l ed H£0 to yield a protein concentration of approximately 3-4 mg/ml. The membranes were quick frozen in acetonp/dry ice and stored at -20°C for future use. - 35 -2.1.3 Liver Membranes Liver membranes were prepared according to a procedure described by T.K. Ray (177). Accordingly, 3.0 gm guinea pig l iver were homogenized in 15 ml 0.5 mM CaCl 2-2 mM DTT-10 mM Tris pH 7.4-7.5 buffer for 30 sec with a Polytron PT-10 homogenizer at 1/4 maximum speed. The homogenate was diluted to 2% (1:10 dilution) with homogenization buffer and f i l tered through a 0.5 2 mm nylon mesh. The f i l tered homogenate was then centrifuged at 1,200 x g for 10 min and the supernatant was discarded. The pellet was resuspended in 20 ml homogenization buffer and homogenized with a Potter-Elvehjem homogen-izer using 5 up and down strokes with a tight f i t t i n g teflon pestle. The homogenate was diluted to 1/2 the previous dilution volume ( i . e . 4% homogen-ate) with homogenization buffer and centrifuged at 500 x g for 10 min. The pellet was resuspended in approximately 30 ml of homogenization buffer and again homogenized 5 strokes with a Potter-Elvehjem homogenizer. This homo-genate was again centrifuged at 500 x g for 10 min and the pellet was resus-pended in 4.0 ml of homogenization buffer. To this was added 14 ml 62% (w/w) sucrose (made up in r^O) which resulted in a final sucrose concen-tration of 48%. 3.0 ml of this sample were layered on the bottom of each of 6 gradient tubes and successive 3.0 ml layers of 45%, 41% and 2.5 ml 37% (w/w) sucrose were added on top. The gradient was centrifuged at 90,000 x g for 2 hrs in a Beckman ultracentrifuge. Liver membranes appeared as a band at the 37-41% sucrose interface. The membranes were collected, washed once in 5 volumes 10 mM Tris pH 7.4, resuspended in double d i s t i l l ed b^ O (3-4 mg/ml protein) and quick frozen in acetone/dry ice. - 36 -2.1.4 Lung Membranes The method used for the preparation of lung membranes was basically the same procedure as the one that was used for the preparation of l iver membranes except the sucrose concentrations in the gradient were changed. Brief ly, a 20% homogenate was prepared by homogenizing 3.0 gm guinea pig lung in 15 ml 0.5 CaCl 2-2 mM DTT-10 mM Tris pH 7.4-7.5 buffer with a Polytron homogenizer for 30 sec at 1/4 maximum speed. The homogenate was diluted to 2%, f i l t e r e d , centrifuged and rehomogenized as described in section 2.1.3 for l i v e r . After the third centrifugation step (500 x g for 10 min), the pellet was made up to 18 ml by the addition of 55% (w/w) sucrose. 3.0 ml of this sample were layered on the bottom of each of 6 gradient tubes. The remainder of the gradient that was layered on top of the 55% sample layer included 3.0 ml of 52%, 48% and 2.5 ml 45% (w/w) sucrose. This gradient was centrifuged at 90,000 x g for 2 hrs in a Beckman ultracentrifuge. Lung membranes, which equilibrated at the 48-52% sucrose interface, were collected, washed twice in 10 mM Tris pH 7.4, resuspended in double d i s t i l l ed h^O and quick frozen in acetone/dry ice. 2.1.5 Preparation of Liver Lysosomes A 10% homogenate of l iver tissue was prepared by homogenizing l iver in 0.25 M sucrose with a Potter-Elvehjem homogenizer using 4 strokes with a tight f i t t ing teflon pestle. The homogenate was centrifuged at 600 x g for 5 min and the pellet was discarded. The supernatant was centrifuged at 3,500 x g for 15 min and the pellet was resuspended in a volume of 0.25 M sucrose that equalled the original homogenate volume. This supension was used as a lysosomal-rich preparation for experiments. - 37 -2.2 Enzyme Assays 2.2.1 Membrane Bound Enzymes 188.8.131.52 Acetylcholinesterase Acetylcholinesterase act ivity in erythrocyte ghost membrane preparations was determined at room temperature by monitoring the hydrolysis of 0.03 M acetylthiocholine in the presence of 0.01 M 5,5'-dithiobis-(2-nitrobenzoic acid) (DTNB) and 0.1 M Tris-HCl buffer pH 8.0. The reaction was carried out in a spectrophotometric cuvette by adding 2.75 ml 0.1 M Tris buffer pH 8.0, 0.1 ml 0.01 M DTNB (made up in Tris buffer), 0.1 ml human erythrocyte ghost membranes (approximately 0.3-0.4 mg protein/ml) and 0.05 ml 0.03 M acetyl-thiocholine. The absorbance at 412 nanometers was recorded every minute for five minutes following the addition of the acetylthiocholine. When acetyl-cholinesterase act ivi ty was assayed in rat erythrocyte ghosts, a more concentrated membrane suspension was used ( i . e . 3.0-4.0 mg protein/ml). 184.108.40.206 Nitrophenylphosphatase (NPPase) Nitrophenylphosphatase act ivity in various membrane preparations was determined by measuring the hydrolysis of p-nitrophenylphosphate in the presence of Mg + + and K + at 37°C and pH 7.4. The reaction mixture consisted of 1.0 ml 0.15 M imidazole buffer pH 7.4, 0.1 ml 0.09 M p-nitro-phenylphosphate, 0.1 ml 0.09 M MgCl 2 , 0.1 ml 0.9 M KCI, 0.2 ml membrane suspension and double d i s t i l l ed h^ O in a final reaction volume of 3.0 ml. The protein concentration of the membrane suspension and the reaction times varied according to the type of membranes being assayed. For example, when the enzyme act ivi ty was assayed in human erythrocyte ghosts, suspensions containing 3.0 to 4.0 mg protein/ml were used and the duration of the reaction was 1 hr whereas suspensions of heart membranes containing approxi-- 38 -mately 1.0 mg protein/ml were reacted for 15 minutes. The reaction was stopped by the addition of 1.0 ml cold 20% TCA and the precipitated protein was centrifuged down at 10,000 x g. A 3.0 ml aliquot of the supernatant was taken and combined with 1.0 ml 1.5 M Tris solution. The absorbance of the resulting solution was measured at 412 nanometers. The K+-stimulated component of the enzyme was determined by subtracting the act ivi ty of the enzyme assayed in the presence of Mg + + (basal activity) from the act ivity ++ + , of the enzyme assayed in the presence of both Mg and K (total activ-i t y ) . 220.127.116.11 Adenosine Triphosphatase: (Na+,K+)-ATPase + + (Na ,K )-ATPase act ivi ty was assayed in various membrane prepara-tions by incubating at 37°C a reaction mixture that consisted of 1.0 ml 165 mM Tris-HCl buffer pH 7.4, 0.1 ml 0.09 M MgCl 2 , 0.1 ml 3 mM EGTA, 0.3 ml 30 mM ATP, 0.1 ml 0.6 M KCI, 0.1 ml 2.4 M NaCl, 0.2 ml membrane suspension and double d i s t i l l e d H^ O to make the volume 3.0 ml. Protein concentra-tions of the membrane suspensions varied from 3.0-4.0 mg/ml for erythrocyte ghosts to 0.7-0.8 mg/ml for preparations of heart membranes. Reaction times also varied from 15 minutes for heart membranes to 60 minutes for erythro-cyte ghosts. The reaction was stopped by the addition of 1.0 ml cold 10% TCA and the mixture was centrifuged at 10,000 x g for 5 min. A 3.0 ml aliquot of the supernatant was taken and to it was added 1.8 ml molybdate solution which was composed of 1.4 ml double d i s t i l l ed H^ O and 0.4 ml 5% ammonium molybdate. Color development for inorganic phosphorus was i n i t i a -ted by adding 0.2 ml Fiske-Subbarow reagent to the mixture and the absor-bance at 660 nanometers was determined spectrophotometrically at 15 minutes. These absorbance readings were compared to the readings observed in prepared - 39 -standards that contained known quantities of inorganic phosphorus. The Na +,K +-stimulated portion of the enzyme was determined by subtracting the act ivity observed in the presence of Mg + + alone from the act ivity seen ++ + + when Mg , Na and K were all present in the reaction mixture. The Fiske-Subbarow reagent was prepared fresh weekly by the addition of 0.25 gm l-amino-2-naphthol-4-sulfonic acid (ANS) with s t i rr ing to 100 ml freshly prepared 15% sodium bisul f i te (anhydrous) followed by the addition of 0.5 gm anhydrous sodium sul f i te . The solution was stored in a dark bottle. 18.104.22.168 Cytochrome c Oxidase Cytochrome c oxidase act ivi ty was assayed in membrane fractions in accordance with the method of Rabinowitz et a l . (191) which is a modifica-tion of the procedure outlined by Cooperstein and Lazarow (192). Basically, a stock solution of 0.03 M phosphate buffer pH 7.4 was prepared by mixing 805 ml 0.03 M Na2HP04 with 195 ml 0.03 M KH 2P0 4 . A solution of reduced cytochrome c was freshly prepared by adding 100 yl of freshly prepared 1.2 M sodium hydrosulfite to a 17 yM solution of cytochrome c (Sigma Type III) dissolved in 0.03 M phosphate buffer. The cytochrome c solution was shaken vigorously for several minutes to remove excess hydro-sul f i te . 3.0 ml of this reduced cytochrome c solution were then transferred to a spectrophotometric cuvette and to this an aliquot of no more than 50 yl of sample was mixed. The absorbance at 550 nanometers was taken at 30 second intervals for 3 minutes. The cytochrome c in the cuvette was then ful ly oxidized by the addition of a few crystals of potassium ferricyanide. The absorbance of this fu l ly oxidized solution of cytochrome c was subtrac-ted from each of the absorbances obtained at the 30 second time intervals - 40 -after the sample was added. The logarithm of the difference of these absor-bance values when plotted against time gave a straight line with a negative slope from which could be determined the amount of cytochrome c (in pinoles) that was oxidized per minute per mg sample protein. 22.214.171.124 5'-Nucleotidase 5'-nucleotidase act ivi ty in membrane preparations was assayed by mixing together in a volume of 0.5 ml a reaction mixture which consisted of 50 mM Tris-HCl buffer pH 9.0, 2.5 mM MgCl 2 , 100 mM KCI and 8.0 mM 5'-AMP (final concentrations). The mixture was preincubated at 37°C for 5 minutes and the reaction was then started with 100 yl membrane suspension (making the total reaction volume 0.5 ml). After an appropriate reaction interval (usually 10 or 15 minutes) while the reaction rate was s t i l l l inear, the reaction was stopped by the addition of 0.5 ml cold 12% TCA. The protein precipitate was centrifuged down in an Eppendorf bench-top centrifuge. A 0.6 ml aliquot of the supernatant was taken and assayed for inorganic phosphorus by the addition of 0.28 ml double d i s t i l l ed H^O, 0.08 ml 5% ammonium molybdate and 0.04 ml amino-naphthol-sulfonic acid (ANS) reagent (see Methods 126.96.36.199). The color was developed at room temperature and the absorbance was read at 660 nanometers at precisely 15 minutes after the ANS was added. Standards which contained known quantities of inorganic phosphorus were also assayed and with reference to these absorbances the results of the samples could be expressed as ymoles Pi liberated/mg protein/hour. 2.2.2 Lysosomal Enzymes 188.8.131.52 Acid Phosphatase Acid phosphatase act ivity was determined as described in Sigma technical bulletin No. 104. Basically, the reaction mixture consisted of 0.5 ml 0.09 M - 41 -ci trate buffer pH 4.8, 0.5 ml p-nitrophenylphosphate solution (4 mg/ml) and 0.2 ml of sample. Samples were appropriately diluted so that absorbance values did not exceed 0.900. For example, samples of rat and guinea pig plasma were diluted 1:3 and 1:5 respectively before they were assayed for acid phosphatase act iv i ty . The reaction was carried out at 37°C for 30 minutes and then stopped with 5.0 ml 0.1 N NaOH which also developed the color . The absorbance of the solution was read at 410 nanometers and divided by a factor of 0.3 to convert the results into Sigma units of acid phosphatase act ivity (determined from a standard curve as described in technical bulletin No. 104). 184.108.40.206 N-Acetyl-e-Glucosaminidase N-Acetyl-e-glucosaminidase act ivi ty was determined by measuring the extent of hydrolysis of p-nitrophenyl-N-acetyl-B-D-glucosaminide substrate to p-nitrophenol under acidic conditions. The reaction mixture consisted of 0.5 ml 0.3 M citrate buffer pH 4.3, 0.5 ml p-nitrophenyl-N-acetyl-B-D-gluco-saminide (3 mg/ml), 0.3 ml double d i s t i l l ed H20 and 0.2 ml sample. Samples were appropriately diluted in order that absorbance values of the reaction did not exceed 0.300, the point after which the relationship between absorbance and enzyme quantity became non-linear. Plasma samples were usually diluted 1:3. The reaction was carried out at 37°C for 30 minutes and then terminated with 1.5 ml cold 3% TCA. The solution was then centrifuged at 10,000 x g for 5 minutes and a 2.0 ml aliquot of the superna-tant was taken. To this al iqot, 1.0 ml bicarbonate buffer consisting of 0.5 M NaHCO^-O^ M Na^CO^ was added to develop the color. The absorbance was read at 420 nanometers. The results were expressed as absorbance/ml plasma or as absorbance/mg protein for other types of samples such as lyso-somal suspensions. - 42 -220.127.116.11 Cathepsin D. Assays for cathepsin D act iv i ty were performed as described by Barrett and Heath (177) with some modifications. The total volume of the reaction mixture was 0.4 ml and consisted of 0.1 ml 1 M formate buffer pH 3.0, 0.1 ml 8% (w/v) haemoglobin substrate (prepared as described below) and 0.2 ml sample (plasma samples were diluted 1:2). The reaction was started with the haemoglobin substrate and incubated at 45°C for 60 minutes, after which it was stopped by the addition of 2.5 ml cold 3% TCA. The reaction mixture was allowed to stand on ice for 30 minutes after the addition of the TCA to complete the precipitation of the proteins and then the tubes were centr i-fuged in a table-top centrifuge at 4 ° C . An aliquot of the supernatant (0.2 to 0.4 ml) was then assayed for small molecular weight peptides by using the standard Lowry protein assay. Blanks for the cathepsin D assay contained double d i s t i l l ed H20 instead of 1 M formate buffer. Results were expres-sed as mg protein liberated/hr/ml plasma. The formate buffer was prepared by mixing equal volumes of 1 M sodium formate with 1 M formic acid and t i t rat ing the resulting solution with concentrated formic acid to pH 3.0. Haemoglobin substrate was prepared from one unit of outdated human blood obtained from the Red Cross blood bank. The blood was centrifuged at 3,500 x g for 10 minutes and both the plasma and buffy coat were discarded. The erythrocytes were resuspended in 500 ml 0.15 M NaCl and recentrifuged. The wash was repeated twice more and each time the supernatants as well as remnants of the buffy coat were discarded. The packed cel ls were then resuspended in a equal volume of double d i s t i l l ed H20 and 0.5 volumes of CC1/,. The suspension was stirred vigorously at 4°C for 15 minutes and - 43 -then centrifuged at 10,000 xg. The clear red supernatant containing haemo-globin was dialyzed against double d i s t i l l ed H^ O at. 4°C and then adjusted to 8.0 (w/v) on the basis of a dry weight determination. Ten ml aliquots were quick frozen using acetone/dry ice and stored at -20"C until reauired. 2.3 Haemolysis Experiments Red blood cell haemolysis experiments were conducted essentially as described by Machleidt et a l . (178) with some modifications. Freshly drawn heparinized blood was used for al l experiments. The blood was centrifuged and both the plasma and buffy coat were discarded. The red cells were resuspended in 4 volumes cold 0.9% NaC1-15 mM Tris pH 7.0 buffer, recentri-fuged and the supernatant, along with remaining portions of buffy coat, were discarded. This wash was repeated and the red cel l s were then resuspended in sufficient 0.9% NaCl-15 mM Tris pH 7.0 buffer to give a haematocrit of approximately 6%. For example, 2.0 ml of packed red blood cel ls made up to 32 ml with the NaCl-Tris buffer gave a haematocrit of approximately 6%. This suspension of red cells was kept on ice and thoroughly mixed when aliquots were removed for haemolysis tests. The haemolysis test was set up in the following manner: 0.2 ml of the red cel l suspension was added to 1.3 ml of 0.9% NaCl-15 ITM Tris pH 7.0 buffer. The mixture was allowed to equilibrate for 5 minutes at the desired temperature of the test, usually at room temperature. The reaction was started by the addition of 0.2 ml 0.9% NaC1-15 mM Tris pH 7.0 buffer which contained the agent to be tested. Control samples contained no drug in the buffer. The suspension was incubated for 15 minutes at the desired tempera-ture and then was subjected to a hypotonic challenge by the addition of 2.3 ml 15 mM Tris pH 7.0. The mixture was further incubated for 10 minutes and - 44 -then centrifuged at 15,000 x g for 1 minute. The supernatant was decanted and the absorbance at 540 nanometers was measured. Control haemolysis (without added drug) was approximately 40% of complete haemolysis, i . e . , 0.2 ml red cel l suspension in 3.8 ml double d i s t i l l ed H^O. Test samples (with drug) were usually expressed as percent of control haemolysis where-upon values over 100% indicated lyt ic act ivity while values less than 100% demonstrated anti-haemolytic ab i l i ty or red blood ce l l s tabi l izat ion. 2.4 Enzyme Treatments of Intact Red Blood Cells 2.4.1 Trypsinization The method for trypsinization of intact red blood cel l s was adapted from Seaman and Uhlenbruck (179). Blood was freshly drawn into a heparinized syringe and divided into two portions. Each portion was centrifuged and the plasma and buffy coat were discarded. The red cel l s were washed twice in 0.9% NaCl-15 mM Tris pH 7.0 buffer as described in section 2.3 of Methods. A 0.1%; (w/v) trypsin enzyme solution was made up in 0.9% NaC 1-15 mM Tris pH 7.0 buffer and 3.0 ml of this enzyme solution were mixed with 1.0 ml packed red blood ce l l s . The other portion of packed red blood cel ls served as a control and was mixed in 3.0 ml buffer containing no enzyme. Both control and treated red cel l suspensions were incubated at 37°C for 30 minutes and then were centrifuged. The supernatants were removed and the red cel l s were again washed twice in 0.9% NaCl-15 mM Tris buffer. After the f inal wash, the red blood cel ls were resuspended in the NaCl-Tris buffer to a haemato-c r i t of approximately 6%. This suspension was then used for haemolysis experiments as described in section 2.3. - 45 -2.4.2 Neuraminidase Treatment The procedure for treating intact erythrocytes with neuraminidase has been adapted from several sources (179,180,181). Freshly drawn heparinized blood was divided into two portions, a control and enzyme treated sample. The blood was centrifuged and the plasma and buffy coat were discarded. The red blood cel ls were then washed twice as described in section 2.3 except that isotonic saline was used instead of 0.9% NaCl-15.mM Tris buffer. After the wash, the red cells were resuspended in 0.145 M NaCl-5.0 mM CaCl 2-0.3 mM NaHCOg pH 7.2 buffer in a ratio of 3.0 ml buffer for each ml of packed erythrocytes. The treated red cel l suspension received 10 units of neurami-nidase enzyme solution (500 units/ml, Behringwerke) for each ml of red blood cell suspension. The control red cell suspension received an equal volume of H 20. The suspensions were incubated at 37°C for 30 minutes and then centrifuged. The supernatants were removed and the red cel ls were washed twice with cold isotonic saline. For haemolysis studies, the red cel l s were washed once more in 0.9% NaCl-15 mM Tris pH 7.0 buffer and then resuspended in this buffer to a haematocrit of approximately 6%. The neuraminidase treatment removed about 70% of the s i a l i c acid residues from the intact erythrocytes. 2.4.3 Phospholipase A 2 Treatment The method for treating intact erythrocytes with phospholipase A 2 was based on the procedure reported by Roelofsen and associates (182). These investigators reported the effectiveness of using phopholipase A 2 prepara-tions from Naja naja venom to hydrolyze phospholipids in intact erythrocytes (68% of the red cel l lecithin was degraded), whereas enzyme preparations from porcine pancreas were ineffective against intact red c e l l s . Therefore, - 46 -phospholipase Ar, preparations from Naj a naj a venom (Sigma) were used in our studies. The enzyme preparation (1,000 units/vial) was dissolved in 1 ml double d i s t i l l ed H O^ to give an act ivity of 1,000 units/ml. This enzyme solution was then heated at 70°C for 10 minutes to destroy proteoly-t i c act ivity and then stored at -20 °C until further required. Erythrocytes were obtained by centrifuging freshly drawn heparinized blood and the red cel l s were then washed twice with cold isotonic saline to remove all traces of the buffy coat. After the second wash, the erythro-cytes were resuspended in 0.9% NaCl-10 mM CaC 12~5 mM Tris pH 8.0 buffer in the ratio of 3.0 ml buffer/ml packed red blood ce l l s . Erythrocyte suspen-sions that were to be reacted with enzyme received 10 units of phospholipase A 2 solution/ml of red cel l suspension. Control and treated suspensions were then incubated at 37°C for 1 hour. The samples were then centrifuged and the supernatants were discarded. The erythrocytes were washed twice in cold isotonic saline and once with 0.9% NaCl-15 mM Tris pH 7.0 buffer. For haemolysis studies, the red cells were then resuspended in 0.9% NaCl-15 mM Tris pH 7.0 buffer to a haematocrit of 6%. 2.4.4 Removal of Cholesterol from Intact Erythrocytes The procedure that was used to extract cholesterol from intact erythro-cytes was based on the method developed by Gottlieb (183). This method requires approximately 50 ml of freshly drawn heparinized blood. The blood was centrifuged and both plasma and red blood cel ls were saved. The erythrocytes were washed twice in cold isotonic saline and then once more with a buffer composed of 0.9% NaCl-4 mM KH2P04-16 mM Na2HP04-0.4% glucose-0.1% adenosine pH 7.4. The red blood cel l s were f ina l ly resuspended in an equal volume of the buffer solution (usually 25 ml) and then stored in this manner at 4°C until the plasma was prepared. - 47 -The plasma (25 ml) was incubated at 37°C for 60-72 hours in order to allow esterification of the lipoprotein cholesterol by the enzyme lec i th in-cholesterol acyltransferase (LCAT) which is present in the plasma. Gentamy-cin sulfate was added to the plasma in a concentration of 200 yg/ml to prevent bacterial contamination. After the incubation period, the large molecular weight plasma proteins and some of the low density lipoproteins were precipitated by the addition of an equal volume of 60% (NH^^SO^. The mixture (50 ml) was stirred at room temperature for 30 minutes and then centrifuged at 10,000 x g for 5 minutes. The supernatant was then dialyzed against 2 l i ters of 0.9% NaCl-20 mM Tris buffer pH 7.4-7.5 for 24 hours at 4 ° C . The dialysate was changed after 12 hours. After dia lys i s , the plasma was diluted 1:6 with the sodium chloride-phosphate-glucose-adenosine buffer described above. This solution (300 ml), which contained the high density lipoproteins required for cholesterol exchange, was incubated with the stored red blood cells after they were centrifuged, in the ratio of 100 ml plasma solution for each 1.5 ml packed red blood ce l l s . The suspension was incubated at 37°C for a total of 18 hours. The plasma solution was changed after 6 and 12 hours by centrifuging the mixture and then resuspending the erythrocytes in another 100 ml aliquot of the plasma solution which was kept stored at 4 ° C . As a control, 1.5 ml packed red blood cel l s were resuspended in 100 ml 0.9% NaCl-4 mM KH2P04-16 mM Na2HP04-0.4% glucose-0.1% adenosine pH 7.4 buffer and then incubated at 37°C for a total of 18 hours. The buffer was changed after 6 and 12 hours of incubation. Following these incubations, the solutions were centrifuged and the erythrocytes were washed once in 0.9% NaCl-15 mM Tris pH 7.0 buffer. The cel l s were then resuspended in the same buffer to a haematocrit of approximately 6% for haemolysis experiments. - 48 -2.5 Compositional Assays 2.5.1 Protein Protein analysis of various preparations was done according to the method described by Lowry et a l . (184). More exp l i c i t ly , the assay was set up in the manner described below: Stock solution A was prepared by dissolving 20 gm Na^O^, 0.2 gm Na + ,K + -tartrate and 4.0 gm NaOH in double d i s t i l l ed H?0 to a volume of one l i t e r . Stock solution A was stored at room temperature. Stock solution B consisted of a 1% (w/v) solution of cupric sulfate (CuS04.5H20) in double d i s t i l l ed H 20. Solution C was prepared freshly by mixing 1 part stock solution B with 99 parts stock solution A. The sample to be assayed was made up to 0.5 ml with double d i s t i l l ed H O^ and to this was added 5.0 ml solution C. The mixture was incubated at room temperature for 10 minutes and then 0.5 ml of Folin-Ciocalteu's phenol reagent, freshly diluted 1:1 with double d i s t i l l ed H 20, was added and the solution was thoroughly mixed. The color was allowed to develop for 30 minutes at room temperature and then the absorbance was read at 750 nanome-ters. The amount of protein in each sample was determined from a standard curve that was prepared by assaying known concentrations of bovine serum albumin. The reaction was linear up to 100 yg protein. Blanks contained double d i s t i l l e d H20 instead of protein. 2.5.2 S ia l ic Acid (N-Acetylneuraminic Acid) Assays for s i a l i c acid were based on the method described by Warren (185). The following solutions were prepared for the assay: - 49 -Stock solution A, consisting of 0.2 M NaI04 (sodium m-periodate) in 9 M phosphoric acid, was prepared by dissolving 4.28 gm of sodium m-periodate in 60 ml concentrated phosphoric acid (15 M) and then diluting the solution to 100 ml with double d i s t i l l ed H 20. Stock solution B consisted of a 10% (w/v) sodium m-arsenite-0.5 M Na2S04-0.1 N H 2S0 4 solution which was prepared by dissolving 5.0 gm sodium m-arsenite and 3.55 gm Na2S04 in 0.1 N H 2S0 4 to a final volume of 50 ml. Solution C was prepared fresh by dissolving 2-thiobarbituric acid in a 0.5 M Na2S04 solution to a concentration of 0.6% (w/v). The solution was heated to enhance solubi l i ty . The sample to be assayed was made up to a total volume of 0.2 ml with double d i s t i l l ed H 20. For example, when membrane suspensions were assayed, a 50 ul aliquot of the suspension (3.0-4.0 ma protein/ml) was added to 150 yl double d i s t i l l ed H 20. Then 20 yl of 1.1 N H 2S0 4 was added to acidify the sample. The sample was then hydrolyzed at 80°C for 30 min-utes. After cooling, 100 yl of stock solution A was added and, following vortexing, the mixture was incubated for 20 minutes at room temperature in the dark. Then 1.0 ml stock solution B was added and the solution was mixed until i t became colorless . To this was added 3.0 ml of solution C and, after vortexing, the solution was placed in a boil ing water bath for 15 minutes. After the samples were cooled, the colored complex was extracted by adding 4.0 ml cyclohexanone, vortexing and centrifuging the samples in a table-top centrifuge to separate the organic and aqueous phases. The top (cyclohexanone) layer, which contained the colored complex, was removed and the absorbance at 549 nanometers was measured. This absorbance reading was - 50 -compared to the absorbance obtained from assaying a known quantity of N-acetylneuraminic acid (usually 5 yg br approximately 20 nanomoles, based on a molecular weight of 267.24). 2.5.3 Cholesterol Analysis The method used for cholesterol analysis was based on the procedure reported by Zak and co-workers (186). An iron stock reagent was prepared by dissolving 2.5 gm FeCl-j^H^O i n 25 ml glacial acetic acid. This stock reagent was stored at - 2 0 ° C . A "working" Zak reagent was prepared by d i lut-ing the iron stock reagent 1:100 with concentrated h^SO^. Sample volumes used were 50 yl or less. 2.0 ml glacial acetic acid were added to the sample and the mixture was allowed to stand at room temperature until clear (approximately 30 min.). Then 1.3 ml of the "working" Zak reagent were carefully added so as to underlay the glacial acetic acid layer. The layers were then mixed thoroughly and the color was allowed to develop for 30 minutes, at which time the absorbance at 565 nanometers was determined. Sample absorbances were compared to a standard absorbance where a known quantity of cholesterol (50 yg or 129.3 nanomoles) was assayed. 2.5.4 Phospholipid Analysis Phospholipid content was determined by assaying for phospholipid phos-phorus as described by Bartlett (187) except that the samples were digested with perchloric acid instead of sulfuric acid. The assay was conducted as follows: Stock solution A (Fiske-Subbarow reagent) was prepared by adding 0.25 gm l-amino-2-naphthol-4-sulfonic acid (ANS) to 100 ml freshly prepared 15% (w/v) sodium bisul f i te (anhydrous) and then adding 0.5 gm anhydrous sodium sul f i te . The solution was f i l tered and then stored in a dark bottle for periods of up to a week. - 51 -Stock solution B consisted of a 5% (w/v) ammonium molybdate solution in water. Sample volumes were kept as small as possible. For example, 50 yl of a membrane suspension that contained 3.0-4.0 mg protein/ml was found to be sufficient for this assay. 1.5 ml of 70% perchloric acid were added to the sample and the mixture was digested at 230°C in a sand bath for 30 minutes. After the samples were cooled, 7.6 ml double d i s t i l l ed h^ O were added and the solution was thoroughly mixed. A 4.5 ml aliquot was taken and color was developed by adding 0.5 ml double d i s t i l l ed ^ 0 , 0.2 ml of stock solution B, 0.2 ml stock solution A, mixing and then placing the mixture in a boil ing water bath for 7 minutes. After cooling, the absorbance of the samples at 830 nanometers was determined. Sample absorbances were compared to absorb-ances that were obtained from a standard solution containing a known quanti-ty of inorganic phosphorus (usually 2.0 yg). To convert ymoles phosphorus/ml to ymoles phospholipid/ml, the conversion factor of 25 and an average molec-ular weight of 700 were used. 2.5.5 Ketodeoxyoctanoic Acid (KD0) Assays for the trisaccharide 2-keto-3-deoxyoctonate in endotoxin prepar-ations were basically conducted according to the procedure of Weissbach and Hurwitz (188) as modified by Osborn (189). In more deta i l , the assay was organized in the following manner: Stock solution A which consisted of 0.025 M NaI04 (sodium m-periodate) was prepared by dissolving 5.35 gm NalO^ in 1 l i t e r 0.125 N h^SO^. Stock solution B consisted of 0.2 M NaAs02 (sodium arsenite) dissolved in 0.5 N HCl. - 52 -0.4 ml of stock solution A was added to 0.2 ml of sample solution which consisted of a 2.0 mg/ml suspension of bacterial endotoxin in water. The solutions were mixed and incubated at room temperature for 20 minutes. Then 0.5 ml stock solution B was added and after mixing, the mixture was allowed to stand for 2 minutes. 2.0 ml freshly prepared 0.3% thiobarbituric acid were then added and after vortexing, the samples were placed in a boiling water bath for 20 minutes. The chromophore was then extracted by adding 2.0 ml cyclohexanone, vortexing and centrifuging the samples to f ac i l i t a te separation of the aqueous and organic phases. The absorbance of the cyclo-hexanone phase was measured at 548 nanometers. The quantity of KDO in each sample was determined from a standard curve that was prepared by assaying 2.5 to 50 nanomoles of KDO (Sigma) as described above for samples. To determine the " tota l " quantity of 2-keto-3-deoxyoctonate in endotoxin preparations, i t was f i r s t necessary to subject the endotoxin to acid hydro-lysis in order to free all bound KDO. This was done by incubating 0.2 ml of a 2 mg/ml endotoxin suspension with 0.2 ml 0.05 N r^SO^ and 0.1 ml double d i s t i l l e d r^O in a boil ing water bath for 20 minutes. After cool-ing the sample, a 0.2 ml aliquot of the acid-hydrolyzed sample was taken and assayed for free KDO as described above. 2.6 Thin Layer Chromatography 2.6.1 Intact Erythrocytes Phospholipids were extracted from intact erythrocytes by either of the following two methods. Method A, a modified procedure, gave results similar to the more standard procedure, Method B. Method A: To 1.0 ml packed red blood ce l l s , 1.0 ml H^ O was added and the mixture was allowed to stand at room temperature for 15 minutes. Then - 53 -4.0 ml methanol were added and after mixing, the solution was allowed to stand for 1 hour. 6.0 ml chloroform were then added and, after mixing, the suspension was allowed to stand for another 1 hour. The extracted red cel ls were then centrifuged at 9,000 x g for 5 minutes and the supernatant was washed 3 times by adding 1.0 ml aliquots of 0.75% (w/v) NaCl, vortexing, incubating for 15 minutes at room temperature and centrifuging the samples in a bench-top centrifuge to enhance phase separation. The top aqueous layer was discarded after each wash. Following the final wash, the organic layer was evaporated to dryness under nitrogen and the residue was redissol-ved in 0.2 ml chloroform:methanol (2:1) just before the total sample was applied to thin layer plates for chromatographic separation. Method B: This method of extracting phospholipids from intact red blood cel l s is the same as that described by Reed and associates (190). 5.0 ml methanol were added to 1.5 ml packed red blood cel ls and the l ipids were extracted for 5 minutes at room temperature. Then 5.0 ml chloroform were added and the red cel ls were extracted for another 5 minutes. The cells were then centrifuged at low speed (3,000 x g) and the supernatant was collected. If the supernatant contained appreciable amounts of haemoglobin, this was then removed by washing the supernatant with 1.5 ml 0.75% (w/v) NaCl. The extraction procedure was repeated twice more and all the superna-tants were pooled. The supernatants were f i l tered through glass wool i f fine particles of red cells were present. The pooled supernatants were evaporated to dryness under nitrogen and the residue was extracted with 2.0 ml chloroform for 5 minutes. This was repeated twice more and the extracts (6.0 ml) were pooled. The CHC1^ was evaporated to dryness (under nitro-gen) and the powder was redissolved in a volume of benzene that yielded a - 54 -c o n c e n t r a t i o n of a p p r o x i m a t e l y 20 mg p h o s p h o l i p i d / m l benzene. Twenty t o 30 y l were then a p p l i e d t o the t h i n l a y e r p l a t e . 2.6.2 Membranes and E n d o t o x i n s P h o s p h o l i p i d s were e x t r a c t e d from v a r i o u s membrane p r e p a r a t i o n s and b a c t e r i a l e n d o t o x i n s by i n c u b a t i n g 0.5 ml membrane s u s p e n s i o n (3.0-4.0 mg p r o t e i n / m l ) or 0.5 ml of a s o l u t i o n of e n d o t o x i n (8.0 mg endotoxin/ml h^O) with a 5.0 ml m i x t u r e of CHCl^:methanol (2:1) at room t e m p e r a t u r e f o r 2 h o u r s . The m i x t u r e was then washed 3 times w i t h 1.0 ml a l i q u o t s of 0.75% NaCl by c e n t r i f u g a t i o n and the aqueous l a y e r was d i s c a r d e d each t i m e . The o r g a n i c l a y e r was then e v a p o r a t e d t o dryness under n i t r o g e n and the phospho-l i p i d s were r e d i s s o l v e d i n 0.2 ml CHCl^MeOH (2:1) j u s t b e f o r e t h e y were a p p l i e d t o the t h i n l a y e r p l a t e . 2.6.3 S e p a r a t i o n of P h o s p h o l i p i d s P h o s p h o l i p i d s were s e p a r a t e d on aluminum s h e e t s p r e - c o a t e d w i t h s i l i c a g e l 60 (Brinkman). These s h e e t s were f i r s t " a c t i v a t e d " by h e a t i n g at 110°C f o r 30 m i n u t e s . The p h o s p h o l i p i d s were s e p a r a t e d i n one dimension u s i n g a s o l v e n t system composed of CHCl^:MeOH:NH^ in the r a t i o o f 14:6:1 (by volume). P h o s p h o l i p i d s such as p h o s p h a t i d y l e t h a n o l a m i n e and p h o s p h a t i d y l -s e r i n e , which c o n t a i n p r i m a r y amino groups, were v i s u a l i z e d w i t h n i n h y d r i n s p r a y reagent (Baker) w h i l e other p h o p h o l i p i d s were i d e n t i f i e d w i t h i o d i n e vapor. The p h o s p h o l i p i d s p o t s were then s c r a p e d , d i g e s t e d i n 1.5 ml 70% p e r c h l o r i c a c i d at 230°C f o r 30 minutes and assayed f o r i n o r g a n i c phosphorus as d e s c r i b e d i n Methods s e c t i o n 2.5.4 t o determine the a u a n t i t y of phospho-l i p i d s i n each s p o t . - 55 -2.7 Detoxification of Endotoxin 2.7.1 Sodium Hydroxide Treatment This method of detoxifying endotoxin is based on the study reported by Neter and co-workers (82). Brief ly , 100 mg of Escherichia co l i lipopolysac-charide (Difco) were dissolved in 6.0 ml 0.25 N NaOH. The solution was then heated at 56°C for 1 hour. After cooling, the solution was neutralized with glacial acetic acid and then the lipopolysaccharide was reprecipitated by mixing the neutralized solution of endotoxin with 180 ml cold absolute ethanol. After the lipopolysaccharide was allowed to precipitate by stand-ing for several hours on ice, the ethanolic solution was centrifuged at 10,000 x g for 10 minutes. The ethanol was discarded and the lipopolysac-charide pellet was resuspended in approximately 15 ml double d i s t i l l ed H^ O and lyophilized. 2.7.2 Sodium Periodate Treatment It was also demonstrated by Neter and associates that endotoxins could be detoxified by reacting them with sodium periodate (82). According to this procedure, 100 mg of Escherichia coli lipopolysaccharide were dissolved in 80 ml double d i s t i l l ed H£0. To this solution 10 ml 1.0 N sodium acet-ate buffer pH 5.0 and 10 ml 0.1 M sodium periodate solution were added. The mixture was incubated at room temperature in the dark for 6 and 1/2 hours during which time the solution was continuously s t irred. The reaction mixture was then dialyzed against d i s t i l l ed H2O at 4°C for 36 hours. The d i s t i l l ed H2O was changed every 12 hours. Following dia lys i s , the endotoxin solution was lyophilized and stored at 4°C for future use. - 56 -2.7.3 Treatment with Hydroxy]amine The method for detoxifying endotoxins by treating with hydroxylamine is based on the procedure described by Mclntire and co-workers (193). An alka-line hydroxylamine solution was prepared by mixing a solution of 2.5% NaOH in ethanol with a solution of 2.5% NH^OH.HCl in ethanol. The NaOH/ethanol solution was prepared by f i r s t dissolving 0.5 gm NaOH in 2.0 ml double d i s t i l l e d H^ O and then this was mixed with 18 ml 95% ethanol. Similarly, 0.375 gm hyroxylamine hydrochloride was f i r s t dissolved in 1.0 ml double d i s t i l l ed H20 and then mixed with 14 ml 95% ethanol. The two ethanol solutions were mixed together on ice and after standing for approximately 20 minutes, the mixture was centrifuged to remove the precipitated NaCl. 20 ml of this freshly prepared alkaline hydroxylamine supernatant were taken and mixed with 100 mg Escherichia co l i lipopolysaccharide. This supension of endotoxin was thoroughly homogenized manually with a ground glass homoceni-zer. The fine suspension of endotoxin in the alkaline hydroxylamine solu-tion was incubated at room temperature under nitrogen for 1 hour with constant s t i r r ing . The solution was then centrifuged at 10,000 x g for 5 minutes. The lipopolysaccharide pellet was washed once with 20 ml cold 95% ethanol, once with 20 ml 0.01 M acetic acid in 95% ethanol and once more in 20 ml 95% ethanol. The final pellet was resuspended in approximately 15 ml double d i s t i l l ed H^O, lyophilized and stored at 4°C for future work. 2.8 Radioactive Labelling of Endotoxin 51 CrCl^ (New England Nuclear) was used to label endotoxin as descri-bed by Braude and associates (194). The endotoxin used for radiolabelling studies was lipopolysaccharide extracted from Escherichia col i 026:B6 by the 51 Boivin method (Difco). A " Cr-buffer solution was prepared by taking 5 - 57 -mCi of CrClg (specific act ivity 50-500 mCi/mg) dissolved in 0.1 ml 0.5 M HCl and adding to this 9.9 ml sodium phosphate buffer pH 7.0 that was composed of 2.6 mM NaH?P04-4.5 mM Na2HP04~2.9 mM NaCl. Approximate-ly 10 mg endotoxin were dissolved in every 1 ml of this ^Cr-buffer solu-tion and the mixture was incubated at room temperature for 36 hours with constant s t i r r ing . The radioactive endotoxin solution was then dialyzed against double d i s t i l l e d H^ O for 36 hours at 4 ° C . The H O^ was changed every 12 hours. After d ia lys i s , the endotoxin was lyophilized and stored at 4°C for future experimentation. The specific act iv i ty of the radioactive endotoxin was approximately 20,000 cpm/yq toxin. 51 2.9 Cr-Endotoxin Binding Studies 2.9.1 Membrane Preparations Most of the radiolabeled endotoxin binding studies were done on human erythrocyte ghost membranes. 2.0 ml of an erythrocyte membrane preparation (3.0-4.0 mg protein/ml) were centrifuged and resuspended twice in 0.9% NaCl-15 mM Tris pH 7.0 buffer. After the second wash, the membranes were resuspended in 8.0 ml 0.9% NaC 1-15 mM Tris buffer pH 7.0 to give a protein concentration of approximately 0.60 mg/ml. The incubation mixture, which totaled 2.0 ml, consisted of 0.5 ml membrane suspension, 0.9% NaC1-15 mM Tris pH 7.0 buffer and the desired volume (usually 0.1 ml) of Cr-endo-51 toxin dissolved (0.5 mg/ml) in NaCl-Tris buffer. The 'Cr-endotoxin solution was always sonicated to insure a homogeneous suspension before it was used for al l binding assays. The membranes were pre-incubated in the NaCl-Tris buffer solution at 37°C for 5 minutes before the radiol abelled toxin was added. After addition of the endotoxin, the mixture was further incubated at 37°C for 15 minutes. The incubation mixture was then centr i-- 58 -fuged and the membrane pellet was washed twice by resuspension in 2.0 ml cold NaCl-Tris buffer and centrifugation. After the second wash, the pellet was resuspended in 1.0 ml NaCl-Tris buffer. A 0.2 ml aliquot of this suspension was assayed for radioactivity by l iquid sc int i l l a t ion counting methods as described in section 2.11.1. 2.9.2 Intact Erythrocytes Red blood cells obtained from freshly drawn blood were ut i l ized for 51 Cr-endotoxin binding experiments. The heparinized blood was centrifuged and the red blood cel l s were washed twice in 4 volumes cold 0.9% NaCl-15 mM Tris buffer pH 7.0 to remove all traces of plasma and buffy coat. After the second wash, the erythrocytes were resuspended in the NaCl-Tris buffer to a final haematocrit of approximately 20% (e.g. 2.0 ml packed red blood cel l s were resuspended in buffer to a total volume of 10 ml). A 0.3 ml aliquot of 51 this red blood cel l suspension was used for Cr-endotoxin binding studies as described for membranes in section 2.9.1. 2.10 Liquid Sc int i l la t ion Counting 51 Cr is a gamma emitting radioisotope, but the frequency of decay by gamma emission is only 9%. Therefore, better counting efficiencies are possible i f l iquid sc in t i l l a t ion counting methods are used to detect the emission of X-rays and Auger electrons, the combination of which amount to 51 80 % of the decay frequency of Cr (195). Since the pulse height spectrum of ^ C r is almost identical with the spectrum of (195), al l ^ C r sc int i l l a t ion counting was done using the tritium channel of the liquid sc int i l l a t ion counter. - 59 -2.11 Preparation of Samples for Sc int i l l a t ion Counting 2.11.1 Plasma Membranes 51 Following the procedure used for the Cr-endotoxin/membrane binding assay (section 2.9.1), a 0.2 ml aliquot of the final membrane suspension was placed directly into a glass s c in t i l l a t ion vial and to this was added 0.6 ml of a 1:2 mixture of Protosol (New England Nuclear) :95% ethanol solution. The contents of the sc int i l l a t ion vial were then incubated at 60°C for 45 minutes. The samples were then cooled, 15 ml Biofluor (New England Nuclear) were added and after thorough mixing, 0.5 ml 0.5 N HC1 was added. The samples were thoroughly mixed once again and placed in a sc int i l l a t ion counter to be counted after 1 hour of dark and temperature (10 °C) adaptation. 2.11.2 Intact Red Blood Cells The procedure for preparing samples containing intact red blood cells for s c in t i l l a t ion counting was identical to the procedure used for membranes except that after the 45 minute incubation at 60°C with Protosol:ethanol, the samples were cooled and 0.3 ml 30% H 20 2 was added dropwise to bleach the samples. The vials were loosely capped and incubated at 60°C for an additional 30 minutes. The samples were then cooled and mixed with 15 ml Biofluor and 0.5 N HC1 as described for membranes in section 2.11.1. 2.11.3 Tissues 51 Tissues obtained from animals treated with Cr-endotoxin were prepar-ed for sc int i l l a t ion counting by adding 1.0 ml Protosol (New England Nucle-ar) to 50 mg blotted tissue in a s c in t i l l a t ion vial and digesting at 55°C until the tissue was completely solubilized (usually several hours). The samples were then cooled and 0.1 ml 30% H 20 2 was added to decolorize the samples. The vials were loosely capped and incubated at 55°C for another 30 - 60 -minutes. The samples were cooled once more and mixed with 10 ml Econofluor (New England Nuclear). The vials were allowed to equilibrate in the sc int i l l a t ion counter for 1 hour before they were counted. 2.11.4 Plasma 51 Blood plasma obtained from Cr-endotoxin treated animals was prepared for sc in t i l l a t ion counting by simply mixing 50-100 ul plasma with 15 ml Biofluor (New England Nuclear) and counting in a sc int i l l a t ion counter using the tritium channel. 2.12 Animal Studies Animals used for endotoxin studies were always anaesthetized and trache-ostomized before each experiment. Urethane (1000 mg/kg) served as an accep-table anaesthetic for both guinea pigs and rats which were the animals most commonly used for these experiments. Both the right jugular vein and the right carotid artery were catheterized. The jugular vein catheter was used for injection purposes while blood samples were taken from the catheter in the carotid artery. At the end of each experiment, the animals were sacr i-ficed by giving an overdose of anaesthetic. - 61 -CHAPTER 3 Results 3.1 Some Physiological Effects of Endotoxin Administration The intravenous administration of bacterial endotoxin into laboratory animals causes marked haemodynamic changes which are dose- and time-related. In general, a sub-lethal dose of endotoxin produces an almost immediate f a l l in blood pressure which returns to normal levels after a short period of time. Larger doses of endotoxin produce greater decreases in blood pressure and after a partial recovery, the mean blood pressure can decline steadily with time to shock levels. Haemodynamic changes can also be demonstrated at the microcirculatory level in various organs following injections of endo-133 toxin by monitoring the clearance of radioisotopes such as Xenon from 133 tissue sites of injection. Xenon, being an inert l ipophi l ic molecule, is highly diffusible and its rate of clearance from an injection site is proportional to the capil lary flow to that area. This blood flow is some-times referred to as "nutr i t ional" or "nutrit ive" flow to distinguish it from blood flow which occurs in larger vessels or arteriovenous shunts. The effect of administering 4.0 mg/kg E. co l i endotoxin to eleven anaesthetized rats on the mean blood pressure and nutritive flow in skeletal muscle of the hind limb is shown in Figure 4. Blood pressure was measured from the right carotid artery while nutritive flow in the quadriceps muscle was determined - 62 -EH BLOOD PRESSURE NUTRITIVE FLOW IO men I hr a hr 3 hr ENDOTOXIN HEMORRHAGE Figure 4. Effect of Endotoxin and Haemorrhage on Blood Pressure and Nutritive Flow in the Rat. Comparison of mean carotid arterial blood pressure and nutritive flow in quadraceps muscle in haemorrhaged rats (n = 10) and rats treated with 4.0 mg/kg E. co l i endotoxin (n = 11). Nutritive flow was measured by i 3 6 X e clearance. Data from haemorrhaged animals were taken at the point of irreversible shock as defined in ref. 247. Results are expressed as % pretreatment ( in i t i a l ) value. Haemorrhagic shock nutritive flow is s i g n i f i -cantly greater (.01 > P > .001) than flow 3 hr post-endotoxin. - 63 -133 by the Xenon clearance technique and calculated according to Lassen et a l . (196). It is evident from Figure 4 that by 10 min after injection of the endotoxin, the mean blood pressure had decreased to an average value of 56% of normal while the nutritive flow was more markedly reduced to an average value of 31% of the control or pre-endotoxin flow rate. It is interesting to note that while the mean blood pressure showed a substantial recovery to an average value of 80% of normal by the third hour after the endotoxin was injected (at which time the experiment was terminated), the nutritive flow rate remained essentially unchanged from the reduced rate that was apparent at 10 min after the endotoxin was given. To further examine the behavior of nutritive flow in a different experimental shock model, haemorrhagic shock was induced in a group of 10 urethane-anaesthe-tized rats by bleeding the heparinized animals from the femoral artery into a pressurized reservoir until their mean blood pressure was 1/3 of their normal pressure. The animals were maintained at this blood pressure by additional bleedings or reinfusions of blood until 30% of the total bled volume was returned to the animals. At this point the remainder of the haemorrhaged blood was reinfused. Nutritive flow determinations in the haemorrhaged group of rats were made at two different time intervals. The f i r s t was performed at the moment when the mean blood pressure had decreased to 1/3 normal pressure and the second when 30% of the bled volume was reinfused to maintain this decreased blood pressure in the face of impending haemorrhagic shock. The measured nutritive flow rates were found to be 22% and 14% of normal respectively (data not shown). Therefore, haemorrhage can markedly affect the nutrit ive flow in skeletal muscle. An attempt was made to measure nutritive flow in - 64 -haemorrhaged rats under conditions more closely approximating those in the endotoxin-treated animals by reinfusing a l l the haemorrhaged blood into the animals following the development of shock in order to make the blood volume and average mean blood pressure similar to that seen three hours after endo-toxin administration. Under these circumstances endotoxaemia showed a greater effect on skeletal muscle nutritive flow than did haemorrhagic shock. The data presented in Figure 4 on haemorrhagic shock show the average mean blood pressure and average nutritive flow in the quadriceps muscle of the rat hind limb determined 5 min after total reinfusion of the blood. In comparison to animals receiving E. coli endotoxin, rats in haemorrhagic shock had considerably higher nutritive flow rates, which were on the aver-age, 62% of the normal flow rate. This rate is s ignif icantly higher (.01 > P > .001) than the 3 hr post-endotoxin mean nutritive flow of 30% of normal. These results serve to emphasize the marked sensi t iv i ty to endotox-in of nutritive flow, a c r i t i c a l determinant of tissue survival in states of circulatory impairment. Tissue injury from a variety of causes would be expected to result in the release of lysosomal hydrolases into the plasma. The increase in plasma acid phosphatase act ivity at various times after 10 mg/kg E. col i endotoxin (026:B6) was administered intravenously to anaesthetized rats is shown in Figure 5. The blood sampling had no effect on plasma acid phosphatase levels in the control animals as indicated in Figure 5. However, rats which received 10 mg/kg E. coli endotoxin (native endotoxin) had significantly higher (P < .001) plasma acid phosphatase levels than control animals by 1 hr following the administration of endotoxin. The levels of this lysosomal enzyme continued to increase dramatically in the plasma with time until the - 65 -NATIVE ENDOTOXIN (0 h 2 D UJ tn < I a U) • I a u NaOH DETOXIFIED' PERIODATE OETOXIFIEDo CONTROL T I M E ( h r s ) Figure 5. Effect of Native and Detoxified E. col i Endotoxins on Plasma Acid Phosphatase Activi ty in the Rat. Blood samples (1.0 ml) for acid phosphatase analysis were taken from each rat at hourly intervals after injection of the native or detoxified endotoxins (10 mg/kg) for a period of 5 hr. Five animals were used in each experimental group except the native endotoxin-treated group (n = 10) and control group (n = 10). - 66 -animals expired, which usually occurred between 5 and 6 hr following the injection of this large dose of endotoxin. In marked contrast to the effects of native endotoxin, E. coli endotoxin which was detoxified by sodium hydroxide or periodate treatment (Methods sections 2.7.1. and 2.7.2) had vir tual ly no effect on plasma acid phosphatase act ivi ty (Figure 5). Also unlike the native endotoxin, the detoxified endotoxins were devoid of lethal effects under these conditions. These results suggest that eleva-tions in the plasma level of this lysosomal hydrolase may provide a useful measure of endotoxin toxici ty in vivo. To determine i f the effect of endotoxaemia on lysosomal hydrolase activ-ity in the plasma was peculiar to the rat and i f lysosomal enzymes other than acid phosphatase could also be found to increase, comparable experi-ments were conducted in anaesthetized guinea pigs with the inclusion of assays for two additional lysosomal enzymes, namely N-acetyl-B-glucosamini-dase and cathepsin D. Under the experimental conditions employed, the animals survived no longer than 5 hr after receiving a 2.0 mg/kg dose of endotoxin. Figure 6 depicts the effects of intravenous E. co l i endotoxin (2.0 mg/kg) on plasma acid phosphatase, N-acetyl-e-glucosaminidase and cathepsin D act ivit ies in guinea pigs. Significant elevations in act ivity of a l l three plasma lysosomal enzymes were apparent by 1 hr following the injection of the endotoxin. Also, the act ivit ies of these enzymes continued to increase as the condition of the animals deteriorated with time. Enzyme levels in control animals remained unaltered throughout the experimental period. Thus, these experiments indicate that guinea pigs, l ike rats, respond to endotoxaemia by displaying elevated plasma lysosomal enzyme act iv i t ies . It is interesting to note that the patterns of increase in - 67 -7 • < a a U) • < j CD a < E 5 • < J a £ < 5 J a _ 3 7 -2 5 SO 15 ACID PHOSPHATASE A-I 5 GLUCOSAMINIDASE T I M E ( hr ) Figure 6. Effect of E. co l i Endotoxin on Plasma Lysosomal Enzyme Activi ty in Guinea Pigs. Blood samples (1.3 ml) were taken from each urethane anaesthetized guinea pig at hourly intervals after the injection of endotoxin (2.0 mg/kg) for a time period of 5 hr. Controls received no endotoxin. Enzyme a c t i v i -ties are as defined in the Methods section. Data represent mean ± S.E.M. of 10 animals. - 68 -act iv i t ies of the three plasma lysosomal enzymes do not coincide. For instance, N-acetyl-e-glucosaminidase act ivity reached a maximum value by 3 hr after the injection of endotoxin whereas the act ivi t ies of acid phospha-tase and cathepsin D continued to increase up to. the point of death (5 hr) . Figure 7 shows the effect of varying doses of endotoxin on the plasma activ-it ies of acid phosphatase, N-acetyl-e-glucosaminidase and cathepsin D in different groups of urethane anaesthetized guinea pigs 2 hr after endotoxin administration. It is apparent that although the levels of al l three lyso-somal enzymes increase in the plasma in response to increasing doses of endotoxin, the relationship between the elevation in plasma enzyme act ivity and endotoxin dose is not the same for a l l the enzymes. Nonetheless, i t is apparent from these experiments that the act ivi t ies of several representa-tive lysosomal enzymes probably can be used to monitor the condition of animals during endotoxaemia. A method such as this would have great poten-t i a l in c l in i ca l situations where the progress of patients through episodes of gram-negative sepsis could be more accurately assessed, provided that humans in gram-negative septicaemia also display elevated plasma concentra-tions of lysosomal enzymes. To explore this poss ib i l i ty , blood samples obtained from patients admitted to the Intensive Care Unit, Vancouver General Hospital (a number of whom werein gram-negative sepsis) were assayed for plasma acid phosphatase, N-acetyl-p-glucosaminidase and cathepsin D act iv i t ie s . Table 1 shows the plasma act ivi t ies of each of these enzymes in the 36 patients studied, along with the normal plasma lysosomal enzyme levels obtained from healthy volunteers. The patient data were placed into either of two categories; those patients in shock (blood pressure < 90/60) and - 69 -• ' • • » 1 I 2 3 4 5 B E N D O T O X I N ( m g / k g ) i i i Effect of Varying Doses of E. co l i Endotoxin on Plasma Lysosomal Enzyme Act ivi ty in Guinea Pigs. Each dose of endotoxin was administered intravenously to guinea pigs and blood samples for lysosomal enzyme analysis were taken at 2 hr following the injection of endotoxin. Each dose-group consisted of 10 animals except for the 6 mg/kg dose-group (n = 5). Data are presented as mean ± S.E.M. TABLE 1. Plasma lysosomal enzyme act iv i t ie s in patients with septicaemia and/or shock Acid Phosphatase Cathepsin D** Glucosaminidase (units/ml plasma) (absorbance/ml plasma) * (a) Controls 0.31 ± .01 0.20 ± .05 0.95 ± .06 (b) Gm -ve septic shock 1.35 ± .49 5.27 ± 1.13 3.30 ± .57 (n = 13; 1 survivor) (c) Other shock 0.60 ± .11 1.04 ± .34 2.24 i .32 (n = 10; 4 survivors) Stat is t ical significance (b v_s c) N.S. .0025 > p > .0005 N.S. (d) Gm -ve sepsis, no shock 0.74 ± .28 1.00 ± .26 2.13 ± .33 (n = 9; 6 survivors) (e) Gm +ve sepsis, no shock 0.57 ± .31 0.48 ± .28 1.60 ± .15 (n = 4; 3 survivors) Stat i s t ical significance (d vs_ e) N.S. N.S. N.S. Values are given as mean ± SEM, n = number of experimental subjects and enzyme act iv i ty units are as defined in the METHODS. S ta t i s t i ca l ly significant increases relative to control were present in a l l groups (b-e) for the three enzymes studied (Student's t test, p values ranged from < 0.05 to < 0.001). P values for inter-group s ta t i s t ica l comparisons are as indicated. N.S. = No significant difference. * n = 15, 27 and 26 for acid phosphatase, cathepsin D, and glucosaminidase data, respectively. ** See F ig . 3 for individual values. bacteraemic patients without shock. It is interesting to note from Table 1 that the majority of the patients in shock (13/23 or 57%) and the patients in sepsis-without-shock (9/13 or 69%) had gram-negative bacteraemia. This is consistent with the known high incidence of gram-negative septicaemia that occurs in North American hospitals today. In our study, bacteraemias were generally identified by blood culturing techniques. However, in some cases where gram-negative septicaemia was strongly suspected due to the manifestation of characteristic c l in ica l signs peculiar to this infection, but could not be confirmed with blood cultures, the diagnosis was assigned on the basis of independent evaluations made by two collaborating physi-cians. Consistent with the results from our animal studies on endotoxaemia, gram-negative septic patients in shock exhibited signif icantly higher plasma act ivi t ies of acid phosphatase (.025 > P > .01), N-acetyl-e-glucosaminidase (.001 > P) and cathepsin D (.001 > P) than controls. Although the plasma act ivi t ies of these enzymes were also elevated in c r i t i c a l l y i l l patients with shock from causes other than gram-negative bacteraemia, the act ivi t ies were less than those seen in patients with gram-negative shock. The activ-ity of cathepsin D, in particular, was s ignif icantly higher in gram-negative shock plasma than in plasma of patients with other types of shock (.0025 > P > .0005). These results suggest that the elevation of lysosomal enzyme activity in plasma of patients in gram-negative shock is not entirely due to compromised tissue perfusion which is present in a l l types of shock. Rather, additional factors, such as the release of endotoxins from gram-neg-ative b a c i l l i , may be responsible for much of the lysosomal enzyme act ivity seen in patients infected with these organisms. The observation that plasma lysosomal enzyme activity is less markedly elevated when gram-negative - 72 -septicaemia is uncomplicated with shock, may indicate a lesser degree of endotoxaemia than when shock is present (Table 1 ) . What is particularly interesting in Table 1 , however, is the indication that cathepsin D can perhaps be used as a f a i r l y specific marker for gram-negative bacteraemia and the act ivity of this protease in the plasma may be useful in aiding the diagnosis or evaluating the progression of a gram-negative septic episode. When the cathepsin D data from al l the patients studied are presented, as in Figure 8, i t is evident that only two of the thirteen patients in gram-negative shock had plasma cathepsin D act iv i t ies in the normal range ( i . e . false-negatives). Interestingly, while blood cultures from both these patients indicated the presence of gram-negative organisms, there was a strong poss ib i l i ty that their shock states may have been due to factors other than gram-negative septicaemia. For example, one patient had endocar-d i t i s from a septic focus involving a prosthetic aortic valve while the other patient had severe gastrointestinal haemorrhage in addition to a meta-static pancreatic carcinoma at the time when blood samples were taken for lysosomal enzyme analysis. With regard to false-positive data, Figure 8 reveals that five of ten "other" shock patients exhibited higher than normal plasma cathepsin D act iv i ty . The highest value ( 3 .16 units) was found in a patient with a septicaemia caused by $-haemolytic streptococcus, an organism which can produce potent toxins (streptolysins) capable of causing cel lular ly s i s . Nonetheless, this patient's plasma cathepsin D activity ( 3 . 16 units) was s t i l l lower than the mean value (5.27 units) seen in the plasma of gram-negative septic shock patients. It is also of interest to note from Figure 8 that of the nine gram-negative septic patients without shock, only one had normal plasma cathepsin D activity while the others had somewhat elevated values. - 73 -(0 < J a I s • IO > r h U < to a ui z r < (J O.B • 0>4 a.2 CONTROL ORAM + VE o CAROIOSENIC GRAM -VE • GRAM +VE J 0 a i-2 Q U I L • RAM -VI SHOCK OTHER SHOCK SEPSIS Nb SHOCK Figure 8. Comparison of Plasma Cathepsin D Activit ies in Patients C r i t i c a l l y 111 with Sepsis and/or Shock. Each point represents the plasma cathepsin D act iv i ty in one patient with sepsis and/or shock. In this study, shock was defined as a blood pres-sure reading of less than 90/60. Inset shows the range of plasma cathepsin D act ivity in 27 healthy volunteers. Thus, while the information from human studies needs to be substantia-ted, i t does suggest that the high plasma concentrations of lysosomal enzymes (particularly cathepsin D) seen in gram-negative septicaemia cannot be simply due to the consequences of a state of compromised tissue perfusion but rather additional factors, such as the release of endotoxins from the invading organism, could also be responsible. Consistent with this hypothe-sis are results from animal studies examining the effect of hypotension, with resultant impairment in tissue perfusion, on the accumulation of lyso-somal hydrolases in plasma. At 3 hours following the administration of endotoxin (2 mg/kg) to guinea pigs, only modest decreases in mean blood pressure (to 75 ± 12% of control, mean ± S.D.) were observed (data not shown) when appreciable elevations in plasma lysosomal enyzme act ivi t ies had occurred (see Figure 6). For comparative purposes, guinea pigs were bled from the carotid artery to 50% of their normal mean blood pressure and main-tained at this pressure for 3 hours by additional withdrawals or reinfusions of blood. Despite the fact that the haemorrhaged animals were subjected to more severe hypotensive conditions, their plasma lysosomal enzyme act ivit ies were significantly lower than those seen in endotoxaemia as indicated in Table 2. Thus i t seems l ikely that endotoxins may exert direct actions on cells in various tissues, resulting in cel lular death and the leakage of lysosomal enzymes into the c irculat ion. As an i n i t i a l approach to determin-ing whether endotoxins are indeed capable of such actions, experiments were devised to examine the ab i l i ty of endotoxins to bind to cel lular membranes. E l 3.2 Binding Studies with Cr-Endotoxin Escherichia coli lipopolysaccharide (026:B6) extracted by the Boivin method were obtained from Difco Laboratories (Detroit) and were radiolabel-TABLE 2. Comparative effects of experimental endotoxaemia and haemorrhage on lysosomal hydrolase act iv i t ies in plasma Plasma lysosomal hydrolase act iv i ty Acid Phosphatase Increase Cathepsin D Increase Glucosaminidase Increase (units/ml plasma) ( control) (units/ml plasma) ( control) (Absorb/ml plasma) ( control) Controls 2.05 ± .11 — 1.28 ± .05 — 14.33 ± .54 (n = 10) Endotoxaemia 4.94 ± . 4 1 a ' b 2.4 4.76 ± . 5 0 a ' b 3.7 25.28 ± 2 . 0 1 a ' b 1.8 (n = 10) Haemorrhage 2.27 ± .13 1.1 2.95 ± .34 a 2.3 17.70 ± 1.01 a 1.2 (n = 6) Endotoxaemia values were obtained 3 hr following injection of E. co l i endotoxin. Measurements under haemorrhagic conditions were made 3 hr after guinea pigs had been bled to produce a 50 percent reduction in mean blood pressure, n = number of experimental animals. A l l values are given as mean * SEM; s ta t i s t ica l significance was assessed by Student's t test. a Signi f icant ly elevated (P < .01) relative to control group. b Signi f icant ly elevated (P < .02) relative to haemorrhage group. led with CrCl^ (New England Nuclear) as described in Methods section 51 2.8. The Cr radionuclide bound tightly to the endotoxin and was associ-ated with the proteolipid components of the toxin. This is i l lustrated in 51 Figure 9 where Cr-endotoxin was disrupted by hydroxylamine treatment (see Methods section 2.7.3) and separated into three fractions by Sepharose 6B chromatography. Al l three fractions, which contained protein and phospholipid (the latter not shown for s impl ic i ty) , contained radioactive 51 Cr, with the component appearing between fractions 40 and 60 being most 51 heavily labelled. The Cr did not bind to a 2-keto-3-deoxyoctonate (KDO)-containing component which appeared as a peak between fractions 60 and 51 80 (not shown in Figure 9). Thus, the Cr seems to preferentially bind to proteolipid components of endotoxin. 51 Experiments were then designed to determine i f the Cr-endotoxin was capable of binding to cells and cel l membranes in a specific manner. Human erythrocytes provided a convenient model system for study. The procedure used for the binding study was as described in Methods section 2.9.1 and 2.9.2. Human erythrocyte ghosts (approximately 300 yg protein) incubated 51 with 25 yg/ml Cr-endotoxin exhibited binding which could be largely prevented by the addition of cold (or unlabelled) toxin (Figure 10). 51 Approximately 75 % of the bound Cr-endotoxin was displaceable under these experimental conditions. The data in Figure 11 (bottom) indicate that haemoglobin-free membranes are capable of binding considerably greater 51 quantities of Cr-endotoxin per mg membrane protein than intact erythro-cytes (1.0 ml of packed red blood cells was considered to be equivalent to 5.0 mg ghost protein). It also can be noted from this Figure that the 51 binding of Cr-endotoxin to red cel l ghosts and intact cel l s appeared to - 77 -• iaa so aa B O B O t o o F R A C T I O N I M O . Figure 9. Chromatography of coli Endotoxin. ^Cr-endotoxin was treated with hydroxylamine (as described in the Methods section) and 5.0 mg was placed on a Sepharose 6B column (qel = 1.6 x 90 cm) and eluted with 0.9% NaCl - 15 mm t r i s -.02% NaN3 (pH 7.6) . 2.0 ml fractions were collected and analyzed spectrophotometrically at 254 nm (closed c i rc le s ) . 50 yl aliquots of each fraction were analyzed for radio-act ivity by sc int i l l a t ion counting (open c i rc le s ) . - 78 -I O O r 1 ' S O i o o I S O B O O UNLABELLED ENDOTOXIN ( u g / m l ) Figure 10. Displacement of Bound ^ C r - E . c oi-j Endotoxin from Human Erythrocyte Membranes by Unlabelled Endotoxin. Effect of varying concentrations of unlabelled E. col i endotoxin on binding of 5 1Cr-endotoxin (25 yg/ml) to human erythrocyte membranes (300 yg protein). Results are expressed as % of 5 1Cr-endotoxin bound in the absence of unlabelled endotoxin. Each value represents the mean ± S.D. of 3 separate experiments done in t r ip l ica te . - 79 -Figure 11. Comparison of 5 1 C r - E . coli Endotoxin Binding to Human Erythrocytes and Erythrocyte Membranes. The lower graph compares the binding of varying concentrations of ^•'•Cr-endotoxin to human erythrocyte membranes (solid line) and red cel l s (dotted l ine) . The binding procedure is described in Methods section 2.9.1 and 2.9.2. 1.0 ml packed red blood cells was considered to be equivalent to 5 mg membrane protein. Each value represents the mean ± S.D. of 3 separate experiments done in t r i p l i c a t e . Upper graphs show Scatchard plots of the binding data. - 80 -involve two classes of binding sites with a third non-saturating component (up to concentrations of 400 yg/ml endotoxin) being present only in the ghosts. This latter component in ghosts may reflect non-specific associa-tion of endotoxin with latent membrane components exposed during the course of the step-wise haemolysis procedure used to prepare the membranes from intact ce l l s . In an attempt to obtain more detailed quantitative informa-tion from these binding data, Scatchard analysis was employed. This confirmed the existence of more than one class of binding s ite in both intact ce l l s and ghosts (Figure 11, upper panels). Binding characteristics (apparent dissociation constant Kp) and maximal binding capacity (B m a x ) were estimated from linear portions of Scatchard plots as summarized in Table 3. It is apparent from this Table that at endotoxin concentrations less than 200 yg/ml, the binding sites on the erythrocyte ghost membrane display-ed a higher aff inity for the radiolabeled E. col i endotoxin than did the sites on intact red c e l l s . Secondly, the endotoxin binding capacity seen in the ghosts was also greater than that seen in the red cel ls for concentra-tions of endotoxin less than 50 yg/ml. As mentioned ear l ier , the nature of the association between erythrocyte ghost membranes and endotoxin at concen-trations of endotoxin exceeding 300 yg/ml is l ike ly a non-specific type of adsorption. With regard to the saturable binding sites of intermediate a f f ini ty , i t seems unlikely that these have physiological relevance because blood concentrations of endotoxin in excess of 50 yg/ml are rarely encoun-tered even in experimental endotoxaemia. The high af f inity binding sites for endotoxin may well be most representative of membrane sites of toxin interaction in vivo. Our earlier observations (Figure 10) that binding of - 81 -TABLE 3. 51 Binding characteristics of Cr-endotoxin in intact human erythrocytes and erythrocyte ghosts. [Data were derived from the Scatchard plots depicted in Figure 13.] KD (Mg/ml) B m , „ (ug/mg membrane protein) [ Cr-endotoxin] (ug/ml) Intact red cells Ghost Intact red cel ls Ghost < 50 65 15 2 6 50-200 720 125 15 16 > 200 2900 145 - 82 -51 Cr-endotoxin may be effectively antagonized by unlabelled toxin is indi-cative of a specific interaction which may involve previously characterized lipopolysaccharide receptors present on red cel l membranes (197). 3.3 Effects of Endotoxin on Red Blood Cell Membranes 3.3.1 Erythrocyte Ghosts It was of interest to examine what functional consequences, i f any, occurred as a result of the interaction of E. co l i endotoxin with plasma membranes. Again, the human erythrocyte ghost was used as a model membrane system. In i t i a l investigations indicated that E. co l i endotoxin had an inhibitory effect on Na + , K+-ATPase act ivity so the ab i l i ty of Na+ and K + to stimulate this enzyme was greatly diminished in the presence of endotoxin. However, the results were rather d i f f i cu l t to interpret because commercially prepared endotoxins (Difco) contain C a + + and this resulted in the activation of a Ca++-dependent ATPase which is also present in eryth-rocyte plasma membranes. To circumvent this ambiguity, the effect of endo-toxin on the act ivi ty of K+-nitrophenylphosphatase (K+-NPPase), which + + + represents the K -stimulated portion of the Na ,K -ATPase enzyme, was examined. The results are shown in Figure 12. It is evident that endotoxin is an effective inhibitor of both the K stimulated component of the enzyme and the basal act ivi ty , i . e . the act ivity of the enzyme in the absence of Mg + + and K + . The act ivit ies of both of these components was inhibited by approximately 50% at low (50 yg/ml) concentrations of endotox-++ in . The Mg -stimulated component of the NPPase enzyme was very much less sensitive to inhibi t ion, so that endotoxin concentrations as high as 500 yg/ml inhibited this act ivity to the extent of only 30%. These results emphasize that endotoxin binding is capable of modifying the functional integrity of plasma membranes. - 83 -Figure 12. Effect of E. co l i Endotoxin on K + - and Mg++-p-Nitrophenylphosphatase Activi ty in Human Erythrocyte Membranes. (Mg+ +,K+)-p-Nitrophenylphosphatase act ivity was determined in human erythrocyte membranes (Methods section 18.104.22.168) when these were incubated oin the absence and presence of varying concentrations of endotoxin at 3 7 ° C . Reaction time was 1 hr. The results are an average of 3 separate experi-ments done in duplicate. - 84 -3.3.2 Intact Red Blood Cells One simple and convenient method of examining functional consequences of endotoxin binding to intact cel ls is to investigate the effect of endotoxin on the s tab i l i ty of red blood cells to hypotonic challenge. The methodology for these experiments is described in Methods 2.3. Figure 13 i l lustrates the marked protective effects of increasing concentrations of E. col i endo-toxin on red cel l haemolysis at 37°C expressed as a percent of haemolysis occurring in the absence of endotoxin. A 50% reduction in haemolysis was Q obtained at an endotoxin concentration of 0.5 mg/10 red c e l l s , with maxi-mal protection occurring at a 4-fold greater endotoxin concentration ( i . e . 2 mg/10^ c e l l s ) . The mechanism of the protective effect afforded by endo-toxin may involve a direct membrane stabi l iz ing action or i t could also be argued that endotoxin is acting by increasing the pre-lyt ic leakage of K + out of the red blood ce l l s , thereby making the cel ls more resistant to hypo-tonic ly s i s . This latter explanation can be ruled out, however, because the release of K+ from red blood cel ls in the early phase of haemolysis (where less than 10% haemoglobin release had occurred) was unaffected by endotoxin (Figure 14). Therefore, endotoxin protects human red blood cel ls from hypotonic lysis by a mechanism other than increasing the pre-lyt ic leakage of K + from the ce l l s . Since studies of the in vivo actions of endotoxin were largely performed using Wistar rats, the effects of endotoxin on the s tab i l i ty of rat erythro-cytes to hypotonic lysis were also investigated. The results summarized in Figure 15, comparing rat and human red cel ls at 2 0 ° C , revealed a predomi-nantly lyt ic action of endotoxin in the rat as contrasted with the s t ab i l i z -ation seen in human erythrocytes. These two dramatically opposing effects - 85 -m g E N D O T O X I N / I O 9 R B C Figure 13. Antihaemolytic Effect of E. col i Endotoxin on Human Red Blood Ce l l s . Varying concentrations of endotoxin were incubated with human red cel l s at 37°C for 15 min before the addition of a hypotonic solution as described in Methods section 2.3. Haemolysis was determined by the absorbance of free haemoglobin at 540 nm. Results are expressed as % haemolysis seen when no endotoxin was present. The data represent the mean of three separate exper-iments done in t r ip l i ca te . - 86 -Figure 14. Effect of E. co l i Endotoxin on Pre-lytic Leakage of K + in Human Erythro-cytes Human erythrocytes were incubated in the presence (open circles) and absence (closed circles) of 500 pg endotoxin under condition identical to those in haemolysis experiments (Methods section 2.3) except that the NaCl concentration was varied such that control haemolysis (no endotoxin) ranged from 1% to 50% of total haemolysis (lysis in d i s t i l l e d H 2 O ) . Supernatants were analyzed spectophotometrically at 540 nM for haemoglobin and by atomic absorption spectrometry for K content. Data represent the mean of 3 experiments assayed in duplicate. - 87 -1 • ' • * s a e 8 i a m g E N D O T O X I N / l O 9 R B C Figure 15. Effects of Increasing Concentrations of E. col i Endotoxin on the Osmotic Stabi l i ty of Rat and Human Red Cel l s . Haemolysis experiments were conducted at 20°C as described in Methods section 2.3. Results are the mean ± S.D. of t r ip l i ca te experiments u t i l i z -ing 3 different blood samples. - 88 -of endotoxin probably occur by virtue of the amphipathic nature of this macromolecular complex in a manner similar to the known effects that small amphipathic molecules (e.g. local anaesthetics) have on red ce l l s t ab i l i ty . To further investigate these actions of endotoxin on human and rat erythro-cytes the effect of varying temperature on the s tabi l iz ing and ly t i c a c t i v i -ties of endotoxin was examined. Figure 16 shows that the degree of s t a b i l i -9 zation produced by a fixed concentration of endotoxin (4 mg/10 red cel l s ) was markedly temperature dependent. It is interesting to note that endotoxin, which enhances the lysis of rat red blood cel ls at lower temperatures, can s ignif icantly stabil ize these cells when the haemolysis test is conducted at 3 7 ° C . A similar trend is seen when endotoxin is incubated with normal human red cel ls in that greater protection is seen with increasing temperature, although no lysis occurred, even at 5 ° C . One possible explanation for this marked effect of temperature on the s tabi l iz ing action of endotoxin in human red cel l s could involve a reduction in toxin binding with decreasing temperature. However, no effect of temperature on endotoxin binding to human erythrocytes could be demon-strated (Table 4). The observation that E. col i endotoxin protected human erythrocytes from hypotonic lysis to a much greater degree than rat red blood cel l s suggested that certain compositional entities of the red ce l l membrane may greatly influence the antihaemolytic action of endotoxin. When plasma membranes from both human and rat erythrocytes were analyzed for phospholipid content by thin layer chromatography, i t became apparent that the rat erythrocyte membranes contained greater quantities of phosphatidylcholine (PC) but lesser amounts of sphingomyelin (Sph) than did human erythrocyte membranes - 89 -200 in > j • 150 • HUMAN /\(l_CAT DEFICIENT) Ui I J • a h Z • U i o o 5 0 RAT H U M A N ( C O N T R O L ) 1 , 1 1 i a sa 30 a o T E M P ( ° C ) Figure 16. Temperature-Dependent Effects of Endotoxin on the Osmotic Stabi l i ty of Erythrocytes. Erythrocytes from normal humans, rats and from a patient with congenital deficiency of plasma lecithin-cholesterol acyltransferase (LCAT) act ivity were incubated with E. col i endotoxin (4 mg/io" red cel l s ) at various temperatures. Data for human and rat erythrocytes represent the average of 3 different blood samples while the LCAT-deficiency patient results are from a single experiment. - 90 -TABLE 4. 51 Effect of temperature on the binding of Cr-labelled lipopolysaccharide (serotype 026:B6, lot number 669176) by human erythrocytes Temperature Vc) Endotoxin binding 5 327 ± 90 10 360 ± 53 15 487 ± 92 25 330 ± 155 37 387 ± 11 The results are the mean ± SD of experiments using erythrocytes from three different healthy volunteers. Concentration of Cr-endo-toxin was 25 yg/ml. * 8 Binding is expressed as ng toxin bound per 10 erythrocytes. as indicated in Table 5. Thus, rat erythrocyte membranes have a higher PC/Sph rat io, which is known to be an important determinant of membrane f lu id i ty , than human red cel l membranes. Interestingly, i t was found that erythrocytes from a patient with a congenital deficiency in lecithin-choles-terol acyltransferase (LCAT) had a membrane PC/Sph ratio similar to that seen in rat red blood c e l l s . When the LCAT patient's erythrocytes were used in the haemolysis experiments, the results of the temperature-dependent effects of endotoxin were completely different from those obtained with normal human erythrocytes, but str ikingly similar to the results seen with rat red blood cel ls as shown in Figure 16. These observations then suggest that the compositional make-up of plasma membranes is a very important determinant of the consequences of endotoxin-red ce l l interaction. In an attempt to further investigate the effects of compositional differences in red blood cel l membranes on the actions of endotoxin, eryth-rocytes from a variety of animal species were subjected to a hypotonic challenge (Methods 2.3) in the presence and absence of endotoxin and at different temperatures. Figure 17 shows the results of these experiments and what is immediately apparent is that of al l the species of red blood cells examined, none were stabilized by endotoxin to the same extent as human red blood cel l s at any temperature. These experiments did reveal, however, that PC/Sph ratio alone does not determine the membrane actions of endotoxin since rabbit and cat erythrocytes, which have PC/Sph ratios approximating that in human cells (198), exhibited temperature profiles quite comparable to those of the rat. Similarly, no obvious correlation was found between the cholesterol content of the erythrocytes studied and the temperature-dependent effects of endotoxin on osmotic f r a g i l i t y . TABLE 5. Thin layer chromatographic analysis of phospholipid profiles from normal human erythrocytes, rat erythrocytes and erythrocytes from a patient with a congenital deficiency of plasma lec i th in : cholesterol acyltransferase (LCAT) Total Phospholipid Phosphorus P Choline ratio Species P Ethanolamine P Serine P Choline Sphingomyelin Sphingomyelin Human (control) 28.6 ± 1.1 11.8 ± 1.0 30.6 ± 1.9 24.8 ± 2.0 1.2 Rat 27.9 ± 0.9 14.1 ± 0.5 43.3 ± 1.3 10.7 ± 1.7 4.0 Human (LCAT deficient) 18.6 ± 0.1 6.7 ± 1.1 55.6 ± 0.1 15.0 ± 0.1 3.7 Results expressed as mean ± SD. - 93 -•J 1 1 l — 10 ea 3 0 ao T E M P ( ° C ) Figure 17. Comparison of the Temperature-Dependent Effects of Endotoxin on the Osmotic Stabi l i ty of Erythrocytes from Various Animal Species. The concentration of E. col i endotoxin used in these experiments was 4 mg/10^ red ce l l s . The data represent the average of 3 different blood samples from each species. - 94 -In further experiments to investigate the role of membrane structural components in influencing the ab i l i ty of endotoxin to protect human erythro-cytes from hypotonic lys i s , red ce l l s were enzymatically modified using neuraminidase, trypsin or phospholipase A (see Methods 2.4.1, 2.4.2, 2.4.3) which altered the membrane carbohydrate, protein and phospholipid components of the erythrocytes, respectively. The results of these experiments are shown in Figure 18. It is apparent that the only enzyme treatment which appreciably affected the ab i l i ty of endotoxin to protect the modified red cel l s from hypotonic lysis was phospholipase A treatment. The effect of phospholipase A treatment is especially apparent at higher temperatures. This treatment reduces membrane PC content by approximately 60%. In contrast, neuraminidase treatment, which removed approximately 75% of the membrane s i a l i c acid, had no influence on the ab i l i ty of endotoxin to protect these cells from hypotonic ly s i s . These results indicate that the integrity of the phospholipid components of the membrane is important in determining how effectively endotoxin stabilizes the erythrocytes to hypo-tonic lys i s . Our results suggest that i t is probably the ab i l i ty of the hydrophobic ( l ipid) component of the endotoxin to interact with the erythro-cyte membrane that has been affected by phospholipase A treatment. To examine the influence of the l i p i d component of endotoxin on red cel l s t ab i l i ty , we undertook experiments using chemically modified endotoxins wherein the l ip id region of the complex was altered by sodium hydroxide or hydroxylamine treatment (Methods 2.7.1, 2.7.3). These treatments are known to hydrolyze the ester-linked fatty acids in the l i p i d A region of the endo-toxin complex. Hydroxylamine treatment is more drastic than sodium hydrox-ide treatment in the sense that i t removes amide as well as ester-linked - 95 -• T R E A T E • 0) in > j • £ UJ I j • a Z • CJ too x U N T R E A T E D N E U R A M I N I D A S E ao 30 T E M P ( ° C ) 4 0 Figure 18. Temperature-Dependent Effects of Endotoxin (4 mg/10^ red cel ls) on Modi-fied Human Erythrocytes. Human erythrocytes wre modified enzymatically with neuraminidase, tryp-sin and phospholipase A as described in Methods section 2.4. Results represent the mean ± S.D. of 3 experiments. - 96 -fatty acids in l i p i d A. We have also modified the carbohydrate regions of endotoxin by sodium periodate (NalO^) treatment (Methods 2 . 7 . 2 ) . The ab i l i ty of these carbohydrate- and 1ipid-modified endotoxins to s tabi l ize human red blood cel ls was then examined at various temperatures and compared to the effects of control (unmodified) endotoxin. The results of these experiments are shown in Figure 19. It is apparent that modification of the l ip id regions of endotoxin with NaOH or NH2OH had a marked effect on the antihaemolytic action of endotoxin, whereas alteration of the carbohydrate portion had no effect whatsoever. Incidently, i t is known that a l l these treatments (NaOH, NH2OH and NalO^) y ie ld endotoxins that are much less toxic than native endotoxin in vivo. Of particular interest was the obser-vation that the NaOH- and NHgOH-treated endotoxins enhanced the lysis of human red blood ce l l s , particularly at low temperatures. However, these endotoxins s t i l l had a ly t ic effect at 3 7 ° C , although to a lesser extent. The hydroxylamine-modified endotoxin had stronger lyt ic actions than the sodium hydroxide-treated toxin particularly at 3 7 ° C . The interpretation of studies with NaOH- and NH20H-modified toxins is complicated by the fact that these substances are not homogeneous but rather consist of at least four different forms that may be resolved by Sepharose 6B chromatography. Unmodified endotoxin, which normally appears as a homogeneous macromolecular complex (Figure 2 0 , top), yields four discernable peaks (bottom graph) of molecular weights ranging from 1 x 1 0 6 (peak I) to 1 x 1 0 4 (peak IV) following sodium hydroxide treatment. Al l peaks consisted of proteolipid material except peak III which was found to be glycolipid in nature. Peaks I and II consisted of protein, l i p id and carbohydrate. Similar results were obtained when E. co l i endotoxin was modified with hydroxylamine (Methods - 97 -PERIOD ATE • CONTROL o DETOXIFIED 60 40 SO (0 Hi > J • I W I J • II h 2 • U 160 140 ISO IOO 80 BO a a SODIUM HYDROXIDE HYDROXYL A MINE •-_ . — — • A CONTROL * DETOXIFIED • CONTROL • DETOXIFIED I O SO 30 a o T E M P ( ° C ) Figure 19. Effects of Detoxified Endotoxins on the Osmotic Stabi l i ty of Human Erythro-cytes as a Function of Temperature. E. col i endotoxin was chemically modified (detoxified) by treatment with periodate, sodium hydroxide and hydroxylamine as described in Methods section 2.7. Toxin concentration in a l l experiments was 4 mg/10^ red ce l l s . Each point represents the mean of 3 experiments performed on 2 different blood samples. - 98 -PROTEIN o-o KDO CONTROL E c If) tu h < Ul L) 2 < to a • cn ID < • OB • • 4 .OS - .3 • .1 .1 S • . 0 8 »oa E c CD If) h < Ul CJ Z < CD a o ul CD < SO 4Q 60 F R A C T I O N IMO. BO Figure 20. Fractionation of Sodium Hydroxide-Detoxified Endotoxin. Sodium hydroxide detoxified E. col i endotoxin (5.0 mg) was fractionated by Sepharose 6B gel f i l t r a t ion chromatography (gel height = 1.6 x 35 cm) using 0.9% NaCl - 0.02% NaN3 - 15 mM Tris (pH 7.0) as eluant. 1.0 ml fractions were collected and analyzed for protein, KDO and phospholipid, the latter being present in al l fractions but not shown for sake of c l a r i ty . Upper graph shows the separation profile of native (control) endotoxin. - 99 -2.7.3) except' that greater quantities of the peak III glycolipid material were formed as shown in Figure 21 (bottom). When a l l four peaks were tested for lyt ic act ivi ty using the haemolysis test, i t was found that only peak III enhanced the lysis of human erythrocytes in hypotonic buffer. The fact that hydroxy!amine-hydrolyzed endotoxin contained a greater proportion of the glycolipid component (peak III) than did sodium hydroxide-treated endo-toxin, probably explains why the hydroxylamine-modified endotoxin had greater ly t i c act ivity than the sodium hydroxide-modified derivative. In conclusion, the studies thus far have indicated that the structural integr i-ty of the phospholipid components of the human erythrocyte membrane is an important determinant of the action of E. coli endotoxin on red ce l l s . Endotoxins are capable of exerting either s tabi l iz ing or ly t ic actions on red cel l s depending on such factors as ambient temperature and the animal species from which the red cel l s were obtained. Furthermore, alteration of the l i p i d portion of the endotoxin complex i t se l f abolishes the antihaemoly-t i c action of endotoxin, possibly as the result of the formation of smaller glycol ipid components which enhance the lysis of human red blood cel ls in hypotonic solution. 51 3.4 Cr-Endotoxin Binding In Vivo The in vitro studies mentioned above have shown that endotoxin binds to plasma membranes and can influence some membrane-dependent activit ies and possibly the physical state of ce l l s . The next step in these investigations 51 was to administer Cr-labelled E. col i endotoxin in vivo in order to determine i f the endotoxin showed preferential binding to any organ or tissue and i f the toxicity of endotoxin in vivo could be correlated with the 51 association of Cr-endotoxin with some target organ. For these experi-- 100 -• — • PROTEIN o -o KDO CONTROL HYDROX YLAMINE DETOXIFIED • .5 .08 .04 0 in h < ui u 2 < m a. o (0 o s o Figure 21. Fractionation of Hydroxylamine-Detoxified Endotoxin Hydroxylamine-detoxified E. coli endotoxin (5.0 mg) was fractionated by Sepharose 6B gel f i l t r a t i o n chromatography (gel height 1.6 x 35 cm) using 0.9% NaCl - 0.02% NaN3 - 15 mM Tris (pH 7.0) as eluant. 1.0 ml fractions were collected and analyzed for protein, KDO and phospholipid, the latter being present in a l l fractions but not shown for sake of c l a r i ty . Upper graph shows the separation profile of native (control) endotoxin. - 101 -51 merits, Cr labelled E. coli endotoxin was given intravenously at three doses (1.0, 3.0 and 6.0 mg/kg) to three separate groups of urethane-anaes-thetized guinea pigs. Blood samples were taken at hourly intervals and the plasma was assayed for acid phosphatase, a lysosomal enzyme whose act ivity in plasma has previously been shown to correlate well with endotoxin t o x i c i -51 ty (see Figures 5, 6 and 7). Three hours following Cr-endotoxin admini-stration, a blood sample was taken, the guinea pigs were sacrificed and portions (approximately 50 mg) of l iver , spleen, kidney, heart and lung tissue were removed, blotted and processed for sc int i l l a t ion counting (Methods 2.11.3). The experimentally-measured ^Cr-endotoxin binding (in ng/mg wet weight) for each tissue was standardized by expressing i t relative to the administered dose of endotoxin and was correlated with the plasma acid phosphatase act ivi ty (expressed as Sigma units/ml plasma) at 3 hr following toxin infusion. In the binding data depicted in Figure 22 for 51 three different organs, only Cr-endotoxin binding to the lung correlated posit ively with plasma acid phosphatase act ivity (r = 0.964). Two other 51 organs examined for Cr-endotoxin binding were l iver and spleen. The results (not depicted graphically here) showed a negative correlation between endotoxin binding to guinea pig l iver and plasma acid phosphatase act ivity which was very similar to that seen for the kidney in Figure 22. The spleen did not exhibit any obvious correlation between ^Cr-endotoxin binding and plasma acid phosphatase t i t e r s . Quantitatively, the l iver and spleen showed the greatest capacity for binding ^Cr-endotoxin (approximately 12-16 ng/mg tissue/dose) compared to 51 the other organs studied. This may reflect the uptake of Cr-endotoxin by macrophages which are abundant in both the l iver and the spleen. Lung - 102 -KIDNEY Ul 0) 0 5 • LUNG Q 3 z • . z E 5 HEART ACID PHOSPHATASE ( UNITS / ml ) Figure 22. Correlation Between Accumulation of 5lCr-Endotoxin in Various Guinea Pig Organs with Toxicity. The amount of Slo-endotoxin (ng/mg wet weight) in kidney, heart and lung tissues 3 hr following, the administration of either 1.0 mg/kg (n = 3), 3.0 mg/kg (n = 3) or 6.0 mg/kg (n = 2) 5 * C r - E . co l i endotoxin to guinea pigs was correlated with plasma acid phosphatase act iv i ty . The contribution of dosage on the tissue content of radiolabelled toxin was corrected for by dividing the amount of toxin bound in the tissues (ng/mg wet weight) by the administration dose of toxin in mg/kg. - 103 -51 tissue also contains macrophages and some of the apparent Cr-endotoxin "binding" may be due to endotoxin uptake by lung macrophages. However, our results strongly suggest that the interaction of endotoxin with lung tissue may be an important determinant of endotoxin toxic i ty in vivo. From this suggestion, one would predict that detoxified endotoxins may differ from native endotoxin in their ab i l i ty to bind to lung tissue. This prediction was tested and the results are shown in Figure 23. A decreased binding relative to native endotoxin was seen with both NaOH-detoxified (0.02 > P > 0.01) and NH20H-detoxified (0.01 > P > 0.005) toxins at 2 hr after these substances were injected intravenously into guinea pigs at a dose of 3.0 mg/kg. However, i t can also be seen from Figure 23 that E. col i endotoxin treated with sodium periodate (which is known to detoxify endotox-in as well) bound to lung tissue as effectively as native endotoxin. It seems, therefore, that binding to lung tissue per se is not sufficient for an endotoxin to have a lethal effect in vivo, but, in addition, the endotox-in must be capable of inducing some functionally relevant perturbation in the target tissue. Since periodate detoxification modifies the antigenic carbohydrate portion of endotoxin, this component of the molecule seems crucial in determining these deleterious consequences of endotoxin binding, including those involving the lung. While alterations in the ab i l i ty of endotoxin to interact with and perturb target tissues are undoubtably important in determining the altered in vivo toxic i ty of chemically-modified toxins, i t was also of interest to assess the effects of modification on the pharmacokinetic properties of these substances. Figure 24 shows the rate of disappearance of 51 Cr-labelled detoxified toxins from plasma following intravenous admini-- 1 0 4 -Ul D (J) (fl ID Z D J CD E Z X 0 a h 0 • z a UJ O) e -Ul > < 2 • Ul X 0 I-ui • I • a 2 1 ui 2 1 < j >• X 0 d • >• I • ui X • I-ui • • Ul X • DC K Ul u a • EIMDOTOXIIM Figure 2 3 . Accumulation of 5 1 Cr-Labelled Native and Detoxified Endotoxins in Guinea Pig Lung Tissue. The accumulation of native (toxic) and chemically detoxified 5 1 C r - l a -belled E. col i endotoxins in lung tissue (ng/mg wet weight ± S.E.M.) was compared in 4 groups of 5 guinea pigs at 2 hours following the administration of radiolabeled endotoxins ( 3 . 0 mg/kg). Significantly lower quantities of NaOH-detoxif ied toxin ( . 0 2 > P > . 0 1 ) and N H J J O H - detoxified toxin ( . 0 1 > P > . 0 0 5 ) accumulated in lung tissue in comparison to native endotoxin. - 105 -I 2 T I M E ( h r s ) Figure 24. Clearance of ^•'•Cr-Labelled Native and Detoxified Endotoxins from Guinea Pig Plasma. Comparison of plasma concentrations of native and chemically detoxified 51cr-E. coli endotoxins (yg/ml plasma ± S.E.M.) in guinea pigs at 1, 30, 60 and 120 min following an intravenous injection of the radiolabelled toxins (3.0 mg/kg). Data represent an average of 5 animals for each group except native toxin (serotype 668735) which consisted of 11 animals. - 106 -stration into anaesthetized guinea pigs. Blood samples were taken at 1.0, 20, 40, 60, and 120 min after injection and the plasma was prepared for sc int i l l a t ion counting. Based on the amount of endotoxin injected (3.0 mg/kg) and the estimate of total blood volumes of the guinea pigs used (25-30 ml), the concentration of endotoxin present in the plasma after complete mixing would be approximately 85 ug/ml. It is interesting to note from Figure 24 that the only endotoxins which approached this concentration at 1.0 min were the NaOH- and NH20H-detoxified endotoxins. The plasma concentrations of these endotoxins at 1.0 min are in sharp contrast to the concentration of the periodate-treated endotoxin (approximately 25 ug/ml) at this same time period. The 1.0 min concentrations of the native endotoxins f e l l somewhere in between the two extremes seen with the detoxified endotox-ins. It is clear from these preliminary experiments that chemical detoxif i-cation may exert different effects on the ab i l i ty of toxins to bind to target tissues, on their intr insic act ivity once bound and on their rate of removal from the c irculat ion. Thus, the biological act ivity of any particu-lar modified toxin is determined by a complex interplay between these va r i -ous properties. 3.5 Study of Drugs as Possible Endotoxin Antagonists 3.5.1 Antagonists to Endotoxin Binding In Vitro The objective of this series of experiments was to test the ab i l i ty of 51 various drugs to antagonize the binding of Cr-labelled E. co l i endotoxin to membranes using the human erythrocyte ghost as an in vitro model. The purpose of these experiments was to find an agent that would be potentially useful in antagonizing the effects of endotoxin in vivo at the cel lular membrane level . - 107 -Figure 25 reveals that lidocaine, a membrane active drug, which like endotoxin stabilizes human erythrocytes against hypotonic ly s i s , does not appreciably affect the binding of E. col i endotoxin to human red cel l ghosts even in concentrations as high as 10 mM. Similarly, methylprednisolone, which has occasionally been used in the management of certain types of shock (including gram-negative sepsis), did not antagonize the binding of the 51 Cr-endotoxin to the erythrocyte membrane preparation under the experi-mental conditions used (Figure 26). However, d,l-propranolol and its quaternary ammonium analogue, pranolium, effectively antagonized endotoxin binding almost to the same degree as unlabelled toxin. When the effect of 0, 0.5 and 2 mM concentrations of propranolol on the binding of E. coli endotoxin to erythrocyte membranes was subjected to double reciprocal plot analysis, propranolol appeared to be acting as a competitive antagonist. Figure 27 shows the effect of propranolol on endotoxin binding at toxin concentrations ranging from 5-50 ug/ml while Figure 28 shows a similar i n h i -bitory action of propranolol at endotoxin concentrations of 50 to 200 ug/ml (see Table 2). It can be seen that in both cases, d, 1-propranolol appears to act as a competitive antagonist of E. co l i endotoxin binding. Estimated K.j values for propranolol at these two classes of sites were approximately 0.4 x 10" 3 M (Figure 27) and 1.0 x 10~3 M (Figure 28). 3.5.2. Effectiveness of Endotoxin Antagonists In Vivo The ab i l i ty of several drugs, including those shown to antagonize the binding of endotoxin to membranes in v i t ro , to counteract the toxic effects of endotoxin in vivo was examined u t i l i z ing elevations in plasma acid phos-phatase levels as a measure of endotoxin toxic i ty in vivo. Figure 29 summarizes the results of these experiments which were performed using male - 108 -a.5 5 7.5 Id LIDO.(mlVl) BO IOO ISO SOOENDO.(ug/ml) C Q N C , Figure 25. Displacement of Bound col i Endotoxin from Human Erythrocyte Membranes with Lidocaine. A b i l i t y of lidocaine to displace 5 1 C r - E . co l i endotoxin from human erythrocyte ghosts is compared to the' displacement profi le obtained with cold (unlabelled) E. co l i endotoxin. Results are expressed as % of the amount of ^Cr-endotoxin bound in the absence of drug or unlabelled toxin (100% bound). Concentration of 5 1Cr-endotoxin was 25 pg/ml. Data repre-sent the mean ± S.D. of 3 separate experiments. - 109 -Figure 26. Effect of Methyl prednisolone, Pranolium and Propranolol on 51rjr-E. co l i Endotoxin Binding to Membranes. Effect of varying concentrations (0-2.0 mM) of methyl prednisolone, pranolium and d,l-propranolol on the displacement of 5*Cr-endotoxin bound to human erythrocyte membranes is compared to unlabelled E. col i endotoxin. Results are expressed as percent of the amount of ^^Cr-endotoxin bound in the absence of drug or unlabelled toxin (100% bound). Concentration of SlCr-endotoxin was 25 yg/ml.. Data represent the mean ± S.D. of 3 separate experiments (S.D. for drugs not shown for sake of c l a r i t y ) . - 110 -Figure 27. Double Reciprocal Plot of ^ C r - E . co-|j Endotoxin Binding to Erythrocyte Membranes in the Presence of Propranolol. The binding of S l p - E . col i endotoxin (5-50 ug/ml) to human red cel l membranes in the presence of 0, 0.5, and 2.0 mM d,l-propranolol are expressed as a double reciprocal plot. Data represent the mean of 3 separate experi-ments. - Ill -Figure 28. Double Reciprocal Plot of ^ C r - E . C0]-j Endotoxin Binding (High Concentra-tions) to Membranes in the Presence of Propranolol. Results of 51Cr-E. coli endotoxin (50-200 yg/ml) binding to human erythrocyte membranes in the presence of 0, 0.5 and 2.0 mM d,l-propranolol are presented as a double reciprocal plot. Data represent the mean of 3 experiments. - 112 -01 h 2 3 O DRUG 01 < I a in • I a a • < C P Z .25 mg/ kg • • _ PRANDLIUM ,S mg/kg HYDROCORTISONE 35 mg/kg I d.PROPRANOLOL.I mg/kg C O N T R O L T I M E ( h r s ) F i g u r e 29. E f f e c t o f V a r i o u s Drugs on Plasma A c i d Phosphatase A c t i v i t y i n E n d o t o x i n -T r e a t e d R a t s . Each of the i n d i c a t e d drugs was g i v e n as a s i n g l e i n t r a v e n o u s i n j e c t i o n t o s e p a r a t e groups of p e n t o b a r b i t a l a n a e s t h e t i z e d r a t s (n = 5 f o r each drug) 10 min b e f o r e E. c o l i e n d o t o x i n (10 mg/kg) was a d m i n i s t e r e d . Plasma from b l o o d samples (1.0 ml) o b t a i n e d at h o u r l y i n t e r v a l s from each r a t was assayed f o r a c i d phosphatase a c t i v i t y . C o n t r o l animals (n = 5) r e c e i v e d no drug or e n d o t o x i n whereas 0 d r u g - t r e a t e d r a t s (n = 10) r e c e i v e d o n l y endo-t o x i n (10 mg/kg). Data r e p r e s e n t mean ± S.E.M. In comparison t o 0 drug group, s i g n i f i c a n t l y lower a c i d phosphatase a c t i v i t i e s were seen w i t h : P r a n o l i u m .05 > P > .025 at 2 hr H y d r o c o r t i s o n e .05 > P > .025 at 5 hr d - p r o p r a n o l o l .025 > P at 2, 3, 4, and 5 h r . - 113 -Wistar rats. Drug-treated animals received a single intravenous injection of the drug, at the dose indicated 10 min before a bolus of 10 mg/kg E. co l i endotoxin was given. It is evident from Figure 29 that drugs such as chlor-promazine, pranolium and hydrocortisone were able to offer some degree of protection from the effects of the endotoxin, particularly during the later stages of endotoxaemia. Animals pretreated with d,l-propranolol (0.1 mg/kg) died shortly after receiving the bolus of endotoxin (data not shown). However, when rats were pretreated with the d-isomer of propranolol, which is devoid of appreciable beta blocking act iv i ty , a highly significant reduc-tion in the ab i l i ty of endotoxin to elevate plasma acid phosphatase levels was seen at 2 hr and for the duration of the experiment. The d-isomer of propranolol was indistinguishable from the racemate as an antagonist for endotoxin binding to red cel l membranes in vitro (data not shown). 51 The effects of various drug pretreatments on Cr-endotoxin binding to guinea pig lung tissue following in vivo administration of toxin were next examined. It was discovered that a parallelism existed between the efficacy of endotoxin antagonists in vitro and their ab i l i ty to decrease the binding of the toxin to lung tissue in vivo. For example, propranolol, the most effective in vitro endotoxin antagonist, s ignificantly (.005 > P > .002) reduced the binding of endotoxin to lungs in vivo and to a greater extent than did any of the other drugs shown in Figure 30. Experimental animals pretreated with pranolium at a dose equivalent to that used for d-proprano-lol (0.1 mg/kg), also exhibited a significant (0.05 > P > 0.02) decrease in 51 pulmonary endotoxin binding. The reduction in Cr-E. col i endotoxin content seen in the methylprednisol one-treated group of animals (Figure 30) did not achieve stat ist ical significance. - 114 -Ul D Hi ID 2 D j Ul E i a -B • _ e X • a • 2 Ul 0) c 2 -J • a h 2 • 1) UJ 2 • • (l) 2 • 111 a a J >• I i-UJ j • 2 < LT a Ul 2 iii • 2 Ul 0 < j • j • 2 < a a • a • 111 2 < (J • • j Figure 30. Effect of Drug Pretreatment on Accumulation of 5 1Cr-Endotoxin in Guinea Pig Lung. Methyl prednisolone (35 mg/kg), pranolium (0.1 mg/kg) and d-propranolol (0.1 mg/kg) were given as a single intravenous injection to urethane anaes-thetized guinea pigs 10 min before 5 1 C r - E . co l i endotoxin (3.0 mg/kg) was administered. Lidocaine (1 mg/kg/hr) and adenosine (50 mg/kg/hr) were given as a continuous intravenous drip. Data ± S.E.M. were obtained at 3 hr following endotoxin injection. In a l l cases, n = 5 except d-propranolol (n = 7) and controls (n = 11). Pranolium (.05 > P > .025) and d-propranolol (.005 > P > .002) were significantly different from controls. - 115 -The effects of a corticosteroid (hydrocortisone), pranolium, and d-propranolol pretreatment on E. co l i endotoxin lethal i ty when injected into albino Swiss mice were examined. The results of this mortality study are shown in Figure 31. It is interesting to note that al l three drug pretreat-ments offered protection against the lethal effects of endotoxin for up to 18 hr, at which time 80% of the untreated group of animals had died. The fact that each of these drugs offered a considerable amount of protection against the lethal effects of a large bolus of endotoxin for 18 hr although their administration was not continued throughout the experimental period was impressive and suggested that with optimal dosing regimens, these drugs might prove more effective for longer periods of time. A f inal series of experiments concerning possible in vivo antagonists to E. col i endotoxin was performed whereby the fea s ib i l i ty of using chemically detoxified endotoxin as antagonists was explored. To this end, rats were injected with either NaOH-detoxified endotoxin or periodate-detoxified endo-toxin prior to the administration of native E. co l i endotoxin (10 mg/kg) and effects on plasma acid phosphatase act ivi ty were examined (Figures 32 and 33). Some degree of protection was obtained with 2.5 mg/kg NaOH-detoxified endotoxin. However, increasing the dose of the detoxified endotoxin to 5.0 mg/kg produced no additional protection, whereas a further increase to 10 mg/kg abolished the protective effect seen at lower doses. In compari-son, periodate-detoxified endotoxin pretreatment also reduced plasma acid phosphatase levels and when the dose of this detoxified endotoxin was increased from 2.5 mg/kg to 5.0 mg/kg, a further reduction in plasma acid phosphatase levels was evident (Figure 33). Thus, using plasma acid phos-phatase act ivity as a measure of endotoxin tox ic i ty , periodate-detoxified - 116 -Figure 31. Effect of Drug Treatment on Mortality Rates in Mice Injected with Endotoxin. Groups of Swiss mice received intraperitoneal injections of pranolium (0.5 mg/kg; n = 10), d-propranolol (0.5 mg/kg; n = 40) or hydrocortisone (35 mg/kg; n = 10) 30 min prior to and at 12 hr after an injection of E. col i endotoxin (40 mg/kg). No drug (n = 50) represents animal s- injected only with endotoxin (40 mg/kg). - 117 -h 2 ui U) < < I a tn • I a r / / I: ii : // / // / a • NaOH DETOXIFIED I O.O mg/kg NO TREATMENT NaOH DETOXIFIED 2 . 5 mg/kg 5 . 0 mg/ kg • U CONTROL TIME ( h r s ) Figure 32. NaOH-Detoxified Endotoxin as an Endotoxin Antagonist in vivo. The effect of pretreating anaesthetized rats with various doses of NaOH-detoxified endotoxin 30 min prior to an injection of native (toxic) E. col i endotoxin (10 mg/kg) on toxicity as determined by plasma acid phospha-tase act ivity (± S.E.M.) is indicated. N = 5 for al l groups except no treatment (native endotoxin only), n = 10. Controls received neither native nor detoxified toxin. Detoxified endotoxin alone had minimal effects on plasma acid phosphatase act ivity (see Figure 5). - 118 -s • h 2 NO T R E A T M E N T Ul U) < < I a • I 2 a • ./I 1 5 3 4 T I M E ( h r s ) P E R I O D A T E D E T O X I F I E D 2 . 5 m g / k g P E R I O D A T E D E T O X I F I E D 5 . O m g / k g C O N T R O L Figure 33. Periodate-Detoxified Endotoxin as an Endotoxin Antagonist in vivo. The effect of pretreating anaesthetized rats with periodate-detoxified endotoxin 30 min prior to an injection of native (toxic) E. co l i endotoxin (10 mg/kg) on toxicity as determined by plasma acid phosphatase act ivity (± S.E.M.) is indicated. N = 5 for all groups except no treatment (native endotoxin only), n = 10. Controls received neither native nor detoxified toxin. Pretreatment with 5.0 mg/kg periodate-detoxified toxin resulted in significantly (.01 > P) lower plasma acid phosphatase act ivi ty at 2 to 5 hr than no treatment (native endotoxin alone). - 119 -endotoxin was a better in vivo antagonist of native endotoxin than was NaOH-detoxified endotoxin. 3.6 Effect of Endotoxin and Gentamycin on Endotoxin Toxicity In Vivo The therapeutic management of gram-negative sepsis c l i n i c a l l y almost invariably involves the use of aminoglycoside antibiot ics . It was therefore of interest to study the possible effect of aminoglycosides on the in vivo toxic i ty of E. co l i endotoxin. Gentamycin was selected as a representative aminoglycoside antibiotic on the basis of i ts frequent use in gram-negative septicaemia. Again, plasma acid phosphatase levels were used as a measure of endotoxin toxicity. The results of these experiments are summarized in Figure 34. It is evident that, when administered acutely at a c l i n i c a l l y relevant dose (1 mg/kg, intravenously), gentamycin did not signficantly modify the effect of endotoxin on plasma enzyme act iv i ty . However, when this same dose of gentamycin was administered intraperitoneally to the animals daily for three days prior to endotoxin challenge (10 mg/kg, intra-venously), a highly significant (.002 > P > .001) elevation in plasma acid phosphatase levels relative to rats receiving endotoxin but no prior amino-glycoside therapy was seen. An apparent synergistic effect with endotoxin could also be. demonstrated when a single larger dose of gentamycin (10 mg/kg) was administered (Figure 34). It should be noted that gentamycin alone had no effect on plasma acid phosphatase act iv i ty . To further inves-tigate the influence of gentamycin on endotoxin toxic i ty , mortality studies were carried out. Swiss albino mice were pretreated with gentamycin (1 mg/kg) for 3 days after which various doses of E. col i endotoxin were administered intraperitoneally and mortalities assessed 12 and 24 hours later (Figure 35). By 12 hr, the gentamycin-pretreated mice displayed a - 120 -Ul U) I a ai • I a £ cn h 2 3 4 --I • a i-2 • U h 2 CD H 2 Ul 0 01 JC E h 2 ui 13 > TJ \ 01 JC 0) E i-2 Ul ID 01 OI E • • E N D O T O X I N IO mg/kg Figure 34. Effect of Gentamycin in Combination with Endotoxin on Toxicity in Rats. Rats were pretreated with gentamycin either acutely (1 mg/kg or 10 mg/kg intravenously, 15 min before E. co l i endotoxin 10 mg/kg) or chronically (1 mg/kg/day for 3 days, given as twice daily intraperitoneal injections of 0.5 mg/kg). E. co l i endotoxin (10 mg/kg) was injected 30 min after the f inal gentamycin injection in the chronically treated animals. The results are expressed as mean ± S.E.M. acid phosphatase act ivity at 3 hr following endo-toxin administration. Controls received no endotoxin or gentamycin whereas 0 gentamycin represents animals that received only endotoxin. N = 10 for a l l groups. Acid phosphatase in chronically (1.0 mg/kg/day) and acutely (10 mg/kg) treated rats was s ignif icantly (.002 > P > .001) elevated over 0 gentamycin group. - 121 -[ | ENDOTOXIN • 0 5 .1 .2 «4 . 6 mg E N D O T O X I N /l5g ' Figure 35. Effect of Chronic Gentamycin Treatment on Mortality in Endotoxin-Treated Mice. Swiss mice were pretreated with gentamycin (1.0 mg/kg/day given as twice daily injections of 0.5 mg/kg) for 3 days after which various doses of E. coli endotoxin were administered intraperitoneally and mortalities were assessed 12 and 24 hr later . Mice receiving only endotoxin were pretreated twice daily with saline. N = 10 for each dose of endotoxin in both groups (total of 50 mice/treatment group). - 1 2 2 -higher mortality at al l doses of endotoxin. At 24 hr, the increase in mortality in the gentamycin-treated animals was s t i l l obvious but only at the lower doses of endotoxin. Thus, these mortality studies substantiated the previous experiments suggesting a synergistic action of gentamycin on endotoxin toxic i ty in vivo. 3.7 Var iab i l i ty in Commercially Available Endotoxin Preparations Investigators studying experimental endotoxaemia are well aware of the variations in results that can be obtained with endotoxins of different serotypes. During the course of the present study, we have obtained evidence indicating that marked differences in biological act ivi ty exist even among commercially obtained E. coli endotoxin preparations differing only in lot number. Attempts to chemically characterize these endotoxins of varying biological potency have yielded a procedure which may be useful in standard-izing these preparations. The method involves the use of the primary amino group-modifying chromophoric probe, trinitrobefeonesulfonic acid (TNBS), whose binding to endotoxin can be measured spectrophotometrically at 335 nanometers. It was found that endotoxin preparations incorporating TNBS to a greater extent in v i t ro , were generally found to be more toxic in vivo than endotoxins which incorporated less TNBS. Figure 36 i l lustrates this point by showing the correlation (r = 0.95) between the extent of TNBS incorporation per mg endotoxin and the toxicity in vivo, expressed in terms of units plasma acid phosphatase act ivi ty , for five E. co l i endotoxins (serotype 026:B6) with the indicated lot numbers. - 123 -Figure 36. Correlation Between Extent of TNBS Incorporation into E. co l i Endotoxin and Toxicity in Rats. The incorporation of trinitrobenzenesulfonic acid (TNBS) into E. coli endotoxin (026:B6) of various lot numbers was determined by incubating 1.0 mg toxin in a 3.0 ml reaction mixture consisting of 1.0 ml 20 mM Tris-HCl pH 8.0, 1.9 ml H20 and 0.1 ml 10 mM TNBS, pH 8.0, for 35 min at 3 7 ° C . The v reaction was init iated by the addition of TNBS and terminated by the addi-tion of 2.0 ml of a 1:1 mixture of 10% SDS and 1 M HCl. The absorbance was read at 335 nm. Toxicity in rats was determined by measuring plasma acid phosphatase act ivi ty 3 hr following an injection of endotoxin (10 mg/kg). Data represent mean ± S.E.M. of 10 animals for each lot number. - 124 -CHAPTER 4 Discussion and Conclusions 4.1 Gram-Negative Septicaemia: A Formidable Medical Problem Recently, the results from a ten year study on gram-negative bacteraemia involving approximately 600 patients have been reported by Kreger and assoc-iates (199,200). One interesting observation made by these investigators was that in spite of the development of potent antimicrobial agents, the incidence of gram-negative bacteraemia has been continually increasing since the 1950's and, as a result , constitutes the majority of cases of bacterae-mia presently seen in the United States (and probably Canada as well) . The high f a ta l i ty rate associated with gram-negative bacteraemia has made i t one of the major causes of death from infection in North American hospitals (201). The increasing frequency of gram-negative bacteraemia appears to be related to a growing proportion of patients that are more aged and have an underlying pathological condition such as granulocytopenia, congestive heart fa i lure , diabetes mellitus, neoplasms or renal insufficiency. Also, the increasing use of manipulative procedures involving the urinary or respira-tory tracts and extensive treatments with antibiot ics , corticosteroids or antimetabolites contribute s ignif icantly to the r i s ing frequency of gram-negative bacteraemia (199,200). In their study, Kreger and co-workers found that the bacteraemias were most frequently caused by Escherichia c o l i . Other causative organisms, found in decreasing order of frequency, were Klebsiella-Enterobacter-Serratia, Pseudomonas, Proteus and Bacteroides. Only 16% of the bacteraemias were polymicrobic. When the source of the bacteraemia could be determined, i t was most frequently found to be the - 125 -urinary tract , followed by the gastrointestinal and respiratory tracts. However, in almost one third (30%) of the patients with underlying host disease, the site of origin of the bacteraemia could not be identified (199). It is well recognized that shock is a common complication of gram-nega-tive bacteraemia (147). In the study reported by Kreger and associates, shock occurred in 441 of the gram-negative septic patients and in these patients, the fa ta l i ty rate was seven times greater than in septic patients who did not develop shock (200). It has been proposed that the cause of shock commonly seen in gram-negative septicaemia is due to the release of endotoxins (168). This postulation was made by Weil and co-workers even before assay techniques were developed to detect the presence of endotoxins in the plasma. Certainly i t was known that most of the c l in ica l manifesta-tions of gram-negative bacteraemia could be reproduced in animals by admini-stering endotoxin isolated from gram-negative b a c i l l i (167). Also, experi-ments involving the parenteral administration of endotoxin into human volun-teers have been conducted and responses similar to those seen c l i n i c a l l y in gram-negative bacteraemia were noted (202). More credence was given to the belief that endotoxins were involved in the sequelae of c l in i ca l gram-nega-tive bacteraemia when Crutchley and Jorgensen demonstrated that viable gram-negative organisms grown in cultures could readily release endotoxin into the surrounding medium (16,19). However, firm proof that endotoxins were released from b a c i l l i in gram-negative septicaemia was not obtained until the Limulus lysate technique was developed, whereby the presence of endotoxin in the plasma could be detected in concentrations as low as 0.1 picograms/ml plasma (203). Indeed, i t has been shown that c l in i ca l endotox-aemia can occur in the absence of a well defined septic focus, particularly - 126 -when the reticuloendothelial system is impaired to the extent that endotox-i n , absorbed from the gastrointestinal tract , can accumulate in the plasma (203,204). Although endotoxins most commonly originate from gram-negative b a c i l l i , some strains of gram-negative cocci , such as Neisseria meningiti- dis, can also liberate endotoxins in amounts which vary from strain to strain (205). Thus, i t has been suggested that this variable release of endotoxin according to strain may account for the different c l in ica l presen-tations of meningococcal infections (205). Therefore, while some investiga-tors may s t i l l question the contention that endotoxin plays a major role in bacteraemic shock in man (206,207), i t is generally accepted by most inves-tigators in this area that these bacterial ce l l wall constituents are largely responsible for the morbidity associated with gram-negative infec-tions (89). As a result , most of the experimental work on various aspects of gram-negative bacteraemia (particularly shock) have ut i l ized purified endotoxin preparations. 4.2 Host Defense Mechanisms in Bacteraemia and Endotoxaemia Certain differences are apparent in the host reaction to injections of" endotoxin as compared to infusions of live bacteria (207,208). One major difference, in particular, is the manner by which the host system strives to eliminate endotoxins and whole bacteria from the c i rculat ion. It seems that bacteria are phagocytized primarily by polymorphonuclear neutrophil leuko-cytes while endotoxin is mainly detoxified by the cel l s of the reticuloendo-thel ia l system and blood monocytes (207,209). As mentioned in the Intro-duction section of this thesis, the importance of granulocytes in determin-ing the outcome of a gram-negative septic episode c l i n i c a l l y cannot be overstated (151,155). Certainly, the prognosis for granulocytopenic - 127 -patients that develop gram-negative bacteraemia can be greatly improved with granulocyte transfusions (210,211). However, i t has also been demonstrated by Helium and Solberg that the bactericidal act ivity of neutrophil granulo-cytes, as measured by the reduction of nitroblue tetrazolium (NBT) dye, can be greatly inhibited in patients with severe bacterial infections (212,213). Secondly, the studies of Cartwright, Galbraith, and co-workers have shown that the ha l f - l i f e of a mature polymorphonuclear neutrophil leukocyte in the circulation is only approximately six or seven hours (214,215). Thus, i t seems that although polymorphonuclear leukocytes are the most immediate phagocytizing cel ls during an infection, they are short-lived and hence the function of macrophages may serve as an important defense against invading organisms (216). Furthermore, when consideration is made of the evidence that gram-negative infections can be complicated by endotoxaemia and, since endotoxins are detoxified primarily by reticuloendothelial cel ls (209), the role of these ce l l s , or macrophages, in the defense of gram-negative infec-tions gains even more significance. Therefore, the successful recovery from a gram-negative bacteraemia depends on the phagocytic functions of c i rculat-ing granulocytes, such as the polymorphonuclear neutrophils, and macropha-ges, which have the added capability of detoxifying endotoxin. Although the macrophages form an extensive network of phagocytic cells throughout the body, col lect ive ly known as the reticuloendothelial system, the Kupffer cel l s of the l iver are the most important in clearing c i rculat-ing endotoxin from, the blood in a l l experimental animals studied (217). It has been demonstrated that l iver poisons, such as carbon tetrachloride, and beryllium phosphate, which is toxic to Kupffer ce l l s , can make animals treated with these hepatotoxins hypersusceptible to the lethal effects of endotoxin (218). Indeed, i t was recognized as early as 1947 by Beeson, that - 128 -blocking the phagocytic act ivity of the reticuloendothelial system with colloidal thorium dioxide (thorotrast) made animals more susceptible to the effects of injected endotoxin (219). Another agent which can impair hepatic phagocytic function is lead acetate, and this compound has also been shown to sensitize animals such as the subhuman primate to the lethal effects of endotoxin (220). Alternatively, one would expect that substances which stimulate phagocytic act ivity of the reticuloendothelial system would make animals resistant to the effects of endotoxin. However, such agents have paradoxically been shown to render animals more sensitive to endotoxin. For example, BC6 and glucan, which is the purified polysaccharide component of the yeast ce l l wall extract zymosan, have both been shown to sensitize animals to the lethal effects of endotoxin (221,222,223). This is rather surprising in view of the fact that these agents, glucan in particular, can enhance non-specific host resistance to a variety of diseases including certain bacterial infections (224) and malignant tumors (225). A possible explanation for this paradoxical hypersensitive response to endotoxin seen in glucan- and BCG-treated animals may be related to the observation that these substances exacerbate the hypoglycaemia normally seen in endotoxic shock (217,226,227). It has long been recognized that resistance to the effects of endotoxin (or "tolerance") can be produced in animals by the daily administration of sublethal doses of endotoxin for several days. Animals made "tolerant" to endotoxin displayed increased phagocytic act ivity of the reticuloendothelial cel ls (217). Thus, i t was believed that the development of "tolerance" to endotoxin was due to enhanced phagocytic act ivi ty of the reticuloendothelial macrophages. However, Starzecki and associates found that the clearance of 51 Cr-endotoxin from the circulation was the same in normal and endotoxin-- 129 -resistant dogs (228). Furthermore, Greisman and co-workers demonstrated that blockade of the reticuloendothelial phagocytic act ivi ty with thorotrast did not appreciably affect tolerance to the pyrogenic effect of endotoxin in endotoxin-resistant rabbits (229). These investigators, in fact, found that normal, thorotrast-treated rabbits could-be made tolerant to endotoxin when these animals were transfused with plasma from endotoxin-resistant rabbits. These results suggested the presence of a humoral factor that was responsi-ble for the development of tolerance to endotoxin in animals. Although s t i l l perhaps controversial, there is good evidence that this humoral resis-tance factor represents antibodies against endotoxin (230). Certainly, a 19S immunoglobulin specific for the "core" or glycolipid region of the endotoxin has been demonstrated to be involved in the transference of the endotoxin-tolerance t ra i t (231,232). Freedman has shown that antibodies formed against the glycolipid portion of the endotoxin complex can provide passive, transferable protection against homologous and heterologous endo-toxins (233). This is in contrast to antibodies directed against the 0-antigenic polysaccharide region of the endotoxin which only provides homo-logous protection, or, in other words, protection against endotoxins of the same serotype (232). Indeed, the c l in ica l value of having high t i ters of anti-endotoxin immunoglobulins to reduce the high frequency of shock and death normally associated with gram-negative septicaemia is becoming more strongly appreciated by investigators (234,235). The mechanism by which endotoxin immunoglobulins, particularly those directed against the core regions of the complex, decrease the toxic i ty of endotoxin is believed to be due to a process of opsonization as well as to an antitoxic effect (234). F ina l ly , other studies have indicated that endotoxin can be inactivated or "detoxified" in the plasma by nonimmunoglobulin factors that are largely - 130 -unidentified at the present (236). Skarnes has proposed that endotoxin can be detoxified in the plasma by two enzymes which are both a-globulins (237,238). One enzyme, a heat-stable esterase, causes disaggregation of the endotoxin complex while another, a heat-labile enzyme, detoxifies the disag-gregated endotoxin (237,238). Other investigators have found only one protein, an a-globulin, in human serum that could cause irreversible disag-gregation of the endotoxin complex (239). This protein, unlike the ct-globu-1 ins reported by Skarnes and co-workers, was neither a lipoprotein nor an esterase (239). Studies by Ulevitch and associates have also indicated that endotoxin is detoxified in the plasma by what is l ikely a disaggregation process which does not seem to be enzymatic in nature but does depend upon plasma l ipids (high density lipoproteins in particular) to occur (240,241). Thus, although there is a consensus that the humoral phase of the blood is also capable of detoxifying endotoxin, the exact nature of this detoxifica-tion process remains to be more ful ly elucidated. Therefore, the host defense mechanisms against gram-negative bacteraemia and .associated endotoxaemia principally involve a cel lular phagocytic system and a humoral defense reaction. The cel lular phagocytes are comprised of circulating polymorphonuclear granulocytes and the macrophages of the reticuloendothelial system. The humoral defense response consists of immunoglobulins and other factors, which may involve enzymes or lipopro-teins, that detoxify endotoxin. Normally, these protective mechanisms are sufficient to prevent the progression of a gram-negative bacteraemia to c r i t i c a l stages exemplified by hypotension and shock. However, in circum-stances where the function of these defense mechanisms may be compromised, such as in very aged patients or patients that are granulocytopenic or are immunosuppressed because of cancer chemotherapy, radiation therapy or - 131 -patients with l iver disease, the development of gram-negative septicaemia could have very serious consequences (234). In circumstances such as these, therefore, specific and effective methods of counteracting the deleterious effects of endotoxin would be required. To determine what treatment measures would be the most appropriate and effective to undertake, i t is necessary to understand the various effects that endotoxins exert in the body that can lead to the development of shock. 4.3 Pathophysiology of Endotoxaemia Shock is a term frequently used c l i n i c a l l y to describe a syndrome consisting of protracted hypotension, pal lor , a cold, moist skin, mental confusion and ol iguria (242). At the turn of this century, the severity of shock was determined by the nature of the pulse. Before World War I, popular use of the sphygmomanometer allowed blood pressure measurements to be a monitor in shock. Now, with refined instrumentation to determine cardiac output and intravascular pressures, i t is apparent that heart rate, blood pressure and cardiac output are only secondary indicators of shock. Rather, the primary defect in shock is a reduction in effective or microcir-culatory blood flow which in turn impairs the transcapillary exchange of essential substrates and oxygen (242). It is an undisputed fact that endotoxin is a potent shock-inducing agent. However, the mechanism by which endotoxin produces shock is s t i l l obscure. The effects of endotoxaemia on the microcirculation have received limited study and considerable controversy exists with regard to the actions of endotoxins on this part of the circulatory system (244). One convenient method of studying the effects of endotoxin on microcirculatory or "nutr i -133 t ive" flow in vivo is to monitor the washout of Xenon from an injection 133 s ite in some particular tissue. Xenon, being an inert and l ipophil ic - 132 -substance, readily diffuses across cel l membranes and therefore, i ts rate of removal from an injection site closely coincides with the degree of c a p i l l -ary blood flow to that area (243). Using this technique, we have found that E. c o l i endotoxin caused a marked reduction in skeletal muscle capi l lary blood flow ten minutes after i t was given intravenously (4.0 mg/kg, L-D.^Q) to rats. Interestingly, this effect on microcirculatory flow persisted for the duration of the experimental time (3 hrs) despite the fact that a substantial recovery in the mean blood pressure had occurred during this same time period (80% of normal 3 hrs after endotoxin), which suggested that the reduced capil lary flow was not simply due to low blood pressure. This observation that microcirculatory flow can be greatly impaired in spite of a reasonable blood pressure supports observations by other investigators that intra-arterial pressure is an unreliable indicator of the severity of shock in bacteraemia (245). The mechanism by which endotoxin impairs micro-circulatory flow does not seem to be due to a direct effect on the microvas-culature but rather to the release of vasoactive substances (147). For example, i t is known that endotoxins increase plasma levels of catechola-mines which are believed to be responsible for the constriction in both precapillary arterial sphincters and postcapillary venular sphincters, caus-ing what has been termed "ischaemic anoxia" (246). Interestingly, i t has been proposed that ischaemic anoxia, which represents the reversible phase of shock, is a manifestation of not only endotoxic, but of all types of shock (246). The irreversible phase of shock has been described as "stagnant anoxia" and is characterized by an increased capi l lary hydrostatic pressure as a result of the relaxation of the precapillary arteriolar sphincters only (246). Irreversible shock can be produced experimentally in a f a i r ly short period of time by haemorrhaging animals to a low blood - 133 -pressure (one-third of normal) and maintaining the animals at this low mean blood pressure by additional bleedings or reinfusions until such time as 30% of the maximal bled volume has been returned to the animals. At this point, irreversible shock ensues, even when the remaining volume of haemorrhaged blood is transfused back to the animals (247). Under these conditions, the microcirculatory blood flow in the skeletal muscle of the rat , as measured 133 by Xenon washout, was found to be approximately two-fold greater than the flow at three hours following endotoxin administration, even though the mean blood pressures were similar in both groups of rats. This implies a greater degree of vasoconstriction exists in- the microcirculation of skele-tal muscle after three hours of endotoxaemia than is seen when irreversible shock is induced by haemorrhage. Alternatively, i t can also be proposed that the reduced microcirculatory flow seen in the endotox in-treated animals could be due to clogging of the capi l lar ies as a result of an increase in blood viscosity and microcoagulation (248,249). However, Weidner and co-workers have shown that the dog forelimb lost weight ten to fifteen minutes after the animal was injected with endotoxin and this observation, along with the noted increase in skeletal muscle vascular resistance, was explained by the authors as evidence for vasoconstriction occurring in the skeletal muscle vasculature (250). As already mentioned, catecholamines have been proposed to be responsi-ble for the reduction in microvascular flow seen in endotoxaemia (246). The plasma levels of adrenaline and noradrenaline can increase from ten- to thirty-fold after the administration of endotoxin (251). In addition to the catecholamines, other vasoconstrictor agents are known to be released into the circulation in response to endotoxin administration. Examples of some of these substances include renin/angiotensin (252,253), 5-hydroxytryptamine - 134 -(254) and prostaglandins, such as F2 a , which have vasoconstrictive actions (255,256). However, i t is also well established that the administration of endotoxin to animals can increase the concentration of substances in the plasma that have potent vasodilatory actions, such as histamine (257,258), bradykinin (259,260) and prostaglandins of the E series (255,256). There-fore, i t is d i f f i cu l t to attribute the effects of endotoxin on peripheral microcirculatory flow solely to the release of vasoactive agents, although these mediators may exert important effects locally in certain organs such as the lung (261). Probably much of the peripheral vasoconstriction seen in endotoxaemia is due to baroreceptor/sympathetic nerve stimulation that occurs in response to the systemic hypotension caused by endotoxin, in order that adequate blood flow is maintained to more v i ta l organs. One may then question the relevance of impaired microcircul atory flow in skeletal muscle to the pathophysiology of endotoxaemia. It can be stated, however, that the haemodynamics of the circulation in skeletal muscle are not responsible for the systemic hypotension that occurs upon the administration of endotoxin to animals. Therefore, to address the original question regarding the precipitation of endotoxic shock, the examination of the microcirculatory status in vital organs such as the heart, lung, kidney, e tc . , would probably provide a good estimate of the severity of shock in endotoxaemia. A technique whereby the degree of shock could be effectively measured would have great potential for c l in i ca l application in gram-negative bacteraemias. However, measurement of microcirculatory blood flow in v i ta l organs is d i f f icu l t to do experimental-ly and not feasible c l i n i c a l l y . To circumvent this technical d i f f i cu l ty of directly assessing the perfusion of organs in shock, i t is possible to do so indirect ly by assaying the plasma for certain substances that are released - 135 -from tissues only during states of compromised perfusion. Lactic acid, a product of anaerobic glycolysis , is an example of one such substance that accumulates in the blood during conditions of systemic oxygen def ic i t (243,245). Weil and A f i f i have shown that the concentration of lactate in arterial blood is a good indicator of the severity of shock induced by haemorrhage (262). Indeed, blood lactate levels are commonly used to assess the extent of circulatory failure in patients at the present time. Other substances which may be used to estimate the severity of shock are certain intracel lular enzymes such as the lysosomal hydrolases (263,264). The release of these enzymes into the plasma occurs only under conditions associated with extensive tissue damage and cel lular death (265). Although plasma levels of lactate or lysosomal enzymes can be used as indicators of the severity of shock, they cannot be used to identify the organs in which the circulation is primarily impaired because lactate and lysosomal enzymes are f a i r l y ubiquitous substances. There are some non-lysosomal enzymes however, which are peculiar to cel ls of certain organs and therefore, the presence of these enzymes in the plasma would identify the particular tissue in which cel lular lysis was occurring. Some examples of these enzymes which have been used to determine circulatory impairment in various organs include: creatine phosphokinase (CPK) which is specific for muscle primar-i l y and nervous tissue (266); ornithine carbamyltransferase which is present in l iver cel ls (267); and alkaline phophatase of intestinal origin which is identified on the basis that this enzyme is speci f ical ly inhibited by L-phenylalanine (268). Of a l l these substances that have been used to determine the severity of shock, i t appears that the plasma concentrations of lysosomal enzymes (cathepsin D in part icular) , provide the best measure of the seriousness of the shock state (269). Although lysosomes are present - 136 -in most types of ce l l s , these organelles are concentrated to the greatest extent in the splanchnic organs, l iver and macrophages (270,271). Therefore the presence of lysosomal enzymes in the plasma during shock states usually is an indication of circulatory impairment in the hepato-splanchnic region. Indeed, the circulation in this region can be deranged to such an extent that, in haemorrhagic shock for instance, the intestinal mucosal barrier breaks down (272), permitting the systemic absorption of endotoxins from the gut to occur (273,274). The work reported in this thesis demonstrated the value of using plasma lysosomal enzyme activity as a measure of the physiological deterioration that occurs progressively in animals from the moment that they are injected with a lethal dose of endotoxin to their time of death. The three represen-tative lysosomal enzymes that were employed in these studies included a phosphatase (acid phosphatase), a glucosidase (N-acetyl-e-glucosaminidase) and a protease (cathepsin D). Al l three enzymes appeared to be rel iable indicators of the toxic effects of endotoxin in vivo as demonstrated by the manner in which the plasma act iv i t ies of these enzymes varied with dose of endotoxin and with time after a lethal dose of endotoxin was administered. Endotoxins detoxified chemically by sodium hydroxide or sodium periodate treatment had v ir tual ly no effect on plasma lysosomal enzyme act ivity during the time period in which an equivalent dose of native endotoxin was lethal to the animals. It was interesting to note that although the plasma ac t iv i -ties of the three lysosomal enzymes under study increased with time and with dosage of endotoxin, the patterns of their rise in act ivity were different. These results then suggest that the lysosomal enzymes in question may be originating from different tissues upon which the effects of endotoxin administration are not the same. It is known that tissues do differ in - 137 -their lysosomal enzyme content. For example, muscle lysosomes predominantly contain cathepsin D and ribonuclease (270), whereas i t has been reported that the intestine is a major source of acid phosphatase (275). Secondly, the lymphatic drainage from various organs may be different and since this is the major route by which lysosomal enzymes enter the systemic circulation (275,276), i t is another factor that can influence the concentration of these enzymes in the plasma. Another possible explanation for the variation in the plasma concentrations of the three enzymes in endotoxaemia is that the clearance mechanisms for these enzymes may di f fer . For example, i t is known that cathepsin D is largely removed from the circulation by the reticuloendothelial system (277) while glucosidases such as N-acetyl -e-gluc-osaminidase are cleared primarily by endothelial cel l s l ining the hepatic sinusoids by a process involving a cel l surface receptor that specif ical ly "recognizes" this glycoprotein (278). It also is believed that the function of the reticuloendothelial system can become impaired in shock states (263,269,279,280) and therefore, a lysosomal enzyme such as cathepsin D could accumulate in the plasma to a greater extent than perhaps N-acetyl -e-glucosaminidase whose plasma clearance does not to ta l ly depend upon re t i cu l -oendothelial function. This may explain our observations in E. col i endotoxin-treated guinea pigs that plasma concentrations of cathepsin D were increased several-fold at a time when N-acetyl-e-glucosaminidase levels had not quite doubled. Thus, on the basis of our findings concerning the effect of E. col i endotoxin on plasma lysosomal enzyme act ivity in rats and guinea pigs, i t can be said that both cathepsin D and acid phosphatase seem to be better indicators of the state of endotoxic shock in these animals than N-acetyl-e-glucosaminidase. - 138 -It was most interesting to compare the lysosomal enzyme data obtained in the laboratory setting u t i l i z ing animals and purified E. coli lipopolysac-charide with the data that were obtained on the same enzymes from patients in gram-negative bacteraemic shock. As in the animal experiments, the act ivit ies of the three lysosomal enzymes studied in the plasma of these patients were significantly elevated over the values normally seen in plasma from healthy human volunteers. Again, as in the animal experiments, there were differences in the extent to which each of these enzymes was elevated in individual patients. On the average, acid phosphatase was elevated four-fold, N-acetyl-B-glucosaminidase also increased four-fold while the average plasma cathepsin D concentration in the gram-negative shock patients was approximately twenty-six times normal. These results may indicate that patients in shock with gram-negative bacteraemia perhaps sustain a greater degree of microcirculatory impairment in peripheral tissue ( i . e . skeletal muscle) than in the hepatosplanchnic region, since the increase in cathepsin D act iv i ty in the plasma of these patients was proportionally greater than were the increases in acid phosphatase or N-acetyl-B-glucosaminidase ac t iv i -t ies . Supporting this explanation are experiments that have examined the effects of endotoxic shock on the circulatory status of sub-human primates. These experiments have demonstrated that the administration of endotoxin to monkeys and baboons does not alter the circulation to the intestine as much as i t does in the dog, for example, an animal in which endotoxin is known to cause splanchnic pooling (281,282). Another factor which probably contrib-utes to the large increases in plasma cathepsin D act iv i ty in gram-negative septic shock patients is the likelihood that the function of the reticuloen-dothelial system is depressed in these patients, many of whom had a severe disease underlying the gram-negative bacteraemia. Thusl both animal and - 139 -patient studies have indicated that plasma lysosomal enzyme activity can be a rel iable measure of the pathophysiological state of either the animals treated with purified endotoxin or patients with gram-negative bacteraemia. In particular, plasma cathepsin D concentration showed the greatest increase in gram-negative septic shock patients while both cathepsin D and acid phosphatase act ivi t ies were greatly elevated in the plasma of endotoxin-treated rats and guinea pigs. From our observations on the effect of endotoxaemia or gram-negative septicaemia on plasma lysosomal enzyme act ivity in animals and patients, the question arises as to whether lysosomal enzymes in the plasma are simply a manifestation of cel lular damage or whether these "free" enzymes contribute in some way to exacerbate shock in experimental endotoxaemia or c l in i ca l gram-negative septicaemia. In other words, does the presence of these enzymes in the plasma signify that a potentially dangerous positive-feedback effect could develop whereby the lysosomal enzymes themselves cause additional tissue and organ damage and subsequently the release of more enzymes ? Although some investigators may have reservations as to whether free lysosomal enzymes in the plasma intensify the pathophysiology of shock (283), much of the evidence tends to support the concept that these enzymes do play an active role in the development of irreversible shock (263,284). There is certainly no question about the fact that lysosomes contain enzymes which can digest a wide variety of substances such as proteins, l i p id s , carbohydrates and nucleic acids. Thus, they have the potential of inducing considerable host damage under favorable conditions such as those encoun-tered in shock where the accumulation of lact ic acid lowers the pH to a level nearer the optimum for lysosomal enzymes (285). It has been demon-strated that infusion of fractions rich in lysosomal enzymes into dogs can - 140 -produce a state of circulatory shock characterized by hypotension, splanch-nic vasoconstriction and haemorrhagic lesions of the bowel (277,286). The effects of these lysosomal enzyme infusions were more pronounced in animals whose reticuloendothelial function was surgically impeded (277). Infusion of lysosomal products has also been shown to affect the microcirculation in rabbit or rat mesentery (287) and can cause endothelial proliferation in isolated perfused cat hearts (272). Some investigators also believe that lysosomal proteases can indirectly exacerbate the f a i l ing circulation in shock by causing the formation of a small molecular weight peptide (800-1000 M.W.) that directly depresses myocardial contract i l i ty and hence referred to as a "myocardial depressant factor" (288,289). The source of the myocardial depressant factor appears to be the pancreas and its formation is consequen-t ia l to pancreatic ischaemia (289). However, other investigators have not been able to detect the presence of any factors in the plasma of endotoxin shocked dogs that depressed myocardial contract i l i ty (290). F ina l ly , lysosomal enzymes may intensify the shock condition by disrupting mitochon-drial function and impairing energy production. In this regard, Mela and associates have shown that most of the changes in mitochondrial function that occur in endotoxic as well as haemorrhagic shock can be reproduced in vitro by incubating lysosomal hydrolases with isolated mitochondria (291). Thus, the general progression of events in endotoxic shock may be presented simply as a reduction in microvascular flow and related oxygen tension which causes accumulation of lact ic acid, depletion of cel lular energy, destabi l i-zation of lysosomes and consequently the release of lysosomal enzymes. These, in turn, can cause a further reduction in blood flow, depletion of energy and extensive tissue injury. - 141 -Virtually every tissue in the body is affected in some way by the actions of administered endotoxin (292). Studies have shown that when radiolabelled endotoxin was injected intravenously into a variety of animals, most of the radioactivity was found to be associated with the l iver (217,292). Significant amounts of radiolabelled endotoxin also accumulated in the lung, spleen and kidney, while no detectable amounts of radioactivity appeared in the brain (292). Virtual ly a l l of the circulating radiolabelled endotoxin was distributed between the plasma and buffy coat fraction (293). Interestingly, i t has also been shown that in the buffy coat, almost a l l of the radiolabelled endotoxin was bound to the platelet fraction and not leukocytes (294). However, since most of the endotoxin administered to animals seems to accumulate in the l iver , i t is believed that a large portion of the toxic and lethal effects of endotoxaemia can be attributed to injury of this organ (see 217). In our study, we have investigated the possible existence of a relat ion-ship between the binding of radiolabelled endotoxin to various organs in anaesthetized guinea pigs and toxici ty , as determined by plasma lysosomal hydrolase act iv i ty . Acid phosphatase was used as a representative lysosomal enzyme while the organs studied included heart, lung, l i ver , spleen and kidney. Also ut i l ized in this study was E. ,coli endotoxin radiolabelled 51 with Cr, as i t has been previously shown that this radionuclide is an appropriate label for endotoxin (194, 294). We have found that three hours following the intravenous injection of ^Cr-endotoxin (3.0 mg/kg) into guinea pigs, the greatest approximate accumulation of toxin was found in the spleen (50 ng/mg) and liver (40 ng/mg) followed by lung (10 ng/mg), kidney 51 (10 ng/mg) and heart (1.0 ng/mg). Presumably, much of the Cr-endotoxin uptake in some of these organs (spleen and l iver in particular) was due to - 142 -phagocytosis by reticuloendothelial ce l l s . Although i t is undisputed that the role of the l iver is important in endotoxaemia, the role of the spleen is unresolved. While investigators believe that the reticuloendothelial cel l s of the spleen play an important role in detoxifying endotoxin (295), others have demonstrated that splenectomy had very l i t t l e influence on the toxic i ty of endotoxin when administered to mice or guinea pigs (296,297). The authors, therefore, concluded that the spleen plays only a minor role in the sequelae of endotoxaemia. The spleen does, however, possess the capabi-l i t y of producing antibodies against endotoxin (230,298), which may be important in the development of resistance to subsequent exposures of endo-toxin (230). But under experimental conditions, where lethal doses of endo-toxin are given as a single bolus injection, antibody production in the spleen would be minimal. The results of our investigations have indicated 51 that the accumulation of Cr-endotoxin in the spleen of guinea pigs correlated poorly (r=+0.24) with plasma acid phosphatase levels at three hours following the administration of varying doses (1.0, 3.0 and 6.0 mg/kg) of the radiolabelled toxin to these animals. (Tissue binding was standard-ized to 1.0 mg/kg dose of endotoxin to eliminate effect of increasing dosage 51 on tissue accumulation of Cr-endotoxin). These observations imply that the spleen is not the primary organ involved in the pathophysiology of endo-toxic shock and support the conclusions that other investigators (296,297) have made regarding the function of this organ in experimental endotoxaemia. The heart was another organ in our study for which we found no correla-51 tion between Cr-endotoxin accumulation and plasma acid phosphatase act iv i ty . Of the f ive organs examined, the heart accumulated the lowest amount of ^Cr-endotoxin per mg tissue (wet weight). It is well documen-ted that myocardial fai lure occurs both in experimental endotoxic shock - 143 -(299,300,301) and in c l in ica l septic shock (302,303). However, controversy exists over what factors are responsible for the cardiac failure seen in endotoxaemia and septic shock. Some investigators believe that coronary hypoperfusion consequent to systemic hypotension (290,299) and shunting of blood from the endocardium (304) are factors which may be responsible for the decreased myocardial contract i l i ty seen in endotoxic shock. Other investigators have demonstrated the presence of humoral substances in endo-toxaemia (305) and in c l in ica l sepsis (306,307) that can depress myocardial contrac t i l i ty . There are also reported studies implicating certain media-tors such as histamine (308) and vasopressin (309) which are released during endotoxaemia and have been shown to impair the normal function of the heart. Additional changes noted in the heart during endotoxaemia include a decreased sensit ivity to calcium or cardiac stimulants such as catechola-mines (310) and changes in energy metabolism (311,312). Since endotoxin does not appear to have any direct cardiac depressant effects (313), i t seems the causes of myocardial failure in endotoxaemia are multiple and 51 complex. This may explain our observation that the Cr-endotoxin binding to the myocardium did not correlate with systemic toxici ty in endotoxin-treated guinea pigs. 51 Our studies relating the distribution of E. co l i Cr-endotoxin in guinea pigs with toxicity have also revealed that plasma acid phosphatase act ivi ty correlated negatively with the amount of radiolabelled toxin present in either kidney (r=-0.83) or l iver (r=-0.76). Stated in another 51 way, our data indicate that the accumulation of Cr-endotoxin in these organs decreased as the severity of shock increased. A possible explanation for these results may be that as the intensity of endotoxic shock increased, the blood flow to the kidney and l iver decreased, hence reducing the deliv-- 144 -51 ery of Cr-endotoxin to these organs. It has long been documented that the administration of endotoxin to animals can cause renal cortical necrosis such as that observed in the generalized Schwartzman reaction (Introduction section 1.4). The cause of the renal necrosis is due to occlusion of the glomerular capi l lar ies by f ibr in aggregates formed by the action of endotox-in on the coagulation system (314). Also, in chronic pyelonephritis, which frequently develops in patients who have a history of urinary tract infec-tions, the cause is believed to be a localized immune reaction to endotoxin bound in renal tissue, giving rise to abacterial nephritis (315). Support-ing this theory are observations that the frequency of chronic pyelonephri-t i s in patients with urinary tract infections is s ignif icantly greater when these patients also display a high t i t e r of ant i - l ip id A antibodies (316). Thus the effects of endotoxin on the kidney both experimentally and c l i n i -ca l ly seem to be the result of indirect actions involving both the coagula-tion and immune systems (292). Our observations with ^Cr-endotoxin distribution to the kidney during endotoxaemia support the concept that blood flow to the kidney is reduced in this condition. However, i t cannot be said whether the reduction in blood flow is primarily due to systemic hypotension or to capi l lary blockage by coagulation products or whether both factors are involved. 51 As mentioned, Cr-endotoxin accumulation in the l iver also correlated negatively with plasma acid-phosphatase act ivity in guinea pigs. This again suggests that hepatic blood flow decreases in accordance with the severity 51 of endotoxaemia in these animals. Although less Cr-endotoxin accumula-ted in the l iver following endotoxin administration, the amount that did accumulate was s t i l l much greater than that found in the other organs studied (except the spleen) at equivalent plasma acid phosphatase levels. - 145 -For example, when guinea pigs were injected with the highest dose of 51 Cr-endotoxin used in this study (6.0 mg/kg), the kidney contained approximately 1 ng/mg tissue after three hours whereas the l iver s t i l l bound 9-10 ng toxin/mg tissue (values standardized to 1 mg/kg dose of endotoxin). 51 The large quantity of Cr-endotoxin taken up by the l iver is not entirely the result of phagocytosis by reticuloendothelial c e l l s . Some investigators have demonstrated that as much as 75% of the endotoxin in l iver can be associated with parenchymal cel l s (317). Perhaps related to this are obser-vations from ultrastructural studies revealing that the l iver is highly susceptible to injury during endotoxaemia (217,222,284,318). Indeed, a manifestation of this injury is systemic hypoglycaemia which has become a hallmark of the latter stages of gram-negative endotoxin shock and c l in i ca l sepsis (319,320,321,322). Some of the factors that are believed to be responsible for the hypoglycaemia include diminished l iver blood flow (320), impaired gluconeogenesis (323), increased peripheral ut i l iza t ion of glucose (322) and the release from macrophages of substances which exert insul in-l ike actions (324). It has been proposed that one consequence of the systemic hypoglycaemia in endotoxic shock is vasomotor fai lure resulting in peripheral pooling of blood and intensification of the shock state (325). Thus, the importance of l iver function in endotoxaemia is undisputable. As was mentioned in connection with the kidney, our observations of decreasing 51 Cr-endotoxin accumulation in the l iver with increasing plasma lysosomal hydrolase act ivi ty support the idea proposed by Manson and co-workers (320) that l iver injury in endotoxaemia may be the result of impaired perfusion of this organ. Our observations may also be explained, however, by the depres-sion in the phagocytic function of the reticuloendothelial cel ls in the l iver , which is known to occur in endotoxaemia (326). Although liver func-- 146 -tion is undoubtably important in endotoxaemia, our studies relating 51 Cr-endotoxin distribution in various organs to toxic i ty have suggested, however, that the tissue primarily affected in endotoxaemia is the lung. It was noted forty years ago by Moon that pulmonary congestion represen-ted a c r i t i c a l feature in shock states (see 327). Today, cl inicians are acutely aware that pulmonary fai lure constitutes a major complication in sepsis, particularly in gram-negative septicaemia (328). McGovern has recently reported a study where lung lesions were found in two-thirds of a l l patients dying of gram-negative septicaemia, which was more than twice the incidence of lesions found in other organs (329). Associated with this high incidence of pulmonary failure in gram-negative septicaemia is an alarming mortality rate that approaches 90% (328). Various terms have been used to describe the pulmonary fai lure associated with shock states, including post-traumatic wet lung (330), adult acute respiratory distress syndrome (328,331) and shock-lung (332). In the early stage, this syndrome consists of a pneumonitis which is characterized by in ter s t i t i a l septal oedema, intravascular congestion, in f i l t r a t ion of leukocytes, diffuse focal alveolar collapse ("focal atelectasis") and some in ter s t i t i a l hemorrhage. This lesion can then develop into bronchopneumonia characterized by massive protein-containing oedema, where the lung weight can be increased as much as 50 to 100% above normal and, by hemorrhage into the intra-alveolar, peribronchial as well as perivascular regions of the lung (333). The shock-Tung condition causes a reduction in the venti1ation/perfusion rat io, or what has been termed a pulmonary "shunt", as a result of the perfusion of unventilated alveoli by venous blood (333). Vaughn and co-workers demon-strated that the pathophysiology of shock-lung seen c l i n i c a l l y could be reproduced experimentally by infusing endotoxin into sub-human primates - 147 -(334). Other investigators have also demonstrated the ab i l i ty to reproduce shock-lung lesions in animals by administering endotoxin (333,335,336). Furthermore, i t has been demonstrated that injuries or functional changes can occur in the lung at endotoxin doses that are too low to affect the systemic circulation or cause lesions to other organs (264,337,338). Thus, i t appears that the lung is sensitive to the effects of endotoxin and is a target organ in the pathophysiology of endotoxaemia. Two effects of endotoxin involved in the development of pulmonary fai lure are an alteration in pulmonary vascular permeability and pulmonary hypertension (336). Several mechanisms have been proposed by which endotox-in may cause these effects in the lung. Mechanical obstruction of the pulmonary microcirculation by thrombi or by aggregates of platelets and leukocytes is one possible mechanism by which endotoxin can increase pulmon-ary vascular resistance (261,333,336). However, while there is l i t t l e doubt that endotoxin can cause aggregation of these blood elements and can activ-ate the coagulation system (Introduction, section 1.4), the consensus appears to be that endotoxin exerts its pulmonary effects predominantly through the formation and/or release of vasoactive substances such as hista-mine, 5-hydroxytryptamine, lysosomal hydrolases, kinins and prostanoids (261,264,333,335,336,338,339). Some of these substances are probably released from leukocytes and platelets upon their sequestration and aggrega-tion in the lung. Of particular importance, perhaps, is the potent vasocon-str ictor prostanoid, thromboxane A^, which may originate from platelets (338). Leukocytes are a rich source of lysosomal proteases such as cathep-sin D (340), which can be very injurious to the lung (330,333,339). A relat ively recent study reported by Demling and co-workers on the effect of E. co l i endotoxin in sheep has indicated that lysosomal hydrolase - 148 -activity in the lung lymph paralleled the extent of pulmonary injury in these animals. Furthermore, these investigators demonstrated that the lung showed signs of damage before any other signs of systemic injury were appar-ent (264). While our observations in guinea pigs essentially support those of Demling and associates with regard to the lung being a major site of damage in endotoxaemia, lysosomal hydrolase act ivity in our study was not measured in lung lymph but rather in the plasma. This would be more l ikely a reflection of generalized cel lular injury rather than an indication of lung damage only. Thus, although endotoxaemia affects many organs and tissues, the toxici ty of endotoxin closely parallels its accumulation in the lung. The question then arises as to how endotoxin accumulates in the lung. 51 Presumably, some of the Cr-endotoxin accumulation is the result of phagocytosis by pulmonary macrophages. Mori and co-workers have shown that the pulmonary macrophages in the rat are much less effective in clearing endotoxin from the circulation than the l iver Kupffer cel l s (341). Also, the phagocytic efficiency of pulmonary macrophages is dependent upon high oxygen tensions which contrasts with the ab i l i ty of polymorphonuclear leuko-cytes and monocytes to phagocytose particles under anaerobic as well as aerobic conditions (216,342). Therefore, i t is quite possible that in conditions such as shock-lung, where the pulmonary oxygen tension is greatly reduced, the phagocytic act ivity of the pulmonary macrophages would be impaired. Thus, one would not expect to see a l inear, positive correlation between the accumulation of endotoxin in the lung and endotoxicity i f the uptake of endotoxin by the lung was entirely due to phagocytosis. One may then inquire as to whether the presence of endotoxin in the lung is partly the result of endotoxin binding to pulmonary endothelial ce l l s . - 149 -We have attempted to answer these questions by studying the interaction 0 T " E. co l i endotoxin with a model tissue c e l l , namely the red blood c e l l , and with plasma membranes isolated from red blood ce l l s , since both are 1 51 easily obtainable in pure form. Ut i l i z ing Cr-labelled E. col i endotox-i n , we could indeed demonstrate that the radiolabelled toxin bound in a specific manner to intact human erythrocytes and human erythrocyte ghost 51 membranes. Approximately 75-80% of the bound Cr-endotoxin was displace-51 able with unlabelled endotoxin providing the concentration of Cr-endo-51 toxin did not exceed 50 ug/ml. When higher concentrations of Cr-endo-toxin were used, particularly those exceeding 200 yg/ml, a large portion of the measured binding was non-displaceable and seemingly unsaturable. A possible explanation for this large increase in apparent non-specific 51 binding is that at high concentrations, Cr-endotoxin macromolecules can interact with each other to form large micellular aggregates which may co-sediment with the cells or membranes during the centrifugation step of the binding assay thereby obscuring any specific binding to the structures in question. This particular technical d i f f i cu l ty in endotoxin binding studies has also been noted and recently reported by other investigators 51 (343). However, at physiologically relevant Cr-endotoxin concentrations (50 pg/ml or less) we have found that the binding characteristics (Kp and B ) of endotoxin were higher for ghosts than for intact red ce l l s . This max' 3 3 probably is understandable considering that the ghost preparation used was not resealed and contained fragmented membranes which would l ike ly possess an increased number of potential binding s i tes . Our results with E. co l i endotoxin binding to human red cells were in close agreement with those reported by Ciznar and Shands who studied the binding of biosynthetically labelled ^C-endotoxin from S. typhi murium to sheep erythrocytes (344). - 150 -Other investigators have also shown that endotoxin can bind to a variety of cel l s in addition to the erythrocyte, including isolated hepatocytes (345), lymphocytes (346), granulocytes (347), macrophages (343), platelets (348,349) and cel lular organelles such as lyso.somes (350). Therefore, the demonstrated ab i l i ty of endotoxins to bind to a variety of cel ls lends credence to the poss ib i l i ty that some of the pulmonary accumulation of 51 Cr-endotoxin which we noted in guinea pigs could be the result of endo-toxin binding to lung endothelial ce l l s . A question which evolves from the endotoxin binding studies i s , can endotoxin affect cel lular function directly? Most of the investigations on endotoxaemia reported in the literature suggest that endotoxins affect organ, function indirect ly through the formation and release of mediators which in turn affect organ perfusion or are themselves directly injurious to tissues. We have shown in our in vitro studies that E. col i endotoxin direct ly inhib-its K -p-nitrophenylphosphatase (K -pNPPase) act ivi ty in human red cel l + membranes. Since this enzyme represents a partial reaction of the Na , + + + K -ATPase, an enzyme crucial in the maintenance of ce l lular Na and K homeostasis (351), this inhibitory action of endotoxin could have a profound effect on cel lular function. Our findings may be relevant to the recent observation by Sayeed that Na + transport was impaired in lung slices obtained from endotoxin-treated rats (336). A recent report by Liu and Onji + + + has also demonstrated that both K -pNPPase and Na ,K -ATPase ac t iv i -ties in isolated dog myocyte membrane preparations are inhibitable by endo-toxin in vitro (352). Furthermore, i t has been shown that endotoxin can impair calcium transport in vascular smooth muscle subcellular membranes, again suggesting that endotoxin can directly influence cel lular integrity (353). A related observation is our finding that the act ivi t ies of acid - 151 -phosphatase, N-acetyl-B-glucosaminidase and cathepsin D were, on the average, higher in the plasma of gram-negative septic shock patients than in patients that were in shock due to other causes. In particular, the mean plasma cathepsin D activity in gram-negative septic shock patients was s ignif icantly elevated (.0025 > P > .0005) over the mean act ivity of this lysosomal protease in the plasma of other patients in shock. Since both groups of patients were in circulatory shock, i t is d i f f i cu l t to explain these observations solely on the basis of impaired circulation and inade-quate tissue perfusion. This is supported by our results comparing the effects of haemorrhage and endotoxaemia on plasma lysosomal enyzme activity in guinea pigs and by the results of a study conducted by other investiga-tors who compared the effects of endotoxic and haemorrhagic shock on pulmon-ary damage in dogs (354). The results of this latter investigation indica-ted that the lung damage tended to be more immediate and more severe in dogs treated with endotoxin than in the haemorrhaged animals. We fee l , there-fore, that our results suggest that endotoxins add to the severity of the shock condition by directly impairing cel lular integrity through their actions on the plasma membrane resulting in lysosomal disruption and cel l death. An interesting poss ib i l i ty suggested by our c l in ica l study is the potential usefulness of measuring plasma cathepsin D activity in the diagno-sis or perhaps even in assessing the prognosis of patients with gram-nega-tive septic shock. 4.4 Therapy of Gram-negative Bacteraemia With the observations of the above mentioned studies in mind, i t is interesting to speculate on an effective mode of therapy for gram-negative bacteraemia. Certainly, antibiotic therapy is necessary and can be effec-tive i f used appropriately (199). However, i t is also recognized that ante-- 152 -cedent antibiotic treatment can be deleterious and has been reported to be associated in some cases with an increased incidence of shock in gram-nega-tive septic patients (199). Interestingly, we have been able to demonstrate that gentamycin, an antibiotic commonly used in the management of gram-nega-tive bacteraemia, can actually enhance the toxic effects of E. co l i endotox-in in rats. The mechanism by which gentamycin enhances the toxic effects of endotoxin is unknown but may possibly involve an action at the level of the mitochondrion. It has recently been reported that gentamycin can interfere with mitochondrial functions, such as calcium uptake (356). Mitochondrial integrity may also be impaired during endotoxaemia in vivo (291,357,358,359). Indeed, McGivney and Bradley have shown that the addition of endotoxin directly to cell cultures can produce deleterious effects on cel lular mito-chondrial function (360). Presumably because of its large size, endotoxin would act, at least i n i t i a l l y , at the plasma membrane, thereby affecting mitochondrial function indirectly whereas gentamycin could interact with and modify cel lular mitochondria direct ly . Therefore, gentamycin and endotoxin might act synergistically on the mitochondrion resulting in the depletion of cel lular energy, interference with cel lular homeostatic processes and, ultimately, lysosomal destabil ization. This proposed synergistic effect of gentamycin and endotoxin may have c l in i ca l relevance since, as already mentioned, poor results are frequently obtained with antecedent antibiotic therapy in gram-negative bacteraemia (199). Antibiotic therapy can also have deleterious side-effects in gram-negative bacteraemia by virtue of the bactericidal actions causing destruction of gram-negative bacteria resulting in the liberation of endotoxin (361). Thus, with regard to the treatment of gram-negative septicaemia, i t is apparent that some form of combination therapy would be desirable. One aspect of the therapy would involve antibi-- 153 -otics to prevent the growth of the invading organisms while the other compon-ent should be directed to the antagonism of the toxic effects of liberated endotoxins. The realization of the importance of endotoxin antagonism in gram-negative septicaemia is not new and some investigators recommend the use of corticosteroids as part of the therapy to combat the effects of endotoxins (359). Others have questioned the use of corticosteroids in gram-negative septic shock (200). No suitable agent which would effectively antagonize the toxic actions of endotoxin in vivo has been discovered to date. We believe that an effective endotoxin antagonist must be one that could prevent the binding of endotoxin to cel lular components and/or attenuate the adverse consequences of endotoxin-cel1 interaction. To determine what kind of agent would be potentially capable of antagonizing the binding of endo-toxin to cel l membranes, a better understanding of the nature of the inter-action between endotoxin and membranes is required. Specific endotoxin receptors, which have been shown to be glycoprotein in nature, are present in red blood cell membranes (197,355). Many investigators believe that endotoxin binds and interacts primarily with phospholipid regions of membranes (344,345,346,347). We have taken a simple approach to the study of endotoxin-cellular interactions by examining the effects of E. co l i endotoxin on red blood cel l haemolysis induced by hypotonic challenge. The results of this portion of the thesis work have recently been published (362). As reported, we have found that E. col i endotoxin can protect human erythrocytes from hypotonic lysis in a concentration-dependent and temper-ature-sensitive manner. Greater protection was seen with increasing concentrations of toxin or with increasing temperature. Most agents which protect red cells from hypotonic lysis are capable of increasing the - 154 -" f lu id i ty " or "disorder" of membrane l ipids (363,364,365). Such antihaemo-ly t i c and f luidiz ing properties are exhibited by a wide variety of amphipa-thic compounds such as local anaesthetics, steroids, anti-psychotic drugs, etc. (365). Since endotoxin is a complex of amphipathic subunits, i t perhaps is not too surprising that i t also can stabil ize red blood cel l s in hypotonic media. However, the marked susceptibi l i ty of endotoxin-induced red ce l l stabilization to temperature changes is not shared by other more conventional amphipathic membrane stabil izers such as lidocaine, propranolol and sucrose (362). The effects of temperature on endotoxin antihaemolysis cannot be attributed to variations in toxin binding but rather, by analogy with the cytotoxic agent r i c i n (366), may reflect a temperature-sensitive membrane alteration by the bound toxin which governs the s tabi l iz ing effect. The importance of membrane l ipids to the antihaemolytic effect of endotoxin was suggested by our studies with phospholipase A-treated human erythro-cytes. The protective effect of endotoxin was reduced in these erythro-cytes, particularly at the higher temperatures (25°C and 3 7 ° C ) . Interest-ingly, modifying human erythrocytes with neuraminidase or trypsin had l i t t l e effect on the ab i l i ty of endotoxin to stabil ize these modified erythrocytes against hypotonic lys i s . A very interesting observation in our studies was that the ab i l i ty of E. co l i endotoxin to protect red blood cel l s from hypotonic lysis was str iking-ly different when erythrocytes from various animals were used instead of human erythrocytes. The degree of protection offered by endotoxin was much less when animal red blood cel ls were used, and at low temperatures (5°C) a lyt ic effect of endotoxin on these ce l l s became apparent. Compositional analyses of the various erythrocyte membranes (362) fai led to provide an adequate explanation for the greater susceptibil i ty of human red cel ls to - 1 5 5 -the s tabi l iz ing actions of endotoxin. However, Aloni and co-workers repor-ted that the effect of temperature on the osmotic f r a g i l i t y of the human erythrocyte was much greater than that seen for other species of red blood cells (367). It was suggested by these authors that differences in the l i p i d matrix of the red cel l membranes were responsible for the observed inverse relationship between temperature and osmotic f r a g i l i t y . Since the antihaemolytic effect of endotoxin also seems to be influenced by the l ip id matrix of the red cel l membrane, i t perhaps can be concluded that the characteristics of this component of the human erythrocyte membrane may govern the greater protective effect of endotoxin in these ce l l s . It is d i f f i cu l t to assess the physicochemical nature of the erythrocyte membrane l i p i d matrix simply on the basis of compositional assays since this is determined by a combination of several factors such as cholesterol, phospha-tidylcholine/sphingomyelin rat io , and degree of acyl-chain saturation (364). However, the differences obtained with human erythrocytes could also be explained by other factors, such as surface area/volume ratio for example, which is known to correlate positively with osmotic resistance (364,368). In regard to the antihaemolytic effect of endotoxin, the importance of the polysaccharide and l ip id components of the toxin complex was also examined by testing the effect of chemically modified E. col i toxins on the osmotic f r ag i l i ty of human erythrocytes. Modification of the polysaccharide portion of endotoxin with sodium periodate did not influence the antihaemo-ly t i c action of endotoxin whereas mild alkaline hydrolysis with either sodium hydroxide or hydroxylamine, both of which alter the l i p i d component, markedly reduced the ab i l i ty of endotoxin to protect human red cells from hypotonic ly s i s . Further, sodium hydroxide- or hydroxylamine-treated endotoxins exhibited haemolytic rather than antihaemolytic effects. Other - 156 -investigators also noted that alkali-treated endotoxins can cause haemolysis of red cel ls (344). However, our results have indicated that alkali-treated endotoxins are not homogeneous but can be separated into four fractions by Sepharose 6B chromatography. Of these four fractions, only one (peak III), which was glycolipid in nature, displayed haemolytic act iv i ty . Thus, direct comparisons between alkali-treated and native endotoxins should be avoided unless a particular fraction of the alkali-treated toxin is specified. Therefore, i t is apparent from our studies that the consequences of endotox-in binding to cel lular membranes are l ike ly mediated by hydrophobic interac-tions involving both the l i p i d A component of endotoxin and the l i p i d matrix of the membrane. This postulate is in agreement with the conclusions of other investigators (344,345,369,370). Therefore, regarding the quest for a suitable endotoxin antagonist, our investigations were directed towards various agents which are known to interact with the l ip id matrix of cel lular membranes and like endotoxin, have been shown to exhibit antihaemolytic effects in v i t ro . The rationale was based on the hypothesis that i f certain pharmacological agents interact with ce l l membranes in a similar manner to endotoxin, these agents may compete with endotoxin for membrane binding sites and thereby act as pharma-cological antagonists. Although a wide variety of structurally diverse compounds are able to protect red blood cel l s from hypotonic lysis (365), we have focussed on those which have been shown to exhibit protective effects in experimental endotoxaemia. Indeed, a wide variety of substances, includ-ing a number of membrane-active agents, have been reported to offer benefi-cial effects in endotoxaemia or septicaemia. These include levan, a poly-fructoside of high molecular weight (371); polymyxin B, a cationic antibiot-ic (175); lidocaine (375,376); naloxone (377); propranolol (378,379); - 157 -nonsteroidal anti-inf1ammatory agents (261,380,381); prostacyclin (382); infusions of glucose-insulin-potassiurn solutions (383); thromboxane synthe-tase inhibitors (384,385); thromboxane antagonists (385); and corticoster-oids (223,359,372,386,387). We have chosen to study lidocaine, propranolol and corticosteroids as possible endotoxin antagonists for our studies since these drugs are known to have membrane stabi l iz ing or antihaemolytic actions like endotoxin (365,388). In addition to antihaemolytic effects, cortico-steroids have also been shown to be effective in protecting red blood cel l s from the ly t i c actions of phospholipase C (389) or sulfhydryl group modify-ing reagents such as N-ethylmaleimide and p-chloromercuribenzoic acid (390). 51 Interestingly, when the effects of these drugs on the binding of Cr-en-dotoxin (E. co l i ) to human erythrocyte ghost membranes were examined, the 51 results indicated that propranolol antagonized the binding of Cr-endo-toxin to the membranes almost as effectively as cold (unlabelled) endotoxin. In contrast, methylprednisolone and lidocaine had v i r tua l ly no effect on the 51 binding of Cr-endotoxin. It was also interesting to note that the dimethyl quaternary derivative of propranolol, pranolium, inhibited the 51 membrane binding of Cr-endotoxin to approximately the same extent as did propranolol. Further, these drugs were capable of modifying the accumula-51 tion of Cr-endotoxin in the lungs of guinea pigs in vivo. Guinea pigs pretreated with pranolium or propranolol (0.1 mg/kg) .had s ignif icantly less 51 Cr-endotoxin present in their lungs than did untreated control animals at three hours following a 3 mg/kg intravenous injection of the radiolabel-led toxin. Although propranolol was more effective than pranolium in reduc-51 ing the pulmonary binding of Cr-endotoxin, i t was found that the d-isomer of propranolol (which is v i r tual ly devoid of e-blocking properties) was better tolerated by endotoxin-treated guinea pigs than the racemate - 158 -(d, 1-propranolol.. Presumably, the deleterious effect of d, 1-propranolol was largely due to the impairment of the response of the heart and lungs to the beneficial effects of sympathetic act ivi ty or endogenously-released catecho-lamines on these organs during endotoxaemia. Both the racemate and the d-isomer of propranolol were equally effective, however, in preventing the 51 binding of Cr-endotoxin to human erythrocyte membranes in v i t ro . In contrast, pretreating guinea pigs with methylprednisolone (35 mg/kg) or infusions of lidocaine (1 mg/kg/hr) or adenosine (0.5 mg/kg/min) had no 51 significant effect on the pulmonary accumulation of Cr-endotoxin. The 51 small decrease in Cr-endotoxin act ivity observed in the lungs of methyl-prednisolone-treated guinea pigs may be a reflection of an inhibitory effect of the steroid on pulmonary macrophage act ivity (391) rather than an action on endotoxin binding to lung tissue. When a different animal species was used, namely the rat, treatment with d-propranolol was again found to be more effective than other drugs in reducing the tox ic i ty of E. co l i endotox-in (10 mg/kg) as judged by plasma acid phosphatase act iv i ty . Drug pretreat-ment resulted in decreased enzyme activity (.02 < p < .05) at two hours following endotoxin administration and levels remained significantly reduced for the remainder of the f ive hour experiment. Pretreatment with pranolium was less effective than d-propranolol but equally effective to hydrocorti-sone or chlorpromazine in lowering plasma acid phosphatase act ivity in the endotoxin-treated rat. Mortality studies in mice have also indicated that injections of d-propranolol can offer some protection for approximately 18 hours against a dose of E. col i endotoxin which caused 80% mortality in non-drug treated animals. Thus, while optimal dosage and scheduling of d-propranolol treatment have not been determined in our study, the results obtained with relat ively small doses (0.1 mg/kg) of drugs in animals given - 159 -large bolus injections of endotoxin were certainly very encouraging. Indeed the drug may be even more effective in situations where smaller levels of endotoxin are introduced to the circulation in a slow, sustained manner over long periods of time, as l ikely would occur in c l in i ca l gram-negative bacteraemi as. The results of our studies then, demonstrate the fea s ib i l i ty of prevent-ing the binding of endotoxins to tissues as an effective means of reducing the toxic actions of these bacterial ce l l wall constituents. The ab i l i ty of certain drugs (such as propranolol) to prevent the accumulation of endotoxin in lung tissue appeared to correlate with their protective effects against endotoxin toxici ty in vivo, as measured by plasma acid phosphatase eleva-tions. It has been shown by other investigators that the tissue concentra-tions of d,l-propranolol, following its intravenous administration, are highest in the lung (392). This selective distribution of propranolol to the lung in vivo coupled with its ab i l i ty to antagonize the binding of endotoxin to membranes are undoubtedly two major factors responsible for the salutary actions of propranolol in experimental endotoxaemia. Drugs such as propranolol, which can directly antagonize the binding of endotoxin at a target organ would have greater potential in the therapy of endotoxaemia than drugs such as steroids or chlorpromazine for example, which may act indirect ly by stabi l iz ing lysosomes (393,394) or adenosine which affects blood flow to organs. Furthermore, the relative lack of potentially de t r i -mental B-adrenergic antagonism in d-propranolol (as compared with the race-mate) should make c l in ica l studies of i ts therapeutic potential in gram-neg-ative septicaemia more feasible. F ina l ly , not a l l the biological actions of endotoxins are harmful to the host. Indeed, some beneficial actions of endotoxins include the enhancement - 160 -of host resistance to infections (395,396), radiation injury (397,398), and tumor growth (399,400). However, the application of these beneficial actions to humans is precluded by the toxic effects of endotoxin. One possible way of circumventing this problem might be through the use of detoxified toxins (401). In this regard, we have examined the poss ibi l i ty of using detoxified toxins as endotoxin antagonists in vivo. Ut i l i z ing plasma acid phosphatase act ivi ty as an indication of endotoxicity in rats, preliminary results have indicated that endotoxins detoxified by sodium periodate treatment can protect against the toxic actions of native endotox-in when the modified toxins were administered thir ty minutes before native endotoxin. The observation that sodium periodate-detoxified toxins were more effective in this regard than sodium hydroxide-modified endotoxins may be related to our observations that sodium periodate-detoxified endotoxin accumulated in the lung to approximately the same extent as did native endo-toxin. These observations may then pose the question, why is the sodium periodate-modified endotoxin devoid of toxici ty i f i t accumulates in lung tissue like native endotoxin? Since sodium periodate treatment removes the 0-antigen polysaccharide chain of endotoxin, this portion of the toxin complex may contribute to the toxic actions of endotoxin, presumably by causing an acute inflammatory-like or anaphylactic reaction to occur in the lung, perhaps in a manner analagous to what is believed to occur in abacter-ial nephritis (315). Thus, i t seems that while the l i p i d component of the endotoxin complex is important for binding to cel lular membranes, the immune reactions initiated by the polysaccharide chain can be toxic particularly i f these are localized and interfere with the function of a v i ta l organ such as the lung. - 1 6 1 -In conclusion, then, the studies reported in this thesis indicate the value of u t i l i z ing plasma lysosomal enzyme act iv i t ies as a measure of endo-toxin tox ic i ty in animal experiments and suggest that plasma levels of lyso-somal hydrolases may also provide a rel iable assessment of the severity of shock in patients with gram-negative septicaemia. Plasma concentrations of cathepsin D, in particular, were found to be s ignif icantly elevated in patients with gram-negative septic shock as compared with other forms of shock and may therefore be of some diagnostic value. The elevated plasma lysosomal hydrolase act ivi t ies in patients with gram-negative septic shock suggest the possible involvement of endotoxins which may have direct ce l lu-lar actions on certain organs. Experimentally, i t was shown that E. col i endotoxins bind to cel l membranes in a specific manner and such binding can modify the functional properties of membranes. It was shown that certain drugs such as d-propranolol could antagonize the binding of E. col i endotox-in to human red cel l ghost membranes in v i t ro , and when administered in vivo, could reduce the accumulation of endotoxin by the lung, a primary organ of pathological involvement in endotoxaemia. The results of our experiments suggest that the use of drugs such as d-propranolol in combina-tion with appropriate antibiotic therapy may greatly improve the effective-ness of currently employed therapeutic modalities in the c l in ica l management of gram-negative septicaemia. - 162 -LIST OF REFERENCES 1. Milner, K .C . , Rudbach, J.A. and Ribi , E . : In "Microbial Toxins", G. Weinbaum, S. Kadis, and S.J. A j l , (eds). Academic Press, New York and London, Vol . 4, Chap. 1, p. 1 (1971). 2. Branham, S .E . : Toxic Products of Bacterium Enterit idis and of Related Micro-organisms. J. Infect. Dis. 37, 291-308 (1925). 3. Morrison, D.'C. and Ulevitch, R . J . : The Effects of Bacterial Endotox-ins on Host Mediation Systems. Am. J. Pathol. 93, 527-617 (1978). 4. Kabir, S., Rosenstreich, D.L. and Mergenhagen, S .E . : In "Bacterial Toxins and Cell Membranes", J. Jeljaszewicz, and T. Wadstrom, (eds). Academic Press, London, New York, San Francisco, Chap. 3, p. 59 (1978). 5. Van Heyningen, W.E.: In "Microbial Toxins", S.J. A j l , S. Kadis, and T.C. Montie, (eds). Academic Press, New York and London, Vol . 1, Chap. 1, pp. 1-3 (1970). 6. Ecker, E . E . : The Pathogenic Effect and Nature of a Toxin Produced by B. paratyphosus B. J. Infect. Dis. 21, 541-553 (1917). 7. Zahl, P.A. and Hutner, S .H. : Action of Bacterial Toxins on Tumors. III. Some Biological Properties of Purified Salmonella typhimurium Endotoxin. Proc. Soc. Exp. B i o l . Med. 52, 116-118 (1943). 8. Hutner, S.H. and Zahl, P .A . : Action of Bacterial Toxins on Tumors. IV. Distribution of Tumor-Hemorrhage Agents Among Bacterial Species. Proc. Soc. Exp. Bio l . Med. 52, 364-368 (1943). 9. Zahl, P .A. , Hutner, S.H. and Cooper, F.S. : Action of Bacterial Toxins on Tumors. V. Immunological Protection Against Tumor Hemorrhage. Proc. Soc. Exp. B i o l . Med. 54, 48-50 (1943). 10. Carey, W.F. and Baron, L .S . : Comparative Immunologic Studies of Cell Structures Isolated from Salmonella typhosa. J. Immunol. 83, 517-520 (1959). : — 11. Rib i , E . , Milner, K.C. and Perrine, T .D . : Endotoxic and Antigenic Fractions from the Cell Wall of Salmonella enter i t id i s . Methods for Separation and Some Biologic Act iv i t i e s . J. Immunol. 82, 75-84 (1959). 12. Rudbach, J . A . , Milner, K.C. and Ribi , E . : Occult Endotoxin in Bacter-ial Protoplasm. J. Bacteriol . 99, 51-56 (1969). 13. Todd, J .P . , Shaw, J . A . M . , Blain, J .A. and Boyle, W.: The Isolation and Comparison of Pyrogenic Factors from Proteus vulgaris. J. Pharm. Pharmacol. 10, 223T-226T (1958). - 163 -14. Taylor, A . , Knox, K.W. and Work, E . : Chemical and Biological Proper-ties of an Extracellular Lipopolysaccharide from Escherichia col i Grown Under Lysine Limiting Conditions. Biochem. J. 99, 53-61 (1966). 15. Crutchley, M. J . , Marsh, D.G. and Cameron, J . : Free Endotoxin. Nature 214, 1052 (1967). 16. Crutchley, M.J . , Marsh, D.G. and Cameron, J . : Biological Studies on Free Endotoxin and a Non-toxic Material from Culture Supernatant Fluids of Escherichia coli 078K80. J. Gen. Microbiol . 50, 413-420 (1968). 17. Roberts, R.S.: Preparation of Endotoxin. Nature 209, 80 (1966). 18. Rogers, D.: Release of a Lipopolysaccharide-Protein Complex from Escherichia coli A by Warm Water Treatment. Biochim. Biophys. Acta 230, 72-81 (1971). 19. Jorgensen, J .H. and Smith, R.F . : Measurement of Bound and Free Endo-toxin by the Limulus Assay. Proc. Soc. Exp. B io l . Med. 146, 1024-1031 (1974). 20. Nikaido, H . : Biosynthesis and Assembly of Lipopolysaccharide and the Outer Membrane Layer of Gram-Negative Cell Wall. In "Bacterial Membranes and Walls". L. Leive (ed). Marcel Dekker, Inc., New York, Vol . 1, Chap. 3, p. 131 (1973). 21. DePetis, S.: Ultrastructure of the Cell Wall of Escherichia col i and Chemical Nature of its Constituent Layers. 0. Ultrastruct. Res. 19, 45-83 (1967). 22. Murray, R .G.E . , Steed, P. and Elson, H.H. : The Location of the Muco-peptide in Sections of the Cell Wall of Escherichia co l i and Other Gram-Negative Bacteria. Can. J. Microbiol . Ut 547-560 (1965). 23. DiRienzo, J .M. , Nakamura, K. and Inouye, M. : The Outer Membrane Proteins of Gram-Negative Bacteria: Biosynthesis, Assembly and Func-tions. Ann. Rev. Biochem. 47, 481-532 (1978). 24. Osborn, M. J . , Gander, J . E . , Par i s i , E. and Carson, J . : Mechanism of Assembly of the Outer Membrane of Salmonella typhimurium. Isolation and Characterization of Cytoplasmic and Outer Membrane. J. B i o l . Chem. 247, 3962-3972 (1972). 25. Glauert, A.M. and Thornley, M.J . : The Topography of the Bacterial Cell Wall. Ann. Rev. Microbiol. 23, 159-198 (1969). 26. Nanninga, N.: Ultrastructure of the Cell Envelope of Escherichia coli B After Freeze-Etching. J. Bacteriol. 101, 297-303 (197UT: - 164 -27. Bayer, M.E. and Remsen, C C : Structure of Escherichia col i After Freeze-Etching. J. B a c t e r i d . 101, 304-313 (1970). 28. Miura, T. and Mizushima, S.: Separation by Density Gradient Centrifu-gation of Two Types of Membranes from Spheroplast Membrane of Escheri- chia col i K12. Biochim. Biophys. Acta _150, 159-161 (1968). 29. B e l l , R.M., Mavis, R.D., Osborn, M.J. and Vagelos, P.R. , : Enzymes of Phospholipid Metabolism: . Localization in the Cytoplasmic and Outer Membrane of the Cell Envelope of Escherichia coli and Salmonel1 a typhimurium. Biochim. Biophys. Acta 249, 628-635 (1971). 30. White, D.A., Albright, F.R. , Lennarz, W.J. and Schnaitman, C .A. : Distribution of Phospholipid-Synthesizing Enzymes in the Wall and Membr.ane Subfractions of the Envelope of Escherichia c o l i . Biochim. Biophys. Acta 249, 636-642 (1971). 31. Schnaitman, C.A. : Examination of the Protein Composition of the Cell Envelope of Escherichia co l i by Polyacrylamide Gel Electrophoresis. J. Bacteriol . 104, 882-889 (1970). 32. Mindich, L . : Synthesis and Assembly of Bacterial Membranes. In "Bacterial Membranes and Walls", L. Leive (ed). Marcel Dekker, Inc., New York, Vol 1, Chap. 1, p. 2 (1973). 33. Luderitz, 0., Staub, A.M. and Westphal, 0.: Immunochemistry of 0 and R Antigens of Salmonella and Related Enterobacteriaceae. Bacteriol. Rev. 30, 192-255 (1966). 34. Palmer, J.W. and Gerlough, T .D . : A Simple Method for Preparing Ant i -genic Substances from the Typhoid Bacil lus. Science 92, 155-156 (1940). ~ ~ 35. Westphal, 0. , Luderitz, 0. and Bister, F . : Uber die Extraktion von Bakterien mit Phenol-Wasser. Z. Naturforsch. _7B, 148-155 (1952). 36. Westphal, 0. and Jann, K.: Bacterial Lipopolysaccharides. Extraction with Phenol-Water and Further Applications of the Procedure. In "Methods in Carbohydrate Chemistry", R.L. Whistler (ed). Academic Press, New York, Vol . 5, pp. 83-91 (1965). 37. Morgan, W.T. J . : The Isolation and Properties of a Specific Antigenic Substance from B. Dysenteriae (Shiga). Biochem. J. 31, 2003-2021 (1937). 38. Leive, L . , Shovlin, V.K. and Mergenhagen, S .E . : Physical, Chemical and Immunological Properties of Lipopolysaccharide Released from Escherichia col i by Ethylenediaminetetraacetate. J. B i o l . Chem. 243, 6384-6391 (1968). - 165 -39. Morrison, D.C. and Leive, L . : Fractions of Lipopolysaccharide from Escherichia col i 0111:B4 Prepared by Two Extraction Procedures. J. B io l . Chem. 250, 2911-2919 (1975). 40. Fukushi, K., Anacker, R .L . , Haskins, W.T., Landy, M. , Milner, K.C. and Ribi , E . : Extraction and Purification of Endotoxin from Enterobacter-iaceae: A Comparison of Selected Methods and Sources. J. B a c t e r i d . 87, 391-400 (1964). 41. Nowotny, A .M. , Thomas, S., Duron, O.S. and Nowotny, A . : Relation of Structure to Function in Bacterial O-Antigens. I. Isolation Methods. J. B a c t e r i d . 85, 418-426 (1963). 42. Nowotny, A . : Relation of Structure to Function in Bacterial O-Anti-gens. II. Fraction of Lipids Present in Boivin-Type Endotoxin of Serratia marcescens. J. Bac te r id . 815, 427-435 (1963). 43. Rib i , E . , Haskins, W.T., Landy, M. and Milner, K . C : Symposium on Bacterial Endotoxins. I. Relationship of Chemical Composition to Biological Act iv i ty . B a c t e r i d . Rev. 25, 427-436 (1961). 44. Burton, A . J . and Carter, H .E . : Purification and Characterization of the Lipid A Component of the Lipopolysaccharide from Escherichia c o l i . Biochemistry 3, 411-418 (1964). 45. Luderitz, 0. , Westphal, 0., Staub, A.M. and Nikaido, H. : Isolation and Chemical and Immunological Characterization of Bacterial Lipopoly-saccharides. In "Microbial Toxins", Weinbaum, G . , Kadis, S. and A j l , S.J. (eds). Academic Press, New York and London, Vol . 4, Chap. 4, pp. 145-233 (1971). 46. Westphal, 0. and Luderitz, 0.: Chemische Erforschung von Lipopolysac-chariden Gram-Negativer Bakterien. Angew. Chem. 66_, 407-417 (1954). 47. Nowotny, A . : Molecular Aspects of Endotoxic Reactions. Bacteriol . Rev. 33, 72-98 (1969). 48. Ta i , C. and Goebel, W.F.: On the Nature of the Toxic Component of the Somatic Antigen of Shigella paradysenteriae Type Z (Flexner). J. Exp. Med. 92 25-34 (1950)"^ 49. Wober, W. and Alaupovic, P.: Studies on the Protein Moiety of Endo-toxin from Gram-Negative Bacteria. Eur. J. Biochem 19^ , 357-367 (1971). 50. Tripodi , D. and Nowotny, A . : Relation of Structure to Function in Bacterial O-Antigens. V. Nature of Active Sites in Endotoxic Lipopo-lysaccharides of Serratia marcescens. Ann. N.Y. Acad. Sci . 133, 604-621 (1966). 51. Kim, Y.B. and Watson, D.W.: Biologically Active Endotoxins from Salmonella Mutants Deficient in 0- and R-Polysaccharides and Heptose. J. Bacteriol . 94, 1320-1326 (1967). - 166 -52. Kasai, N. and Nowotny, A . : Endotoxic Glycolipid from a Heptoseless Mutant of Salmonella minnesota. J. Bacteriol 94, 1824-1836 (1967). 53. Haskins, W.T., Landy, M. , Milner, K.C. and Ribi , E . : Biological Properties of Parent Endotoxins and Lipoid Fractions with a Kinetic Study of Acid-Hydrolyzed Endotoxin. J. Exp. Med. 114, 665-684 (1961). 54. Ribi , E . , Haskins, W.T., Landy, M. and Milner, K . C : Preparation and Host-Reactive Properties of Endotoxin with Low Content of Nitrogen and Lip id . J. Exp. Med. 114, 647-663 (1961). 55. Morgan, W.T.J, and Partridge, S.M.: An Examination of the O-Antigenic Complex of Bact. typhosum. Br. J. Exp. Pathol. 23_, 151-165 (1942). 56. Landy, M., Johnson, A . G . , Webster, M.E. and Sagin, J .F . : Studies on the O-Antigen of Salmonella typhosa. II. Immunological Properties of the Purified Antigen. J. Immunol. 74, 466-478 (1955). 57. Ribi , E . , Hoyer, B .H. , Milner, K . C , Perrine, T . O . , Larson, C L . and Goode, G. : Physical and Chemical Analysis of Endotoxin from Salmonel- la enter i t id i s . J. Immunol 84, 32-47 (1960). 58. Webster, M.E . , Sagin, J .F . , Landy, M. and Johnson, A . G . : Studies on the 0-Antigen of Salmonella typhosa. I. Purif ication of the Antigen. J. Immunol. 74, 455-465 (1955)'. 59. Sultzer, B.M. and Goodman, G.W.: Endotoxin Protein: A B-Cell Mitogen and Polyclonal Activator of C3H/HeJ Lymphocytes. J. Exp. Med. 144, 821-827 (1976). 60. Wober, W. and Alaupovic, P.: Studies on the Protein Moiety of Endo-toxin from Gram-Negative Bacteria. Eur. J. Biochem. 19, 340-356 (1971). 61. Wu, M.C. and Heath, E .C . : Isolation and Characterization of Lipopoly-saccharide Protein from Escherichia c o l i . Proc. Natl. Acad. Sci . 70, 2572-2576 (1973). 62. Morrison, D . C , Betz, S.J. and Jacobs, D.M.: Isolation of a Lipid A Bound Polypeptide Responsible for "LPS-Initiated" Mitogenesis of C3H/HeJ Spleen Cel l s . J. Exp. Med. 144, 840-846 (1976). 63. Freedman, H.H. , Fox, A . E . , W i l l i s , S. and Schwartz, B.S.: Role of Protein Component in Endotoxin Modification of Host Reactivity. Proc. Soc. Exp. B io l . Med. 125, 1316-1320 (1967). 64. Nikaido, H. and Nakae, T . : The Outer Membrane of Gram-Negative Bacteria. In "Advances in Microbial Physiology", A.H. Rose and J .G. Morris (eds). Academic Press, Vol . 20, p. 168 (1979). - 167 -65. Kauffmann, F . : C. The Classification of Salmonella-Species. I. Gener-al Serology. In "Classification of Bacteria", K. Kauffmann (ed). Munksgaard Publ . , p. 17 (1975). 66. Kauffmann, F . : D. The Classification of Escherichia-Species. I. General Serology. In "Classification of Bacteria", K. Kauffmann (ed). Munksgaard Publ . , p.132 (1975). 67. Edstrom, R.D. and Heath, E . C . : The Biosynthesis of Cell Wall Lipopo-lysaccharide in Escherichia c o l i . VII. Studies on the Structure of the O-Antigenic Polysaccharide. J. B io l . Chem. 242, 4125-4133 (1967). 68. Jann, B., Reske, K. and Jann, K.: Heterogeneity of Lipopolysacchar-ides. Analysis of Polysaccharide Chain Lengths by Sodium Dodecylsulf-ate-Polyacrylamide Gel Electrophoresis. Eur. J. Biochem. 60, 239-246 (1975). ~ ~ 69. Gmeiner, J . : The Isolation of Two Different Lipopolysaccharide Frac-tions from Various Proteus mirabilis Strains. Eur. J. Biochem. 58, 621-626 (1975). 70. Galanos, C , Luderitz, 0., Rietschel, E.T. and Westphal, 0.: Newer Aspects of the Chemistry and Biology of Bacterial Lipopolysaccharides, with Special Reference to their Lipid A Component. In "International Review of Biochemistry", Vol . 14, Chap. 6, Biochemistry of Lipids II. T.W. Goodwin (ed). University Park Publ. pp. 239-335 (1977). 71. Nikaido, H . , Takeuchi, Y. , Ohnishi, S . I . , and Nakae, T . : Outer Membrane of Salmonella typhi murium Electron Spin Resonance Studies. Biochim. Biophys. Acta 465, 152^164"(1977). 72. Luderitz, 0., Galanos, C , Lehmann, V . , Nurminen, M., Rietschel, E .T . , Rosenfelder, G . , Simon, M. and Westphal, 0 . : Lipid A: Chemical Structure and Biological Act iv i ty . J. Infect. Dis. 128, Suppl. 17 (1973). 73. Edwards, J.R. and Hayashi, J .A . : Structure of a Rhamnolipid from Pseudomonas aeruginosa. Arch. Biochem. Biophys. I l l , 415-421 (1965). 74. Si lbert , D.F . , Ulbright, T.M. and Honegger, J . L . : Ut i l izat ion of Exogenous Fatty Acids for Complex Lipid Biosynthesis and its Effect on de Novo Fatty Acid Formation in Escherichia col i K-12. Biochemistry 12, 164-171 (1973). 75. Hannecart-Pokorni, E . , Dekegel, D. and Depuydt, F . : Macromolecular Structure of Lipopolysaccharides from Gram-Negative Bacteria. Eur. J. Biochem. 38, 6-13 (1973). 76. Beer, H . , Staehelin, T . , Douglas, H. and Braude, I .A. : Relationship Between Particle Size and Biological Act ivi ty of E. co l i Boivin Endo-toxin. J. C l i n . Invest. 44, 592-602 (1965). - 168 -77. Neter, E . , Westphal, 0., Luderitz, 0., Gorzynski, E.A. and Eichenber-ger, E . : Studies on Enterobacterial Lipopolysaccharides. Effect of Heat and Chemicals on Erythrocyte Modifying, Antigenic, Toxic and Pyrogenic Properties. J. Immunol. 76, 377-385 (1956). 78. 01 ins, A.L . and Warner, R . C : Physiocochemical Studies on a Lipopoly-saccharide from the Cell Wall of Azotobacter vinelandi i . J . B io l . Chem. 242, 4994-5001 (1967). 79. Malchow, D., Luderitz, 0. , Kickhofen, B., Westphal, 0. and Gerish, G . : Polysaccharides in Vegetative and Aggregation-Competent Amoebae of Dictyostelium discoideum. Eur. J. Biochem. 7_, 239-246 (1969). 80. Romeo, D., Girard, A. and Rothfield, L . : Reconstitution of a Func-tional Membrane Enzyme System in a Monomolecular Film. 1. Formation of a Mixed Monolayer of Lipopolysaccharide and Phospholipid. J . Mol. B io l . 53. 475-490 (1970). 81. Osborn, M.J . : Structure and Biosynthesis of the Bacterial Cell Wall. Ann. Rev. Biochem. 38, 501-538 (1969). 82. Muhlradt, P.: Biosynthesis of Salmonella Lipopolysaccharide. The Iji Vitro Transfer of Phosphate to the Heptose Moiety of the Core. Eur. J. Biochem. 11, 241-248 (1969). 83. Muhlradt, P .F . , Wray, V. and Lehmann, V . : A 3 1P-Nuclear-Magnetic-Resonance Study of the Phosphate Groups in Lipopolysaccharide and Lipid A from Salmonella. Eur. J . Biochem. 81, 193-203 (1977). 84. Ng, A-K., Butler, R . C , Chen, C-L.H. and Nowotny, A . : Relationship of Structure to Function in Bacterial Endotoxins. IX. Differences in the Lipid Moiety of Endotoxic Glycolipids. J. Bacteriol . 126, 511-515 (1976). 85. Galanos, C. and Luderitz, 0.: Electrodialysis of Lipopolysaccharides and Their Conversion to Uniform Salt Forms. Eur. J. Biochem. 54, 603-610 (1975). 86. Reichgott, M. J . , Melmon, K .L . , Forsyth, R.P. and Greineder, D.: Cardiovascular and Metabolic Effects of Whole or Fractionated Gram-Negative Bacterial Endotoxin in the Unanesthetized Rhesus Monkey. C i rc . Res. 33, 346-352 (1973). 87. Kuida, H . , Gilbert , R.P., Hinshaw, L .B . , Brunson, J.G. and Visscher, M.B.: Species Differences in Effect of Gram-Negative Endotoxin on Circulat ion. Am. J. Physiol. 200, 1197-1202 (1961). 88. Greisman, S.E. and Hornick, R.B.: Comparative Pyrogenic Reactivity of Rabbit and Man to Bacterial Endotoxin. Proc. Soc. Exp. B i o l . Med. 131, 1154-1158 (1969). - 169 -89. E l i n , R.J. and Wolff, S.M.: Biology of Endotoxin. Ann. Rev. Med. 27, 127-141 (1976). 90. Wood, W.B. Jr . : Studies on the Cause of Fever. N. Engl. J. Med. 258, 1023-1031 (1958). 91. Snell , E.S. and Atkins, E . : Interactions of Gram-Negative Bacterial Endotoxin with Rabbit Blood In Vitro . Am. J. Physiol . 212, 1103-1112 (1968). 92. Murphy, P.A. : Quantitative Aspects of the Release of Leukocyte Pyro-gen from Rabbit Blood Incubated with Endotoxin. J. Exp. Med. 126, 763-770 (1967). 93. Haeseler, F . , Bode!, P. and Atkins, E . : Characteristics of Pyrogen Production by Isolated Kupffer Cells In Vi t ro . J. Reticuloendothel. Soc. 22, 569-581 (1977). 94. Moore, D.M., Cheuk, S.F. , Morton, J .D. , Ber l in , R.D. and Wood, W.B. J r . : Studies on the Pathogenesis of Fever. XVIII. Activation of Leukocytes for Pyrogen Production. J. Exp. Med. 131, 179-188 (1970). 95. Nordlund, J . J . , Root, R.K. , and Wolff, S.M.: Studies on the Origin of Human Leukocyte Pyrogen. J. Exp. Med. 131, 727-743 (1970). 96. Moore, D.M., Murphy, P .A. , Chesney, P.J . and Wood, W.B. J r . : Synthe-sis of Endogenous Pyrogen by Rabbit Leukocytes. J. Exp. Med. 137, 1263-1274 (1973) 97. Rafter, G.W.: Stimulation of Pyrogen Production by Lipopolysacchar-ide-Leukocyte Interactions. In"Microbiology 1980", D. Schlessinger (ed). American Soc. Microbiol. Publ. , pp. 154-157 (1980). 98. Dinarello, C .A . : Production of Endogenous Pyrogen. Fed. Proc. 38, 52-56 (1979). 99. Bodel, P.: Spontaneous Pyrogen Production by Mouse Histiocytic and Myelmonocytic Tumor Cell Lines In Vi t ro . J. Exp. Med. 147, 1503-1516 (1978). 100. Dinarello, C A . , Goldin, N.P. and Wolff, S.M.: Demonstration and Characterization of Two Distinct Human Leukocyte Pyrogens. J. Exp. Med. 139, 1369-1381 (1974). 101. Dinarello, C A . and Wolff, S.M.: Partial Purification of Human Leuko-cytic Pyrogen. Inflammation 2, 179-189 (1977). 102. Murphy, P .A. , Chesney, J. and Wood, W.B. Jr . : Further Purification of Rabbit Leukocyte Pyrogen. J. Lab. C l i n . Med. 83, 310-322 (1974). - 170 -103. Cranston, W.I . : Mechanisms of Production of Fever by Endogenous Pyrogens. In "Microbiology 1980", D. Schlessinger (ed). Am. Soc. Microbiol. Publ. pp. 147-149 (1980). 104. Skarnes, R.C. and McCracken, J .A. : Mediating Role of Prostaglandin E in Endotoxin Fever. In "Microbiology 1980", D. Schlessinger (ed). Am. Soc. Microbiol . Publ . , pp. 162-165 (1980). 105. Shwartzman, G . : Studies of Bacillus typhosus Toxic Substances. I. The Phenomenon of Local Skin Reactivity to B^_ typhosus Culture F i l t r a te . J. Exp. Med. 48, 247-268 (1928). 106. Horn, R.G. and Col l ins , R.D.: Studies on the Pathogenesis of the Generalized Shwartzman Reaction. The Role of Granulocytes. Lab. Invest. 18, 101-107 (1968). 107. Robinson, A . J . , Rapaport, S.I. and Brown, S .F. : Procoagulant Act iv i ty of Peritoneal Leukocytes: Effects of Cortisone and Endotoxin. Am. J. Physiol. 235, H333-H337 (1978). 108. Margaretten, W.: Some Cl inical Correlations of the Generalized Shwartzman Reaction. Rev. Can. B io l . 25_, 69-80 (1966). 109. Taub, R.N. , Rodriguez-Erdmann, F. and Dameshek, W.: Intravascular Coagulation, the Shwartzman Reaction and the Pathogenesis of T .T .P . Blood 24, 775-779 (1964). 110. Rapaport, S .I . , Tatter, D., Coeur-Baron, N. and Hjort, P .F . : Pseudo- monas Septicemia with Intravascular Clotting Leading to the General-ized Shwartzman Reaction. N. Engl. J. Med. 271, 80-84 (1964). 111. Starzl , T . E . , Lerner, R.A. , Dixon, F . J . , Groth, C . G . , Brettschneider, L. and Terasaki, P . I . : Shwartzman Reaction After Human Renal Homo-transplantation. N. Engl. J . Med. 278, 642-648 (1968). 112. Fine, D.P.: Activation of the Classic and Alternate Complement Path-ways by Endotoxin. J. Immunol. 112, 763-769 (1974). 113. Ward, P.A. and Johnson, K . J . : Effects of Endotoxins on Serum Media-tors (Complement Kinin and Clotting Systems). In "Microbiology 1975", D. Schlessinger (ed). Am. Soc. Microbiol. Publ . , pp. 327-329 (1975). 114. Jensen, J . A . : Anaphylatoxin in its Relation to the Complement System. Science 155, 1122-1123 (1967). 115. Vogt, W.: Activation, Act ivi t ies and Pharmacologically Active Products of Complement. Pharmacol. Rev. 26, 125-169 (1974). 116. Naff, G.B.: Role of the Properdin System. In "Microbiology 1975", D. Schlessinger (ed). Am. Soc. Microbiol. Publ . , pp. 219-222 (1975). - 171 -117. Mergenhagen, S .E . , Snyderman, R., Gewurz, H. and Shin, H.S. : S ign i f i -cance of Complement to the Mechanism of Action of Endotoxin. Curr. Top. Microbiol. Immunol. 50, 37-77 (1969). 118. McCabe, W.R.: Serum Complement Levels in Bacteremia Due to Gram-Nega-tive Organisms. N. Engl. J. Med. 288, 21-23 (1973). 119. Fearon, D.T. , Ruddy, S., Schur, P.H. and McCabe, W.R.: Activation of the Properdin Pathway of Complement in Patients with Gram-Negative Bacteremia. N. Engl. J. Med. 292, 937-940 (1975). 120. Yoshikawa, T . , Tanaka, K.R. and Guze, L .B . : Infection and Dissemina-ted Intravascular Coagulation. Medicine 50, 237-258 (1971). 121. Rodriguez-Erdmann, F . : Bleeding Due to Increased Intravascular Blood Coagulation. Hemorrhagic Syndromes Caused by Consumption of Blood-Clotting Factors (Consumption-Coagulopathies). N. Engl. J . Med. 273, 1370-1378 (1965). 122. Rodriguez-Erdmann, F. : Studies on the Pathogenesis of the Generalized Shwartzman Reaction. III. Trigger Mechanism for the Activation of the Prothrombin Molecule. Thromb. Diath. Haem. 12, 471-483 (1964). 123. Nies, A.S. and Melman, K . L . : Variation in Endotoxin-induced Kinin Production and Effect Between the Rabbit and the Rhesus Monkey. Am. J. Physiol. 225, 230-233 (1973). 124. Margaretten, W.: Local Tissue Damage in Disseminated Intravascular Clotting. Am. J. Cardiol . 20, 185-190 (1967). 125. Merskey, C , Johnson, A . J . , Kleiner, G.J. and Wohl, H . : The Defibrin-ation Syndrome: Cl in ica l Features and Laboratory Diagnosis. Br. J. Haemat. 13, 528-549 (1967). 126. Ph i l l ip s , L . L . : Alterations in the Blood Clotting System in Dissemin-ated Intravascular Coagulation. Am. J. Cardiol . 20, 174-184 (1967). 127. Beller, F.K. and Graeff, H. : Deposition of Glomerular Fibrin in the Rabbit After Infusion with Endotoxin. Nature 215_, 295-296 (1967). 128. Beller , F .K. , Graeff, H. and Gorstein, F. : Disseminated Intravascular Coagulation During the Continuous Infusion of Endotoxin in Rabbits. Morphologic and Physiologic Studies. Am. J. Obstet. Gynecol. 103, 544-554 (1969). 129. Graeff, H. and Beller , F.K. : Hematological Studies During Endotoxin Infusion in Rabbits. Thromb. Diath. Haem. 20, 420-429 (1968). 130. Morrison, D.C. and Cochrane, C.G. : Direct Evidence for Hageman Factor (Factor XII) Activation by Bacterial Lipopolysaccharides (Endotoxins). J. Exp. Med. 140, 797-811 (1974). - 172 -131. Kaplan, A . P . , Goetzl, F . J . and Austen, K .F . : The Fibr inolyt ic Pathway of Human Plasma. II. The Generation of Chemotactic Act iv i ty by Activa-tion of Plasminogen Proactivator. J. C l i n . Invest. 52, 2591-2595 (1973). 132. Ratnoff, O.D.: Some Relationships Among Hemostasis, F ibrinolyt ic Phenomena, Immunity and the Inflammatory Response. Adv. Immunol. 1 0 , 145-227 (1969). 133. Lerner, R.G., Rapaport, S.I. and Spitzer, J .M. : Endotoxin-Induced Intravascular Clott ing: the Need for Granulocytes. Thromb. Diath. Haem. 20, 430-443 (1968). 134. Garner, R. and Evensen, S.A.: Endotoxin-Induced Intravascular Coagu-lation and Shock in Dogs: The Role of Factor VII. Br. J . Haematol. 27, 655-668 (1974). 135. Horowitz, H . I . , Des Prez, R.M. and Hook, E.W.: Effects of Bacterial Endotoxin on Rabbit Platelets. II. Enhancement of Platelet Factor 3 Act ivi ty In Vitro and In Vivo. J. Exp. Med. 116, 619-633 (1962). 136. McKay, D.G. : Vessel Wall and Thrombogenesis - Endotoxin. Thromb. Diath. Haem. 29, 11-26 (1973). 137. Horn, R.G.: Evidence for Participation of Granulocytes in the Patho-genesis of the Generalized Shwartzman Reaction: A Review. In "Bacterial Lipopolysaccharides", E.H. Kass and S.M. Wolff (eds). Univ. of Chicago Press Publ . , Chicago and London, pp. 126-135 (1973). 138. Bohn, E. and Muller-Berghaus, G.: The Effect of Leucocyte and Plate-let Transfusion on the Activation of Intravascular Coagulation by Endotoxin in Granulocytopenic and Thrombocytopenic Rabbits. Am. J. Pathol. 84, 239-258 (1976). 139. Ream, V . J . , Deykin, D., Gurewich, V. and Wessler, S.: The Aggregation of Human Platelets by Bacterial Endotoxin. J. Lab. C l i n . Med. 66, 245-252 (1965). — 140. Des Prez, R.M., Horowitz, H.I. and Hook, E.W.: Effects of Bacterial Endotoxin on Rabbit Platelets. I. Platelet Aggregation and Release of Platelet Factors In Vitro . J. Exp. Med. 114, 857-874 (1961). 141. Muller-Berghaus, G . : Trigger Mechanism of Endotoxin-Induced Intravas-cular Coagulation. In "Gram-Negative Bacterial Infections and Mode of Endotoxin Actions", B. Urbaschek, R. Urbaschek and E. Neter (eds). Springer-Verlag Publ. pp. 362-365 (1973). 142. Rivers, R.P.A. , Hathaway, W.E. and Weston, W.L. : The Endotoxin Induced Coagulant Act ivi ty of Human Monocytes. Br. J. Haematol. 30, 311-316 (1975). — - 173 -143. H i l l e r , E . , Saal, J .G . , Ostendorf, P. and Gr i f f i ths , G.W.: The Procoagulant Act ivi ty of Human Granulocytes, Lymphocytes and Monocytes Stimulated by Endotoxin. K l in . Wochenschr. ^5, 751-757 (1977). 144. S t i l l , W.J.S. and Scott, G.B.D.: An Electron Microscopic Study of the Endothelial Changes and the Nature of the "Fibr inoid" Produced in the Generalized Shwartzman Reaction. Exp. Mol. Pathol. 5_, 118-124 (1966). 145. McGrath, J .M. and Stewart, G. J . : The Effects of Endotoxin on Vascular Endothelium. J. Exp. Med. 129, 833-848 (1969). 146. Stewart, G.J. and Anderson, M.J . : An Ultrastructural Study of Endo-toxin Induced Damage in Rabbit Mesenteric Arteries . Br. J. Exp. Pathol. 52, 75-80 (1971). 147. Christy, J . H . : Pathophysiology of Gram-Negative Shock. Am. Heart J. 81, 694-701 (1971). 148. Waisbren, B.A.: Bacteremia Due to Gram-Negative Bac i l l i Other than the Salmonella. Arch. Int. Med. 88, 467-488 (1951). 149. Borden, C.W. and Hal l , W.H.: Fatal Transfusion Reactions From Massive Bacterial Contamination of Blood. N. Engl. J . Med. 245, 760-765 (1951). 150. McCabe, W.R.: Gram-Negative Bacteremia. Adv. Int. Med. 19, 135-158 (1974). — 151. Young, L .S . , Martin, W.J. , Meyer, R.D., Weinstein, R.J. and Anderson, E . T . : Gram-Negative Rod Bacteremia: Microbiologic, Immunologic and Therapeutic Considerations. Ann. Int. Med. 86, 456-471 (1977). 152. Slack, J .M. and Snyder, I.S.: Opportunistic Enteric Bacteria. In "Bacteria and Human Disease", Year Book Medical Publishers, Inc. Chicago and London, p. 273 (1978). 153. McCabe, W.R.: Immunoprophylaxis of Gram-Negative Bacil lary Infec-tions. Ann. Rev. Med. 27, 335-341 (1976). 154. Freid, M.A. and Vosti , K .L . : The Importance of Underlying Disease in Patients with Gram-Negative Bacteremia. Arch. Intern. Med. 121, 418-423 (1968). 155. Bodey, G.P. , Buckley, M. , Sathe, Y.S. and Freire ich, E . J . : Quantita-tive Relationships Between Circulating Leukocytes and Infection in Patients with Acute Leukemia. Ann. Intern. Med. 64, 328-340 (1966). 156. Root, R.K., Rosenthal, A.S. and Balestra, D . J . : Abnormal Bacterici-dal , Metabolic and Lysosomal Functions of Chediak-Higashi Syndrome Leukocytes. J . C l i n . Invest. 51, 649-665 (1972). - 174 -157. Lehrer, R.I. and Cline, M.J . : Leukocyte Myeloperoxidase Deficiency and Disseminated Candidiasis: The Role of Myeloperoxidase in Resis-tance to Candida Infection. J. C l i n . Invest. 48, 1478-1488 (1969). 158. Lehrer, R.I. and Cline, M.J . : Leukocyte Candidacidal Act iv i ty and Resistance to Systemic Candidiasis in Patients with Cancer. Cancer 27, 1211-1217 (1971). 159. Rosner. F. Valmont, I., Kozinn, P.J. and Caroline, L . : Leukocyte Function in Patients with Leukemia. Cancer ^5, 835-842 (1970). 160. Sbarra, A . J . , Shirley, W. and Baumstark, J.S.: Effect of Osmolarity on Phagocytosis. J . Bacteriol. 85, 306-313 (1963). 161. Bybee, J.D. and Rogers, D.E . : The Phagocytic Act iv i ty of Polymorpho-nuclear Leukocytes Obtained from Patients with Diabetes Mellitus. J. Lab. C l i n . Med. 64, 1-13 (1964). 162. Brandt, L. and Hedberg, H. : Impaired Phagocytosis by Peripheral Blood Granulocytes in Systemic Lupus Erythematosus. Scand. J . Haematol. 6_, 348-353 (1969). 163. Wenger, M.E. and Giles , G.B.: Nitroblue Tetrazolium Dye Reduction by Peripheral Leukocytes from Rheumatoid Ar thr i t i s and Systemic Lupus Erythematosus Patients Measured by a Histochemical and Spectrophoto-metric Method. J. Lab. C l i n . Med. 82, 513-521 (1973). 164. Alexander, J.W., Hegg, M. and Altemeier, W.A.: Neutrophil Function in Selected Surgical Disorders. Ann. Surg. 168, 447-458 (1968). 165. Alexander, J.W. and Wixson, D.: Neutrophil Dysfunction and Sepsis in Burn Injury. Surg. Gynecol. Obstet. 130, 431-438 (1970). 166. Shubin, H . , Weil, M. and Nishijima, H. : C l in ica l Features in Shock Associated with Gram-Negative Bacteremia. In "Gram-Negative Bacterial Infections and Mode of Endotoxin Actions", B. Urbaschek, R. Urbaschek, and E. Neter (eds). Springer-Verlag Publ. pp. 411-417 (1973). 167. McCabe, W.R. and Jackson, G.G.: Gram-Negative Bacteremia. II. C l i n i -ca l , Laboratory and Therapeutic Observations. Arch. Intern. Med. 110, 856-864 (1962). 168. Weil, M.H., Shubin, H. and Biddle, M.: Shock Caused by Gram-Negative Microorganisms. Analysis of 169 Cases. Ann. Intern. Med. 60, 384-400 (1964). " ~ 169. Robson, H.G.: Gram-Negative Bacteremia and Septic Shock: Pathophy-siology and Cl inica l Diagnosis. In "Current Concepts in the Manage-ment of Gram-Negative Bacterial Infections", A.R. Ronald (ed). Amer. Elsevier Publ. Co. , New York, pp. 67-71 (1974). - 175 -170. S c h i m p f f , S., S a t t e r l e e , W., Young, V.M. and S e r p i c k , A.: E m p i r i c Therapy w i t h C a r b e n i c i l l i n and Gentamicin f o r F e b r i l e P a t i e n t s with Cancer and G r a n u l o c y t o p e n i a . N. E n g l . J . Med. 284, 1061-1065 ( 1 9 7 1 ) . 171. Z i n n e r , S.H. and McCabe, W.R.: E f f e c t s of IgM and IgG A n t i b o d y i n P a t i e n t s w i t h Bacteremia Due t o Gram-Negative B a c i l l i . J . I n f e c t . D i s . 133, 37-45 (1976). 172. Young, L.S., Meyer, R.D. and Armstrong, D.: Pseudomonas a e r u g i n o s a V a c c i n e i n Cancer P a t i e n t s . Ann. I n t e r n . Med. 79. 518-527 (1973)'. 173. McCabe, W.R., K r e g e r , B.E. and Johns, M.: T y p e - s p e c i f i c and C r o s s - R e -a c t i v e A n t i b o d i e s i n Gram-Negative B a c t e r e m i a . N. E n g l . J . Med. 287, 261-267 (1972). 174. M o r r i s o n , D.C. and J a c o b s , D.M.: B i n d i n g of Polymyxin B t o the L i p i d A P o r t i o n o f B a c t e r i a l L i p o p o l y s a c c h a r i d e s . Immunochemistry 13_, 813-818 (1976). 175. Palmer, J.D. and R i f k i n d , D.: N e u t r a l i z a t i o n of the Hemodynamic E f f e c t s of E n d o t o x i n by Polymyxin B. Surg. G y n e c o l . O b s t e t . 138, 755-759 (1974). 176. Godin, D.V. and S c h r i e r , S.L.: Mechanism of I n a c t i v a t i o n of E r y t h r o -c y t e Membrane Adenosine T r i p h o s p h a t a s e by C a r b o d i i m i d e s . B i o c h e m i s t r y 9, 4068-4077 (1970). 177. B a r r e t t , A . J . and Heath, M.F.: Lysosomal Enzymes. In "Lysosomes, a L a b o r a t o r y Handbook", J.T. D i n g l e ( e d ) . N o r t h - H o l l a n d P u b l . Co., Amsterdam, New York, O x f o r d , Chap. 2, pp. 124-126 (1977). 178. M a c h l e i d t , H., Roth, S. and Seeman, P.: The Hydrophobic Expansion of E r y t h r o c y t e Membranes by the Phenol A n e s t h e t i c s . B i o c h i m . B i o p h y s . A c t a 255, 178-189 (1972). 179. Seaman, G.V.F., and Uhlenbruck, G.: The S u r f a c e S t r u c t u r e of E r y t h r o -c y t e s from Some Animal S o u r c e s . A r c h . Biochem. B i o p h y s . 100, 493-502 (1963). 180. Simon-Reuss, I., Cook, G.M.W., Seaman, G.V.F. and Heard, D.H.: E l e c -t r o p h o r e t i c S t u d i e s on Some Types of Mammalian T i s s u e C e l l . Cancer Res. 24, 2038-2043 (1964). 181. Seaman. G.V.F., Ja c k s o n , L . J . and Uhlenbru c k , G.: A c t i o n of a-amylase P r e p a r a t i o n s and Some P r o t e a s e s on the S u r f a c e of Mammalian E r y t h r o -c y t e s . A r c h . Biochem. Bio p h y s . 122, 605-613 (1967). 182. R o e l o f s e n , B., Zwaal, R.F.A., C o m f u r i u s , P., Woodward, C B . and van Deenen, L.L.M.: A c t i o n of Pure P h o s p h o l i p a s e A£ and P h o s p h o l i p a s e C on Human E r y t h r o c y t e s and G h o s t s . B i o c h i m . B i o p h y s . A c t a 241, 925-929 (1971). - 176 -183. Gottlieb, M.H.: A Method for the Increased Extraction of Cholesterol from Human Red Blood Cells by Modified Plasma Lipoprotein. Prep. Biochem. 8, 463-469 (1978). 184. Lowry, O.H. , Rosebrough, N.J . , Farr, A.L. and Randall, R.J . : Protein Measurement with the Folin Phenol Reagent. J . B io l . Chem. 193, 265-275 (1951). 185. Warren, L . : The Thiobarbituric Acid Assay of S ia l ic Acids. J. B io l . Chem. 234, 1971-1975 (1959). 186. Zak, B., Dickenmen, R . C , White, E .G. , Burnett, H. and Cherney, P . J . : Rapid Estimation of Free and Total Cholesterol. Am. J. C l i n . Pathol. 24, 1307-1315 (1954). 187. Bartlett , G.R.: Phosphorus Assay in Column Chromatography. J. B io l . Chem. 234, 466-468 (1959). 188. Weissbach, A. and Hurwitz, 0.: The Formation of 2-Keto-3-Deoxyhepton-ic Acid in Extracts of Escherichia co l i B. J. B io l . Chem. 234, 705-709 (1959). 189. Osborn, M.J . : Studies on the Gram-Negative Cell Wall. I. Evidence for the Role of 2-Keto-3-Deoxyoctonate in the Lipopolysaccharide of Salmonella typhi murium. Biochemistry 50, 499-506 (1963). 190. Reed, C . F . , Swisher, S .N. , Marinetti, G.V. and Eden, E .G . : Studies of the Lipids of the Erythrocyte. I. Quantitative Analysis of the Lipids of Normal Human Red Blood Cel l s . J. Lab. C l i n . Med. 56, 281-289 (1960). ~ ~ 191. Rabinowitz, M. , Getz, G.S., Casey, J. and Swift, H . : Synthesis of Mitochondrial and Nuclear DNA in Anerobically Grown Yeast During the Development of Mitochondrial Function in Response to Oxygen. J. Mol. B io l . 41, 381-400 (1969). 192. Cooperstein, S.J. and Lazarow, A . : A Microspectrophotometric Method for the Determination of Cytochrome Oxidase. J . B io l . Chem. 189, 665-670 (1951). 193. Mclntire, F . C , Sievert, H.W., Barlow, G .H. , Finley, R.A. and Lee, A . Y . : Chemical, Physical and Biological Properties of a Lipopolysac-charide from Es^hej2id]ij_c^^ Biochemistry 6, 2363-2372 (1967). 194. Braude, A . I . , Carey, F . J . , Sutherland, D. and Zalesky, M.: Studies with Radioactive Endotoxin. I. The Use of ^ C r to Label Endotoxin of Escherichia c o l i . J. C l i n . Invest. 34, 850-857 (1955). 195. Sheppard, G. and Marlow, C .G. : The Simultaneous Measurement of 5 * C r and 1 4 C by Liquid Scint i l la t ion Counting. Int. J. Appl. Radiat. Isot. 22, 125-127 (1971). - 177 -196. Lassen, N.A., Lindbjerg, J . and Munck, 0. : Measurement of Blood-Flow Through Skeletal Muscle by Intramuscular Injection of Xenon-133. Lancet 1, 686-689 (1964). 197. Springer, G.F . , Adye, J . C , Bezkorovainy, A. and Jirgensons, B.: Properties and Act ivi ty of the Lipopolysaccharide-Receptor from Human Erythrocytes. Biochemistry 13, 1379-1389 (1974). 198. Nelson, G . J . : Lipid Composition of Erythrocytes in Various Mammalian Species. Biochim. Biophys. Acta 144, 221-232 (1967). 199. Kreger, B .E . , Craven, D.E . , Carling, P.C. and McCabe, W.R.: Gram-Negative Bacteraemia. III. Reassessment of Etiololgy, Epidemiol-ogy and Ecology in 612 Patients. Am. J. Med. 68, 332-343 (1980). 200. Kreger, B .E . , Craven, D.E. and McCabe, W.R.: Gram-Negative Bacterae-mia. IV. Re-evaluation of Cl inical Features and Treatment in 612 Patients. Am. J. Med. 68, 344-355 (1980). 201. McCabe, W.R.: Gram-Negative Bacteraemia. Disease-a-Month. Year Book Medical Publ. , Inc., Chicago. (1973). 202. Wolff, S.M.: Biological Effects of Bacterial Endotoxins in Man. J . Infect. Dis. 128 (suppl.), S259-S264 (1973). 203. Caridis , D.T. , Reinhold, R.B. , Woodruff, P.W.H. and Fine, J . : Endo-toxaemia in Man. Lancet U 1381-1386 (1972). 204. Woodruff, P.W.H., 0 'Carro l l , D . I . , Koizumi, S. and Fine, J . : Role of the Intestinal Flora in Major Trauma. J. Infect. Dis. 128 (suppl.), S290-S294 (1973). 205. Andersen, B.M., Skjorten, F. and Solberg, 0. : Electron Microscopical Study of Neisseria meningitidis Releasing Various Amounts of Free Endotoxin. Acta Pathol. Microbiol. Scand., Sect. B. 87, 109-115 (1979). 206. Waisbren, B.A.: Gram-Negative Shock and Endotoxin Shock. Am. J. Med. 36, 819-824 (1964). 207. Postel, J . , Schloerb, P.R. and Furtado, D.: Pathophysiologic Altera-tions During Bacterial Infusions for the Study of Bacteremic Shock. Surg. Gynecol. Obstet. 141, 683-692 (1975). 208. Hinshaw, L .B . , Solomon, L . A . , Holmes, D.D. and Greenfield, L . J . : Comparison of Canine Responses to Escherichia col i Organisms and Endotoxin. Surg. Gynecol. Obstet. 127, 981-988 (1968). 209. F i lk in s , J .P . : Comparison of Endotoxin Detoxification by Leukocytes and Macrophages. Proc. Soc. Exp. B i o l . Med. 137, 1396-1400 (1971). - 178 -210. Love, L . J . , Schimpff, S .C . , Schitter, C.A. and Wiernik, P . H . : Improved Prognosis for Granulocytopenic Patients with Gram-Negative Bacteremia. Am. J . Med. 68, 643-648 (1980). 211. Herzig, B .H. , Herzig, G.P. , Graw , R.G. J r . , B u l l , M.I. and Ray, K.K.: Successful Granulocyte Transfusion Therapy for Gram-Negative Septicemia. N. Engl. J. Med. 296, 701-705 (1977). 212. Solberg, C O . and Helium, K.B. : Neutrophil Granulocyte Function in Bacterial Infections. Lancet 2, 727-730 (1972). 213. Helium, K.B. and Solberg, C O . : Granulocyte Function in Bacterial Infections in Man. Acta Pathol. Microbiol. Scand. 85, 1-9 (1977). 214. Cartwright, G .E . , Athens, J.W. and Wintrobe, M.M.: The Kinetics of Granulopoiesis in Normal Man. Blood 24, 780-803 (1964). 215. Galbraith, P.R., Valberg, L.S. and Brown, M. : Patterns of Granulocyte Kinetics in Health, Infection and in Carcinoma. Blood 25, 683-692 (1965). ~ 216. Oren, R., Farnham, A . E . , Saito, K., Milofsky, E. and Karnovsky, M.L. : Metabolic Patterns in Three Types of Phagocytizing Ce l l s . J. Cell B io l . 17, 487-501 (1963). 217. Ut i i i , R., Abernathy, C O . and Zimmerman, H. J . : Endotoxin Effects on the Liver. Life Sc i . 20, 553-568 (1977). 218. Vacher, J . , Delevallee, F. and Deraedt, R.: Mechanism of the Hyper-susceptibil i ty to the Lethal Effect of Endotoxin (ET) Induced in Mice by Injection of Beryllium Phosphate. Toxicol . Appl. Pharmacol. 40, 99-108 (1977). 219. Beeson, P.B. : Effect of Reticuloendothelial Blockade on Immunity to the Schwartzman Phenomenon. Proc. Soc. Exp. B io l . Med. 64, 146-149 (1947). " ~ 220. Holper, K., Trejo, R.A., Brettschneider, L. and DiLuzio, N.R.: Enhancement of Endotoxin Shock in the Lead-Sensitized Subhuman Primate. Surg. Gynecol. Obstet. 136, 593-601 (1973). 221. Saito, K. and Suter, E . : Lysosomal Acid Hydrolases and Hyperreactiv-i ty to Endotoxin in Mice Infected With BCG. J. Exp. Med. 121, 739-749 (1965). 222. Shands, J.W. and Senterfitt, V . C : Endotoxin Induced Hepatic Damage in BCG-Infected Mice. Am. J. Pathol. 67, 23 (1972). 223. DiLuzio, N.R. and Crafton, C.G. : A Consideration of the Role of the Reticuloendothelial System (RES) in Endotoxin Shock. In "Shock: Biomedical, Pharmacological and Cl inical Aspects", A. Bertel l i and N. Black (eds). Plenum Press, New York. pp. 27-57 (1970). - 179 -224. Kokoshis, P . L . , Williams, D.L. , Cook, J .A. and DiLuzio, N.R.: Increased Resistance to Staphylococcus aureus Infection and Enhance-ment in Serum Lysozyme Act ivi ty by Glucan. Science 199, 1340 (1978). 225. DiLuzio, N.R., Hoffmann, E.O. , Cook, J .A . , Browder, W. and Mansell, P.W.A.: Glucan-Induced Enhancement in Host Resistance to Experimental Tumors. In "Control of Neoplasia by Modulation of the Immune System", M.A. Chirigos (ed). Raven Press, New York. p. 475 (1977). 226. Senterfitt , V.C. and Shands, J.W.: Endotoxin Induced Metabolic Alter-ations in BCG Infected (Hyperactive) Mice. Proc. Soc. Exp. B i o l . Med. 159, 69-74 (1978). 227. Cook, J .A . , Dougherty, W.J. and Holt, T .M. : Enhanced Sensit ivity to Endotoxin Induced by the RE Stimulant, Glucan. C i r c . Shock 7_, 225-238 (1980). 228. Starzecki, B., Reddin, J . L . , Gran, A. and Spink, W.W.: Distribution of Endotoxin -^Cr in Normal and Endotoxin-Resistant Dogs. Am. J. Physiol. 213, 1065-1071 (1967). 229. Greisman, S .E . , Carozza, F.A. and H i l l s , J .D. : Mechanisms of Endotox-in Tolerance I. Relationship Between Tolerance and Reticuloendothelial System Phagocytic Act ivi ty in the Rabbit. J. Exp. Med. 117, 663-674 (1963). 230. Greisman, S .E . , Young, E . J . , Workman, J .B. , Ollodart, R.M. and Hornick, R.B.: Mechanisms of Endotoxin Tolerance. The Role of the Spleen. J. C l i n . Invest. 56, 1597-1607 (1975). 231. Kim, Y.B. and Watson, D.W.: Modification of Host Responses to Bacter-ial Endotoxins. II. Passive Transfer of Immunity to Bacterial Endotox-in with Fractions Containing 19S Antibodies. J. Exp. Med. 121, 751-759 (1965). 232. Braude, A . I . and Douglas, H. : Passive Immunization Against the Local Shwartzman Reaction. J. Immunol. 108, 505-512 (1972). 233. Freedman, H.H. : Passive Transfer of Protection Against Lethality of Homologous and Heterologous Endotoxins. Proc. Soc. Exp. B io l . Med. 102, 504-506 (1959). 234. Braude, A . I . , Ziegler, E . J . , Douglas, H. and McCutchan, J .A . : Antibody to Cell Wall Glycolipid of Gram-Negative Bacteria. Induction of Immunity to Bacteremia and Endotoxemia. J. Infect. Dis. 136 (suppl.), 167-173 (1977). 235. Pollack, M. and Young, L.S . : Protective Act iv i ty of Antibodies to Exotoxin A and Lipopolysaccharide at the Onset of Pseudomonas aerugi-nosa Septicemia in Man. J. C l i n . Invest. 63, 276-286 (1979). - 180 -236. F i lk ins , J .P . : Blood Endotoxin Inactivation After Trauma and Endotox-icosis . Proc. Soc. Exp. B io l . Med. 151, 89-92 (1976). 237 . Skarnes, R . C : The Inactivation of Endotoxin After Interaction with Certain Proteins of Normal Serum. Ann. N.Y. Acad. Sc i . 133, 644-662 (1966). 238. Skarnes, R . C : Host Defense Against Bacterial Endotoxemia: Mechanism in Normal Animals. J. Exp. Med. 132, 300-316 (1970). 239. Johnson, K . J . , Ward, P .A. , Goralnick, S. and Osborn, M.J . : Isolation from Human Serum of an Inactivator of Bacterial Lipopolysaccharide. Am. J. Pathol. 88, 559-574 (1977). 240. Ulevitch, R.J. and Johnston, A.R. : The Modification of Biophysical and Endotoxic Properties of Bacterial Lipopolysaccharides by Serum. J. C l i n . Invest. 62, 1313-1324 (1978). 241. Ulevitch, R . J . , Johnston, A.R. and Weinstein, D.B.: New Function for High Density Lipoproteins. Their Participation in Intravascular Reac-tions of Bacterial Lipopolysaccharides. J. C l i n . Invest. 64, 1516-1524 (1979). — 242. Weil, M.H. , Shubin, H. and Carlson, R.: Treatment of Circulatory Shock. Use of Sympathomimetic and Related Vasoactive Agents. J . Am. Med. Assoc. 231, 1280-1286 (1975). 243. Mi l ler , T . A . , Lindenauer, S.M. and Pozderac, R .V. : 1 3 3 Xenon Clear-ance in the Diagnosis of Arterial Occlusive Disease. J. Surg. Res. 16, 412-417 (1974). 244. Baker, C .H . : Effects of Endotoxin on Microvascular Flow Velocity and Indicator Dispersion. C i rc . Shock 7 387-398 (1980). 245. Nishijima, H. , Weil, M.H., Shubin, H. and Cavanilles, J . : Hemodynamic and Metabolic Studies on Shock Associated with Gram-Negative Bactere-mia. Medicine 52, 287-294 (1973). 246. L i l l e h e i , R . C , Longerbeam, J .K. , Bloch, J .H. and Manax, W.G.: Nature of Irreversible Shock: Experimental and Cl in ica l Observations. Ann. Surg. j JO, 682-708 (1964). 247. Bacalzo, L . V . , Cary, A . L . , Mi l l e r , L.D. and Parkins, W.M.: Methods and Cr i t i ca l Uptake Volume for Hemorrhagic Shock in Rats. Surgery 70, 555-560 (1971). 248. Litwin, M.S.: Blood Viscosity in Shock. Am. J. Surg. 110, 313-316 (1965). 249. Hardaway, R.M.: Microcoagulation in Shock. Am. J. Surg. 110, 298-301 (1965). - 181 -250. Weidner, W.J . , Grega, G.J. and Haddy, F . J . : Changes in Forelimb Weight and Vascular Resistances During Endotoxin Shock. Am. J. Physiol. 221, 1229-1237 (1971). 251. L i l l e h e i , R . C , Longerbeam, J.K., Bloch, J .H. and Manax, W.G.: The Modern Treatment of Shock Based On Physiologic Principles . C l i n . Pharmacol. Ther. 5, 63-101 (1964). 252. Ha l l , R.C. and Hodge, R .L . : Vasoactive Hormones in Endotoxin Shock: A Comparative Study in Cats and Dogs. J. Physiol . 213, 69-84 (1971). 253. Isakson, P . C , Shofer, F . , McKnight, R . C , Feldhaus, R.A. , Raz, A. and Needleman, P.: Prostaglandins and the Renin-Angiotensin System in Canine Endotoxemia. J . Pharmacol. Exp. Ther. 200, 614-622 (1977). 254. Davis, R.B., Meeker , W.R. Jr. and Bailey, W.L. : Serotonin Release by Bacterial Endotoxin. Proc. Soc. Exp. B i o l . Med. 108, 774-776 (1961). 255. Anderson, F . L . , Jubiz, W., Tsagaris, T . J . and Kuida, H . : Endotoxin-induced Prostaglandin E and F Release in Dogs. Am. J. Physiol. 228, 410-414 (1975). 256. Fletcher, J.R., Ramwell, P.W. and Herman, C M . : Prostaglandins and the Hemodynamic Course of Endotoxin Shock. J . Surg. Res. 20, 589-594 (1976). 257. Hinshaw, L .B . , Vick, J .A . , Carlson, J.A. and Fan, Y. : Role of Hista-mine in Endotoxin Shock. Proc. Soc. Exp. B io l . Med. 104, 379-381 (1960). 258. Vick, J .A . , Mehlman, B. and Heiffer, M.H.: Early Histamine Release and Death Due to Endotoxin. Proc. Soc. Exp. B io l . Med. 137, 902-906 (1971). 259. Thai, A.P. and Sardesai, V .M. : Shock and the Circulating Polypeptides. Am. J. Surg. U 0 , 308-312 (1965). 260. Kimball, H.P. , Melmon, K.L. and Wolff, S.M.: Endotoxin-induced Kinin Production in Man. Proc. Soc. Exp. B i o l . Med. 139, 1078-1082 (1972). 261. Parratt, J.R. and Sturgess, R.M.: The Possible Roles of Histamine, 5-Hydroxytryptamine, and Prostaglandin F? a as Mediators of the Acute Pulmonary Effects of Endotoxin. Br. J . Pharmacol. 60, 209-219 (1977). 262. Weil, M.H. and A f i f i , A . A . : Experimental and Cl in ica l Studies on Lactate and Pyruvate as Indicators of the Severity of Acute Circula-tory Failure (Shock). Circulation 41, 989-1001 (1970). 263. Lefer, A .M. : The Role of Lysosomes in Circulatory Shock. A Minire-view. Life Sci . 19, 1803-1810 (1976). - 182 -264. Demi ing, R.H. , Proctor, R., Grossman, J . , Duy, N. and Starl ing, J . : Comparison of the Systemic and Pulmonary Vascular Response to Endotox-in with Plasma and Lung Lymph Lysosomal Enzyme Release: Effect of Steroid Pretreatment. C i r c . Shock 7, 317-331 (1980). 265. Lewis, G.P.: Intracellular Enzymes in Local Lymph as a Measure of Cellular Injury. J. Physiol. 191, 591-607 (1967). 266. Hess, J .M. , MacDonald, R.P., Frederic, R . J . , Jones, R.N. , Neely, J. and Gross, D. : Serum Creatine Phophokinase (CPK) Act iv i ty in Disor-ders of Heart and Skeletal Muscle. Ann. Int. Med. 61, 1015-1028 (1964). — 267. Fronek, A . , Rosen, H. and Witzel, T . : Ornithine Carbamoyl Transferase as an Index of Severity in Hemorrhagic Shock. Proc. Soc. Exp. B io l . Med. 145, 411-414 (1974). 268. Bounous, G. and Hugon, J . : Serum Levels of Intestinal Alkaline Phos-phatase in Relation to Shock. Surgery 69, 238-245 (1971). 269. Artman, R . C , Loegering, D.J. and Smith, J . J . : Changes in Plasma Levels of Lysosomal and Non-Lysosomal Enzymes During Hemorrhagic Hypo-tension. Proc. Soc. Exp. B io l . Med. 149, 1029-1031 (1975). 270. Shibko, S., Caldwell, K .A . , Sawant, P.L. and Tappel, A . L . : Distribu-tion of Lysosomal Enzymes in Animal Tissues. J. C e l l . Comp. Physiol. 61, 85-92 (1963). 271. DeDuve, C. and Wattiaux, R.: Functions of Lysosomes. Ann. Rev. Physiol. 28, 435-492 (1966). 272. Sutherland, N.G. , Bounous, G. and Gurd, F . N . : Role of Intestinal Mucosal Lysosomal Enzymes in the Pathogenesis of Shock. J. Trauma 8, 350-377 (1968). 273. Ravin, H.A. and Fine, J . : Biological Implications of Intestinal Endo-toxins. Fed. Proc. 21, 65-68 (1962). 274. Cuevas, P., de la Maza, L .M. , Gilbert , J. and Fine, J . : The Lung Lesion in Four Different Types of Shock in Rabbits. Arch. Surg. 104, 319-322 (1972). 275. Clermont, H.G. , Adams, J .T. and Williams, J.S.: Source of a Lysosomal Enzyme Acid Phosphatase in Hemorrhagic Shock. Ann. Surg. J75, 19-25 (1972). 276. Berman, I.R., Moseley, R.V. , Lamborn, P.B. and Sleeman, K.: Thoracic Duct Lymph in Shock: Gas Exchange and Balance and Lysosomal Enzymes in Hemorrhagic and Endotoxin Shock. Ann. Surg. 169, 202-209 (1969). - 183 -277. Glenn, T .M. , Lefer, A .M. , Reardsley, A . C . , Ferguson, W.W., Lopez-Rasi, A . M . , Serate, T .S . , Morris, J.R. and Wagensteen, S .L . : Circulatory Responses to Splanchnic Lysosomal Hydrolases in the Dog. Ann. Surg. 176, 120-127 (1972). 278. Steer, C . J . , Kusiak, J.W., Brady, R.O. and Jones, E .A . : Selective Hepatic Uptake of Human ^-Hexosaminidase A by a Specific Glycoprotein Recognition System on Sinusoidal Cel l s . Proc. Natl . Acad. Sci . 76, 2774-2778 (1979). 279. Altura, B.M. and Hershey, S.G.: Sequential Changes in Reticuloendo-thel ia l System Function After Acute Hemorrhage. Proc. Soc. Exp. B io l . Med. 139, 935-939 (1972). 280. Loegering, D. J . , Kaplan, J .E. and Saba, T . M . : Correlation of Plasma Lysosomal Enzymes Levels with Hepatic Reticuloendothelial Function After Trauma. Proc. Soc. Exp. B io l . Med. 152, 42-46 (1976). 281. Brobman, G .F . , Ulano, H.B. , Hinshaw, L.B. and Jacobson, E.D. : Mesen-ter ic Vascular Responses to Endotoxin in the Monkey and Dog. Am. J. Physiol. 219, 1464-1467 (1970). 282. Swan, K .G. , Barton, R.W. and Reynolds, D.G.: Mesenteric Hemodynamics During Endotoxemia in the Baboon. Gastroenterol. 61^ , 872-876 (1971). 283. Alho, A . : Lysosomal Enzymes in Murine Tourniquet Shock. Acta Chir . Scand. 136, 555-560 (1970). 284. Rangel, D.M., Byfield, J . E . , Adomian, G .E . , Stevens, G.H. and Folkal-srud, E.W.: Hepatic Ultrastructral Response to Endotoxin Shock. Surgery 68, 503-511 (1970). 285. Schumer, W.: Lactic Acid as a Factor in the Production of Irreversi-b i l i t y in Oligohaemic Shock. Nature 212, 1210-1212 (1966). 286. Gazzaniga, A.B. and O'Connor, N.E . : Effects of Intravenous Infusion of Autologus Kidney Lysosomal Enzymes in the Dog. Ann. Surg. 172, 804-812 (1970). 287. Janoff, A. and Zweifach, B.W.: Production of Inflammatory Changes in the Microcirculation by Cationic Proteins Extracted from Lysosomes. J. Exp. Med. 120, 747-762 (1964). 288. Wangensteen, S .L . , Geissinger, W.T., Lovett, W.L. , Glenn, T.M. and Lefer, A .M. : Relationship Between Splanchnic Blood Flow and a Myocar-dial Depressant Factor in Endotoxin Shock. Surgery 69, 410-418 (1971). 289. Glenn, T.M. and Lefer, A .M. : Significance of Splanchnic Proteases in the Production of a Toxic Factor in Hemorrhagic Shock. C i r c . Res. 29, 338-349 (1971). - 184 -290. Hinshaw, L .B . : Myocardial Function in Endotoxin Shock. Circ Shock (suppl.) 1, 43-51 (1979). 291. Mela, L . , M i l l e r , L .D. , Bacalzo, L.V. Jr . , Olofsson, K. and White, R.R.: Role of Intracellular Variations of Lysosomal Enzyme Activity and Oxygen Tension in Mitochondrial Impairment in Endotoxemia and Hemorrhage in the Rat. Ann. Surg. 178, 727-735 (1973). 292. Bradley, S.G.: Cellular and Molecular Mechanisms of Action of Bacter-ial Endotoxins. Ann. Rev. Microbiol. 33_, 67-94 (1979). 293. Braude, A . I . , Carey, F.J . and Zalesky, M.: Studies with Radioactive Endotoxin. II. Correlation of Physiologic Effects with Distribution of Radioactivity in Rabbits Injected with Lethal Doses of E. col i Endo-toxin Labelled with Radioactive Sodium Chromate. J . C l i n . Invest. 34, 858-866 (1955). 294. Herring, W.B., Herion, J . C , Walker, R.I. and Palmer, J .G . : Distr ibu-tion and Clearance of Circulating Endotoxin. J. C l i n . Invest. 42 79-82 (1963). 295. Rutenburg, S., Skarnes, R., Palmerio, C. and Fine, J . : Detoxification of Endotoxin by Perfusion of Liver and Spleen. Proc. Soc. Exp. B io l . Med. 125, 455-459 (1967). 296. Farrar, W.E. Jr . : Endotoxin Detoxification by Guinea Pig Tissue Homogenates and Possible Significance of this Reaction In Vivo. Proc. Soc. Exp. B io l . Med. 118, 218-221 (1965). 297. Agarwal, M.K., Parant, M. and Parant, F . : Role of Spleen in Endotoxin Poisoning and Reticuloendothelial Function. Br. J. Exp. Pathol. 53, 485-491 (1972). — 298. Gormus, B.J. and Shands, J.W. Jr . : Endotoxin-Stimulated Spleen Ce l l s : Characterization of the Responding Ce l l s . J . Immunol. 115, 118-123 (1975). 299. Archer, L . T . , Benjamin, B.A., Beller-Todd, B.K., Brackett, D. J . , Wilson, M.F. and Hinshaw, L . B . : Does L D ^ Q O E . co l i Shock Cause Myocardial Fai lure? Circ . Shock 9, 7-16 (1982). 300. Hess, M .L . : Concise Review: Subcellular Function in the Acutely Fai l ing Myocardium. C i rc . Shock 6, 119-136 (1979). 301. Mi l l e r , T . H . , Priano, L . L . , Jorgensen, J.J. and Traber, D . L . : Cardio-respiratory Effects of Psuedomonas and E. co l i Endotoxins in the Awake Dog. Am. J. Physiol. 232, H682-H689 ( 1 9 7 7 7 1 302. Woo, P., Carpenter, M.A. and Trunkey, D. : Ionized Calcium: The Effect of Septic Shock in the Human. J. Surg. Res. 26, 605-610 (1979). - 185 -303. Weisel, R.D., Vito, L . , Dennis, R.C. , Valer i , C.R. and Hechtman, H.B.: Myocardial Depression During Sepsis. Am. J. Surg. 133, 512-521 (1977). 304. Kleinman, W.M., Krause, S.H. and Hess, M.L . : Cardiogenic Endotoxin Shock: Role of Coronary Flow and Subendocardial Ischemia. C i r c . Shock 6, 197-198 (1979). 305. Lefer, A .M. : Mechanisms of Cardiodepression in Endotoxin Shock. C i rc . Shock (suppl.) U 1-8 (1979). 306. McConn, R., Greineder, J .K. , Wasserman, F. and Clowes, G.H.A. J r . : Is There a Humoral Factor That Depresses Ventricular Function in Sepsis? Circ . Shock (suppl.) 1, 9-22 (1979). 307. Maksad, A . K . , Cha, C- J . , Stuart, R . C , Brosco, F.A. and Clowes, G.H.A. J r . : Myocardial Depression in Septic Shock: Physiologic and Metabo-l i c Effects of a Plasma Factor on an Isolated Heart. C i rc . Shock (suppl.) 1 35-42 (1979). 308. Hess, M.L. , Krause, S.M. and Kornwatana, P.: Myocardial Failure and Excitation-Contraction Uncoupling in Canine Endotoxin Shock: Role of Histamine and the Sarcoplasmic Reticulum. C i rc . Shock 7, 277-287 (1980). 309. Wilson, M.F. , Brackett, D.J . , Tompkins, P., Benjamin, B., Archer, L.T. and Hinshaw, L . B . : Elevated Plasma Vasopressin Concentrations During Endotoxin and E. co l i Shock. Adv. Shock Res. (5, 15-26 (1981). 310. McCaig, D.J. and Parratt, J .R.: Reduced Myocardial Response to C a l c i -um During Endotoxin Shock in the Cat. C i rc . Shock 7_, 23-30 (1980). 311. Sardesai, V.M. and Walt, A . J . : Myocardial Enzyme Patterns and Energy Metabolism in Endotoxin Shock. Biochem Med. 5_, 440-446 (1971). 312. Kuttner, R .E . , Apantaku, F.0. and Schumer, W.: Glycolytic Intermedi-ates in Rat Heart After Endotoxin Treatment. C i rc . Shock 7, 405-411 (1980). 313. Hinshaw, L .B . , Archer, L . T . , Greenfield, L . J . , M i l l e r , J .H. and Guenter, C .A . : Effect of Endotoxin on Myocardial Performance. J. Trauma 12, 1056-1062 (1972). 314. Braude, A . I . , Douglas, H. and Davis, C . E . : Treatment and Prevention of Intravascular Coagulation with Antiserum to Endtotoxin. J. Infect. Dis. 128 (suppl.), S157-S164 (1973). 315. Westenfelder, M. and Galanos, C : Experimental Lipid A-Induced Nephritis in the Dog. II. A Possible Role of Lipid A in the Pathogene-sis of Abacterial Pyelonephritis. Infection 2, 174-177 (1974). - 186 -316. Westenfelder, M. , Galanos, C , Madsen, P.O. and Marget, W . : Patholog-ical Act ivi t ies of Lipid A: Experimental Studies in Relation to Chronic Pyelonephritis. In "Microbiology 1977", D. Schlessinger (ed). Am. Soc. Microbio l . , Washington, D.C. , pp. 277-279 (1979). 317. Zlydaszyk, J.D. and Moon, R. J . : Fate of 5 1Cr-Labeled Lipopolysac-charide in Tissue Culture Cells and Livers of Normal Mice. Infect. Immun. 14, 100-105 (1976). 318. White, R.R., Mela, L . , Bacalzo, L . V . , Olofsson, K. and M i l l e r , L .D . : Hepatic Ultrastructure in Endotoxemia, Hemorrhage, and Hypoxia: Emphasis on Mitochondrial Changes. Surgery 73. 525-534 (1973). 319. Hinshaw, L . B . : Concise Review: The Role of Glucose in Endotoxin Shock. Ci rc . Shock 3_, 1-10 (1976). 320. Munson, N.H. , Wendel, T . H . , Eaton, L.R. and Hess, M.L . : Hemodynamics, Hypoglycemia, and Hepato-Pancreatic Pathology During the Course of Endotoxin Shock. Adv. Shock Res. 6, 107-120 (1981). 321. Cameron, D .E . , Chaudry, I .H. , Schleck, S. and Baue, A . E . : Hepatocel-lular Dysfunction in Early Sepsis Despite Increased Hepatic Blood Flow. Adv. Shock Res. 6 65-74 (1981). 322. Mi l l e r , S . I . , Wallace, R.J. J r . , Musher, D.M., Septimus, E . J . , Kohl, S. and Baughn, R .E . : Hypoglycemia as a Manifestation of Sepsis. Am. J. Med. 68, 649-654 (1980). 323. F i lk ins , J.P. and Cornell , R.P.: Depression of Hepatic Gluconeogene-sis and the Hypoglycemia of Endotoxin Shock. Am. J. Physiol. 227, 778-781 (1974). 324. F i lk ins , J .P. : Role of the RES in the Pathogenesis of Endotoxic Hypo-glycemia. C i rc . Shock 9 269-280 (1982). 325. Altura, B.M. and Altura, B.T. : Differential Effects of Anoxia, Exogenous Glucose, and Metabolic Inhibitors on Drug- and Hormone-In-duced Contractions of Arterial Smooth Muscle. C i rc . Shock 3_, 169-189 (1976). 326. Nakatani, T . , Tanaka, J . , Sato, T . , Isoyama, T . , Jones, R.T. , Cowley, R.A. and Trump, B .F . : The Pathophysiology of Septic Shock: Studies of Reticulendothelial System Function and Liver High-Energy Metabolism in Rats Following Sublethal and Lethal Escherichia co l i Injection. Adv. Shock Res. 7, 147-157 (1982). 327. Lewin, I. , Weil, M.H., Shubin, H. and Sherwin, R.: Pulmonary Failure Associated with Cl inical Shock States. J. Trauma 11, 22-35 (1971). 328. Kaplan, R .L . , Sahn, S.A. and Petty, T . L . : Incidence and Outcome of the Respiratory Distress Syndrome in Gram-Negative Sepsis. Arch. Intern. Med. 139, 867-869 (1979). - 187 -329. McGovern, V . J . : The Pathophysiology of Gram-Negative Septicaemia. Pathology 4, 265-271 (1972). 330. Wilson, J.W.: Treatment or Prevention of Pulmonary Cellular Damage with Pharmacologic Doses of Corticosteroid. Surg. Gynecol. Obstet. 134, 675-681 (1972). 331. Ashbaugh, D., Bigelow, D.B., Petty, T .L . and Levine, B .E . : Acute Respiratory Distress in Adults. Lancet _2, 319-323 (1967). 332. Hardway, R.M., James, P.M. Jr . , Anderson, R.W., Bredenberg, C E . and West, R .L . : Intensive Study and Treatment of Shock in Man. J. Am. Med. Assoc. 199, 779-790 (1967). 333. Clowes, G.H.A. Jr . : The Pulmonary Response to Sepsis and Endotoxin: Cl in ica l and Experimental Observations. In "Gram-Negative Bacterial Infections and Mode of Endotoxin Actions. Pathophysiological, Immuno-logical and Cl in ica l Aspects", B. Urbaschek, R. Urbaschek and E. Neter (eds). Springer-Verlag, New York, Vienna, pp. 419-428 (1975). 334. Vaughn, D.L . , Guenter, C A . and Stookey, J . L . : Endotoxin Shock in Primates. Surg. Gynecol. Obstet. ^26, 1309-1317 (1968). 335. Weir, E .K. , Mlczoch, J . , Reeves, J .T. and Grover, R .F . : Endotoxin and Prevention of Hypoxic Pulmonary Vasoconstriction. J. Lab. C l i n . Med. 88, 975-983 (1976). 336. Sayeed, M.M.: Pulmonary Cellular Dysfunction in Endotoxin Shock: Metabolic and Transport Derangements. C i r c . Shock 9, 335-355 (1982). 337. Reeves, J .T. and Grover, R.F. : Blockade of Acute Hypoxic Pulmonary Hypertension by Endotoxin. J. Appl. Physiol. 36, 328-332 (1974). 338. Hales, C .A. , Sonne, L . , Peterson, M. , Kong, D., M i l l e r , M. and Wat-kins, W.D.: Role of Thromboxane and Prostacyclin in Pulmonary Vasomo-tor Changes after Endotoxin in Dogs. J. C l i n . Invest. 68, 497-505 (1981). — 339. Kux, M., Coalson, J . J . , Massion, W.H. and Guenter, C A . : Pulmonary Effects of E. co l i Endotoxin: Role of Leukocytes and Platelets . Ann. Surg. 175, 26-34 (1972). 340. Wasi, S., Murray, R.K., Macmorine, D.R.L. and Movat, H.Z. : The Role of PMN-Leucocyte Lysosomes in Tissue Injury, Inflammation and Hyper-sensit ivity. II. Studies on the Proteolytic Act iv i ty of PMN-Leucocyte Lysosomes of the Rabbit. Br. J. Exp. Pathol. 47, 411-423 (1966). 341. Mori, K., Matsumoto, K. and Gans, H. : On the In Vivo Clearance and Detoxification of Endotoxin by Lung and Liver. Ann. Surg. 177, 159-163 (1973). - 188 -342. Cohen, A.B. and Cl ine, M.J . : The Human Alveolar Macrophage: Isola-t ion, Cultivation In Vitro , and Studies of Morphologic and Functional Characteristics. J . C l i n . Invest. 50, 1390-1398 (1971). 343. Haef fner-Cavail Ion, N. , Chaby, R., Cavaillon, J-M. and Szabo, L . : Lipopolysaccharide Receptor on Rabbit Peritoneal Macrophages. 1. Bind-ing Characteristics. J. Immunol 128, 1950-1954 (1982). 344. Ciznar, I. and Shands, J.W. J r . : Effect of Alkali-Treated Lipopoly-saccharide on Erythrocyte Membrane Stabi l i ty . Infect. Immunol. 4_, 362-367 (1971). 345. Pagani, R., Portoles, M.T. and Municio, A .M. : The Binding of Escheri- chia co l i Endotoxin to Isolated Rat Hepatocytes. FEBS Letters 131, 103-106 (1977). 346. Kabir, S. and Rosenstreich, D.L. : Binding of Bacterial Endotoxin to Murine Spleen Lymphocytes. Infect. Immunol. ^5_, 156-164 (1977). 347. Springer, G.F. and Adye, J . C : Endotoxin-Binding Substances from Human Leukocytes and Platelets. Infect. Immunol. 12, 978-986 (1975). 348. Hawiger, J . , Hawiger, A. and Timmons, S.: Endotoxin-Sensitive Membrane Component of Human Platelets. Nature 256, 125-127 (1975). 349. Hawiger, J . , Hawiger, A . , Steckley, S., Timmons, S., and Cheng, C : Membrane Changes in Human Platelets Induced by Lipopolysaccharide Endotoxin. Br. J. Haematol. 35, 285-299 (1977). 350. Bona, C , Chedid, L. and Lamensans, A . : In Vitro Attachment of Radio-active Endotoxins to Lysosomes. Infect. Immunol. 4, 532-536 (1971). 351. Ahmed, K. and Judah, J .D . : Preparation of Lipoproteins Containing Cation-Dependent ATPase. Biochim. Biophys. Acta 93, 603-613 (1964). 352. L iu , M-S. and Onji , T . : Effects of Endotoxin in the Myocardial (Na ,K )-ATPcise Enzyme System: Involvement of Lipids in Endotox-in-Induced Changes in Enzyme Act iv i t i e s . Biochem. Med. 25, 315-326 (1981). ~ ~ 353. Soulsby, M.E . , Ford, G.D., Davis, J .A . , Wincklhofer, K. and Hess, M.L . : The Effect of Gram-Negative Endotoxin on Calcium Accumulation by Subcellular Fractions of Vascular Smooth Muscle. C i rc . Shock 3, 325-335 (1976). 354. Alho, A . , Motsay, G.J. and L i l l e h e i , R . C : Comparison of Lung Changes in Haemorrhagic and Endotoxin Shock. Ann. Chir . Gynaecol. Fenniae 60, 202-206 (1971). 355. Yokoyama, K., Terao, T. and Osawa, T . : Membrane Receptors of Human Erythrocytes for Bacterial Lipopolysaccharide (LPS). Jap. J. Exp. Med. 48, 511-517 (1978). - 189 -356. Sastrasinh, M. , Weinberg, J.M. and Humes, H.D.: The Effect of Genta-mycin on Calcium Uptake by Renal Mitochondria. Life Sc i . 30, 2309-2315 (1982). 357. Nicholas, G.G. , Mela, L.M. and Mi l l e r . L .D . : Early Alterations in Mitochondrial Membrane Transport during Endotoxemia. J . Surg. Res. 16, 375-383 (1974). 358. Jones, G.R.N, and Kramer, R.: Translocase Inhibition in Endotoxic Shock: A Late Step in the Sequence of Lethal Events. Biochem. Pharmacol. 27, 2931-2937 (1978). 359. Mela, L . : Reversibi l i ty of Mitochondrial Metabolic Response to Circu-latory Shock and Tissue Ischemia. C i rc . Shock (suppl.) 1, 61-67 (1979). 360. McGivney, A. and Bradley, S.G.: Action of Bacterial Endotoxin and Lipid A on Mitochondrial Enzyme Act iv i t ies of Cells in Culture and Subcellular Fractions. Infect. Immun. 25, 664-671 (1979). 361. Altemeier, W.A., Todd, J .C. and Inge, W.W.: Gram-Negative Septicemia: A Growing Threat. Ann. Surg. 166, 530-542 (1967). 362. Godin, D.V. , Tuchek, J .M. and Garnett, M.E. : Studies on the Interac-tion of Escherichia col i Endotoxin with Erythrocyte Membranes. Can. J . Physiol. Pharmacol. 60, 977-985 (1982). 363. Lenaz, G . , Curatola, G. , Mazzanti, L. and Parenti-Castel1i, G. : Biophysical Studies on Agents Affecting the State of Membrane Lipids: Biochemical and Pharmacological Implications. Mol. C e l l . Biochem. 22, 3-32 (1978). 364. Cooper, R.A.: Abnormalities of Cell-Membrane F lu id i ty in the Patho-genesis of Disease. N. Engl. J . Med. 297, 371-377 (1977). 365. Seeman, P.: The Membrane Actions of Anesthetics and Tranquilizers. Pharmacol. Rev. 24, 583-603 (1972). 366. Hughes, R . C : How Do Toxins Penetrate Cel ls? Nature 281 526-527 (1979). 367. Aloni , B. Eitan, A. and Livne, A . : The Erythrocyte Membrane Site for the Effect of Temperature on Osmotic F rag i l i ty . Biochim. Biophys. Acta 465, 46-53 (1977). 368. Shohet, S.B.: Hemolysis and Changes in Erythrocyte Membrane Lipids (Second of Two Parts). N. Engl. J . Med. 286, 638-644 (1972). 369. Hammerling, U. and Westphal, 0.: Synthesis and Use of 0-Stearoyl Polysaccharides in Passive Hemagglutination and Hemolysis. Eur. J. Biochem. 1 46-50 (1967). - 190 -370. Benedetto, D.A., Shands, J.W. Jr. and Shah, D.O.: The Interaction of Bacterial Lipopolysaccharide with Phospholipid Bilayers and Monolay-ers. Biochim. Biophys. Acta 298, 145-157 (1973). 371. Michowitz, M. , Leibovici , J . and Wolman, M.: Protective Effect of Native Levan on Endotoxin Toxicity in Mice and Rats. Experientia 35, 804-805 (1979). 372. Greisman, S .E . , DuBuy, J.B. and Woodward, C . L . : Experimental Gram-Negative Bacterial Sepsis: Prevention of Mortality Not Prevent-able by Antibiotics Alone. Infect. Immun. 25, 538-557 (1979). 373. P h i l l i p s , S.J. and Vick, J .A . : The Pretreatment of E. co l i Endotoxin Shock with WR2823: A New Alpha-Adrenergic Blocking Agent. Surgery 69, 510-514 (1971). 374. Anderson, R.W., James, P.M., Bredenberg, C E . and Hardaway, R.M.: Phenoxybenzamine in Septic Shock. Ann. Surg. 165, 341-350 (1967). 375. Fletcher, J.R. and Ramwell, P.W.: Lidocaine or Indomethacin Improves Survival in Baboon Endotoxin Shock. J . Surg. Res. 24, 154-160 (1978). 376. Fletcher, J.R. and Ramwell, P.W.: E. col i Endotoxin Shock in the Dog; Treatment With Lidocaine or Indomethacin. Br. J. Pharmacol. 64, 185-191 (1978). 377. Reynolds, D.G., Gur l l , N.J . , Vargish, T . , Lechner, R.B. , Faden, A . I . and Holaday, J.W.: Blockade of Opiate Receptors With Naloxone Improves Survival and Cardiac Performance in Canine Endotoxic Shock. Circ . Shock 7, 39-48 (1980). 378. Berk, J . L . , Hagen, J .F . , Beyer, W.H., Gerber, M.J. and Dochat, G.R.: The Treatment of Endotoxin Shock by Beta Adrenergic Blockade. Ann. Surg. 169, 74-81 (1969). 379. Berk, J . L . , Hagen, J .F . , Maly, G. and Koo, R.: The Treatment of Shock with Beta Adrenergic Blockade. Arch. Surg. _104, 46-51 (1972). 380. Parratt, J.R. and Sturgess, R.M.: The Effect of a New Anti-Inf 1 amma-tory Drug, Flurbiprofen, on the Respiratory, Haemodynamic, and Metabo-l i c Responses to E. col i Endotoxin Shock in the Cat. Br. J. Pharmacol. 58, 547-551 (1976"T 381. Goto, F . , Fuj i ta , T . , Otani, E. and Yamamuro, M.: The Effect of Indo-methacin and Adrenergic Receptor Blocking Agents on Rat and Canine Responses to Endotoxin. Circ . Shock 413-424 (1980). 382. Fletcher, J.R. and Ramwell, P.W.: The Effects of Prostacyclin (PGI2) On Endotoxin Shock and Endotoxin-Induced Platelet Aggregation in Dogs. C i rc . Shock 7, 299-308 (1980). - 191 -383. Bronsveld, W., van Lambalgen, A . A . , van den Bos, G . C , Teule, G.J.J , and Thi j s , L . G . : Ventricular Function, Hemodynamics, and Oxygen Consumption During Infusions of Blood and Glucose-Insulin-Potassium (GIK) in Canine Endotoxin Shock. C i rc . Shock 9, 145-156 (1982). 384. Casey, L . C , Fletcher, J.R., Zmudka, M.I. and Ramwell, P.W.: Preven-tion of Endotoxin-induced Pulmonary Hypertension in Primates by the Use of a Selective Thromboxane Synthetase Inhibitor, OXY 1581. J. Pharmacol. Exp. Ther. 222, 441-446 (1982). 385. Cook, J .A . , Wise, W.C. and Halushka, P .V. : Elevated Thromboxane Levels in the Rat During Endotoxic Shock. Protective Effects of Imidazole, 13-Azaprostanoic Acid, or Essential Fatty Acid Deficiency. J . C l i n . Invest. 65, 227-230 (1980). 386. Clermont, H.G. , Williams, J.S. and Adams, J .T . : Steroid Effect on the Release of the Lysosomal Enzyme Acid Phosphatase. Ann. Surg. 179, 917-921 (1974). 387. Schumer, W.: Steroids in the Treatment of Cl inical Septic Shock. Ann. Surg. 184, 333-339 (1976). 388. Godin, D.V. , Au, T. and Garnett, M.E. : Membrane Perturbational Effects of Antiarrhythmic Drugs. J . Mol. C e l l . Cardiol . 11, 261-274 (1979). 389. Eaglstein, W.H., Reasot, H.S. and Hsia, S .L . : Membrane Stabi l izat ion: Effect of Hydrocortisone Sodium Succinate on Phospholipase C Treated Sheep Erythrocytes. Proc. Soc. Exp. B io l . Med. 132, 919-923 (1969). 390. Agarwal, K.N. and Garby, L . : Inhibition by Corticosteroids of Red Cell Lysis In Vi t ro . Acta Endocrinol. Suppl. 93, 3-27 (1964). 391. Gudewicz, P.W.: The Effects of Cortisone Therapy on Lung Macrophage Host Defense Function and Glucose Metabolism. Circ . Shock 8, 95-103 (1981). 392. Schneck, D.W., Pritchard, J .F. and Hayes, H.H. Jr . : Studies on the Uptake and Binding of Propranolol by Rat Tissues. J . Pharmacol. Exp. Ther. 203, 621-629 (1977). 393. Guth, P.S., Amaro, J . , Sellinger, O.Z. and Elmer, L . : Studies In Vitro and In Vivo of the Effects of Chlopromazine on Rat Liver Lyso-somes. Biochem. Pharmacol. 14, 769-775 (1965). 394. Weissmann, G . : The Effects of Corticosteroids, Especially on Lysosomes and Biomembranes. In "Rheumatoid A r t h r i t i s , Pathogenetic Mechanisms and Consequences in Therapeutics". W. Muller, H.G. Harwerth, and K. Fehr, (eds.). Academic Press, London and New York, pp 577-584 (1971). - 192 -395. Cluff, L . E . : Effects of Endotoxins on Susceptibil ity to Infections. J. Infect. Dis. 122, 205-215 (1970). 396. H i l l . A.W., Hibbitt , K.G. and Shears, A . : Endotoxin Tolerance and the Induction of Nonspecific Resistance to Salmonella dub!in and strepto-coccus agalactiae Infections in Mice. Br. J. Exp. Pathol. 55, 269-274 (1974). ~~" 397. Prigal , S.J. : Protection of Neonates Against Whole-Body Radiation by the Administration of a Single Emulsified Injection of a Lipopolysac-charide During Pregnancy. Proc. Soc. Exp. B i o l . Med. 131, 159-163 (1969). 398. Smith, W.W., Brecher, G . , Budd, R.A. and Fred, S.: Effects of Bacter-ial Endotoxin on the Occurrence of Spleen Colonies in Irradiated Mice. Radiat. Res. 27, 369-374 (1966). 399. Bober, L .A . , Kranepool, M.J. and Hollander, V . P . : Inhibitory Effect of Endotoxin on the Growth of Plasma Cell Tumor. Cancer Res. 36, 927-929 (1976). 400. Carswell, E .A . , Old, L . J . , Kassel, R.L. , Green, S., Fiore, N. and Williamson, B.: An Endotoxin-Induced Serum Factor That Causes Necro-sis of Tumors. Proc. Natl. Acad. Sci U.S.A. 72, 3666-3670 (1975). 401. Snyder, S .L . , Walker, R . I . , MacVittie, T . J . and Sheil , J .M. : Biologic Properties of Bacterial Lipopolysaccharides Treated with Chromium Chloride. Can. J . Microbiol. 24, 495-501 (1978). PUBLICATIONS Godin O.V., Ng. T. Wan and Tuchek, J.M. (1976) Studies on the interaction of propranolol with erythrocyte membranes. Biochim. Biophys. Acta 436: 757-773. Godin, D.V., Tuchek, J.M. and Moore, M. (1980) Membrane alterations in actue myocardial ischemia. Can. J . Biochem. 58: 777-786. Godin, D.V., Tuchek, J.M. and Garnett, M.E. (1982) Studies on the interaction of Escherichia co l i endotoxin with erythrocyte membranes. Uah. J . Physiol. Pharmacol. 60: 977-985. Godin, D.V. and Tuchek, J.M. (1983) Plasma acid phosphatase levels in endotoxaemia: modification by drugs and chemically detoxified toxins. Br i t . J . Pharmacol. 79: 421-428. Godin, D.V., Wright, J.M., Tuchek, J.M and Scudamore, C H . (1983) Plasma lysosomal enzymes in experimental and c l in ica l endotoxaemia. C l in . Invest. Med. (in press). Godin, D.V., Tuchek, J.M. and Garnett, M.E. (1983) In vivo and in v itro studies of bacterial emJotoxin-membrane interactions and the effects of membrane-active agents. Br i t . J . Pharmacol, (submitted). Wilson, T.W., McCauley, F.A., Tuchek, J.M. (1983) Determination of 6-keto-prostglandin F i * and thromoboxane B2 In urine by high performance l iquid chromotgraphy and radiioimmunosassay. J . Chromatogr. (in press).
UBC Theses and Dissertations
Gram-negative endotoxaemia Tuchek, John Michael 1983
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.
- 831-UBC_1984_A1 T82.pdf [ 9.96MB ]
- JSON: 831-1.0096559.json
- JSON-LD: 831-1.0096559-ld.json
- RDF/XML (Pretty): 831-1.0096559-rdf.xml
- RDF/JSON: 831-1.0096559-rdf.json
- Turtle: 831-1.0096559-turtle.txt
- N-Triples: 831-1.0096559-rdf-ntriples.txt
- Original Record: 831-1.0096559-source.json
- Full Text
Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url: